1 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
4 all elements for both endian flavors.
6 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
9 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
10 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
11 and SI_REG for 64bit SYSV ABI targets.
13 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
15 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
16 to multi_reg_return. Clarify that we are skipping USEs of multiple
17 return registers. Use bool type where appropriate.
19 2013-10-18 Jan Hubicka <jh@suse.cz>
21 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
23 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
24 (X86_TUNE_PUSH_MEMORY): Likewise.
25 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
26 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
27 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
29 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
30 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
32 (ix86_option_override_internal): Update to use tune features instead
35 2013-10-18 Cong Hou <congh@google.com>
37 PR tree-optimization/58508
38 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
39 statement that contains data refs with zero-step.
41 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
43 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
44 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
45 * gimple-low.c (gimple_check_call_arg,
46 gimple_check_call_matching_types): Move to cgraph.c.
47 * gimple-low.h: Remove prototype.
48 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
49 Relocate from gimple-low.c.
50 * cgraph.h: Add prototype. Don't include basic-block.h.
51 * gimplify.c: Add gimple-low to include list.
52 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
53 * tree-eh.c: Add gimple-low to include list.
54 * tree-nested.c: Likewise.
55 * cfgexpand.c: Add tree-ssa-address.h to include list.
57 * gimple-fold.c: Likewise.
58 * gimple-ssa-strength-reduction.c: Likewise.
59 * trans-mem.c: Likewise.
60 * tree-mudflap.c: Likewise.
61 * tree-ssa-loop-ivopts.c: Likewise.
62 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
63 (degenerate_phi_result): Move to tree-phinodes.c.
64 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
65 * tree-ssa-threadedge.c: Likewise.
66 * tree-vrp.c: Likewise.
67 * tree-phinodes.c (degenerate_phi_result): Relocate here.
68 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
69 * tree-phinodes.h (degenerate_phi_result): Add prototype.
70 * tree-ssa-copy.c: Include tree-ssa-dom.h.
71 * tree-ssa-forwprop.c: Likewise.
72 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
73 pass_data_cleanup_cfg_post_optimizing,
74 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
75 * tree-optimize.c: Delete File.
76 * graphite.c: Include tree-cfgcleanup.h.
78 * tree-cfg.c: Likewise.
79 * tree-profile.c: Likewise.
80 * tree-ssa-dse.c: Likewise.
81 * tree-ssa-loop-ivcanon.c: Likewise.
82 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
83 * tree-outof-ssa.c: Include sbitmap.h.
84 * tree-ssa-live.c: Likewise.
85 * tree-ssa-propagate.c: Likewise.
86 * tree-ssa-structalias.c: Likewise.
87 * tree-stdarg.c: Likewise.
88 * Makefile.in (OBJS): Delete tree-optimize.o.
89 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
90 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
91 * varasm.c: Include basic-block.h.
92 * cfgloop.h: Include function.h instead of basic-block.h
93 (bb_loop_depth): Move to cfgloop.c.
94 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
96 2013-10-18 Teresa Johnson <tejohnson@google.com>
98 * predict.c (probably_never_executed): Compare frequency-based
99 count to number of training runs.
100 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
102 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
104 * config/arm/arm.c (cortexa9_extra_costs): New table.
105 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
107 2013-10-18 Jeff Law <law@redhat.com>
109 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
111 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
112 (register_jump_thread): Add "registered_jump_thread" debug
114 * dbgcnt.def (registered_jump_thread): New debug counter.
116 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
118 * config/rs6000/rs6000.c: Include cgraph.h.
120 2013-10-18 Teresa Johnson <tejohnson@google.com>
122 * tree-ssa-tail-merge.c (replace_block_by): Update edge
123 weights during merging.
125 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
127 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
128 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
129 * tree-ssa-operands.h: Remove prototype.
130 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
131 * gimple.c (virtual_operand_p): Relocate from gimple.c.
132 * gimple.h: Add prototype.
133 * gimple-ssa.h: Include tree-ssa-operands.h.
134 * tree-dump.c: Add tree-cfg.h to include list.
135 * tree-ssa-alias.c: Add ipa-reference.h to include list.
136 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
137 * config/i386/i386.c: Don't include tree-flow.h.
138 * config/rs6000/rs6000.c: Likewise.
140 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
142 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
144 2013-10-18 Richard Biener <rguenther@suse.de>
146 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
147 or TYPE_UNSIGNED of integral types.
148 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
149 NULL_TREE for zero-precision integral types.
151 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
153 * config/aarch64/arm_neon.h
154 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
156 2013-10-17 Sriraman Tallam <tmsriram@google.com>
158 * opth-gen.awk: Define target_flags_explicit.
160 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
162 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
163 fields to the reg_addr array that describes the valid addressing
164 mode for any register, general purpose registers, floating point
165 registers, and Altivec registers.
166 (FIRST_RELOAD_REG_CLASS): Likewise.
167 (LAST_RELOAD_REG_CLASS): Likewise.
168 (struct reload_reg_map_type): Likewise.
169 (reload_reg_map_type): Likewise.
170 (RELOAD_REG_VALID): Likewise.
171 (RELOAD_REG_MULTIPLE): Likewise.
172 (RELOAD_REG_INDEXED): Likewise.
173 (RELOAD_REG_OFFSET): Likewise.
174 (RELOAD_REG_PRE_INCDEC): Likewise.
175 (RELOAD_REG_PRE_MODIFY): Likewise.
176 (reg_addr): Likewise.
177 (mode_supports_pre_incdec_p): New helper functions to say whether
178 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
179 (mode_supports_pre_modify_p): Likewise.
180 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
181 print the valid address mode bits for each mode.
182 (rs6000_debug_print_mode): Likewise.
183 (rs6000_debug_reg_global): Likewise.
184 (rs6000_setup_reg_addr_masks): New function to set up the address
185 mask bits for each type.
186 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
187 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
188 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
189 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
190 PRE_MODIFY are supported.
191 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
192 registers, instead of {src,dest}_av_p.
193 (rs6000_print_options_internal): Tweak the debug output slightly.
195 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
197 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
200 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
202 * tree-flow.h (struct omp_region): Move to omp-low.c.
203 Remove omp_ prototypes and variables.
204 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
205 (copy_var_decl): Relocate prototype from tree-flow.h.
206 * gimple.c (copy_var_decl): Relocate from omp-low.c.
207 * tree.h: Move prototype to omp-low.h.
208 * omp-low.h: New File. Relocate prototypes here.
209 * omp-low.c (struct omp_region): Make local here.
210 (root_omp_region): Make static.
211 (copy_var_decl) Move to gimple.c.
212 (new_omp_region): Make static.
213 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
214 * tree-cfg.c: Include omp-low.h.
215 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
216 * gimplify.c: Include omp-low.h.
217 * tree-parloops.c: Likewise.
219 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
221 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
222 operands are in memory, prefer to force non-matched operand 1 to
225 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
228 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
229 restrict TImode addresses to single indirect registers if both
230 -mquad-memory and -mvsx-timode are used.
231 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
232 we should emit load/store quad. Remove using %y for quad memory
235 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
236 constraints to allow load/store quad on machines where TImode is
237 not allowed in VSX registers. Use 'n' instead of 'F' constraint
238 for TImode to load integer constants.
240 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
242 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
244 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
246 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
247 handling of STACK_REG.
249 2013-10-17 Richard Biener <rguenther@suse.de>
251 PR tree-optimization/58143
252 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
254 (rewrite_to_defined_overflow): Likewise.
255 (move_computations_dom_walker::before_dom): Rewrite stmts
256 with undefined signed overflow that are not always executed
257 into unsigned arithmetic.
259 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
262 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
263 explicit isa flag to be an options variable, instead of using
264 global_options_set. Remove define from rs6000.h.
265 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
267 * config/rs6000/rs6000.c (rs6000_option_override_internal):
268 Initialize rs6000_isa_flags_explicit.
269 (rs6000_function_specific_save): Add gcc_options* parameter, so
270 that the powerpc builds after the 2013-10-15 changes.
271 (rs6000_function_specific_restore): Likewise.
273 2013-10-16 DJ Delorie <dj@redhat.com>
275 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
276 op is a REG before checking REGNO.
277 (rl78_alloc_physical_registers): Verify pattern is a SET before
280 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
282 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
284 (vec_unpacks_lo_v4sf): Likewise.
285 (vec_unpacks_float_hi_v4si): Likewise.
286 (vec_unpacks_float_lo_v4si): Likewise.
287 (vec_unpacku_float_hi_v4si): Likewise.
288 (vec_unpacku_float_lo_v4si): Likewise.
290 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
292 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
293 (vsx_concat_v2sf): Likewise.
295 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
297 * config/aarch64/aarch64.md
298 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
300 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
302 PR tree-optimization/58697
303 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
304 estimated_loop_iterations_int.
305 (max_stmt_executions_int): Call get_max_loop_iterations_int.
306 (get_max_loop_iterations_int): New. HWINT version of
307 get_max_loop_iterations.
308 * cfgloop.h: Add prototypes.
309 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
310 * loop-unroll.c (decide_peel_once_rolling): Call
311 get_estimated_loop_iterations_int.
312 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
313 * tree-ssa-loop-niter.h: Tweak prototypes.
315 2013-10-16 David Malcolm <dmalcolm@redhat.com>
317 * gengtype-parse.c (struct_field_seq): Ignore access-control
318 keywords ("public:" etc).
320 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
322 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
323 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
325 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
327 * config/arm/arm.opt (mlra): New option.
328 * config/arm/arm.c (arm_lra_p): New function.
329 (TARGET_LRA_P): Define.
331 2013-10-16 Paulo Matos <pmatos@broadcom.com>
333 * tree-core.h (tree_code_name): Remove.
334 * tree.h (get_tree_code_name): New prototype.
335 * tree.c (tree_code_name): Make static.
336 (get_tree_code_name): New function.
337 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
338 tree_class_check_failed, tree_range_check_failed,
339 tree_not_class_check_failed, omp_clause_check_failed,
340 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
341 wrapper get_tree_code_name instead of calling tree_code_name directly.
342 * tree-vrp.c (dump_asserts_for): Likewise.
343 * tree-dump.c (dequeue_and_dump): Likewise.
344 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
345 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
346 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
347 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
348 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
349 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
350 dump_gimple_omp_for): Likewise.
351 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
352 * tree-ssa-pre.c (print_pre_expr): Likewise.
353 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
354 * print-tree.c (print_node_brief, print_node): Likewise.
355 * gimple.c (gimple_check_failed): Likewise.
356 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
357 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
358 * config/mep/mep.c (mep_validate_vliw): Likewise.
359 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
360 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
362 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
364 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
367 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
369 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
370 (MULTILIB_MATCHES): Add multilib for -march=v8.
372 2013-10-15 Sriraman Tallam <tmsriram@google.com>
375 * optc-save-gen.awk: Add extra parameter to the save and restore
377 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
378 * tree.c (build_optimization_node): New parameter. Add extra parameter
379 to call to cl_optimization_save.
380 (build_target_option_node): New parameter. Add extra parameter
381 to call to cl_target_option_save.
382 * tree.h (build_optimization_node): New parameter.
383 (build_target_option_node): New parameter.
384 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
385 build_optimization_node and build_target_option_node.
386 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
387 (handle_pragma_push_options): Ditto.
388 * toplev.c (process_options): Ditto.
389 * opts.c (init_options_struct): Check for opts_set non-null.
390 * target.def (target_option.save): New parameter.
391 (target_option.restore): New parameter.
393 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
394 (ix86_pragma_target_parse): Ditto.
395 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
397 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
398 to build_optimization_node and build_target_option_node.
399 (rs6000_valid_attribute_p): Ditto.
400 (rs6000_pragma_target_parse): Ditto.
401 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
403 * config/i386/i386.h:
404 TARGET_64BIT_P: New Macro
405 TARGET_MMX_P: New Macro.
406 TARGET_3DNOW_P: New Macro.
407 TARGET_3DNOW_A_P: New Macro.
408 TARGET_SSE_P: New Macro.
409 TARGET_SSE2_P: New Macro.
410 TARGET_SSE3_P: New Macro.
411 TARGET_SSSE3_P: New Macro.
412 TARGET_SSE4_1_P: New Macro.
413 TARGET_SSE4_2_P: New Macro.
414 TARGET_AVX_P: New Macro.
415 TARGET_AVX2_P: New Macro.
416 TARGET_AVX512F_P: New Macro.
417 TARGET_AVX512PF_P: New Macro.
418 TARGET_AVX512ER_P: New Macro.
419 TARGET_AVX512CD_P: New Macro.
420 TARGET_FMA_P: New Macro.
421 TARGET_SSE4A_P: New Macro.
422 TARGET_FMA4_P: New Macro.
423 TARGET_XOP_P: New Macro.
424 TARGET_LWP_P: New Macro.
425 TARGET_ABM_P: New Macro.
426 TARGET_BMI_P: New Macro.
427 TARGET_BMI2_P: New Macro.
428 TARGET_LZCNT_P: New Macro.
429 TARGET_TBM_P: New Macro.
430 TARGET_POPCNT_P: New Macro.
431 TARGET_SAHF_P: New Macro.
432 TARGET_MOVBE_P: New Macro.
433 TARGET_CRC32_P: New Macro.
434 TARGET_AES_P: New Macro.
435 TARGET_PCLMUL_P: New Macro.
436 TARGET_CMPXCHG16B_P: New Macro.
437 TARGET_FSGSBASE_P: New Macro.
438 TARGET_RDRND_P: New Macro.
439 TARGET_F16C_P: New Macro.
440 TARGET_RTM_P: New Macro.
441 TARGET_HLE_P: New Macro.
442 TARGET_RDSEED_P: New Macro.
443 TARGET_PRFCHW_P: New Macro.
444 TARGET_ADX_P: New Macro.
445 TARGET_FXSR_P: New Macro.
446 TARGET_XSAVE_P: New Macro.
447 TARGET_XSAVEOPT_P: New Macro.
448 TARGET_LP64_P: New Macro.
449 TARGET_X32_P: New Macro.
450 TARGET_FPMATH_DEFAULT_P: New Macro.
451 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
452 * config/i386/i386.c (ix86_option_override_internal): New parameters.
454 Change ix86_tune_string to access opts->x_ix86_tune_string.
455 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
456 Change ix86_arch_string to access opts->x_ix86_arch_string.
457 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
458 Change ix86_pmode to access opts->x_ix86_pmode.
459 Change ix86_abi to access opts->x_ix86_abi.
460 Change ix86_cmodel to access opts->x_ix86_cmodel.
461 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
462 Change ix86_isa_flags_explicit to access
463 opts->x_ix86_isa_flags_explicit.
464 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
465 Change ix86_regparm to access opts->x_ix86_regparm.
466 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
467 Change ix86_preferred_stack_boundary_arg to access
468 opts->x_ix86_preferred_stack_boundary_arg.
469 Change ix86_force_align_arg_pointer to access
470 opts->x_ix86_force_align_arg_pointer.
471 Change ix86_incoming_stack_boundar_arg to access
472 opts->x_ix86_incoming_stack_boundar_arg.
473 Change ix86_fpmath to access opts->x_ix86_fpmath.
474 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
475 Change ix86_recip_name to access opts->x_ix86_recip_name.
476 Change ix86_stack_protector_guard to access
477 opts->x_ix86_stack_protector_guard.
478 Change ix86_tune_memcpy_strategy to access
479 opts->x_ix86_tune_memcpy_strategy.
480 Change ix86_tune_memset_strategy to access
481 opts->x_ix86_tune_memset_strategy.
482 Change global_options to access opts.
483 Change global_options_set to access opts_set.
484 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
485 Change TARGET_MMX to TARGET_MMX_P (opts->...).
486 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
487 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
488 Change TARGET_SSE to TARGET_SSE_P (opts->...).
489 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
490 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
491 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
492 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
493 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
494 Change TARGET_AVX to TARGET_AVX_P (opts->...).
495 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
496 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
497 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
498 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
499 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
500 Change TARGET_FMA to TARGET_FMA_P (opts->...).
501 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
502 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
503 Change TARGET_XOP to TARGET_XOP_P (opts->...).
504 Change TARGET_LWP to TARGET_LWP_P (opts->...).
505 Change TARGET_ABM to TARGET_ABM_P (opts->...).
506 Change TARGET_BMI to TARGET_BMI_P (opts->...).
507 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
508 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
509 Change TARGET_TBM to TARGET_TBM_P (opts->...).
510 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
511 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
512 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
513 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
514 Change TARGET_AES to TARGET_AES_P (opts->...).
515 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
516 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
517 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
518 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
519 Change TARGET_F16C to TARGET_F16C_P (opts->...).
520 Change TARGET_RTM to TARGET_RTM_P (opts->...).
521 Change TARGET_HLE to TARGET_HLE_P (opts->...).
522 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
523 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
524 Change TARGET_ADX to TARGET_ADX_P (opts->...).
525 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
526 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
527 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
528 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
529 Change TARGET_X32 to TARGET_X32_P (opts->...).
530 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
531 Change TARGET_FLOAT_RETURNS_IN_80387 to
532 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
533 (ix86_function_specific_save): New parameter. Use opts-> fields
534 to replace global fields.
535 (ix86_function_specific_restore): Ditto.
536 (ix86_valid_target_attribute_inner_p): New parameters.
538 Fix call to ix86_handle_option and set_option.
539 (ix86_valid_target_attribute_tree): New parameters.
540 Change global_options to access opts.
541 Change global_options_set to access opts_set.
542 Fix call to ix86_valid_target_attribute_inner_p.
543 Change ix86_tune_string to access opts->x_ix86_tune_string.
544 Change ix86_arch_string to access opts->x_ix86_arch_string.
545 Change ix86_fpmath to access opts->x_ix86_fpmath
546 Fix call to ix86_option_override_internal.
547 Fix call to ix86_add_new_builtins.
548 Fix calls to build_optimization_node and build_target_option_node.
549 (ix86_valid_target_attribute_p): Remove access to global_options.
550 Use new gcc_options structure func_options.
551 Fix call to ix86_valid_target_attribute_tree.
552 Fix call to build_optimization_node.
553 (get_builtin_code_for_version): Fix call to
554 ix86_valid_target_attribute_tree.
556 2013-10-15 David Malcolm <dmalcolm@redhat.com>
558 * Makefile.in (PICFLAG): New.
559 (enable_host_shared): New.
560 (INTERNAL_CFLAGS): Use PICFLAG.
561 (LIBIBERTY): Use pic build of libiberty.a if configured with
562 --enable-host-shared.
563 * configure.ac: Add --enable-host-shared, setting up new
565 * configure: Regenerate.
566 * doc/install.texi (--enable-shared): Add note contrasting it with ...
567 (--enable-host-shared): New option.
569 2013-10-15 Richard Biener <rguenther@suse.de>
571 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
572 for built-in functions.
574 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
576 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
577 (optimize_range_tests_1): New function,
578 extracted from optimize_range_tests.
579 (optimize_range_tests_xor): Similarly.
580 (optimize_range_tests_diff): New function.
581 (optimize_range_tests): Use optimize_range_tests_1.
583 2013-10-15 Cong Hou <congh@google.com>
585 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
586 requirement of the reduction pattern so that one operand of the
587 reduction operation can come from outside of the loop.
589 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
591 * config/arm/neon-schedgen.ml: Remove.
592 * config/arm/cortex-a9-neon.md: Remove comment regarding
595 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
597 * config/arm/types: Remove old neon types.
599 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
601 * config/arm/cortex-a7.md
602 (cortex_a7_neon_type): New.
603 (cortex_a7_neon_mul): Update for new types.
604 (cortex_a7_neon_mla): Likewise.
605 (cortex_a7_neon): Likewise.
607 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
609 * config/arm/cortex-a15-neon.md
610 (cortex_a15_neon_type): New,
612 (cortex_a15_neon_int_1): Remove.
613 (cortex_a15_neon_int_2): Likewise.
614 (cortex_a15_neon_int_3): Likewise.
615 (cortex_a15_neon_int_4): Likewise.
616 (cortex_a15_neon_int_5): Likewise.
617 (cortex_a15_neon_vqneg_vqabs): Likewise.
618 (cortex_a15_neon_vmov): Likewise.
619 (cortex_a15_neon_vaba): Likewise.
620 (cortex_a15_neon_vaba_qqq): Likewise.
621 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
622 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
623 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
625 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
626 (cortex_a15_neon_mla_qqq_8_16): Likewise.
627 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
628 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
629 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
630 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
631 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
632 (cortex_a15_neon_shift_1): Likewise.
633 (cortex_a15_neon_shift_2): Likewise.
634 (cortex_a15_neon_shift_3): Likewise.
635 (cortex_a15_neon_vshl_ddd): Likewise.
636 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
637 (cortex_a15_neon_vsra_vrsra): Likewise.
638 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
639 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
640 (cortex_a15_neon_bp_3cycle): Likewise.
641 (cortex_a15_neon_ldm_2): Likewise.
642 (cortex_a15_neon_stm_2): Likewise.
643 (cortex_a15_neon_mcr): Likewise.
644 (cortex_a15_neon_mrc): Likewise.
645 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
646 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
647 (cortex_a15_neon_fp_vmul_ddd): Likewise.
648 (cortex_a15_neon_fp_vmul_qqd): Likewise.
649 (cortex_a15_neon_fp_vmla_ddd): Likewise.
650 (cortex_a15_neon_fp_vmla_qqq): Likewise.
651 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
652 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
653 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
654 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
655 (cortex_a15_neon_bp_simple): Likewise.
656 (cortex_a15_neon_bp_2cycle): Likewise.
657 (cortex_a15_neon_bp_3cycle): Likewise.
658 (cortex_a15_neon_vld1_1_2_regs): Likewise.
659 (cortex_a15_neon_vld1_3_4_regs): Likewise.
660 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
661 (cortex_a15_neon_vld2_4_regs): Likewise.
662 (cortex_a15_neon_vld3_vld4): Likewise.
663 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
664 (cortex_a15_neon_vst1_3_4_regs): Likewise.
665 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
666 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
667 (cortex_a15_neon_vst3_vst4): Rename to...
668 (cortex_a15_neon_vst4): This, update for new attributes.
669 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
670 (cortex_a15_neon_vld3_vld4_lane): Likewise.
671 (cortex_a15_neon_vst1_vst2_lane): Likewise.
672 (cortex_a15_neon_vst3_vst4_lane): Likewise.
673 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
674 (cortex_a15_neon_ldm_2): Likewise.
675 (cortex_a15_neon_stm_2): Likewise.
676 (cortex_a15_neon_mcr): Likewise.
677 (cortex_a15_neon_mcr_2_mcrr): Likewise.
678 (cortex_a15_neon_mrc): Likewise.
679 (cortex_a15_neon_mrrc): Likewise.
681 (cortex_a15_neon_abd): New.
682 (cortex_a15_neon_abd_q): Likewise.
683 (cortex_a15_neon_aba): Likewise.
684 (cortex_a15_neon_aba_q): Likewise.
685 (cortex_a15_neon_acc): Likewise.
686 (cortex_a15_neon_acc_q): Likewise.
687 (cortex_a15_neon_arith_basic): Likewise.
688 (cortex_a15_neon_arith_complex): Likewise.
689 (cortex_a15_neon_multiply): Likewise.
690 (cortex_a15_neon_multiply_q): Likewise.
691 (cortex_a15_neon_mla): Likewise.
692 (cortex_a15_neon_mla_q): Likewise.
693 (cortex_a15_neon_sat_mla_long): Likewise.
694 (cortex_a15_neon_shift_acc): Likewise.
695 (cortex_a15_neon_shift_imm_basic): Likewise.
696 (cortex_a15_neon_shift_imm_complex): Likewise.
697 (cortex_a15_neon_shift_reg_basic): Likewise.
698 (cortex_a15_neon_shift_reg_basic_q): Likewise.
699 (cortex_a15_neon_shift_reg_complex): Likewise.
700 (cortex_a15_neon_shift_reg_complex_q): Likewise.
701 (cortex_a15_neon_fp_negabs): Likewise
702 (cortex_a15_neon_fp_arith): Likewise
703 (cortex_a15_neon_fp_arith_q): Likewise
704 (cortex_a15_neon_fp_cvt_int): Likewise
705 (cortex_a15_neon_fp_cvt_int_q): Likewise
706 (cortex_a15_neon_fp_cvt_16): Likewise
707 (cortex_a15_neon_fp_mul): Likewise
708 (cortex_a15_neon_fp_mul_q): Likewise
709 (cortex_a15_neon_fp_mla): Likewise
710 (cortex_a15_neon_fp_mla_q): Likewise
711 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
712 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
713 (cortex_a15_neon_bitops): Likewise.
714 (cortex_a15_neon_bitops_q): Likewise.
715 (cortex_a15_neon_from_gp): Likewise.
716 (cortex_a15_neon_from_gp_q): Likewise.
717 (cortex_a15_neon_tbl3_tbl4): Likewise.
718 (cortex_a15_neon_zip_q): Likewise.
719 (cortex_a15_neon_to_gp): Likewise.
720 (cortex_a15_neon_load_a): Likewise.
721 (cortex_a15_neon_load_b): Likewise.
722 (cortex_a15_neon_load_c): Likewise.
723 (cortex_a15_neon_load_d): Likewise.
724 (cortex_a15_neon_load_e): Likewise.
725 (cortex_a15_neon_load_f): Likewise.
726 (cortex_a15_neon_store_a): Likewise.
727 (cortex_a15_neon_store_b): Likewise.
728 (cortex_a15_neon_store_c): Likewise.
729 (cortex_a15_neon_store_d): Likewise.
730 (cortex_a15_neon_store_e): Likewise.
731 (cortex_a15_neon_store_f): Likewise.
732 (cortex_a15_neon_store_g): Likewise.
733 (cortex_a15_neon_store_h): Likewise.
734 (cortex_a15_vfp_to_from_gp): Likewise.
736 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
738 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
740 (cortex_a9_neon_vshl_ddd): Remove.
741 (cortex_a9_neon_vst3_vst4): Likewise.
742 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
744 (cortex_a9_neon_bit_ops_q): New.
746 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
747 (cortex_a9_neon_int_2): Likewise.
748 (cortex_a9_neon_int_3): Likewise.
749 (cortex_a9_neon_int_4): Likewise.
750 (cortex_a9_neon_int_5): Likewise.
751 (cortex_a9_neon_vqneg_vqabs): Likewise.
752 (cortex_a9_neon_vmov): Likewise.
753 (cortex_a9_neon_vaba): Likewise.
754 (cortex_a9_neon_vaba_qqq): Likewise.
755 (cortex_a9_neon_shift_1): Likewise.
756 (cortex_a9_neon_shift_2): Likewise.
757 (cortex_a9_neon_shift_3): Likewise.
758 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
759 (cortex_a9_neon_vsra_vrsra): Likewise.
760 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
761 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
762 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
764 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
765 (cortex_a9_neon_mla_qqq_8_16): Likewise.
766 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
768 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
769 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
770 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
771 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
772 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
773 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
774 (cortex_a9_neon_fp_vsum): Likewise.
775 (cortex_a9_neon_fp_vmul_ddd): Likewise.
776 (cortex_a9_neon_fp_vmul_qqd): Likewise.
777 (cortex_a9_neon_fp_vmla_ddd): Likewise.
778 (cortex_a9_neon_fp_vmla_qqq): Likewise.
779 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
780 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
781 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
782 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
783 (cortex_a9_neon_bp_simple): Likewise.
784 (cortex_a9_neon_bp_2cycle): Likewise.
785 (cortex_a9_neon_bp_3cycle): Likewise.
786 (cortex_a9_neon_ldr): Likewise.
787 (cortex_a9_neon_str): Likewise.
788 (cortex_a9_neon_vld1_1_2_regs): Likewise.
789 (cortex_a9_neon_vld1_3_4_regs): Likewise.
790 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
791 (cortex_a9_neon_vld2_4_regs): Likewise.
792 (cortex_a9_neon_vld3_vld4): Likewise.
793 (cortex_a9_neon_vld1_vld2_lane): Likewise.
794 (cortex_a9_neon_vld3_vld4_lane): Likewise.
795 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
796 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
797 (cortex_a9_neon_vst1_3_4_regs): Likewise.
798 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
799 (cortex_a9_neon_vst1_vst2_lane): Likewise.
800 (cortex_a9_neon_vst3_vst4_lane): Likewise.
801 (cortex_a9_neon_mcr): Likewise.
802 (cortex_a9_neon_mcr_2_mcrr): Likewise.
803 (cortex_a9_neon_mrc): Likewise.
804 (cortex_a9_neon_mrrc): Likewise.
806 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
808 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
810 (cortex_a8_neon_vshl_ddd): Remove.
811 (cortex_a8_neon_vst3_vst4): Likewise.
812 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
814 (cortex_a8_neon_bit_ops_q): New.
816 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
817 (cortex_a8_neon_int_2): Likewise..
818 (cortex_a8_neon_int_3): Likewise.
819 (cortex_a8_neon_int_5): Likewise.
820 (cortex_a8_neon_vqneg_vqabs): Likewise.
821 (cortex_a8_neon_int_4): Likewise.
822 (cortex_a8_neon_vaba): Likewise.
823 (cortex_a8_neon_vaba_qqq): Likewise.
824 (cortex_a8_neon_shift_1): Likewise.
825 (cortex_a8_neon_shift_2): Likewise.
826 (cortex_a8_neon_shift_3): Likewise.
827 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
828 (cortex_a8_neon_vsra_vrsra): Likewise.
829 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
830 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
831 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
833 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
834 (cortex_a8_neon_mla_qqq_8_16): Likewise.
835 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
837 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
838 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
839 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
840 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
841 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
842 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
843 (cortex_a8_neon_fp_vsum): Likewise.
844 (cortex_a8_neon_fp_vmul_ddd): Likewise.
845 (cortex_a8_neon_fp_vmul_qqd): Likewise.
846 (cortex_a8_neon_fp_vmla_ddd): Likewise.
847 (cortex_a8_neon_fp_vmla_qqq): Likewise.
848 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
849 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
850 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
851 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
852 (cortex_a8_neon_bp_simple): Likewise.
853 (cortex_a8_neon_bp_2cycle): Likewise.
854 (cortex_a8_neon_bp_3cycle): Likewise.
855 (cortex_a8_neon_ldr): Likewise.
856 (cortex_a8_neon_str): Likewise.
857 (cortex_a8_neon_vld1_1_2_regs): Likewise.
858 (cortex_a8_neon_vld1_3_4_regs): Likewise.
859 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
860 (cortex_a8_neon_vld2_4_regs): Likewise.
861 (cortex_a8_neon_vld3_vld4): Likewise.
862 (cortex_a8_neon_vld1_vld2_lane): Likewise.
863 (cortex_a8_neon_vld3_vld4_lane): Likewise.
864 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
865 (cortex_a8_neon_vst1_3_4_regs): Likewise.
866 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
867 (cortex_a8_neon_vst1_vst2_lane): Likewise.
868 (cortex_a8_neon_vst3_vst4_lane): Likewise.
869 (cortex_a8_neon_mcr): Likewise.
870 (cortex_a8_neon_mcr_2_mcrr): Likewise.
871 (cortex_a8_neon_mrc): Likewise.
872 (cortex_a8_neon_mrrc): Likewise.
874 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
876 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
878 * config/aarch64/aarch64-simd.md (neon_type): Remove.
879 (aarch64_simd_dup<mode>): Add "type" attribute.
880 (aarch64_dup_lane<mode>): Likewise.
881 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
882 (*aarch64_simd_mov<mode>): Likewise.
883 (aarch64_simd_mov_from_<mode>low): Likewise.
884 (aarch64_simd_mov_from_<mode>high): Likewise.
885 (orn<mode>3): Likewise.
886 (bic<mode>3): Likewise.
887 (add<mode>3): Likewise.
888 (sub<mode>3): Likewise.
889 (mul<mode>3): Likewise.
890 (*aarch64_mul3_elt<mode>): Likewise.
891 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
892 (*aarch64_mul3_elt_to_128df): Likewise.
893 (*aarch64_mul3_elt_to_64v2df): Likewise.
894 (neg<mode>2): Likewise.
895 (abs<mode>2): Likewise.
896 (abd<mode>_3): Likewise.
897 (aba<mode>_3): Likewise.
898 (fabd<mode>_3): Likewise.
899 (*fabd_scalar<mode>3): Likewise.
900 (and<mode>3): Likewise.
901 (ior<mode>3): Likewise.
902 (xor<mode>3): Likewise.
903 (one_cmpl<mode>2): Likewise.
904 (aarch64_simd_vec_set<mode>): Likewise.
905 (aarch64_simd_lshr<mode>): Likewise.
906 (aarch64_simd_ashr<mode>): Likewise.
907 (aarch64_simd_imm_shl<mode>): Likewise.
908 (aarch64_simd_reg_sshl<mode): Likewise.
909 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
910 (aarch64_simd_reg_shl<mode>_signed): Likewise.
911 (aarch64_simd_vec_setv2di): Likewise.
912 (aarch64_simd_vec_set<mode>): Likewise.
913 (aarch64_mla<mode>): Likewise.
914 (*aarch64_mla_elt<mode>): Likewise.
915 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
916 (aarch64_mls<mode>): Likewise.
917 (*aarch64_mls_elt<mode>): Likewise.
918 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
919 (<su><maxmin><mode>3): Likewise.
920 (move_lo_quad_<mode>): Likewise.
921 (aarch64_simd_move_hi_quad_<mode>): Likewise.
922 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
923 (vec_pack_trunc_<mode>): Likewise.
924 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
925 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
926 (*aarch64_<su>mlal_lo<mode>): Likewise.
927 (*aarch64_<su>mlal_hi<mode>): Likewise.
928 (*aarch64_<su>mlsl_lo<mode>): Likewise.
929 (*aarch64_<su>mlsl_hi<mode>): Likewise.
930 (*aarch64_<su>mlal<mode>): Likewise.
931 (*aarch64_<su>mlsl<mode>): Likewise.
932 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
933 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
934 (add<mode>3): Likewise.
935 (sub<mode>3): Likewise.
936 (mul<mode>3): Likewise.
937 (div<mode>3): Likewise.
938 (neg<mode>2): Likewise.
939 (abs<mode>2): Likewise.
940 (fma<mode>4): Likewise.
941 (*aarch64_fma4_elt<mode>): Likewise.
942 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
943 (*aarch64_fma4_elt_to_128df): Likewise.
944 (*aarch64_fma4_elt_to_64v2df): Likewise.
945 (fnma<mode>4): Likewise.
946 (*aarch64_fnma4_elt<mode>): Likewise.
947 (*aarch64_fnma4_elt_<vswap_width_name><mode>
948 (*aarch64_fnma4_elt_to_128df): Likewise.
949 (*aarch64_fnma4_elt_to_64v2df): Likewise.
950 (<frint_pattern><mode>2): Likewise.
951 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
952 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
953 (vec_unpacks_lo_v4sf): Likewise.
954 (aarch64_float_extend_lo_v2df): Likewise.
955 (vec_unpacks_hi_v4sf): Likewise.
956 (aarch64_float_truncate_lo_v2sf): Likewise.
957 (aarch64_float_truncate_hi_v4sf): Likewise.
958 (aarch64_vmls<mode>): Likewise.
959 (<su><maxmin><mode>3): Likewise.
960 (<maxmin_uns><mode>3): Likewise.
961 (reduc_<sur>plus_<mode>): Likewise.
962 (reduc_<sur>plus_v2di): Likewise.
963 (reduc_<sur>plus_v2si): Likewise.
964 (reduc_<sur>plus_<mode>): Likewise.
965 (aarch64_addpv4sf): Likewise.
966 (clz<mode>2): Likewise.
967 (reduc_<maxmin_uns>_<mode>): Likewise.
968 (reduc_<maxmin_uns>_v2di): Likewise.
969 (reduc_<maxmin_uns>_v2si): Likewise.
970 (reduc_<maxmin_uns>_<mode>): Likewise.
971 (reduc_<maxmin_uns>_v4sf): Likewise.
972 (aarch64_simd_bsl<mode>_internal): Likewise.
973 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
974 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
975 (aarch64_get_lane<mode>): Likewise.
976 (*aarch64_combinez<mode>): Likewise.
977 (aarch64_combine<mode>): Likewise.
978 (aarch64_simd_combine<mode>): Likewise.
979 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
980 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
981 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
982 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
983 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
984 (aarch64_<sur>h<addsub><mode>): Likewise.
985 (aarch64_<sur><addsub>hn<mode>): Likewise.
986 (aarch64_<sur><addsub>hn2<mode>): Likewise.
987 (aarch64_pmul<mode>): Likewise.
988 (aarch64_<su_optab><optab><mode>): Likewise.
989 (aarch64_<sur>qadd<mode>): Likewise.
990 (aarch64_sqmovun<mode>): Likewise.
991 (aarch64_<sur>qmovn<mode>): Likewise.
992 (aarch64_s<optab><mode>): Likewise.
993 (aarch64_sq<r>dmulh<mode>): Likewise.
994 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
995 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
996 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
997 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
998 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
999 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1000 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1001 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1002 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1003 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1004 (aarch64_sqdmull<mode>): Likewise.
1005 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1006 (aarch64_sqdmull_n<mode>): Likewise.
1007 (aarch64_sqdmull2<mode>_internal): Likewise.
1008 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1009 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1010 (aarch64_<sur>shl<mode>): Likewise.
1011 (aarch64_<sur>q<r>shl<mode>
1012 (aarch64_<sur>shll_n<mode>): Likewise.
1013 (aarch64_<sur>shll2_n<mode>): Likewise.
1014 (aarch64_<sur>shr_n<mode>): Likewise.
1015 (aarch64_<sur>sra_n<mode>): Likewise.
1016 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1017 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1018 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1019 (aarch64_cm<optab><mode>): Likewise.
1020 (aarch64_cm<optab>di): Likewise.
1021 (aarch64_cm<optab><mode>): Likewise.
1022 (aarch64_cm<optab>di): Likewise.
1023 (aarch64_cmtst<mode>): Likewise.
1024 (aarch64_cmtstdi): Likewise.
1025 (aarch64_cm<optab><mode>): Likewise.
1026 (*aarch64_fac<optab><mode>): Likewise.
1027 (aarch64_addp<mode>): Likewise.
1028 (aarch64_addpdi): Likewise.
1029 (sqrt<mode>2): Likewise.
1030 (vec_load_lanesoi<mode>): Likewise.
1031 (vec_store_lanesoi<mode>): Likewise.
1032 (vec_load_lanesci<mode>): Likewise.
1033 (vec_store_lanesci<mode>): Likewise.
1034 (vec_load_lanesxi<mode>): Likewise.
1035 (vec_store_lanesxi<mode>): Likewise.
1036 (*aarch64_mov<mode>): Likewise.
1037 (aarch64_ld2<mode>_dreg): Likewise.
1038 (aarch64_ld2<mode>_dreg): Likewise.
1039 (aarch64_ld3<mode>_dreg): Likewise.
1040 (aarch64_ld3<mode>_dreg): Likewise.
1041 (aarch64_ld4<mode>_dreg): Likewise.
1042 (aarch64_ld4<mode>_dreg): Likewise.
1043 (aarch64_tbl1<mode>): Likewise.
1044 (aarch64_tbl2v16qi): Likewise.
1045 (aarch64_combinev16qi): Likewise.
1046 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1047 (aarch64_st2<mode>_dreg): Likewise.
1048 (aarch64_st2<mode>_dreg): Likewise.
1049 (aarch64_st3<mode>_dreg): Likewise.
1050 (aarch64_st3<mode>_dreg): Likewise.
1051 (aarch64_st4<mode>_dreg): Likewise.
1052 (aarch64_st4<mode>_dreg): Likewise.
1053 (*aarch64_simd_ld1r<mode>): Likewise.
1054 (aarch64_frecpe<mode>): Likewise.
1055 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1056 (aarch64_frecps<mode>): Likewise.
1058 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1060 * config/arm/iterators.md (V_elem_ch): New.
1062 (VQH_type): Likewise.
1063 * config/arm/arm.md (is_neon_type): New.
1064 (conds): Use is_neon_type.
1065 (anddi3_insn): Update type attribute.
1066 (xordi3_insn): Likewise.
1067 (one_cmpldi2): Likewise.
1068 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1069 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1070 (*movmisalign<mode>_neon_store): Likewise.
1071 (*movmisalign<mode>_neon_load): Likewise.
1072 (vec_set<mode>_internal): Likewise.
1073 (vec_set<mode>_internal): Likewise.
1074 (vec_setv2di_internal): Likewise.
1075 (vec_extract<mode>): Likewise.
1076 (vec_extract<mode>): Likewise.
1077 (vec_extractv2di): Likewise.
1078 (*add<mode>3_neon): Likewise.
1079 (adddi3_neon): Likewise.
1080 (*sub<mode>3_neon): Likewise.
1081 (subdi3_neon): Likewise.
1082 (fma<VCVTF:mode>4): Likewise.
1083 (fma<VCVTF:mode>4_intrinsic): Likewise.
1084 (*fmsub<VCVTF:mode>4): Likewise.
1085 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1086 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1087 (ior<mode>3): Likewise.
1088 (and<mode>3): Likewise.
1089 (orn<mode>3_neon): Likewise.
1090 (orndi3_neon): Likewise.
1091 (bic<mode>3_neon): Likewise.
1092 (bicdi3_neon): Likewise.
1093 (xor<mode>3): Likewise.
1094 (one_cmpl<mode>2): Likewise.
1095 (abs<mode>2): Likewise.
1096 (neg<mode>2): Likewise.
1097 (negdi2_neon): Likewise.
1098 (*umin<mode>3_neon): Likewise.
1099 (*umax<mode>3_neon): Likewise.
1100 (*smin<mode>3_neon): Likewise.
1101 (*smax<mode>3_neon): Likewise.
1102 (vashl<mode>3): Likewise.
1103 (vashr<mode>3_imm): Likewise.
1104 (vlshr<mode>3_imm): Likewise.
1105 (ashl<mode>3_signed): Likewise.
1106 (ashl<mode>3_unsigned): Likewise.
1107 (neon_load_count): Likewise.
1108 (ashldi3_neon_noclobber): Likewise.
1109 (ashldi3_neon): Likewise.
1110 (signed_shift_di3_neon): Likewise.
1111 (unsigned_shift_di3_neon): Likewise.
1112 (ashrdi3_neon_imm_noclobber): Likewise.
1113 (lshrdi3_neon_imm_noclobber): Likewise.
1114 (<shift>di3_neon): Likewise.
1115 (widen_ssum<mode>3): Likewise.
1116 (widen_usum<mode>3): Likewise.
1117 (quad_halves_<code>v4si): Likewise.
1118 (quad_halves_<code>v4sf): Likewise.
1119 (quad_halves_<code>v8hi): Likewise.
1120 (quad_halves_<code>v16qi): Likewise.
1121 (reduc_splus_v2di): Likewise.
1122 (neon_vpadd_internal<mode>): Likewise.
1123 (neon_vpsmin<mode>): Likewise.
1124 (neon_vpsmax<mode>): Likewise.
1125 (neon_vpumin<mode>): Likewise.
1126 (neon_vpumax<mode>): Likewise.
1127 (*ss_add<mode>_neon): Likewise.
1128 (*us_add<mode>_neon): Likewise.
1129 (*ss_sub<mode>_neon): Likewise.
1130 (*us_sub<mode>_neon): Likewise.
1131 (neon_vadd<mode>_unspec): Likewise.
1132 (neon_vaddl<mode>): Likewise.
1133 (neon_vaddw<mode>): Likewise.
1134 (neon_vhadd<mode>): Likewise.
1135 (neon_vqadd<mode>): Likewise.
1136 (neon_vaddhn<mode>): Likewise.
1137 (neon_vmul<mode>): Likewise.
1138 (neon_vfms<VCVTF:mode>): Likewise.
1139 (neon_vmlal<mode>): Likewise.
1140 (neon_vmls<mode>): Likewise.
1141 (neon_vmlsl<mode>): Likewise.
1142 (neon_vqdmulh<mode>): Likewise.
1143 (neon_vqdmlal<mode>): Likewise.
1144 (neon_vqdmlsl<mode>): Likewise.
1145 (neon_vmull<mode>): Likewise.
1146 (neon_vqdmull<mode>): Likewise.
1147 (neon_vsub<mode>_unspec): Likewise.
1148 (neon_vsubl<mode>): Likewise.
1149 (neon_vsubw<mode>): Likewise.
1150 (neon_vqsub<mode>): Likewise.
1151 (neon_vhsub<mode>): Likewise.
1152 (neon_vsubhn<mode>): Likewise.
1153 (neon_vceq<mode>): Likewise.
1154 (neon_vcge<mode>): Likewise.
1155 (neon_vcgeu<mode>): Likewise.
1156 (neon_vcgt<mode>): Likewise.
1157 (neon_vcgtu<mode>): Likewise.
1158 (neon_vcle<mode>): Likewise.
1159 (neon_vclt<mode>): Likewise.
1160 (neon_vcage<mode>): Likewise.
1161 (neon_vcagt<mode>): Likewise.
1162 (neon_vtst<mode>): Likewise.
1163 (neon_vabd<mode>): Likewise.
1164 (neon_vabdl<mode>): Likewise.
1165 (neon_vaba<mode>): Likewise.
1166 (neon_vabal<mode>): Likewise.
1167 (neon_vmax<mode>): Likewise.
1168 (neon_vmin<mode>): Likewise.
1169 (neon_vpaddl<mode>): Likewise.
1170 (neon_vpadal<mode>): Likewise.
1171 (neon_vpmax<mode>): Likewise.
1172 (neon_vpmin<mode>): Likewise.
1173 (neon_vrecps<mode>): Likewise.
1174 (neon_vrsqrts<mode>): Likewise.
1175 (neon_vqabs<mode>): Likewise.
1176 (neon_vqneg<mode>): Likewise.
1177 (neon_vcls<mode>): Likewise.
1178 (clz<mode>2): Likewise.
1179 (popcount<mode>2): Likewise.
1180 (neon_vrecpe<mode>): Likewise.
1181 (neon_vrsqrte<mode>): Likewise.
1182 (neon_vget_lane<mode>_sext_internal): Likewise.
1183 (neon_vget_lane<mode>_zext_internal): Likewise.
1184 (neon_vdup_n<mode>): Likewise.
1185 (neon_vdup_n<mode>): Likewise.
1186 (neon_vdup_nv2di): Likewise.
1187 (neon_vdup_lane<mode>_interal): Likewise.
1188 (*neon_vswp<mode>): Likewise.
1189 (neon_vcombine<mode>): Likewise.
1190 (float<mode><V_cvtto>2): Likewise.
1191 (floatuns<mode><V_cvtto>2): Likewise.
1192 (fix_trunc<mode><V_cvtto>2): Likewise.
1193 (fixuns_trunc<mode><V_cvtto>2
1194 (neon_vcvt<mode>): Likewise.
1195 (neon_vcvt<mode>): Likewise.
1196 (neon_vcvtv4sfv4hf): Likewise.
1197 (neon_vcvtv4hfv4sf): Likewise.
1198 (neon_vcvt_n<mode>): Likewise.
1199 (neon_vcvt_n<mode>): Likewise.
1200 (neon_vmovn<mode>): Likewise.
1201 (neon_vqmovn<mode>): Likewise.
1202 (neon_vqmovun<mode>): Likewise.
1203 (neon_vmovl<mode>): Likewise.
1204 (neon_vmul_lane<mode>): Likewise.
1205 (neon_vmul_lane<mode>): Likewise.
1206 (neon_vmull_lane<mode>): Likewise.
1207 (neon_vqdmull_lane<mode>): Likewise.
1208 (neon_vqdmulh_lane<mode>): Likewise.
1209 (neon_vqdmulh_lane<mode>): Likewise.
1210 (neon_vmla_lane<mode>): Likewise.
1211 (neon_vmla_lane<mode>): Likewise.
1212 (neon_vmlal_lane<mode>): Likewise.
1213 (neon_vqdmlal_lane<mode>): Likewise.
1214 (neon_vmls_lane<mode>): Likewise.
1215 (neon_vmls_lane<mode>): Likewise.
1216 (neon_vmlsl_lane<mode>): Likewise.
1217 (neon_vqdmlsl_lane<mode>): Likewise.
1218 (neon_vext<mode>): Likewise.
1219 (neon_vrev64<mode>): Likewise.
1220 (neon_vrev32<mode>): Likewise.
1221 (neon_vrev16<mode>): Likewise.
1222 (neon_vbsl<mode>_internal): Likewise.
1223 (neon_vshl<mode>): Likewise.
1224 (neon_vqshl<mode>): Likewise.
1225 (neon_vshr_n<mode>): Likewise.
1226 (neon_vshrn_n<mode>): Likewise.
1227 (neon_vqshrn_n<mode>): Likewise.
1228 (neon_vqshrun_n<mode>): Likewise.
1229 (neon_vshl_n<mode>): Likewise.
1230 (neon_vqshl_n<mode>): Likewise.
1231 (neon_vqshlu_n<mode>): Likewise.
1232 (neon_vshll_n<mode>): Likewise.
1233 (neon_vsra_n<mode>): Likewise.
1234 (neon_vsri_n<mode>): Likewise.
1235 (neon_vsli_n<mode>): Likewise.
1236 (neon_vtbl1v8qi): Likewise.
1237 (neon_vtbl2v8qi): Likewise.
1238 (neon_vtbl3v8qi): Likewise.
1239 (neon_vtbl4v8qi): Likewise.
1240 (neon_vtbl1v16qi): Likewise.
1241 (neon_vtbl2v16qi): Likewise.
1242 (neon_vcombinev16qi): Likewise.
1243 (neon_vtbx1v8qi): Likewise.
1244 (neon_vtbx2v8qi): Likewise.
1245 (neon_vtbx3v8qi): Likewise.
1246 (neon_vtbx4v8qi): Likewise.
1247 (*neon_vtrn<mode>_insn): Likewise.
1248 (*neon_vzip<mode>_insn): Likewise.
1249 (*neon_vuzp<mode>_insn): Likewise.
1250 (neon_vld1<mode>): Likewise.
1251 (neon_vld1_lane<mode>): Likewise.
1252 (neon_vld1_lane<mode>): Likewise.
1253 (neon_vld1_dup<mode>): Likewise.
1254 (neon_vld1_dup<mode>): Likewise.
1255 (neon_vld1_dupv2di): Likewise.
1256 (neon_vst1<mode>): Likewise.
1257 (neon_vst1_lane<mode>): Likewise.
1258 (neon_vst1_lane<mode>): Likewise.
1259 (neon_vld2<mode>): Likewise.
1260 (neon_vld2<mode>): Likewise.
1261 (neon_vld2_lane<mode>): Likewise.
1262 (neon_vld2_lane<mode>): Likewise.
1263 (neon_vld2_dup<mode>): Likewise.
1264 (neon_vst2<mode>): Likewise.
1265 (neon_vst2<mode>): Likewise.
1266 (neon_vst2_lane<mode>): Likewise.
1267 (neon_vst2_lane<mode>): Likewise.
1268 (neon_vld3<mode>): Likewise.
1269 (neon_vld3qa<mode>): Likewise.
1270 (neon_vld3qb<mode>): Likewise.
1271 (neon_vld3_lane<mode>): Likewise.
1272 (neon_vld3_lane<mode>): Likewise.
1273 (neon_vld3_dup<mode>): Likewise.
1274 (neon_vst3<mode>): Likewise.
1275 (neon_vst3qa<mode>): Likewise.
1276 (neon_vst3qb<mode>): Likewise.
1277 (neon_vst3_lane<mode>): Likewise.
1278 (neon_vst3_lane<mode>): Likewise.
1279 (neon_vld4<mode>): Likewise.
1280 (neon_vld4qa<mode>): Likewise.
1281 (neon_vld4qb<mode>): Likewise.
1282 (neon_vld4_lane<mode>): Likewise.
1283 (neon_vld4_lane<mode>): Likewise.
1284 (neon_vld4_dup<mode>): Likewise.
1285 (neon_vst4<mode>): Likewise.
1286 (neon_vst4qa<mode>): Likewise.
1287 (neon_vst4qb<mode>): Likewise.
1288 (neon_vst4_lane<mode>): Likewise.
1289 (neon_vst4_lane<mode>): Likewise.
1290 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1291 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1292 (neon_vec_<US>mult_lo_<mode>): Likewise.
1293 (neon_vec_<US>mult_hi_<mode>): Likewise.
1294 (neon_vec_<US>shiftl_<mode>): Likewise.
1295 (neon_unpack<US>_<mode>): Likewise.
1296 (neon_vec_<US>mult_<mode>): Likewise.
1297 (vec_pack_trunc_<mode>): Likewise.
1298 (neon_vec_pack_trunc_<mode>): Likewise.
1299 (neon_vabd<mode>_2): Likewise.
1300 (neon_vabd<mode>_3): Likewise.
1302 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1304 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1305 (load_pair): Update type attribute.
1306 (store_pair): Update type attribute.
1307 * config/aarch64/iterators.md (q): New.
1309 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1311 * config/arm/types.md: Add new types for Neon insns.
1313 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1314 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1315 Sergey Lega <sergey.s.lega@intel.com>
1316 Anna Tikhonova <anna.tikhonova@intel.com>
1317 Ilya Tocar <ilya.tocar@intel.com>
1318 Andrey Turetskiy <andrey.turetskiy@intel.com>
1319 Ilya Verbin <ilya.verbin@intel.com>
1320 Kirill Yukhin <kirill.yukhin@intel.com>
1321 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1323 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1324 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1325 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1327 (srcp14<mode>): Ditto.
1328 (rsqrt14<mode>): Ditto.
1329 (rsqrt14<mode>): Ditto.
1330 (avx512f_vmscalef<mode>): Ditto.
1331 (avx512f_scalef<mode>): Ditto.
1332 (avx512f_getexp<mode>): Ditto.
1333 (avx512f_sgetexp<mode>): Ditto.
1334 (avx512f_fixupimm<mode>): Ditto.
1335 (avx512f_sfixupimm<mode>): Ditto.
1336 (avx512f_rndscale<mode>): Ditto.
1337 (*avx512er_exp2<mode>): Ditto.
1338 (*avx512er_rcp28<mode>): Ditto.
1339 (avx512er_rsqrt28<mode>): Ditto.
1340 (avx512f_getmant<mode>): Ditto.
1341 (avx512f_getmant<mode>): Ditto.
1342 (avx512f_rndscale<mode>): Fix formatting.
1344 2013-10-15 Martin Jambor <mjambor@suse.cz>
1346 * ipa-utils.h (ipa_edge_within_scc): Declare.
1347 * ipa-cp.c (edge_within_scc): Moved...
1348 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1350 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1351 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1352 Sergey Lega <sergey.s.lega@intel.com>
1353 Anna Tikhonova <anna.tikhonova@intel.com>
1354 Ilya Tocar <ilya.tocar@intel.com>
1355 Andrey Turetskiy <andrey.turetskiy@intel.com>
1356 Ilya Verbin <ilya.verbin@intel.com>
1357 Kirill Yukhin <kirill.yukhin@intel.com>
1358 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1360 * config/i386/predicates.md (const_8_to_15_operand): New.
1361 (const_16_to_31_operand): Ditto.
1362 * config/i386/sse.md (V8FI): New.
1364 (reduc_splus_v8df): Ditto.
1365 (reduc_splus_v16sf): Ditto.
1366 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1367 (vec_extract_hi_<mode>): Ditto.
1368 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1369 (vec_set_lo_<mode>): Ditto.
1370 (vec_set_hi_<mode>): Ditto.
1371 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1372 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1373 (avx512f_pshufd_1): Ditto.
1374 (avx512f_broadcast<mode>): Ditto.
1375 (avx512f_broadcast<mode>): Ditto.
1376 (define_split): Split vec_extract_lo into move.
1377 (ssequartermode): Ditto.
1378 (ssedoublemode): Extened with wider modes.
1379 (vec_extract_lo_<mode>): Ditto.
1381 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1382 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1383 Sergey Lega <sergey.s.lega@intel.com>
1384 Anna Tikhonova <anna.tikhonova@intel.com>
1385 Ilya Tocar <ilya.tocar@intel.com>
1386 Andrey Turetskiy <andrey.turetskiy@intel.com>
1387 Ilya Verbin <ilya.verbin@intel.com>
1388 Kirill Yukhin <kirill.yukhin@intel.com>
1389 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1391 * config/i386/predicates.md (register_or_constm1_operand): New.
1392 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1393 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1394 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1395 UNSPEC_SCATTER_PREFETCH
1397 (avx512f_ucmp<mode>3): Ditto.
1398 (avx512f_vternlog<mode>): Ditto.
1399 (avx512f_align<mode>): Ditto.
1400 (<shift_insn><mode>3): Ditto.
1401 (avx512f_<rotate>v<mode>): Ditto.
1402 (avx512f_<rotate><mode>): Ditto.
1403 (avx512f_eq<mode>3): Ditto.
1404 (avx512f_eq<mode>3_1): Ditto.
1405 (avx512f_gt<mode>3): Ditto.
1406 (avx512f_testm<mode>3): Ditto.
1407 (avx512f_testnm<mode>3): Ditto.
1408 (avx512pf_gatherpf<mode>): Ditto.
1409 (*avx512pf_gatherpf<mode>_mask): Ditto.
1410 (*avx512pf_gatherpf<mode>): Ditto.
1411 (avx512pf_scatterpf<mode>): Ditto.
1412 (*avx512pf_scatterpf<mode>_mask): Ditto.
1413 (*avx512pf_scatterpf<mode>): Ditto.
1414 (avx512f_vec_dup_gpr<mode>): Ditto.
1415 (clz<mode>2): Ditto.
1416 (conflict<mode>): Ditto.
1417 (REDUC_SMINMAX_MODE): Extened with wider modes.
1418 (reduc_<code>_<mode>): Ditto.
1419 (vlshr<mode>3): Ditto.
1420 (vashl<mode>3): Ditto.
1422 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1423 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1424 Sergey Lega <sergey.s.lega@intel.com>
1425 Anna Tikhonova <anna.tikhonova@intel.com>
1426 Ilya Tocar <ilya.tocar@intel.com>
1427 Andrey Turetskiy <andrey.turetskiy@intel.com>
1428 Ilya Verbin <ilya.verbin@intel.com>
1429 Kirill Yukhin <kirill.yukhin@intel.com>
1430 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1432 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1435 (avx512fmaskmode): Ditto.
1436 (bcstscalarsuff): Ditto.
1437 (avx512f_blendm<mode>): Ditto.
1438 (cmp_imm_predicate): Ditto.
1439 (avx512f_cmp<mode>3): Ditto.
1440 (avx512f_vec_dup<mode>): Ditto.
1441 (avx512f_vec_dup_mem<mode>): Ditto.
1442 (avx512f_vpermi2var<mode>3): Ditto.
1443 (avx512f_vpermt2var<mode>3): Ditto.
1444 (vec_init<mode>): Ditto.
1445 (avx512f_gathersi<mode>): Ditto.
1446 (*avx512f_gathersi<mode>): Ditto.
1447 (*avx512f_gathersi<mode>_2): Ditto.
1448 (avx512f_gatherdi<mode>): Ditto.
1449 (*avx512f_gatherdi<mode>): Ditto.
1450 (*avx512f_gatherdi<mode>_2): Ditto.
1451 (avx512f_scattersi<mode>): Ditto.
1452 (*avx512f_scattersi<mode>): Ditto.
1453 (avx512f_scatterdi<mode>): Ditto.
1454 (*avx512f_scatterdi<mode>): Ditto.
1455 (sseintprefix): Extened with wider modes.
1456 (VEC_GATHER_IDXSI): Ditto.
1457 (VEC_GATHER_IDXDI): Ditto.
1458 (VEC_GATHER_SRCDI): Ditto.
1460 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1461 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1463 * config/arm/t-aprofile: New file.
1464 * config.gcc: Handle --with-multilib-list option.
1466 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1468 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1471 2013-10-15 Richard Biener <rguenther@suse.de>
1473 * tree-loop-distribution.c (build_empty_rdg): Inline into
1475 (rdg_flag_vertex): Inline into single user.
1476 (rdg_flag_vertex_and_dependent): Likewise.
1477 (build_rdg_partition_for_vertex): Remove processed bitmap.
1478 (rdg_build_partitions): Simplify.
1480 2013-10-15 Richard Biener <rguenther@suse.de>
1482 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1483 Restructure forwarding through conversions and copies to
1484 avoid performing copy-propagation the wrong way. Adjust
1485 recursion invocations.
1486 (forward_propagate_addr_expr): Add argument stating if we
1487 are recursing from a single-use.
1488 (ssa_forward_propagate_and_combine): Adjust.
1490 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1492 * dumpfile.h (gcc::dump_manager): New class, to hold state
1493 relating to dumpfile management.
1494 (get_dump_file_name): Remove in favor of method of dump_manager.
1495 (dump_initialized_p): Likewise.
1496 (dump_start): Likewise.
1497 (dump_finish): Likewise.
1498 (dump_switch_p): Likewise.
1499 (dump_register): Likewise.
1500 (get_dump_file_info): Likewise.
1501 * context.c (gcc::context::context): Construct the dump_manager
1503 * context.h (gcc::context::get_dumps): New.
1504 (gcc::context::m_dumps): New.
1505 * coverage.c (coverage_init): Port to dump_manager API.
1506 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1507 (extra_dump_files_in_use): Likewise.
1508 (extra_dump_files_alloced): Likewise.
1509 (gcc::dump_manager::dump_manager): New.
1510 (dump_register): Convert to...
1511 (gcc::dump_manager::dump_register): ...method, replacing
1512 function-static next_dump with m_next_dump field.
1513 (get_dump_file_info): Convert to...
1514 (gcc::dump_manager::get_dump_file_info): ...method.
1515 (get_dump_file_name): Convert to...
1516 (gcc::dump_manager::get_dump_file_name): ...method.
1517 (dump_start): Convert to...
1518 (gcc::dump_manager::dump_start): ...method.
1519 (dump_finish): Convert to...
1520 (gcc::dump_manager::dump_finish): ...method.
1521 (dump_begin): Replace body with...
1522 (gcc::dump_manager::dump_begin): ...new method.
1523 (dump_phase_enabled_p): Convert to...
1524 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1525 (dump_phase_enabled_p): Convert to...
1526 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1527 (dump_initialized_p): Convert to...
1528 (gcc::dump_manager::dump_initialized_p): ...method.
1529 (dump_flag_name): Replace body with...
1530 (gcc::dump_manager::dump_flag_name): ...new method.
1531 (dump_enable_all): Convert to...
1532 (gcc::dump_manager::dump_enable_all): ...new method.
1533 (opt_info_enable_passes): Convert to...
1534 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1535 (dump_switch_p_1): Convert to...
1536 (gcc::dump_manager::dump_switch_p_1): ...new method.
1537 (dump_switch_p): Convert to...
1538 (gcc::dump_manager::dump_switch_p): ...new method.
1539 (opt_info_switch_p): Port to dump_manager API.
1540 (enable_rtl_dump_file): Likewise.
1541 * opts-global.c (handle_common_deferred_options): Port to new
1543 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1544 (pass_manager::register_one_dump_file): Likewise.
1545 (pass_manager::register_pass): Likewise.
1546 (pass_init_dump_file): Likewise.
1547 (pass_fini_dump_file): Likewise.
1548 * statistics.c (statistics_early_init): Likewise.
1550 2013-10-14 Richard Biener <rguenther@suse.de>
1552 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1553 iterative_hash_canonical_type, gimple_canonical_type_hash,
1554 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1555 gimple_register_canonical_type, print_gimple_types_stats,
1556 free_gimple_type_tables): Move to lto/lto.c
1557 (gt-gimple.h): Do not include.
1558 * gimple.h (gimple_register_canonical_type,
1559 print_gimple_types_stats, free_gimple_type_tables): Remove.
1560 * Makefile.in (GTFILES): Remove gimple.c.
1562 2013-10-14 Travis Snoozy <quandary@remstate.com>
1565 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1566 scanning for msp430x targets.
1568 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1571 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1572 (registers_ok_for_ldd_peep): Move around.
1573 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1574 * config/sparc/sparc.md (widening peepholes): Use it.
1576 2013-10-14 Richard Biener <rguenther@suse.de>
1580 * gimple.c (iterative_hash_canonical_type): Make sure to
1581 record the hash into the correct hashtable slot.
1583 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1585 PR rtl-optimization/58662
1586 * combine.c (try_combine): Take into account death nodes on I2 when
1587 splitting a PARALLEL of two independent SETs. Fix dump message.
1589 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1592 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1593 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1594 * config/sh/t-sh (sh_treg_combine.o): New entry.
1595 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1596 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1597 (register_sh_passes): New function. Register sh_treg_combine pass.
1598 (sh_option_override): Invoke it.
1599 (sh_canonicalize_comparison): Handle op0_preserve_value.
1600 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1601 opportunities. Canonicalize branch condition.
1602 (nott): Allow only if pseudos can be created for non-SH2A.
1604 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1607 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1608 (ix86_expand_movmem): Replace copy_addr_to_reg with
1609 ix86_copy_addr_to_reg.
1610 (ix86_expand_setmem): Likewise.
1612 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1614 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1615 provided by ix86_fp_compare_mode instead of CCFPUmode.
1617 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1619 * config/aarch64/arm_neon.h
1620 (vtbx<1,3>_<psu>8): Fix register constriants.
1622 2013-10-11 Jeff Law <law@redhat.com>
1624 PR tree-optimization/58640
1625 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1626 threading paths that cross over two loop entry points.
1628 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1630 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1631 handle vector float as well.
1632 (*vsx_le_perm_load_v4si): Likewise.
1633 (*vsx_le_perm_store_v2di): Likewise.
1634 (*vsx_le_perm_store_v4si): Likewise.
1636 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1638 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1639 directly to circumvent subtract from splat{31} workaround.
1640 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1642 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1643 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1644 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1645 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1646 separate big and little endian logic.
1647 (*altivec_vperm_<mode>_internal): New define_insn.
1648 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1649 separate big and little endian logic.
1650 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1651 (vec_permv16qi): Add little endian logic.
1653 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1655 * doc/extend.texi (returns_nonnull): Remove arguments.
1657 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1658 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1659 Sergey Lega <sergey.s.lega@intel.com>
1660 Anna Tikhonova <anna.tikhonova@intel.com>
1661 Ilya Tocar <ilya.tocar@intel.com>
1662 Andrey Turetskiy <andrey.turetskiy@intel.com>
1663 Ilya Verbin <ilya.verbin@intel.com>
1664 Kirill Yukhin <kirill.yukhin@intel.com>
1665 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1667 * config/i386/sse.md (VI48F_256_512): New.
1668 (avx2_permvar<mode>): Change to ...
1669 (<avx2_avx512f>_permvar<mode>): This.
1671 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1672 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1673 Sergey Lega <sergey.s.lega@intel.com>
1674 Anna Tikhonova <anna.tikhonova@intel.com>
1675 Ilya Tocar <ilya.tocar@intel.com>
1676 Andrey Turetskiy <andrey.turetskiy@intel.com>
1677 Ilya Verbin <ilya.verbin@intel.com>
1678 Kirill Yukhin <kirill.yukhin@intel.com>
1679 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1681 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1682 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1683 * config/i386/sse.md (VI4_AVX): New.
1684 (sf2simodelower): Ditto.
1685 (sse2_cvtps2dq): Change to ...
1686 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1688 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1689 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1690 Sergey Lega <sergey.s.lega@intel.com>
1691 Anna Tikhonova <anna.tikhonova@intel.com>
1692 Ilya Tocar <ilya.tocar@intel.com>
1693 Andrey Turetskiy <andrey.turetskiy@intel.com>
1694 Ilya Verbin <ilya.verbin@intel.com>
1695 Kirill Yukhin <kirill.yukhin@intel.com>
1696 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1698 * config/i386/sse.md (V_512): New.
1700 (vcond<V_512:mode><VF_512:mode>): Ditto.
1701 (vcond<V_512:mode><VI_512:mode>): Ditto.
1702 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1704 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1705 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1706 Sergey Lega <sergey.s.lega@intel.com>
1707 Anna Tikhonova <anna.tikhonova@intel.com>
1708 Ilya Tocar <ilya.tocar@intel.com>
1709 Andrey Turetskiy <andrey.turetskiy@intel.com>
1710 Ilya Verbin <ilya.verbin@intel.com>
1711 Kirill Yukhin <kirill.yukhin@intel.com>
1712 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1714 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1715 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1717 (fma<mode>4): Removed condition.
1718 (fms<mode>4): Ditto.
1719 (fnma<mode>4): Ditto.
1720 (fnms<mode>4): Ditto.
1721 (fma4i_fmadd_<mode>): Ditto.
1722 (*fma_fmadd_<mode>): Ditto.
1723 (*fma_fmsub_<mode>): Ditto.
1724 (*fma_fnmadd_<mode>): Ditto.
1725 (*fma_fnmsub_<mode>): Ditto.
1726 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1727 (*fma_fmaddsub_<mode>): Ditto.
1728 (*fma_fmsubadd_<mode>): Ditto.
1729 (*fmai_fmadd_<mode>): Ditto.
1730 (*fmai_fmsub_<mode>): Ditto.
1731 (*fmai_fnmadd_<mode>): Ditto.
1732 (*fmai_fnmsub_<mode>): Ditto.
1734 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1735 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1736 Sergey Lega <sergey.s.lega@intel.com>
1737 Anna Tikhonova <anna.tikhonova@intel.com>
1738 Ilya Tocar <ilya.tocar@intel.com>
1739 Andrey Turetskiy <andrey.turetskiy@intel.com>
1740 Ilya Verbin <ilya.verbin@intel.com>
1741 Kirill Yukhin <kirill.yukhin@intel.com>
1742 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1744 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1745 (VI124_256): Changed to ...
1746 (VI124_256_48_512): This.
1747 (ssepackmode): Extended with wider modes.
1748 (<code><mode>3): Changed iterator.
1749 (*avx2_<code><mode>3): Ditto.
1750 (vec_pack_trunc_<mode>): Ditto.
1752 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1753 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1754 Sergey Lega <sergey.s.lega@intel.com>
1755 Anna Tikhonova <anna.tikhonova@intel.com>
1756 Ilya Tocar <ilya.tocar@intel.com>
1757 Andrey Turetskiy <andrey.turetskiy@intel.com>
1758 Ilya Verbin <ilya.verbin@intel.com>
1759 Kirill Yukhin <kirill.yukhin@intel.com>
1760 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1762 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1763 (VI8F_256_512): Ditto.
1764 (abs<mode>2): Changed iterator.
1765 (avx2_perm<mode>): Changed to ...
1766 (<avx2_avx512f>_perm<mode>): This.
1767 (avx2_perm<mode>_1): Changed to ...
1768 (<avx2_avx512f>_perm<mode>_1): This.
1770 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1771 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1772 Sergey Lega <sergey.s.lega@intel.com>
1773 Anna Tikhonova <anna.tikhonova@intel.com>
1774 Ilya Tocar <ilya.tocar@intel.com>
1775 Andrey Turetskiy <andrey.turetskiy@intel.com>
1776 Ilya Verbin <ilya.verbin@intel.com>
1777 Kirill Yukhin <kirill.yukhin@intel.com>
1778 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1780 * config/i386/sse.md (VI48_AVX512F): New.
1781 (VI48_AVX2): Changed to ...
1782 (VI48_AVX2_48_AVX512F): This.
1783 (avx2_ashrv<mode>): Changed to ...
1784 (<avx2_avx512f>_ashrv<mode>): This.
1785 (avx2_<shift_insn>v<mode>): Changed to ...
1786 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1788 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1789 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1790 Sergey Lega <sergey.s.lega@intel.com>
1791 Anna Tikhonova <anna.tikhonova@intel.com>
1792 Ilya Tocar <ilya.tocar@intel.com>
1793 Andrey Turetskiy <andrey.turetskiy@intel.com>
1794 Ilya Verbin <ilya.verbin@intel.com>
1795 Kirill Yukhin <kirill.yukhin@intel.com>
1796 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1798 * config/i386/sse.md (VI4_AVX512F): New.
1799 (VI8_AVX2_AVX512F): Ditto.
1800 (mul<mode>3): Extended with wider modes.
1801 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1802 (mul<mode>3): Ditto.
1803 (vec_widen_<s>mult_odd_<mode>): Ditto.
1805 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1806 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1807 Sergey Lega <sergey.s.lega@intel.com>
1808 Anna Tikhonova <anna.tikhonova@intel.com>
1809 Ilya Tocar <ilya.tocar@intel.com>
1810 Andrey Turetskiy <andrey.turetskiy@intel.com>
1811 Ilya Verbin <ilya.verbin@intel.com>
1812 Kirill Yukhin <kirill.yukhin@intel.com>
1813 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1815 * config/i386/sse.md (VI2_AVX512F): New.
1816 (VI124_AVX512F): Ditto.
1817 (sseunpackmode): Extended with wider modes.
1818 (sseunpackfltmode): Ditto.
1819 (vec_unpacks_float_hi_<mode>): Ditto.
1820 (vec_unpacks_float_lo_<mode>): Ditto.
1821 (vec_unpacku_float_hi_<mode>): Ditto.
1822 (vec_unpacku_float_lo_<mode>): Ditto.
1823 (vec_unpacks_lo_<mode>): Ditto.
1824 (vec_unpacks_hi_<mode>): Ditto.
1825 (vec_unpacku_lo_<mode>): Ditto.
1826 (vec_unpacku_hi_<mode>): Ditto.
1828 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1829 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1830 Sergey Lega <sergey.s.lega@intel.com>
1831 Anna Tikhonova <anna.tikhonova@intel.com>
1832 Ilya Tocar <ilya.tocar@intel.com>
1833 Andrey Turetskiy <andrey.turetskiy@intel.com>
1834 Ilya Verbin <ilya.verbin@intel.com>
1835 Kirill Yukhin <kirill.yukhin@intel.com>
1836 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1838 * config/i386/i386.md (multdiv): New.
1839 (multdiv_mnemonic): Ditto.
1840 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1841 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1842 (<sse>_vmdiv<mode>3): Removed.
1844 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1845 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1846 Sergey Lega <sergey.s.lega@intel.com>
1847 Anna Tikhonova <anna.tikhonova@intel.com>
1848 Ilya Tocar <ilya.tocar@intel.com>
1849 Andrey Turetskiy <andrey.turetskiy@intel.com>
1850 Ilya Verbin <ilya.verbin@intel.com>
1851 Kirill Yukhin <kirill.yukhin@intel.com>
1852 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1854 * config/i386/sse.md (V): Extended with wider modes.
1856 (ssehalfvecmode): Ditto.
1858 (ssepackfltmode): Ditto.
1859 (avx_vec_concat<mode>): Ditto.
1860 (V_256_512): New iterator.
1861 (VF2_512_256): Ditto.
1862 (si2dfmode): New attribute.
1863 (si2dfmodelower): Ditto.
1865 (concat_tg_mode): Ditto.
1866 (floatv4siv4df2): Changed to ...
1867 (float<si2dfmodelower><mode>2): This.
1868 (avx_cvtps2pd256): Changed to ...
1869 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1870 (vec_pack_trunc_v4df): Changed to ...
1871 (vec_pack_trunc_<mode>): This.
1872 (avx_vpermil<mode>): Changed to ...
1873 (<sse2_avx_avx512f>_vpermil<mode>): This.
1874 (<fixsuffix>fix_truncv8dfv8si2): New.
1875 (vec_pack_sfix_trunc_v8df): Ditto.
1876 (avx512f_rndscale<mode>): Ditto.
1877 (avx512f_roundpd512): Ditto.
1878 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1880 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1881 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1882 Sergey Lega <sergey.s.lega@intel.com>
1883 Anna Tikhonova <anna.tikhonova@intel.com>
1884 Ilya Tocar <ilya.tocar@intel.com>
1885 Andrey Turetskiy <andrey.turetskiy@intel.com>
1886 Ilya Verbin <ilya.verbin@intel.com>
1887 Kirill Yukhin <kirill.yukhin@intel.com>
1888 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1890 * config/i386/i386.md (any_fix): New iterator.
1891 (fixsuffix): New attribute.
1892 * config/i386/sse.md (VF1): Extened with wider modes.
1896 (sseintvecmodelower): Ditto.
1897 (ssescalarmode): Ditto.
1898 (ssescalarnum): Ditto.
1900 (ssexmmmode): Ditto.
1901 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1902 (<sse>_rcp<mode>2): Change iterator.
1903 (rsqrt<mode>2): Ditto.
1904 (<sse>_rsqrt<mode>2): Ditto.
1905 (avx2_vec_dup<mode>): Ditto.
1906 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1907 (round<mode>2_sfix): Ditto.
1908 (avx2_pbroadcast<mode>): Ditto.
1909 (*andnot<mode>3): Handle XI mode.
1910 (*<code><mode>3): Ditto.
1911 (AVXTOSSEMODE): Removed.
1912 (avx_vpermil<mode>): Changed to ...
1913 (<sse2_avx_avx512f>_vpermil<mode>): This.
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 (<sse2>_movnt<mode>): Update constraint to "v".
1926 (<sse>_comi): Ditto.
1927 (<sse>_ucomi): Ditto.
1928 (sse_cvtss2siq_2): Ditto.
1929 (sse2_cvtsd2si): Ditto.
1930 (sse2_cvtsd2siq): Ditto.
1931 (sse2_cvttsd2si): Ditto.
1932 (sse2_cvttsd2siq): Ditto.
1933 (<shift_insn><mode>3): Ditto.
1934 (sse2_cvtsi2sdq): Update constraint and prefix.
1935 (sse_cvtsi2ss): Update prefix.
1936 (sse_cvtsi2ssq): Ditto.
1938 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1940 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1941 ignore internal calls.
1943 2013-10-11 Richard Biener <rguenther@suse.de>
1945 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1946 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1947 TREE_OVERFLOW is set.
1949 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1951 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1953 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1954 * gimple.h: Likewise.
1956 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1958 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1961 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1963 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1964 * configure: Regenerate.
1966 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1968 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1969 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1970 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1971 similarly pp_right_paren.
1972 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1973 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1974 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1975 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1977 (omp_declare_simd_clauses_equal,
1978 omp_remove_redundant_declare_simd_attrs): New functions.
1979 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1980 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1981 * tree.h (OMP_LOOP_CHECK): Define.
1982 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1983 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1984 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1985 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1986 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1987 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1988 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1989 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1990 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1991 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1992 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1993 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1994 (omp_remove_redundant_declare_simd_attrs): New prototype.
1995 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1996 GIMPLE_OMP_TEAMS): New codes.
1997 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1998 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2000 (target_nesting_level): New variable.
2001 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2002 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2003 for collapse > 1 static schedule unless ordered.
2004 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2005 (determine_parallel_type): Adjust get_ws_args_for caller.
2006 (install_var_field): Handle mask & 4 for double indirection.
2007 (scan_sharing_clauses): Ignore shared clause on teams construct.
2008 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2009 (create_omp_child_function): If inside target or declare target
2010 constructs, set "omp declare target" attribute on the child function.
2011 (find_combined_for): New function.
2012 (scan_omp_parallel): Handle combined loops.
2013 (scan_omp_target, scan_omp_teams): New functions.
2014 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2015 restrictions and set ctx->cancellable for cancellable constructs.
2016 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2017 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2018 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2019 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2020 (omp_clause_aligned_alignment): New function.
2021 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2022 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2023 on teams constructs. Handle user defined reductions and new
2025 (lower_reduction_clauses): Don't set placeholder to address of ref
2026 if it has already the right type.
2027 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2028 (expand_parallel_call): Use the new non-_start suffixed builtins,
2029 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2030 and GOMP_parallel_end after the call.
2031 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2032 (expand_omp_for_init_counts): Handle combined loops.
2033 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2035 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2036 end of cancellable loops.
2037 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2038 Likewise. Handle collapse > 1 loops.
2039 (expand_omp_simd): Handle combined loops.
2040 (expand_omp_for): Add inner_stmt argument, adjust callers of
2041 expand_omp_for* functions, use expand_omp_for_static*chunk even
2042 for collapse > 1 unless ordered.
2043 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2044 of cancellable sections.
2045 (expand_omp_single): Remove need_barrier variable, just rely on
2046 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2047 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2048 (expand_omp_atomic_load, expand_omp_atomic_store,
2049 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2050 (expand_omp_target): New function.
2051 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2052 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2053 (build_omp_regions_1): Immediately close region for
2054 GF_OMP_TARGET_KIND_UPDATE.
2055 (maybe_add_implicit_barrier_cancel): New function.
2056 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2058 (lower_omp_single): Likewise. Add clobber after the barrier.
2059 (lower_omp_taskgroup): New function.
2060 (lower_omp_for): Handle combined loops. Adjust
2061 lower_rec_input_clauses caller. Handle cancellation.
2062 (lower_depend_clauses): New function.
2063 (lower_omp_taskreg): Lower depend clauses. Adjust
2064 lower_rec_input_clauses caller. Add clobber after the call. Handle
2066 (lower_omp_target, lower_omp_teams): New functions.
2067 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2068 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2069 and GOMP_cancellation_point calls.
2070 (lower_omp): Fold stmts inside of target region.
2071 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2072 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2073 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2074 (BT_FN_VOID_OMPFN_PTR_UINT,
2075 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2076 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2077 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2078 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2079 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2080 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2081 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2082 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2083 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2084 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2085 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2086 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2087 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2088 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2089 BUILT_IN_GOMP_PARALLEL_END.
2090 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2091 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2092 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2093 GF_OMP_FOR_KIND_DISTRIBUTE.
2094 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2095 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2096 (dump_gimple_omp_return): Print lhs if it has any.
2097 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2098 gimple_omp_atomic_seq_cst_p.
2099 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2100 and GIMPLE_OMP_TEAMS.
2101 * langhooks.c (lhd_omp_mappable_type): New function.
2102 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2103 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2105 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2106 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2107 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2108 (struct gimplify_omp_ctx): Add combined_loop field.
2109 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2110 on stmts inside of target region.
2111 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2112 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2113 ORT_TARGET and ORT_TARGET_DATA.
2114 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2116 (omp_notice_threadprivate_variable): Complain about threadprivate
2117 variables in target region.
2118 (omp_notice_variable): Complain about vars with non-mappable type
2119 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2120 (omp_check_private): Ignore ORT_TARGET* regions.
2121 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2122 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2123 (find_combined_omp_for): New function.
2124 (gimplify_omp_for): Handle gimplification of combined loops.
2125 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2127 (gimplify_omp_target_update): New function.
2128 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2129 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2130 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2131 (gimplify_body): If fndecl has "omp declare target" attribute, add
2132 implicit ORT_TARGET context around it.
2133 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2134 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2135 * tree-nested.c (convert_nonlocal_reference_stmt,
2136 convert_local_reference_stmt, convert_gimple_call): Handle
2137 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2138 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2139 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2140 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2141 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2142 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2143 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2144 BUILT_IN_GOMP_LOOP_END_CANCEL,
2145 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2146 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2147 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2148 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2149 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2150 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2151 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2152 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2153 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2154 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2155 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2156 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2157 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2158 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2159 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2160 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2161 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2162 gimple_build_omp_teams): New functions.
2163 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2164 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2165 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2166 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2167 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2168 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2169 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2170 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2171 GF_OMP_ATOMIC_SEQ_CST): New.
2172 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2173 gimple_build_omp_teams): New prototypes.
2174 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2175 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2176 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2177 GIMPLE_OMP_SINGLE as end of range.
2178 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2179 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2180 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2181 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2182 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2183 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2184 gimple_omp_target_kind, gimple_omp_target_set_kind,
2185 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2186 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2187 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2188 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2189 gimple_omp_teams_set_clauses): New inlines.
2190 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2191 and GIMPLE_OMP_TASKGROUP.
2192 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2193 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2194 enum omp_clause_proc_bind_kind): New.
2195 (union omp_clause_subcode): Add depend_kind, map_kind and
2196 proc_bind_kind fields.
2197 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2198 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2199 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2200 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2201 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2203 2013-10-10 Teresa Johnson <tejohnson@google.com>
2205 * predict.c (tree_estimate_probability): Add new parameter
2206 for estimate_bb_frequencies.
2207 (estimate_bb_frequencies): Add new parameter to force estimation.
2208 (rebuild_frequencies): When max frequency in function is small,
2209 recompute counts from frequencies.
2210 * predict.h (estimate_bb_frequencies): New parameter.
2212 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2214 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2215 optimizations are disabled.
2217 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2219 * coverage.c (coverage_finish): Fix leak of da_file_name.
2221 2013-10-10 Jan Hubicka <jh@suse.cz>
2223 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2224 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2227 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2230 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2231 if any labels are in FALLTHRU_BB, use a special label emitted
2232 immediately after the asm goto insn rather than label_rtx
2234 (expand_asm_stmt): Adjust caller.
2235 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2236 edge if the last insn in predecessor is a jump with single successor,
2237 but it isn't simplejump_p.
2239 2013-10-10 Richard Biener <rguenther@suse.de>
2241 PR tree-optimization/58656
2242 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2244 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2246 * gimplify.c: Include expr.h and tm_p.h for targets with special
2247 va-arg padding requirements.
2249 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2251 * tree-flow.h: Move some prototypes to gimple.h.
2252 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2253 * gimple.h: Relocate some prototypes from tree-flow.h
2254 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2256 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2257 (build_va_arg_indirect_ref): Relocate and make static.
2258 (std_gimplify_va_arg_expr): Relocate here.
2259 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2260 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2262 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2264 * doc/md.texi: Document the mnemonic attribute.
2266 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2269 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2270 (set_attr_alternative x ...) when searching for user defined
2273 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2275 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2277 2013-10-09 Easwaran Raman <eraman@google.com>
2279 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2280 * cfgexpand.c (defer_stack_allocation): ...use here
2281 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2283 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2285 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2286 (operand_equal_for_value_replacement): New function, extracted from
2287 value_replacement and enhanced to catch more cases.
2288 (value_replacement): Use operand_equal_for_value_replacement.
2290 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2292 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2293 get_max_loop_iterations.
2295 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2297 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2298 Place comment above function.
2300 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2302 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2303 are not related to tree-cfg.c.
2304 * tree-ssa-address.h: New file. Relocate prototypes.
2305 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2306 (addr_for_mem_ref): New. Combine call to get_address_description and
2307 return addr_for_mem_ref.
2308 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2309 * tree-ssa-live.h: Adjust prototypes.
2310 * passes.c: Include tree-ssa-live.h.
2311 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2312 * graphite.c (graphite_transform_loops): Make static.
2313 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2314 make_pass_graphite, pass_data_graphite_transforms,
2315 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2316 * ipa-pure-const.c (warn_function_noreturn): Make static.
2317 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2318 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2319 Relocate from tree-cfg.c
2320 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2322 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2323 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2324 Move to ipa-pure-const.c.
2325 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2326 Relocate from tree-optimize.c.
2327 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2328 make_pass_fixup_cfg): Move to tree-cfg.c.
2329 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2330 Relocate some prototypes.
2331 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2332 * tree-dump.c (dump_function_to_file): Remove prototype.
2333 Add tree-flow.h to the include file.
2334 * tree-dump.h: Remove prototype.
2335 * tree-parloops.h: New File. Add prototypes.
2336 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2337 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2338 from tree-ssa-loop.c.
2339 * tree-predcom.c (run_tree_predictive_commoning,
2340 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2341 Relocate here from tree-ssa-loop.c.
2342 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2343 ssa_name_values.release ().
2344 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2345 (ssa_name_values): Relocate from tree-flow.h.
2346 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2347 * tree-ssa-loop.c (run_tree_predictive_commoning,
2348 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2349 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2350 make_pass_graphite, pass_data_graphite_transforms,
2351 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2352 tree_parallelize_loops, pass_data_parallelize_loops,
2353 make_pass_parallelize_loops): Move to other files.
2354 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2356 * tree.h: Remove prototypes from tree-address.c.
2358 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2360 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2361 (struct int_tree_map): Move to tree-hasher.h
2362 (SCALE, LABEL, PERCENT): Move to gimple.h
2363 * tree-flow-inline.h: Delete. Move functions to other files.
2364 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2365 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2366 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2367 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2368 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2369 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2370 * tree-hasher.h: Don't include tree-flow.h.
2371 (struct int_tree_map): Relocate from tree-flow.h.
2372 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2373 tree-flow-inline.h and make static.
2374 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2376 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2377 tree-flow-inline.h and make static.
2378 * tree.h (is_global_var, may_be_aliased): Relocate from
2380 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2381 * value-prof.c: No longer include tree-flow-inline.h.
2382 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2384 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2386 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2387 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2388 (enum move_pos): Move to tree-ssa-loop-im.h
2389 * cfgloop.h: Move some prototypes.
2390 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2391 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2392 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2393 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2394 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2395 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2396 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2397 (enum move_pos): Relocate here.
2398 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2400 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2401 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2402 make_pass_lim): Relocate here from tree-ssa-loop.c.
2403 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2405 (loop_edge_to_cancel, unloop_loops): Make static.
2406 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2407 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2408 (tree_complete_unroll, gate_tree_complete_unroll,
2409 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2410 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2411 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2412 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2413 (stmt_invariant_in_loop_p): Remove unused function.
2414 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2415 * tree-ssa-loop-manip.h: New file. Add prototypes.
2416 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2417 (gcov_type_to_double_int): Move to cfgloop.h.
2418 (double_int_cmp, bound_index,
2419 estimate_numbers_of_iterations_loop): Make static.
2420 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2421 (max_loop_iterations): Factor out get_max_loop_iterations.
2422 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2424 * tree-ssa-loop-niter.h: New file. Add prototypes.
2425 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2426 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2427 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2428 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2429 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2430 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2431 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2432 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2433 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2434 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2435 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2436 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2437 pass_data_complete_unroll, make_pass_complete_unroll,
2438 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2439 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2440 tree-ssa-loop-ivcanon.c.
2441 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2442 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2443 tree-ssa-loop-prefetch.c.
2444 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2446 (get_lsm_tmp_name): Relocate and add suffix parameter.
2447 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2448 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2449 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2450 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2451 (get_estimated_loop_iterations): Factor out accessor from
2452 estimated_loop_iterations in tree-ssa-loop-niter.c.
2453 (get_max_loop_iterations): Factor out accessor from
2454 _max_loop_iterations in tree-ssa-niter.c.
2455 * loop-unroll.c (decide_unroll_constant_iterations,
2456 decide_unroll_runtime_iterations, decide_peel_simple,
2457 decide_unroll_stupid): Use new get_* accessors.
2459 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2461 PR tree-optimization/20318
2462 * doc/extend.texi (returns_nonnull): New function attribute.
2463 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2465 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2466 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2468 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2471 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2472 false if both components are bitfields.
2474 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2476 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2477 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2478 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2479 * config/aarch64/aarch64.h
2480 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2481 * config/aarch64/aarch64-simd-builtins.def
2482 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2484 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2486 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2488 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2490 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2492 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2494 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2495 (vneg_f64): New intrinsic.
2496 (vneg_s8): Asm replaced with C.
2497 (vneg_s16): Likewise.
2498 (vneg_s32): Likewise.
2499 (vneg_s64): New intrinsic.
2500 (vnegq_f32): Asm replaced with C.
2501 (vnegq_f64): Likewise.
2502 (vnegq_s8): Likewise.
2503 (vnegq_s16): Likewise.
2504 (vnegq_s32): Likewise.
2505 (vnegq_s64): Likewise.
2507 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2509 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2511 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2513 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2514 packed stack special handling.
2515 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2516 back to fixed stack slots for FPRs saved due to stdarg.
2518 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2520 * config/s390/s390.c (s390_frame_info): Restructure function.
2522 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2524 * config/s390/s390.c (struct s390_frame_layout): New field
2526 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2527 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2528 regs_ever_clobbered to char*.
2529 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2530 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2531 (s390_register_info): Enable FPR save slots. Move/Copy some
2532 functionality into ...
2533 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2534 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2536 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2537 not imply a stack frame.
2538 (s390_init_frame_layout): Remove variable clobbered_regs.
2539 (s390_update_register_info): Remove function.
2540 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2541 cannot be used for register renaming.
2542 (s390_hard_regno_scratch_ok): New function.
2543 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2544 (s390_initial_elimination_offset): Change offset calculation of
2545 the return address pointer.
2546 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2547 (restore_gprs): Set frame related flag.
2548 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2549 (s390_emit_prologue): Call s390_register_info instead of
2550 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2551 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2552 (s390_optimize_prologue): Call s390_optimize_register_info.
2553 Try to remove also FPR slot save/restore INSNs. Remove frame
2554 related flags from restore INSNs.
2556 2013-10-08 DJ Delorie <dj@redhat.com>
2558 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2561 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2562 avoid conflict with the MI use of %c.
2563 * config/rl78/rl78-real.md: change %c to %C throughout.
2564 * config/rl78/rl78-virt.md: Likewise.
2566 2013-10-08 Jan Hubicka <jh@suse.cz>
2568 * config/i386/i386.c (ix86_option_override_internal): Switch
2569 to SSE math for -ffast-math when target ISA supports SSE2.
2571 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2573 * tree-flow.h: Remove some prototypes.
2574 * tree.h: Remove some protypes, add a couple.
2575 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2576 using_eh_for_cleanups_p): Add interface routines for front ends.
2577 * tree-eh.h: New file. Add protoptyes.
2578 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2579 (add_stmt_to_eh_lp_fn): Make static.
2580 (lower_try_finally): Use new using_eh_for_cleanups_p.
2581 * emit-rtl.c: Include tree-eh.h.
2582 * gimple.h: Include tree-eh.h.
2584 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2586 PR tree-optimization/58480
2587 * tree-vrp.c (infer_nonnull_range): New function.
2588 (infer_value_range): Call infer_nonnull_range.
2590 2013-10-08 Dehao Chen <dehao@google.com>
2592 PR tree-optimization/58619
2593 * tree-inline.c (copy_phis_for_bb): Combine location data
2596 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2599 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2600 RTX_FRAME_RELATED_P on INSN.
2602 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2605 (altivec_expand_vec_perm_const): Call it.
2607 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2609 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2610 sequences for LE VSX loads and stores at expand time.
2611 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2613 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2614 (rs6000_gen_le_vsx_permute): New.
2615 (rs6000_gen_le_vsx_load): New.
2616 (rs6000_gen_le_vsx_store): New.
2617 (rs6000_gen_le_vsx_move): New.
2618 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2619 (*vsx_le_perm_load_v4si): New.
2620 (*vsx_le_perm_load_v8hi): New.
2621 (*vsx_le_perm_load_v16qi): New.
2622 (*vsx_le_perm_store_v2di): New.
2623 (*vsx_le_perm_store_v4si): New.
2624 (*vsx_le_perm_store_v8hi): New.
2625 (*vsx_le_perm_store_v16qi): New.
2626 (*vsx_xxpermdi2_le_<mode>): New.
2627 (*vsx_xxpermdi4_le_<mode>): New.
2628 (*vsx_xxpermdi8_le_V8HI): New.
2629 (*vsx_xxpermdi16_le_V16QI): New.
2630 (*vsx_lxvd2x2_le_<mode>): New.
2631 (*vsx_lxvd2x4_le_<mode>): New.
2632 (*vsx_lxvd2x8_le_V8HI): New.
2633 (*vsx_lxvd2x16_le_V16QI): New.
2634 (*vsx_stxvd2x2_le_<mode>): New.
2635 (*vsx_stxvd2x4_le_<mode>): New.
2636 (*vsx_stxvd2x8_le_V8HI): New.
2637 (*vsx_stxvd2x16_le_V16QI): New.
2639 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2641 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2643 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2645 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2646 loop to work also for 31bit ABI.
2647 Save the stack pointer for frame_size > 0.
2649 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2651 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2652 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2653 constraint letters from expanders.
2654 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2655 retry count to general_operand.
2656 ("tabort"): Give operand 0 a mode.
2657 ("tabort_1"): Add mode and constraint letter for operand 0.
2658 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2660 2013-10-04 Jeff Law <law@redhat.com>
2662 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2664 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2666 (thread_across_edge): Here. Call it.
2668 2013-10-04 Cary Coutant <ccoutant@google.com>
2670 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2671 discarding a location list expression (or a piece of one).
2673 2013-10-03 Jan Hubicka <jh@suse.cz>
2675 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2676 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2677 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2679 2013-10-03 Jan Hubicka <jh@suse.cz>
2681 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2682 accumulate-outgoing-args when producing unwind info.
2684 2013-10-03 Wei Mi <wmi@google.com>
2686 * lra-constraints.c (insert_move_for_subreg): New function
2687 extracted from simplify_operand_subreg.
2688 (simplify_operand_subreg): Add reload for paradoxical subreg.
2690 2013-10-03 Rong Xu <xur@google.com>
2692 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2693 the candidate of builtin_expect such that we should fix the
2694 size/time estimation.
2695 (estimate_function_body_sizes): Do the acutally size/time fix-up
2698 2013-10-03 Rong Xu <xur@google.com>
2700 * predict.c (tree_predict_by_opcode): Get the probability
2701 for builtin_expect from param builtin_expect_probability.
2702 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2703 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2704 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2706 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2709 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2710 * calls.c (alloca_call_p): Use get_callee_fndecl.
2711 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2712 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2714 (vrp_visit_stmt): Remove duplicated code.
2716 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2718 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2719 ceildf2, btruncdf2, instead of vsx_* name.
2721 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2722 iterators to only do V2DF and V4SF here. Move the DF code to
2723 rs6000.md where it is combined with SF mode. Replace <VSv> with
2724 just 'v' since only vector operations are handled with these insns
2725 after moving the DF support to rs6000.md.
2726 (vsx_sub<mode>3): Likewise.
2727 (vsx_mul<mode>3): Likewise.
2728 (vsx_div<mode>3): Likewise.
2729 (vsx_fre<mode>2): Likewise.
2730 (vsx_neg<mode>2): Likewise.
2731 (vsx_abs<mode>2): Likewise.
2732 (vsx_nabs<mode>2): Likewise.
2733 (vsx_smax<mode>3): Likewise.
2734 (vsx_smin<mode>3): Likewise.
2735 (vsx_sqrt<mode>2): Likewise.
2736 (vsx_rsqrte<mode>2): Likewise.
2737 (vsx_fms<mode>4): Likewise.
2738 (vsx_nfma<mode>4): Likewise.
2739 (vsx_copysign<mode>3): Likewise.
2740 (vsx_btrunc<mode>2): Likewise.
2741 (vsx_floor<mode>2): Likewise.
2742 (vsx_ceil<mode>2): Likewise.
2743 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2744 (vsx_sminsf3): Likewise.
2745 (vsx_fmadf4): Likewise.
2746 (vsx_fmsdf4): Likewise.
2747 (vsx_nfmadf4): Likewise.
2748 (vsx_nfmsdf4): Likewise.
2749 (vsx_cmpdf_internal1): Likewise.
2751 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2752 simpler to select whether a target has SPE or traditional floating
2753 point support in iterators.
2754 (TARGET_DF_SPE): Likewise.
2755 (TARGET_SF_FPR): Likewise.
2756 (TARGET_DF_FPR): Likewise.
2757 (TARGET_SF_INSN): Macros to say whether floating point support
2758 exists for a given operation for expanders.
2759 (TARGET_DF_INSN): Likewise.
2761 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2762 combining of SF/DF mode operations, using both traditional and VSX
2770 (abs<mode>2): Combine SF/DF modes using traditional floating point
2771 instructions. Add support for using the upper DF registers with
2772 VSX support, and SF registers with power8-vector support. Update
2773 expanders for operations supported by both the SPE and traditional
2774 floating point units.
2775 (abs<mode>2_fpr): Likewise.
2776 (nabs<mode>2): Likewise.
2777 (nabs<mode>2_fpr): Likewise.
2778 (neg<mode>2): Likewise.
2779 (neg<mode>2_fpr): Likewise.
2780 (add<mode>3): Likewise.
2781 (add<mode>3_fpr): Likewise.
2782 (sub<mode>3): Likewise.
2783 (sub<mode>3_fpr): Likewise.
2784 (mul<mode>3): Likewise.
2785 (mul<mode>3_fpr): Likewise.
2786 (div<mode>3): Likewise.
2787 (div<mode>3_fpr): Likewise.
2788 (sqrt<mode>3): Likewise.
2789 (sqrt<mode>3_fpr): Likewise.
2790 (fre<Fs>): Likewise.
2791 (rsqrt<mode>2): Likewise.
2792 (cmp<mode>_fpr): Likewise.
2793 (smax<mode>3): Likewise.
2794 (smin<mode>3): Likewise.
2795 (smax<mode>3_vsx): Likewise.
2796 (smin<mode>3_vsx): Likewise.
2797 (negsf2): Delete SF operations that are merged with DF.
2804 (fmasf4_fpr): Likewise.
2805 (fmssf4_fpr): Likewise.
2806 (nfmasf4_fpr): Likewise.
2807 (nfmssf4_fpr): Likewise.
2808 (sqrtsf2): Likewise.
2809 (rsqrtsf_internal1): Likewise.
2810 (smaxsf3): Likewise.
2811 (sminsf3): Likewise.
2812 (cmpsf_internal1): Likewise.
2813 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2814 (negdf2): Delete DF operations that are merged with SF.
2816 (nabsdf2): Likewise.
2822 (rsqrtdf_internal1): Likewise.
2823 (fmadf4_fpr): Likewise.
2824 (fmsdf4_fpr): Likewise.
2825 (nfmadf4_fpr): Likewise.
2826 (nfmsdf4_fpr): Likewise.
2827 (sqrtdf2): Likewise.
2828 (smaxdf3): Likewise.
2829 (smindf3): Likewise.
2830 (cmpdf_internal1): Likewise.
2831 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2832 (btrunc<mode>2): Delete separate expander, and combine with the
2833 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2834 (btrunc<mode>2_fpr): Likewise.
2835 (ceil<mode>2): Likewise.
2836 (ceil<mode>2_fpr): Likewise.
2837 (floor<mode>2): Likewise.
2838 (floor<mode>2_fpr): Likewise.
2839 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2840 Add support for using the upper registers with VSX and
2841 power8-vector. Move insns to be closer to the define_expands. On
2842 VSX systems, prefer the traditional form of FMA over the VSX
2843 version, since the traditional form allows the target not to
2844 overlap with the inputs.
2845 (fms<mode>4_fpr): Likewise.
2846 (nfma<mode>4_fpr): Likewise.
2847 (nfms<mode>4_fpr): Likewise.
2849 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2850 Richard Earnshaw <richard.earnshaw@arm.com>
2852 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2853 (struct mult_cost_table): Likewise.
2854 (struct mem_cost_table): Likewise.
2855 (struct fp_cost_table): Likewise.
2856 (struct vector_cost_table): Likewise.
2857 (cpu_cost_table): Likewise.
2858 * config/arm/arm.opt (mold-rts-costs): New option.
2859 (mnew-generic-costs): Likewise.
2860 * config/arm/arm.c (generic_extra_costs): New table.
2861 (cortexa15_extra_costs): Likewise.
2862 (arm_slowmul_tune): Use NULL as new costs.
2863 (arm_fastmul_tune): Likewise.
2864 (arm_strongarm_tune): Likewise.
2865 (arm_xscale_tune): Likewise.
2866 (arm_9e_tune): Likewise.
2867 (arm_v6t2_tune): Likewise.
2868 (arm_cortex_a5_tune): Likewise.
2869 (arm_cortex_a9_tune): Likewise.
2870 (arm_v6m_tune): Likewise.
2871 (arm_fa726te_tune): Likewise.
2872 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2873 (arm_cortex_tune): Use generict_extra_costs.
2874 (shifter_op_p): New function.
2875 (arm_unspec_cost): Likewise.
2876 (LIBCALL_COST): Define.
2877 (arm_new_rtx_costs): New function.
2878 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2879 table is available. Use old costs otherwise unless mnew-generic-costs
2881 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2882 (cpu_cost_table): Declare.
2884 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2887 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2888 (*subs_mul_imm_<mode>)
2889 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2890 (*sub_<shift>_<mode>)
2891 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2892 Remove k constraint.
2894 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2896 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2898 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2900 2013-10-02 Teresa Johnson <tejohnson@google.com>
2902 * predict.c (probably_never_executed): New function.
2903 (probably_never_executed_bb_p): Invoke probably_never_executed.
2904 (probably_never_executed_edge_p): Ditto.
2905 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2906 Treat profile insanities conservatively.
2908 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2910 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2912 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2914 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2915 use it. Use smaller increase for scratch. Don't increase reject
2916 for early clobber scratch.
2917 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2918 setting eliminated regs except setting fp from hfp.
2919 (lra_eliminate): Check lra_insn_recog_data on NULL.
2921 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2924 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2925 setting -mvsx-timode by default until the underlying problem is fixed.
2926 (RS6000_CPU, power7 defaults): Likewise.
2928 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2930 * config/x-linux (host-linux.o): Remove header dependencies.
2931 Use $(COMPILE) and $(POSTCOMPILE).
2932 * config/t-linux-android (linux-android.o): Ditto.
2934 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2936 * Makefile.in (expmed.o-warn): Remove.
2938 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2940 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2941 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2943 2013-10-02 Teresa Johnson <tejohnson@google.com>
2945 * dojump.c (do_jump_1): Divide probability between
2946 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2948 2013-10-02 Tom Tromey <tromey@redhat.com>
2950 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2952 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2954 * tree-flow.h: Remove some prototypes.
2955 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2956 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2957 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2958 mark_virtual_phi_result_for_renaming): Relocate here.
2959 * tree-into-ssa.h: Add prototypes.
2960 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2961 single_pred_before_succ_order.
2962 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2963 (nonfreeing_call_p) Move to gimple.c.
2964 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2966 * basic-block.h (single_pred_before_succ_order): Add prototype.
2967 * gimple.c (nonfreeing_call_p): Relocate here.
2968 * gimple.h: Add prototype.
2969 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2970 * tree-ssa-dom.h: New file. Relocate prototypes here.
2971 * tree-ssa.h: Include tree-ssa-dom.h.
2973 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2975 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2976 Use $(COMPILE) and $(POSTCOMPILE).
2978 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2980 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2982 * tree-flow.h: Remove some prototypes.
2983 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2984 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2985 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2986 propagate_tree_value*): Move from here to...
2987 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2988 propagate_tree_value*): Relocate here.
2989 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2990 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2991 * gimple-fold.c: Remove gimple-fold.h from include list.
2992 * tree-vrp.c: Remove gimple-fold.h from include list.
2993 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2994 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2995 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2996 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2997 * sese.c: Add tree-ssa-propagate.h to include list.
2999 2013-10-02 Richard Biener <rguenther@suse.de>
3001 * tree-loop-distribution.c: Include tree-vectorizer.h for
3003 (enum partition_kind): Remove PKIND_REDUCTION.
3004 (struct partition_s): Remove has_writes member, add reduction_p member.
3005 (partition_alloc): Adjust.
3006 (partition_builtin_p): Likewise.
3007 (partition_has_writes): Remove.
3008 (partition_reduction_p): New function.
3009 (partition_merge_into): Likewise.
3010 (generate_code_for_partition): Commonize builtin partition
3012 (rdg_cannot_recompute_vertex_p): Remove.
3013 (already_processed_vertex_p): Likewise.
3014 (rdg_flag_vertex): Do not set has_writes.
3015 (classify_partition): Adjust.
3016 (rdg_build_partitions): Do not set has_writes, treat all
3017 partitions as useful.
3018 (distribute_loop): Record number of library calls generated. Adjust.
3019 (tree_loop_distribution): Report number of loops and library
3020 calls generated as opt-info.
3022 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3024 * tree-flow.h: Include new .h files. Move prototypes.
3025 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3026 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3027 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3028 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3029 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3030 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3031 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3032 * tree.h (get_ref_base_and_extent): Move prototype out.
3033 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3035 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3036 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3037 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3038 * tree-scalar-evolution.c: Include tree.h.
3039 * sese.c: Include tree.h.
3040 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3041 * dwarf2out.c: Include tree-dfa.h.
3042 * tree-chrec.c: Include tree.h.
3043 * tree-data-ref.c: Include tree.h.
3045 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3047 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3050 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3052 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3053 Use $(COMPILE) and $(POSTCOMPILE).
3054 (sol2-cxx.o): Likewise.
3055 (sol2-stubs.o): Likewise.
3057 * config/x-solaris (host-solaris.o): Likewise.
3059 * config/sparc/t-sparc (sparc.o): Remove.
3060 (sparc-c.o): Remove header dependencies.
3061 Use $(COMPILE) and $(POSTCOMPILE).
3062 * config/sparc/x-sparc: Likewise.
3064 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3066 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3067 * config/arc/arc700.md: Likewise.
3068 * config/arc/arc-modes.def: Likewise.
3069 * config/arc/arc-simd.h: Likewise.
3070 * config/arc/t-arc-uClibc: Likewise.
3071 * config/arc/t-arc-newlib: Likewise.
3073 2013-10-02 Renlin Li <renlin.li@arm.com>
3075 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3077 (aarch64_expand_epilogue): Likewise.
3079 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3080 Yufeng Zhang <yufeng.zhang@arm.com>
3082 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3084 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3085 'base_in' represent a conversion and legal_cast_p_1 holds; set
3086 'base_in' with the returned value from get_unwidened.
3088 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3090 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3091 plus_constant is not used.
3093 2013-10-01 Wei Mi <wmi@google.com>
3095 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3096 * config/i386/i386.md: Add define_peephole2 to
3097 break partial reg stall for cvtss2sd/cvtsd2ss.
3099 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3101 * config/arc/arc.c (pass_arc_ifcvt::clone):
3102 Update for ctxt_ -> m_ctxt change.
3104 2013-10-01 Jeff Law <law@redhat.com>
3106 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3107 outgoing_edge and intermediate_edge fields. Instead store the path.
3108 (redirection_data::hash): Hash on the last edge's destination index.
3109 (redirection_data::equal): Check the entire thread path.
3110 (lookup_redirectio_data): Corresponding changes.
3111 (create_edge_and_update_destination_phis): Likewise.
3112 (thread_single_edge): Likewise.
3114 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3115 Diego Novillo <dnovillo@google.com>
3117 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3118 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3119 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3120 versions of these patterns.
3122 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3123 (__builtin_arc_aligned): Likewise.
3125 * config/arc/arc.md: Expand adc_0 comment stating the intended
3126 purpose and why it isn't ready.
3127 Replace commented out call_value_via_label_mixed with a
3128 plain comment about bl_s.
3130 * config/arc/arc.c (stdio.h): Don't include directly.
3131 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3132 Remove if (1) condition.
3133 (arc_encode_section_info): Fix comment.
3135 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3137 * config/arc/arc.c (arc_conditional_register_usage):
3138 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3139 Also set reg_alloc_order for DMA config regs.
3141 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3142 Jeremy Bennett <jeremy.bennett@embecosm.com>
3144 * doc/install.texi (--with-cpu): Mention ARC.
3145 (arc-*-elf32): New paragraph.
3146 (arc-linux-uclibc): Likewise.
3147 * doc/md.texi (Machine Constraints): Add ARC part.
3148 * doc/invoke.texi: (menu): Add ARC Options.
3149 (Machine Dependent Options) <ARC Options>: Add synopsis.
3150 (node ARC Options): Add.
3151 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3152 (ARC Built-in Functions): New section defining
3153 generic ARC built-in functions.
3154 (ARC SIMD Built-in Functions): New section defining SIMD specific
3156 (Declaring Attributes of Functions): Extended
3157 description of short_call and long_call attributes for ARC and
3158 added index entries.
3160 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3161 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3162 Joern Rennecke <joern.rennecke@embecosm.com>
3163 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3164 Brendan Kehoe <brendan@zen.org>
3165 Michael Eager <eager@eagercon.com>
3166 Simon Cook <simon.cook@embecosm.com>
3167 Jeremy Bennett <jeremy.bennett@embecosm.com>
3169 * config/arc, common/config/arc: New directories.
3171 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3172 Brendan Kehoe <brendan@zen.org>
3173 Simon Cook <simon.cook@embecosm.com>
3175 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3177 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3179 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3180 * tree-ssa-coalesce.h: New. Move prototype to here.
3181 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3182 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3183 (gimple_can_coalesce_p): Move to...
3184 * gimple.c (gimple_can_coalesce_p): Here.
3186 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3188 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3189 (dump_decl_set): Move to gimple.c.
3190 * gimple.h: Don't include tree-ssa-operands.h.
3191 (dump_decl_set): Add prototype.
3192 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3193 Move to gimple-ssa.h.
3194 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3195 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3196 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3197 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3198 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3200 (dump_decl_set): Relocate here.
3201 * gimple-ssa.h: New file.
3202 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3203 Relocate from gimple.h.
3204 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3205 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3206 swap_tree_operands and remove non-ssa path.
3207 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3208 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3210 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3211 vect_is_simple_reduction_1): Use swap_ssa_operands.
3212 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3213 (enum need_phi_state): Move to tree-into-ssa.c.
3214 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3215 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3216 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3217 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3218 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3219 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3220 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3221 num_imm_uses): Move to ssa-iterators.h.
3222 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3223 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3225 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3226 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3227 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3228 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3229 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3230 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3231 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3232 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3233 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3234 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3235 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3236 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3237 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3238 (set_phi_nodes): Move to tree-phinodes.h.
3239 * tree-ssa-operands.h (enum ssa_op_iter_type,
3240 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3241 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3242 (dump_decl_set): Remove prototype.
3243 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3244 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3245 (set_phi_nodes): Relocate from tree-flow-inline.h.
3246 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3248 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3249 include list. Temporarily add gimple.h to include list.
3250 * ssa-iterators.h: New file.
3251 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3252 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3253 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3254 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3255 Relocate from tree-ssa-operands.h.
3256 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3257 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3258 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3259 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3260 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3261 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3262 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3263 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3264 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3265 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3266 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3267 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3268 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3269 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3270 Relocate from tree-flow-inline.h.
3271 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3273 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3276 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3277 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3278 Insert '\t' to output template.
3279 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3280 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3281 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3282 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3284 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3286 * doc/install.texi (Host/target specific installation notes for GCC):
3287 Put @anchor before @heading.
3288 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3289 Use @email for email addresses.
3291 2013-10-01 Jeff Law <law@redhat.com>
3293 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3294 a vec. Only delete the path if we create one without successfully
3295 registering a jump thread.
3296 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3298 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3299 (paths): New vector of jump threading paths.
3300 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3301 (THREAD_PATH): New accessor macro for the entire thread path.
3302 (lookup_redirection_data): Get intermediate and final outgoing edge
3303 from the thread path.
3304 (create_edge_and_update_destination_phis): Copy the threading path.
3305 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3306 jump threading path.
3307 (ssa_redirect_edges): Get edges and block types from the jump threading
3308 path. Free the path vector.
3309 (thread_block): Get edges from the jump threading path. Look at the
3310 entire path to see if we thread to a loop exit. If we cancel a jump
3311 thread request, then free the path vector.
3312 (thread_single_edge): Get edges and block types from the jump threading
3313 path. Free the path vector.
3314 (thread_through_loop_header): Get edges and block types from the jump
3315 threading path. Free the path vector.
3316 (mark_threaded_blocks): Iterate over the vector of paths and store
3317 the path on the appropriate edge. Get edges and block types from the
3318 jump threading path.
3319 (mark_threaded_blocks): Get edges and block types from the jump
3320 threading path. Free the path vector.
3321 (thread_through_all_blocks): Use the vector of paths rather than
3322 a vector of 3-edge sets.
3323 (register_jump_thread): Accept pointer to a path vector rather
3324 than the path vector itself. Store the path vector for later use.
3327 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3328 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3331 * config/s390/s390.c (s390_split_branches): Modify check for table
3333 (s390_chunkify_start): Rearrange table jump insn check in order to
3334 deal with compare and branch insns correctly.
3336 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3340 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3341 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3342 define_insn_and_split.
3343 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3344 (shiftsi3_compare): New pattern.
3346 * config/arm/unspecs.md (UNSPEC_RRX): New.
3348 2013-10-01 Alan Modra <amodra@gmail.com>
3350 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3351 casing inout operands.
3353 2013-10-01 Richard Biener <rguenther@suse.de>
3355 PR tree-optimization/58553
3356 * tree-loop-distribution.c (struct partition_s): Add niter member.
3357 (classify_partition): Populate niter member for the partition
3358 and properly identify whether the relevant store happens before
3359 or after the loop exit.
3360 (generate_memset_builtin): Use niter member from the partition.
3361 (generate_memcpy_builtin): Likewise.
3363 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3365 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3366 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3368 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3370 * basic-block.h (edge_list): Prefix member names with "m_".
3371 * context.h (context): Likewise.
3372 * domwalk.h (dom_walker): Likewise.
3373 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3374 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3375 * hash-table.h (hash_table): Likewise.
3376 * machmode.h (bit_field_mode_iterator): Likewise.
3377 * pass_manager.h (pass_list): Likewise.
3378 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3379 * tree-pass.h (pass_data): Likewise.
3380 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3381 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3382 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3383 * asan.c (pass_data_asan): Update accordingly.
3384 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3385 (control_dependences::control_dependences): Likewise.
3386 (control_dependences::~control_dependences): Likewise.
3387 (control_dependences::~control_dependences): Likewise.
3388 (control_dependences::get_edges_dependent_on): Likewise.
3389 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3390 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3391 * context.c (gcc::context::context): Likewise.
3392 * cprop.c (pass_rtl_cprop::clone): Likewise.
3393 * domwalk.c (dom_walker::walk): Likewise.
3394 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3395 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3396 * mode-switching.c (pass_mode_switching::clone): Likewise.
3397 * passes.c (opt_pass::opt_pass): Likewise.
3398 (pass_manager::pass_manager): Likewise.
3399 * predict.c (pass_strip_predict_hints::clone): Likewise.
3400 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3401 (pass_split_all_insns::clone): Likewise.
3402 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3404 (bit_field_mode_iterator::next_mode): Likewise.
3405 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3406 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3407 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3408 * tree-complex.c (pass_lower_complex::clone): Likewise.
3409 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3410 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3411 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3412 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3413 (pass_fold_builtins::clone): Likewise.
3414 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3415 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3416 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3417 (pass_cd_dce::clone): Likewise.
3418 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3419 (pass_phi_only_cprop::clone): Likewise.
3420 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3421 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3422 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3423 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3424 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3425 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3426 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3427 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3428 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3429 * tree-vrp.c (pass_vrp::clone): Likewise.
3430 * tsan.c (pass_tsan::clone): Likewise.
3432 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3435 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3436 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3439 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3440 optimization, punt if sign_bit_p looked through any zero extension.
3442 2013-09-30 Teresa Johnson <tejohnson@google.com>
3444 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3445 Update redirected out edge count in joiner case.
3446 (ssa_redirect_edges): Common the joiner and non-joiner cases
3447 so that joiner case gets profile updates.
3449 2013-09-30 Richard Biener <rguenther@suse.de>
3451 PR tree-optimization/58554
3452 * tree-loop-distribution.c (classify_partition): Require
3453 unconditionally executed stores for memcpy and memset recognition.
3454 (tree_loop_distribution): Calculate dominance info.
3456 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3458 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3459 (NO_PROFILE_COUNTERS): Likewise.
3460 (PROFILE_HOOK): Likewise.
3461 (FUNCTION_PROFILER): Likewise.
3462 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3464 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3466 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3467 calls and defines in TARGET_MACHO conditional.
3468 (load_macho_picbase_di): Likewise.
3469 (reload_macho_picbase): Likewise.
3470 (reload_macho_picbase_si): Likewise.
3471 (reload_macho_picbase_di): Likewise.
3472 (nonlocal_goto_receiver): Likewise.
3474 2013-09-30 Nick Clifton <nickc@redhat.com>
3476 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3477 that use the MSP430X ISA.
3478 (msp430_option_override): Scan -mmcu command line option for any
3479 MCU name that supports the MSP430X ISA.
3480 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3481 -mmcu options which enable the MSP430X ISA.
3483 2013-09-30 Richard Biener <rguenther@suse.de>
3486 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3487 before looking for setjmp-like calls.
3489 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3492 * config/darwin-protos.h (machopic_get_function_picbase): New.
3493 * config/darwin.c (machopic_get_function_picbase): New.
3494 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3495 label for a new func. (load_macho_picbase_di): Likewise.
3496 (reload_macho_picbase): New expand.
3497 (reload_macho_picbase_si): New insn.
3498 (reload_macho_picbase_di): New insn.
3499 (nonlocal_goto_receiver): New define and split.
3500 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3501 (unspecv enum): Add UNSPECV_NLGR.
3503 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3505 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3506 that altivec registers are correctly sized on Darwin.
3508 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3510 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3511 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3512 * config/x-darwin (host-darwin.o): Likewise.
3513 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3514 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3515 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3517 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3519 * doc/invoke.texi: Fix usage of @tie{} command.
3521 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3523 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3525 2013-09-28 Jan Hubicka <jh@suse.cz>
3527 * config/i386/x86-tune.def: Add documentation for each of the options;
3530 2013-09-28 Jan Hubicka <jh@suse.cz>
3532 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3534 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3535 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3536 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3537 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3538 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3540 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3542 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3543 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3544 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3545 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3546 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3547 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3548 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3549 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3550 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3551 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3552 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3553 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3554 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3555 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3556 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3557 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3558 gimple.h, godump.c, graphite-clast-to-gimple.c,
3559 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3560 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3561 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3562 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3563 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3564 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3565 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3566 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3567 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3568 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3569 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3570 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3571 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3572 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3573 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3574 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3575 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3576 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3577 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3578 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3579 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3580 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3581 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3582 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3583 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3584 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3585 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3586 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3587 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3588 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3589 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3590 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3591 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3592 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3593 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3594 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3595 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3596 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3597 whitespace before "(".
3599 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3601 * expr.h (extract_bit_field): Remove packedp parameter.
3602 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3603 from forward declaration.
3604 (store_split_bit_field): Remove packedp arg from calls to
3605 extract_fixed_bit_field.
3606 (extract_bit_field_1): Remove packedp parameter and packedp
3607 argument from recursive calls and calls to extract_fixed_bit_field.
3608 (extract_bit_field): Remove packedp parameter and corresponding
3609 arg to extract_bit_field_1.
3610 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3612 (extract_split_bit_field): Remove packedp arg from call to
3613 extract_fixed_bit_field.
3614 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3615 (copy_blkmode_from_reg): Likewise.
3616 (copy_blkmode_to_reg): Likewise.
3617 (read_complex_part): Likewise.
3618 (store_field): Likewise.
3619 (expand_expr_real_1): Likewise.
3620 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3621 to extract_bit_field.
3622 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3623 call to extract_bit_field.
3624 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3625 call to extract_bit_field.
3626 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3627 and special packedp behavior from -fstrict-volatile-bitfields
3630 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3632 * lra-eliminations.c (init_elim_table): Guard value_p.
3634 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3636 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3637 DFmode, DImode, and SFmode in the upper VSX registers based on the
3638 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3639 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3640 of the VSX register handling.
3642 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3644 (zero_extendsidi2_lfiwzx): Add support for loading into the
3645 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3646 only do VSX memory options on Altivec registers.
3647 (extendsidi2_lfiwax): Likewise.
3648 (extendsfdf2_fpr): Likewise.
3649 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3650 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3651 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3652 (movdi_internal64): Likewise.
3654 2013-09-27 Xinliang David Li <davidxl@google.com>
3656 * opts.c (finish_options): Adjust parameters
3657 according to vect cost model.
3658 (common_handle_option): Set dynamic vect cost
3660 targhooks.c (default_add_stmt_cost): Compute stmt cost
3662 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3663 Use helper function.
3664 * tree-vectorizer.h (unlimited_cost_model): New function.
3665 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3666 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3668 (vect_enhance_data_refs_alignment): Ditto.
3669 * flag-types.h: New enum.
3670 * common/config/i386/i386-common.c (ix86_option_init_struct):
3671 No need to initialize vect_cost_model flag.
3672 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3675 2013-09-27 Diego Novillo <dnovillo@google.com>
3677 * gimple.h (enum ssa_mode): Remove.
3679 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3681 * cfgloop.h (number_of_loops): Fix typo in check for null.
3683 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3686 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3687 are children of outermost saved_cfun's loop, and set it up to
3688 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3689 if loop != loop0 and SESE region contains bbs that belong to loop0.
3691 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3693 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3694 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3695 (set_address_segment, set_address_base, set_address_index)
3696 (set_address_disp): Accept the argument unconditionally.
3697 (baseness): Remove must_be_base_p and must_be_index_p checks.
3698 (decompose_normal_address): Classify as much as possible in the
3701 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3703 * cse.c (count_reg_usage): Handle INT_LIST.
3704 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3705 * reginfo.c (reg_scan_mark_refs): Likewise.
3706 * reload1.c (eliminate_regs_1): Likewise.
3708 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3711 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3714 2013-09-27 Richard Biener <rguenther@suse.de>
3716 PR tree-optimization/58459
3717 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3718 restriction not propagating into loops.
3720 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3722 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3723 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3724 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3726 2013-09-26 Richard Biener <rguenther@suse.de>
3728 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3730 2013-09-26 Richard Biener <rguenther@suse.de>
3732 * alias.h (component_uses_parent_alias_set): Rename to ...
3733 (component_uses_parent_alias_set_from): ... this.
3734 * alias.c (component_uses_parent_alias_set): Rename to ...
3735 (component_uses_parent_alias_set_from): ... this and return
3737 (reference_alias_ptr_type_1): Use the result from
3738 component_uses_parent_alias_set_from instead of stripping
3739 components one at a time.
3740 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3742 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3744 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3745 Move prototypes to...
3746 * tree-ssa-ter.h: New File. Move prototypes here.
3747 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3748 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3749 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3750 from is_replaceable_p.
3751 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3752 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3753 (process_replaceable): Use ter_is_replaceable_p.
3754 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3755 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3756 newly refactored ssa_is_replaceable_p.
3757 * cfgexpand.c: Include tree-outof-ssa.h.
3758 * ssaexpand.h: Delete.
3760 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3762 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3763 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3765 (create_gimple_tmp): Delete.
3766 (get_expr_type, build_assign, build_type_cast): Move to...
3767 * gimple-builder.c: New File.
3768 (get_expr_type): Relocate from gimple.c.
3769 (build_assign, build_type_cast): Change to only create ssanames.
3770 * gimple.h: Move prototypes to...
3771 * gimple-builder.h: New File. Here.
3772 * tree-ssa.h: And here.
3773 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3774 count_uses_and_derefs): Relocate from gimple.c.
3775 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3776 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3777 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3778 gimple_replace_ssa_lhs.
3779 * asan.c: Include gimple-builder.h.
3780 * Makefile.in: Add gimple-builder.o.
3782 2013-09-26 Richard Biener <rguenther@suse.de>
3784 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3785 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3786 (loe_visit_block): Use gcc_checking_assert.
3787 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3788 gimple_assign_ssa_name_copy_p.
3789 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3792 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3794 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3795 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3797 2013-09-26 Richard Biener <rguenther@suse.de>
3799 PR tree-optimization/58539
3800 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3801 the fact that debug statements are not taking part in loop-closed
3804 2013-09-26 Nick Clifton <nickc@redhat.com>
3806 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3807 time warning message.
3808 (msp430_print_operand_raw): Delete unused letter parameter.
3809 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3810 (msp430_print_operand_address): New function.
3811 (msp430_print_operand): Move address printing code from here to
3813 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3815 (zero_extendpsisi2): Likewise.
3816 (extendpsisi2): New pattern.
3817 (andneghi3): New pattern.
3819 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3821 * config/aarch64/aarch64.opt (mlra): New option.
3822 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3823 (TARGET_LRA_P): Define.
3825 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3827 * expr.c (expand_assignment): Remove obsolete comment.
3829 2013-09-25 Jeff Law <law@redhat.com>
3831 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3832 tree-ssa-threadupdate.h.
3833 (register_jump_thread): Similarly.
3834 * tree-ssa-threadupdate.h: New header file.
3835 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3836 * tree-vrp.c: Likewise.
3837 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3838 (thread_around_empty_blocks): Change type of path vector argument to
3839 an edge,type pair from just an edge. Initialize both elements when
3840 appending to a jump threading path. Tweak references to elements
3842 (thread_across_edge): Similarly. Release memory for the elements
3844 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3845 (dump_jump_thread_path): New function broken out from
3846 register_jump_thread.
3847 (register_jump_thread): Use dump_jump_thread_path. Change type of
3848 path vector entries. Search the path for NULL edges and dump
3849 the path if one is found. Tweak the conversion of path to 3-edge
3850 form to use the block copy type information embedded in the path.
3852 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3854 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3856 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3857 Vladimir Makarov <vmakarov@redhat.com>
3859 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3860 from the least significant bit.
3861 (strip_address_mutations): Add bitfield operations handling.
3862 (must_be_index_p): Add shifting and rotate operations handling.
3863 (set_address_base): Use must_be_base_p predicate.
3864 (set_address_index): Use must_be_index_p predicate.
3866 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3867 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3868 Sergey Lega <sergey.s.lega@intel.com>
3869 Anna Tikhonova <anna.tikhonova@intel.com>
3870 Ilya Tocar <ilya.tocar@intel.com>
3871 Andrey Turetskiy <andrey.turetskiy@intel.com>
3872 Ilya Verbin <ilya.verbin@intel.com>
3873 Kirill Yukhin <kirill.yukhin@intel.com>
3874 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3876 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3878 (ix86_expand_vector_move_misalign): Support new unaligned load and
3879 stores and use new names.
3880 (CODE_FOR_sse2_storedqu): Rename to ...
3881 (CODE_FOR_sse2_storedquv16qi): ... this.
3882 (CODE_FOR_sse2_loaddqu): Rename to ...
3883 (CODE_FOR_sse2_loaddquv16qi): ... this.
3884 (CODE_FOR_avx_loaddqu256): Rename to ...
3885 (CODE_FOR_avx_loaddquv32qi): ... this.
3886 (CODE_FOR_avx_storedqu256): Rename to ...
3887 (CODE_FOR_avx_storedquv32qi): ... this.
3888 * config/i386/i386.md (fpint_logic): New.
3889 * config/i386/sse.md (VMOVE): Extend for AVX512.
3893 (VI_UNALIGNED_LOADSTORE): Ditto.
3894 (sse2_avx_avx512f): Ditto.
3895 (sse2_avx2): Extend for AVX512.
3896 (sse4_1_avx2): Ditto.
3897 (avx2_avx512f): New.
3898 (sse): Extend for AVX512.
3901 (avxsizesuffix): Ditto.
3902 (sseintvecmode): Ditto.
3904 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3905 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3906 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3907 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3908 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3909 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3910 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3911 (STORENT_MODE): Extend for AVX512.
3912 (*absneg<mode>2): Replace constraint "x" with "v".
3913 (*mul<mode>3): Ditto.
3914 (*ieee_smin<mode>3): Ditto.
3915 (*ieee_smax<mode>3): Ditto.
3916 (avx_cmp<mode>3): Replace VF with VF_128_256.
3917 (*<sse>_maskcmp<mode>3_comm): Ditto.
3918 (<sse>_maskcmp<mode>3): Ditto.
3919 (<sse>_andnot<mode>3): Extend for AVX512.
3920 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3921 (<code><mode>3, fpint_logic): New.
3922 (*<code><mode>3): Extend for AVX512.
3923 (avx512flogicsuff): New.
3924 (avx512f_<logic><mode>): Ditto.
3925 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3927 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3928 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3929 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3930 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3931 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3932 (xop_vpermil2<mode>3): Ditto.
3933 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3934 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3935 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3936 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3938 2013-09-25 Tom Tromey <tromey@redhat.com>
3940 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3941 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3942 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3943 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3944 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3945 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3946 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3947 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3948 (GRAPHITE_HTAB_H): Remove.
3950 2013-09-25 Tom Tromey <tromey@redhat.com>
3952 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3954 2013-09-25 Tom Tromey <tromey@redhat.com>
3956 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3958 2013-09-25 Tom Tromey <tromey@redhat.com>
3960 * config/i386/t-i386 (i386.o): Remove.
3961 (i386-c.o): Use COMPILE and POSTCOMPILE.
3963 2013-09-25 Tom Tromey <tromey@redhat.com>
3965 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3967 2013-09-25 Tom Tromey <tromey@redhat.com>
3969 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3970 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3971 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3972 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3973 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3974 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3975 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3976 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3977 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3978 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3979 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3980 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3981 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3982 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3983 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3984 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3985 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3986 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3987 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3988 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3989 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3990 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3991 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3992 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3993 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3994 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3995 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3996 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3997 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3998 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3999 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4000 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4001 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4002 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4003 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4004 (graphite-clast-to-gimple.o, graphite-dependences.o)
4005 (graphite-interchange.o, graphite-poly.o)
4006 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4007 (graphite-optimize-isl.o, tree-vect-loop.o)
4008 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4009 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4010 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4011 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4012 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4013 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4014 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4015 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4016 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4017 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4018 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4019 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4020 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4021 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4022 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4023 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4024 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4025 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4026 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4027 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4028 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4029 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4030 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4031 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4032 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4033 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4034 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4035 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4036 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4037 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4038 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4039 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4040 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4041 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4042 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4043 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4044 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4045 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4046 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4047 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4048 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4049 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4050 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4051 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4052 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4053 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4054 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4055 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4056 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4057 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4058 (gcov-dump.o): Remove.
4059 (default-c.o): Use COMPILE and POSTCOMPILE.
4060 (CFLAGS-gcc.o): New variable.
4061 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4063 2013-09-25 Tom Tromey <tromey@redhat.com>
4065 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4066 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4067 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4068 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4069 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4070 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4071 (c-family/array-notation-common.o, c-family/stub-objc.o)
4072 (c-family/c-ubsan.o): Remove.
4074 2013-09-25 Tom Tromey <tromey@redhat.com>
4076 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4078 2013-09-25 Tom Tromey <tromey@redhat.com>
4080 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4081 to add -DENABLE_SHARED_LIBGCC.
4082 (gcc.o): Don't use subshell.
4084 2013-09-25 Tom Tromey <tromey@redhat.com>
4086 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4087 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4088 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4089 * configure, config.in: Rebuild.
4091 2013-09-25 Tom Tromey <tromey@redhat.com>
4093 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4094 (COMPILE, POSTCOMPILE): New variables.
4095 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4096 (DEPFILES): New variable.
4097 Include ".Po" files.
4098 * configure.ac: Add checks for dependency checking.
4099 * configure, aclocal.m4: Regenerate.
4101 2013-09-25 Tom Tromey <tromey@redhat.com>
4103 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4104 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4106 2013-09-25 Tom Tromey <tromey@redhat.com>
4108 * Makefile.in (generated_files): Add options.h,
4109 target-hooks-def.h, insn-opinit.h,
4110 common/common-target-hooks-def.h, pass-instances.def,
4111 c-family/c-target-hooks-def.h.
4113 2013-09-25 Jeff Law <law@redhat.com>
4115 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4116 than foo[foo.length () - 1] to access last member in a vec.
4117 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4119 2013-09-25 Richard Biener <rguenther@suse.de>
4122 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4124 2013-09-25 Jan Hubicka <jh@suse.cz>
4126 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4129 2013-09-25 Marek Polacek <polacek@redhat.com>
4132 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4133 when determining the type name.
4135 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4137 * config/sh/sh.md: Fix formatting.
4139 2013-09-24 Xinliang David Li <davidxl@google.com>
4141 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4142 max peel iterations parameter.
4143 * param.def: New parameter.
4144 * doc/invoke.texi: Document New parameter.
4146 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4148 * gimple-pretty-print.c: Various whitespace tweaks.
4149 * tree-core.h: Likewise.
4150 * tree-pretty-print.c: Likewise.
4151 * tree-ssa-alias.c: Likewise.
4152 * tree-ssa-copy.c: Likewise.
4153 * tree-ssanames.c: Likewise.
4154 * tree-ssanames.h: Likewise.
4155 * tree-vrp.c: Likewise.
4157 2013-09-24 Alan Modra <amodra@gmail.com>
4161 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4162 for output operands that disallow regs. Don't use EXPAND_WRITE on
4165 2013-09-24 Richard Biener <rguenther@suse.de>
4168 * tree.c (reference_alias_ptr_type): Move ...
4169 * alias.c (reference_alias_ptr_type): ... here and implement
4170 in terms of the new reference_alias_ptr_type_1.
4171 (ref_all_alias_ptr_type_p): New helper.
4172 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4173 use ref_all_alias_ptr_type_p.
4174 (get_deref_alias_set): Add flag_strict_aliasing check here.
4175 (reference_alias_ptr_type_1): New function, split out from ...
4176 (get_alias_set): ... here.
4177 (alias_ptr_types_compatible_p): New function.
4178 * alias.h (reference_alias_ptr_type): Declare.
4179 (alias_ptr_types_compatible_p): Likewise.
4180 * tree.h (reference_alias_ptr_type): Remove.
4181 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4182 to compare MEM_REF alias types.
4184 2013-09-24 Richard Biener <rguenther@suse.de>
4186 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4188 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4190 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4191 reload helper function arrays into a single array reg_addr.
4192 (reload_fpr_gpr): Likewise.
4193 (reload_gpr_vsx): Likewise.
4194 (reload_vsx_gpr): Likewise.
4195 (struct rs6000_reg_addr): Likewise.
4196 (reg_addr): Likewise.
4197 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4198 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4199 (rs6000_init_hard_regno_mode_ok): Likewise.
4200 (rs6000_secondary_reload_direct_move): Likewise.
4201 (rs6000_secondary_reload): Likewise.
4203 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4204 constraints: wu, ww, and wy. Repurpose wv constraint added during
4205 power8 changes. Put wg constraint in alphabetical order.
4207 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4208 for future work to add ISA 2.07 VSX single precision support.
4209 (-mvsx-scalar-double): Change default from -1 to 1, update
4210 documentation comment.
4211 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4212 (-mupper-regs-df): New debug switch to control whether DF values
4213 can go in the traditional Altivec registers.
4214 (-mupper-regs-sf): New debug switch to control whether SF values
4215 can go in the traditional Altivec registers.
4217 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4219 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4220 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4221 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4222 (rs6000_debug_legitimate_address_p): Print if we are running
4223 before, during, or after reload.
4224 (rs6000_secondary_reload): Add a comment.
4225 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4227 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4228 constraints. Update documentation string.
4229 (wd constraint): Likewise.
4230 (wf constraint): Likewise.
4231 (wg constraint): Likewise.
4232 (wn constraint): Likewise.
4233 (ws constraint): Likewise.
4234 (wt constraint): Likewise.
4235 (wx constraint): Likewise.
4236 (wz constraint): Likewise.
4237 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4239 (ww constraint): Likewise.
4240 (wy constraint): Likewise.
4241 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4242 in the previous submissions.
4243 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4245 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4247 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4248 (REG_BR_PROB): Say that the probability is stored in an int_list.
4249 * reg-notes.def: Update commentary to mention INT_LIST.
4250 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4251 (INT_LIST): New rtx.
4252 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4253 * rtlanal.c (int_reg_note_p): New function.
4254 (alloc_reg_note): Assert that the note does not have an int argument.
4255 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4256 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4257 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4258 rather than an INSN_LIST. Handle INT_LIST.
4259 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4260 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4261 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4262 Manipulate them as ints rather than rtxes.
4263 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4264 * regmove.c (copy_src_to_dest): Likewise.
4265 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4267 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4268 into the cases that need it.
4269 * config/arm/arm.c (arm_unwind_emit): Likewise.
4271 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4272 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4273 * loop-doloop.c (add_test, doloop_modify): Likewise.
4274 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4275 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4276 * predict.c (combine_predictions_for_insn): Likewise.
4277 * print-rtl.c (print_rtx): Handle INT_LIST.
4278 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4279 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4280 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4281 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4282 (ix86_split_fp_branch, predict_jump): Likewise.
4283 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4284 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4285 * config/spu/spu.c (ea_load_store_inline): Likewise.
4287 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4288 value of a REG_BR_PROB note.
4289 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4290 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4291 * emit-rtl.c (try_split): Likewise.
4292 * predict.c (br_prob_note_reliable_p): Likewise.
4293 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4294 * reorg.c (mostly_true_jump): Likewise.
4295 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4296 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4297 * config/i386/i386.c (ix86_print_operand): Likewise.
4298 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4299 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4300 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4301 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4302 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4303 * config/sparc/sparc.c (output_cbranch): Likewise.
4304 * config/spu/spu.c (get_branch_target): Likewise.
4305 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4306 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4308 2013-09-23 Jan Hubicka <jh@suse.cz>
4310 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4312 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4313 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4315 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4316 (gimple_fold_call): Dump inconsistent devirtualizations; add
4317 sanity check for type based devirtualizations.
4318 * ipa-prop.c: Include ipa-utils.h
4319 (ipa_intraprocedural_devirtualization): Add sanity check.
4320 (try_make_edge_direct_virtual_call): Likewise.
4322 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4324 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4325 assignment statements.
4327 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4329 * gimple-pretty-print.c (dump_ssaname_info): New function.
4330 (dump_gimple_phi): Call it.
4331 (pp_gimple_stmt_1): Likewise.
4332 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4333 (range_info_def): Declare.
4334 * tree-pretty-print.c (pp_double_int): New function.
4335 (dump_generic_node): Call it.
4336 * tree-pretty-print.h (pp_double_int): Declare.
4337 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4338 * tree-ssanames.h (range_info_def): New structure.
4339 (value_range_type): Move definition here.
4340 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4342 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4344 (set_range_info): New function.
4345 (get_range_info): Likewise.
4346 (duplicate_ssa_name_range_info): Likewise.
4347 (duplicate_ssa_name_fn): Check pointer type and call
4348 duplicate_ssa_name_range_info.
4349 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4350 * tree-vrp.c (value_range_type): Remove definition, now in
4352 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4353 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4354 (SSA_NAME_RANGE_INFO): New macro.
4356 2013-09-23 Richard Biener <rguenther@suse.de>
4358 PR tree-optimization/58464
4359 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4360 (phi_trans_add): Change to add conditionally on being not present.
4361 (phi_translate_1): Remove recursion detection here.
4362 (phi_translate): Pre-seed the cache with NULL to catch
4363 recursion here in a more generic way.
4364 (bitmap_find_leader): Adjust comment.
4365 (get_representative_for): Dump value-numbers.
4366 (create_expression_by_pieces): Likewise.
4367 (insert_into_preds_of_block): Likewise.
4369 2013-09-23 Christian Bruel <christian.bruel@st.com>
4372 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4373 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4375 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4378 2013-09-20 Renlin Li <renlin.li@arm.com>
4380 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4382 (aarch64_expand_epilogue): Likewise.
4383 (aarch64_legitimize_reload_address): Likewise.
4385 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4387 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4388 NULL_TREE before pushing them onto the vector. Likewise for labels.
4390 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4392 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4393 comparison operators when -fno-trapping-math is in effect.
4394 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4395 comparison operators in TFmode and assert that unsupported operators
4397 (ia64_print_operand): Likewise.
4399 2013-09-21 Jan Hubicka <jh@suse.cz>
4401 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4402 (sse_typeless_stores): Enable for core
4403 (sse_load0_by_pxor): Likewise.
4404 (four_jump_limit): Disable for core.
4405 (pad_returns): Likewise.
4406 (avoid_vector_decode): Likewise.
4407 (fuse_cmp_and_branch): Enable for cores.
4408 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4410 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4413 * config/pa/pa.c (pa_option_override): Disable auto increment and
4414 decrement instructions until reload is completed.
4416 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4417 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4418 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4420 2013-09-20 DJ Delorie <dj@redhat.com>
4421 Nick Clifton <nickc@redhat.com>
4423 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4424 (need_to_save): Save bank 0 on interrupts.
4425 (characterize_address): Strip far address wrappers.
4426 (rl78_as_legitimate_address): Likewise.
4427 (transcode_memory_rtx): Likewise.
4428 (rl78_peep_movhi_p): Disable this peephole after devirt.
4429 (rl78_propogate_register_origins): Forget all origins when a
4431 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4432 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4434 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4435 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4436 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4438 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4440 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4442 * config/pa/pa.md: In "scc" insn patterns, change output template to
4443 handle const0_rtx in reg_or_0_operand operands.
4445 2013-09-20 Martin Husemann <martin@NetBSD.org>
4448 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4449 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4451 2013-09-20 Richard Biener <rguenther@suse.de>
4454 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4455 remembering SSA name version and block index.
4456 (new_scev_info_str): Adjust.
4457 (hash_scev_info): Likewise. Also hash the block index.
4458 (eq_scev_info): Adjust.
4459 (find_var_scev_info): Likewise.
4460 (struct instantiate_cache_entry): Remove.
4461 (struct instantiate_cache_type): Use a htab to map name, block
4463 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4464 (get_instantiated_value_entry): Likewise.
4465 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4466 (instantiate_scev_name): Adjust.
4468 2013-09-20 Jeff Law <law@redhat.com>
4470 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4472 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4474 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4475 Call aarch64_simd_expand_args to update op[argc].
4477 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4479 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4482 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4484 * gengtype.c (file_rules): Added rule for *.cc files.
4485 (get_output_file_with_visibility): Give fatal message when no
4488 2013-09-20 Renlin Li <renlin.li@arm.com>
4490 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4491 (aarch64_expand_epilogue): Likewise.
4492 (aarch64_legitimize_reload_address): Likewise.
4494 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4497 * expr.c (expand_assignment): Remove misalignp code path.
4499 2013-09-20 Marek Polacek <polacek@redhat.com>
4502 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4503 TYPE_PRECISION. Add asserts.
4505 2013-09-20 Richard Biener <rguenther@suse.de>
4507 PR tree-optimization/58453
4508 * tree-loop-distribution.c (distribute_loop): Apply the cost
4509 model for -ftree-loop-distribute-patterns, too.
4511 2013-09-20 Richard Biener <rguenther@suse.de>
4514 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4515 make type comparison less strict.
4517 2013-09-20 Alan Modra <amodra@gmail.com>
4519 * configure: Regenerate.
4520 * aclocal.m4: Regenerate.
4522 2013-09-20 Marek Polacek <polacek@redhat.com>
4525 * doc/extend.texi: Document that attribute used is meant to be used
4526 on variables with static storage duration.
4528 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4530 PR tree-optimization/58472
4531 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4532 simd_lane_access set inv_p = false.
4533 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4534 the simduid magic VAR_DECL.
4536 2013-09-19 Jan Hubicka <jh@suse.cz>
4538 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4540 2013-09-17 Jeff Law <law@redhat.com>
4542 * tree-ssa-dom.c (record_temporary_equivalences): New function
4543 split out of dom_opt_dom_walker::after_dom_children.
4544 (dom_opt_dom_walker::thread_across_edge): Move common code
4545 in here from dom_opt_dom_walker::after_dom_children.
4546 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4548 2013-09-19 Jan Hubicka <jh@suse.cz>
4550 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4551 (TARGET_GENERIC): Use PROCESOR_GENERIC
4552 (enum processor_type): Unify generic32 and 64.
4553 * i386.md (cpu): Likewise.
4554 * x86-tune.def (use_leave): Enable for generic32.
4555 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4556 * athlon.md: Change generic64 to generic in all occurences.
4557 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4558 (ix86_target_macros_internal): Likewise.
4559 * driver-i386.c (host_detect_local_cpu): Likewise.
4560 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4562 (generic_memcpy, generic_memset, generic_cost): This one.
4563 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4564 (m_GENERIC32, m_GENERIC64): Remove.
4565 (m_GENERIC): Turn into one flag.
4566 (processor_target): Unify generic tunnings.
4567 (ix86_option_override_internal): Replace generic32/64 by generic.
4568 (ix86_issue_rate): Likewise.
4569 (ix86_adjust_cost): Likewise.
4571 2013-09-19 Jan Hubicka <jh@suse.cz>
4573 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4574 of speculative flag.
4576 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4578 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4579 GOMP_sections_start, even if !exit_reachable.
4581 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4583 * lra-constraints.c (need_for_all_save_p): Use macro
4584 HARD_REGNO_CALL_PART_CLOBBERED.
4585 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4586 set up pseudo conflict hard regs.
4588 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4591 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4594 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4596 PR rtl-optimization/58438
4597 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4598 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4599 from previous subpasses.
4601 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4603 * arm.c (arm_get_frame_offsets): Validate architecture supports
4604 LDRD/STRD before accepting the tuning preference.
4605 (arm_expand_prologue): Likewise.
4606 (arm_expand_epilogue): Likewise.
4608 2013-09-18 Richard Biener <rguenther@suse.de>
4610 PR tree-optimization/58417
4611 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4612 have chrecs with symbols defined in the loop as operands.
4613 (chrec_fold_multiply): Likewise.
4614 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4615 parameters before folding binary operations.
4616 (struct instantiate_cache_entry_hasher): Remove.
4617 (struct instantiate_cache_type): Use a pointer-map.
4618 (instantiate_cache_type::instantiate_cache_type): New function.
4619 (instantiate_cache_type::get): Likewise.
4620 (instantiate_cache_type::set): Likewise.
4621 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4622 (get_instantiated_value_entry): Likewise.
4623 (global_cache): New global.
4624 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4625 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4626 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4627 (instantiate_scev_name): Adjust.
4628 (instantiate_scev): Construct global instead of local cache.
4629 (resolve_mixers): Likewise.
4631 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4632 Paolo Carlini <paolo.carlini@oracle.com>
4635 * doc/implement-cxx.texi: Fix references to the C++ standards.
4637 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4639 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4640 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4641 simd_lane_access drs, update also DR_ALIGNED_TO.
4643 2013-09-18 Marek Polacek <polacek@redhat.com>
4646 * doc/extend.texi: Document no_sanitize_undefined attribute.
4647 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4648 no_sanitize_undefined attribute.
4650 2013-09-18 Nick Clifton <nickc@redhat.com>
4652 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4653 (ASM_DECLARE_FUNCTION_NAME): Define.
4655 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4657 * compare-elim.c (find_comparison_dom_walker): New class
4658 (find_comparisons_in_bb): Rename to
4659 find_comparison_dom_walker::before_dom_children
4660 (find_comparisons): Adjust
4661 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4663 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4664 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4665 (init_walk_dominator_tree): Remove declaration.
4666 (fini_walk_dominator_tree): Remove declaration.
4667 * fwprop.c (single_def_use_dom_walker): New class
4668 (single_def_use_enter_block): Convert to
4669 single_def_use_dom_walker::before_dom_children.
4670 (single_def_use_leave_block): Convert to
4671 single_def_use_dom_walker::after_dom_children.
4672 (build_single_def_use_links): Adjust.
4673 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4675 (find_candidates_in_block): Convert to
4676 find_candidates_dom_walker::before_dom_children.
4677 (execute_strength_reduction): Adjust.
4678 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4679 (sese_dom_walker): New class.
4680 (sese_dom_walker::sese_dom_walker): New constructor.
4681 (sese_dom_walker::~sese_dom_walker): New destructor.
4682 (build_sese_conditions_before): Convert to
4683 sese_dom_walker::before_dom_children.
4684 (build_sese_conditions_after): Convert to
4685 sese_dom_walker::after_dom_children.
4686 (build_poly_scop): Adjust
4687 * tree-into-ssa.c (rewrite_dom_walker): New class
4688 (rewrite_enter_block): Convert to
4689 rewrite_dom_walker::before_dom_children.
4690 (rewrite_leave_block): Convert to
4691 rewrite_dom_walker::after_dom_children.
4692 (rewrite_update_dom_walker): New class.
4693 (rewrite_update_enter_block): Convert to
4694 rewrite_update_dom_walker::before_dom_children.
4695 (rewrite_update_leave_block): Convert to
4696 rewrite_update_dom_walker::after_dom_children.
4697 (rewrite_blocks, rewrite_into_ssa): Adjust.
4698 (mark_def_dom_walker): New class.
4699 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4700 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4701 (mark_def_sites_blocks): Convert to
4702 mark_def_dom_walker::before_dom_children.
4703 (mark_def_site_blocks): Remove.
4704 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4705 (tree_ssa_dominator_optimize): Adjust.
4706 (dom_thread_across_edge): Convert to method
4707 dom_opt_dom_walker::thread_across_edge.
4708 (dom_opt_enter_block): Convert to member function
4709 dom_opt_dom_walker::before_dom_children.
4710 (dom_opt_leave_block): Convert to member function
4711 dom_opt_dom_walker::after_dom_children.
4712 * tree-ssa-dse.c (dse_dom_walker): New class.
4713 (dse_enter_block): Convert to member function
4714 dse_dom_walker::before_dom_children.
4715 (tree_ssa_dse): Adjust.
4716 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4717 (determine_invariantness_stmt): Convert to method
4718 invariantness_dom_walker::before_dom_children.
4719 (determine_invariantness): Remove
4720 (move_computations_dom_walker): New class.
4721 (move_computations_stmt): Convert to method
4722 move_computations_dom_walker::before_dom_children.
4723 (move_computations, tree_ssa_lim): Adjust.
4724 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4725 (nt_init_block): Make method
4726 notrappping_dom_walker::before_dom_children.
4727 (nt_fini_block): Make
4728 method nontrapping_dom_walker::after_dom_children.
4729 (get_non_trapping): Adjust.
4730 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4731 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4732 (eliminate_leave_block): Make method.
4733 eliminate_dom_walker::after_dom_children.
4735 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4736 (strlen_enter_block): Make method
4737 strlen_dom_walker::before_dom_children.
4738 (strlen_leave_block): Make
4739 method strlen_dom_walker::after_dom_children.
4740 (tree_ssa_strlen): Adjust.
4741 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4742 (tree_ssa_uncprop): Adjust.
4743 (uncprop_leave_block): Make method
4744 uncprop_dom_walker::after_dom_children.
4745 (uncprop_leave_block): Make method
4746 uncprop_dom_walker::before_dom_children.
4748 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4750 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4751 branch in same basic block. Check both src and dest of the move insn.
4753 2013-09-17 Nick Clifton <nickc@redhat.com>
4755 * config/rl78/rl78-real.md (bf): New pattern.
4757 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4758 (rl78_print_operand): Do not put a # before a %B.
4759 * config/rl78/rl78.opt: Tweak doc strings.
4761 2013-09-17 DJ Delorie <dj@redhat.com>
4763 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4764 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4765 (rl78_option_override): Likewise, if -mallregs.
4766 (is_virtual_register): Likewise.
4767 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4768 (REGNO_OK_FOR_BASE_P): Likewise.
4769 * config/rl78/rl78.opt (-mallregs): New.
4771 2013-09-17 Nick Clifton <nickc@redhat.com>
4773 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4774 For interrupt functions: save all call clobbered registers if the
4775 interrupt handler is not a leaf function.
4776 (rl78_expand_prologue): Always recompute the frame information.
4777 For interrupt functions: only select bank 0 if one of the bank 0
4778 registers is going to be psuhed.
4780 2013-09-17 DJ Delorie <dj@redhat.com>
4782 * config/rl78/constraints.md: For each W* constraint, rename to C*
4783 and create a W* constraint that checks for an optional ES: prefix
4785 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4786 (es_addr): New. Used to wrap far addresses.
4787 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4788 (rl78_es_base): New.
4789 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4790 wrapped far addresses.
4791 (rl78_print_operand_1): Unwrap far addresses before processing.
4792 (rl78_lo16): Wrap far addresses in unspecs.
4793 (rl78_es_addr): New.
4794 (rl78_es_base): New.
4795 (insn_ok_now): Check for not-yet-wrapped far addresses.
4796 (transcode_memory_rtx): Properly re-wrap far addresses.
4798 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4800 * config/sparc/t-rtems: Add leon3 multilibs.
4802 2013-09-17 Cong Hou <congh@google.com>
4804 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4805 when checking the dot production pattern. The type of rhs operand
4806 of multiply is now checked correctly.
4808 2013-09-17 Jeff Law <law@redhat.com>
4810 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4811 edge implied equivalences into successor phis.
4812 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4813 here from tree-ssa-threadedge.c.
4814 (mark_threaded_blocks): When threading through a joiner, if both
4815 successors of the joiner's clone reach the same block, verify the
4816 PHI arguments are equal. If not, cancel the jump threading request.
4817 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4818 tree-ssa-threadupdate.c
4819 (thread_across_edge): Don't check PHI argument equality when
4820 threading through joiner block here.
4822 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4824 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4825 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4826 (warn_uninit, warn_uninitialized_vars,
4827 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4828 Move to tree-ssa-uninit.c.
4829 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4830 (has_undefined_value_p): New. Pass dependant parts of
4831 ssa_undefined_value_p.
4832 (uninit_undefined_value_p): Use has_undefined_value_p.
4833 (warn_uninit, warn_uninitialized_vars,
4834 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4835 Move from tree-ssa.c.
4836 * tree-ssa.h: Adjust prototypes.
4838 2013-09-17 Jan Hubicka <jh@suse.cz>
4841 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4842 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4843 FUNCTION_NOT_OPTIMIZED.
4844 * ipa-inline-analysis.c (compute_inline_parameters): Function
4845 not optimized is not inlinable unless it is alwaysinline.
4846 (inline_analyze_function): Force calls in not optimized
4847 function not inlinable.
4849 2013-09-17 Jan Hubicka <jh@suse.cz>
4852 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4854 * ipa.c (function_and_variable_visibility): Likewise.
4855 * ipa-profile.c (ipa_profile): Likewise.
4857 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4860 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4861 attribute, and don't inline the resolver in this case.
4863 2013-09-17 Teresa Johnson <tejohnson@google.com>
4865 * coverage.c (get_coverage_counts): Add missing newline.
4867 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4869 PR tree-optimization/58088
4870 * fold-const.c (mask_with_trailing_zeros): New function.
4871 (fold_binary_loc): Make sure we don't recurse infinitely
4872 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4873 Use mask_with_trailing_zeros where appropriate.
4875 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4877 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4878 of 'prev' var to get better distance estimation.
4880 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4882 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4883 * tree-inline.c (is_parameter_of): New predicate.
4884 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4885 a parameter has been remapped.
4886 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4887 (optimize_inline_calls): Initialize transform_parameter.
4888 (copy_gimple_seq_and_replace_locals): Likewise.
4889 (tree_function_versioning): Likewise.
4890 (maybe_inline_call_in_expr): Likewise.
4892 2013-09-17 Nick Clifton <nickc@redhat.com>
4894 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4895 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4897 (is_attr_func): New function.
4898 (msp430_is_interrupt_func): New function.
4899 (is_naked_func): New function.
4900 (is_reentrant_func): New function.
4901 (is_critical_func): New function.
4902 (msp430_start_function): Add annotations for function attributes.
4903 (msp430_attr): New function.
4904 (msp430_attribute_table): New.
4905 (msp430_function_section): New function.
4906 (TARGET_ASM_FUNCTION_SECTION): Define.
4907 (msp430_builtin): New enum.
4908 (msp430_init_builtins): New function.
4909 (msp430_builtin_devl): New function.
4910 (msp430_expand_builtin): New function.
4911 (TARGET_INIT_BUILTINS): Define.
4912 (TARGET_EXPAND_BUILTINS): Define.
4913 (TARGET_BUILTIN_DECL): Define.
4914 (msp430_expand_prologue): Add support for naked, interrupt,
4915 critical and reentrant functions.
4916 (msp430_expand_epilogue): Likewise.
4917 (msp430_print_operand): Handle 'O' character.
4918 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4920 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4921 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4922 (pushm): Use a 'n' rather than an 'i' constraint.
4923 (msp_return): Add generation of the interrupt return instruction.
4924 (disable_interrupts): New pattern.
4925 (enable_interrupts): New pattern.
4926 (push_intr_state): New pattern.
4927 (pop_intr_state): New pattern.
4928 (bic_SR): New pattern.
4929 (bis_SR): New pattern.
4930 * doc/extend.texi: Document MSP430 function attributes and builtin
4933 2013-09-17 Richard Biener <rguenther@suse.de>
4935 PR tree-optimization/58432
4936 * tree-loop-distribution.c (tree_loop_distribution): Also
4937 scan PHIs for outside loop uses and seed a partition from them.
4939 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4941 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4942 (restructure_reference): Call backtrace_base_for_ref.
4944 2013-09-17 Alan Modra <amodra@gmail.com>
4947 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4950 2013-09-16 DJ Delorie <dj@redhat.com>
4952 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4953 vregs location for RL78/G10.
4954 (rl78_expand_prologue): Avoid SEL on G10.
4955 (rl78_expand_epilogue): Likewise.
4956 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4957 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4958 __RL78_G10__ when appropriate.
4959 (ASM_SPEC): Pass -mg10 along to the assembler.
4960 * config/rl78/rl78.md (sel_rb): Disable for G10.
4961 * config/rl78/rl78.opt: Add -mg10 option.
4962 * config/rl78/t-rl78: Add -mg10 multilib.
4964 2013-09-16 Xinliang David Li <davidxl@google.com>
4966 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4967 * omp-low.c (omp_max_vf): Ditto.
4968 (expand_omp_simd): Ditto.
4969 * tree-vectorizer.c (vectorize_loops): Ditto.
4970 (gate_vect_slp): Ditto.
4971 (gate_increase_alignment): Ditto.
4972 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4973 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4974 (gate_tree_vectorize): Name change.
4975 (tree_vectorize): Ditto.
4976 (pass_vectorize::gate): Call new function.
4977 (pass_vectorize::execute): Ditto.
4978 * opts.c: O3 default setting change.
4979 (finish_options): Check new flag.
4980 * doc/invoke.texi: Document new flags.
4981 * common.opt: New flags.
4983 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4985 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4986 (Condition Code): Don't reference it.
4987 * doc/tm.texi: Regenerate.
4989 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4992 * lra-constraints.c (undo_optional_reloads): Consider all optional
4993 reload even if it did not get a hard reg.
4995 2013-09-16 Teresa Johnson <tejohnson@google.com>
4997 * dumpfile.c (dump_loc): Remove newline emission.
4998 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4999 emission to dump_printf_loc calls where missing.
5000 (vect_mark_for_runtime_alias_test): Ditto.
5001 (vect_analyze_data_ref_dependence): Ditto.
5002 (vect_analyze_data_ref_dependences): Ditto.
5003 (vect_slp_analyze_data_ref_dependence): Ditto.
5004 (vect_slp_analyze_data_ref_dependences): Ditto.
5005 (vect_compute_data_ref_alignment): Ditto.
5006 (vect_update_misalignment_for_peel): Ditto.
5007 (vect_verify_datarefs_alignment): Ditto.
5008 (vector_alignment_reachable_p): Ditto.
5009 (vect_get_data_access_cost): Ditto.
5010 (vect_enhance_data_refs_alignment): Ditto.
5011 (vect_find_same_alignment_drs): Ditto.
5012 (vect_analyze_data_refs_alignment): Ditto.
5013 (vect_analyze_group_access): Ditto.
5014 (vect_analyze_data_ref_access): Ditto.
5015 (vect_analyze_data_ref_accesses): Ditto.
5016 (vect_prune_runtime_alias_test_list): Ditto.
5017 (vect_analyze_data_refs): Ditto.
5018 (vect_create_addr_base_for_vector_ref): Ditto.
5019 (vect_create_data_ref_ptr): Ditto.
5020 (vect_grouped_store_supported): Ditto.
5021 (vect_grouped_load_supported): Ditto.
5022 * value-prof.c (check_counter): Ditto.
5023 (check_ic_target): Ditto.
5024 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5025 (vect_recog_widen_mult_pattern): Ditto.
5026 (vect_recog_widen_sum_pattern): Ditto.
5027 (vect_recog_over_widening_pattern): Ditto.
5028 (vect_recog_widen_shift_pattern): Ditto.
5029 (vect_recog_rotate_pattern): Ditto.
5030 (vect_recog_vector_vector_shift_pattern): Ditto.
5031 (vect_recog_divmod_pattern): Ditto.
5032 (vect_recog_mixed_size_cond_pattern): Ditto.
5033 (vect_recog_bool_pattern): Ditto.
5034 (vect_pattern_recog_1): Ditto.
5035 (vect_pattern_recog): Ditto.
5036 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5037 (vect_is_simple_iv_evolution): Ditto.
5038 (vect_analyze_scalar_cycles_1): Ditto.
5039 (vect_get_loop_niters): Ditto.
5040 (vect_analyze_loop_1): Ditto.
5041 (vect_analyze_loop_form): Ditto.
5042 (vect_analyze_loop_operations): Ditto.
5043 (vect_analyze_loop_2): Ditto.
5044 (vect_analyze_loop): Ditto.
5045 (report_vect_op): Ditto.
5046 (vect_is_slp_reduction): Ditto.
5047 (vect_is_simple_reduction_1): Ditto.
5048 (vect_get_known_peeling_cost): Ditto.
5049 (vect_estimate_min_profitable_iters): Ditto.
5050 (vect_model_reduction_cost): Ditto.
5051 (vect_model_induction_cost): Ditto.
5052 (get_initial_def_for_induction): Ditto.
5053 (vect_create_epilog_for_reduction): Ditto.
5054 (vectorizable_reduction): Ditto.
5055 (vectorizable_induction): Ditto.
5056 (vectorizable_live_operation): Ditto.
5057 (vect_loop_kill_debug_uses): Ditto.
5058 (vect_transform_loop): Ditto.
5059 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5060 (vect_stmt_relevant_p): Ditto.
5061 (process_use): Ditto.
5062 (vect_mark_stmts_to_be_vectorized): Ditto.
5063 (vect_model_simple_cost): Ditto.
5064 (vect_model_promotion_demotion_cost): Ditto.
5065 (vect_model_store_cost): Ditto.
5066 (vect_get_store_cost): Ditto.
5067 (vect_model_load_cost): Ditto.
5068 (vect_get_load_cost): Ditto.
5069 (vect_init_vector_1): Ditto.
5070 (vect_get_vec_def_for_operand): Ditto.
5071 (vect_finish_stmt_generation): Ditto.
5072 (vectorizable_call): Ditto.
5073 (vectorizable_conversion): Ditto.
5074 (vectorizable_assignment): Ditto.
5075 (vectorizable_shift): Ditto.
5076 (vectorizable_operation): Ditto.
5077 (vectorizable_store): Ditto.
5078 (vectorizable_load): Ditto.
5079 (vectorizable_condition): Ditto.
5080 (vect_analyze_stmt): Ditto.
5081 (vect_transform_stmt): Ditto.
5082 (vect_is_simple_use): Ditto.
5083 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5084 (vect_can_advance_ivs_p): Ditto.
5085 (vect_update_ivs_after_vectorizer): Ditto.
5086 (vect_do_peeling_for_loop_bound): Ditto.
5087 (vect_gen_niters_for_prolog_loop): Ditto.
5088 (vect_update_inits_of_drs): Ditto.
5089 (vect_create_cond_for_alias_checks): Ditto.
5090 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5091 (vect_build_slp_tree_1): Ditto.
5092 (vect_supported_load_permutation_p): Ditto.
5093 (vect_analyze_slp_instance): Ditto.
5094 (vect_analyze_slp): Ditto.
5095 (vect_make_slp_decision): Ditto.
5096 (vect_detect_hybrid_slp): Ditto.
5097 (vect_bb_vectorization_profitable_p): Ditto.
5098 (vect_slp_analyze_bb_1): Ditto.
5099 (vect_update_slp_costs_according_to_vf): Ditto.
5100 (vect_get_mask_element): Ditto.
5101 (vect_transform_slp_perm_load): Ditto.
5102 (vect_schedule_slp_instance): Ditto.
5103 (vect_schedule_slp): Ditto.
5104 (vect_slp_transform_bb): Ditto.
5105 * profile.c (read_profile_edge_counts): Ditto.
5106 (compute_branch_probabilities): Ditto.
5107 * coverage.c (get_coverage_counts): Ditto.
5109 2013-09-16 Diego Novillo <dnovillo@google.com>
5111 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5113 2013-09-16 Jan Hubicka <jh@suse.cz>
5115 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5116 abstract functions; for static functions check the presence of body.
5118 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5120 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5121 * config/aarch64/aarch64-simd.md
5122 (aarch64_mla_elt<mode>): New.
5123 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5124 (aarch64_mls_elt<mode>): Likewise.
5125 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5126 (aarch64_fma4_elt<mode>): Likewise.
5127 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5128 (aarch64_fma4_elt_to_128v2df): Likewise.
5129 (aarch64_fma4_elt_to_64df): Likewise.
5130 (fnma<mode>4): Likewise.
5131 (aarch64_fnma4_elt<mode>): Likewise.
5132 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5133 (aarch64_fnma4_elt_to_128v2df): Likewise.
5134 (aarch64_fnma4_elt_to_64df): Likewise.
5135 * config/aarch64/iterators.md (VDQSF): New.
5136 * config/aarch64/arm_neon.h
5137 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5138 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5140 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5142 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5143 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5144 (aarch64_mul3_elt_to_128df): Likewise.
5145 (aarch64_mul3_elt_to_64v2df): Likewise.
5146 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5148 (VMUL_CHANGE_NLANES) Likewise.
5151 * config/aarch64/arm_neon.h
5152 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5154 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5156 * config/aarch64/arm_neon.h
5157 (vcvtx_high_f32_f64): Fix parameters.
5159 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5160 Uros Bizjak <ubizjak@gmail.com>
5162 * config/alpha.c: Include tree-ssanames.h.
5164 2013-09-16 Richard Biener <rguenther@suse.de>
5166 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5167 (dot_rdg_1): Handle control_dd.
5168 (create_edge_for_control_dependence): New function.
5169 (create_rdg_edges): Add control dependences if asked for.
5170 (build_rdg): Likewise.
5171 (generate_loops_for_partition): If there are not necessary
5172 control stmts remove all their dependencies.
5173 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5174 (distribute_loop): Pass on control dependences.
5175 (tree_loop_distribution): Compute control dependences and remove
5176 restriction on number of loop nodes.
5178 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5180 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5183 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5185 * cse.c (try_const_anchors): Punt on CC modes.
5187 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5189 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5191 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5194 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5197 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5200 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5202 (darwin_objc1_section): Likewise.
5203 (darwin_file_end): Emit Image Info section when required.
5205 2013-09-14 Jan Hubicka <jh@suse.cz>
5207 * tree-into-ssa.c (gate_into_ssa): New.
5208 (pass_data_build_ssa): Use it.
5209 * cgraph.h (expand_thunk): Update prototype.
5210 * cgraphunit.c (analyze_function): Expand thunks early.
5211 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5212 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5213 set lowered flag; do not add new function.
5214 (assemble_thunks_and_aliases): Update.
5215 * tree-ssa.c (gate_init_datastructures): New gate.
5216 (pass_data_init_datastructures): Use it.
5218 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5221 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5222 xmm register set described in the psABI.
5224 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5226 * dwarf2out.c (should_emit_struct_debug): Add check
5227 for type_decl variable is not NULL.
5229 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5231 * config.gcc: Use new winnt-c.c target hooks
5232 * config/t-winnt: New file
5233 * config/winnt-c.c: New file
5234 * doc/tm.texi.in: Document new hook
5235 * doc/tm.texi: Regenerated
5237 2013-09-13 Jan Hubicka <jh@suse.cz>
5240 * ipa-inline.c (check_callers): New function.
5241 (check_caller_edge): Remove.
5242 (want_inline_function_to_all_callers_p): Also permit alises that are
5244 (inline_to_all_callers): Terminate the walk when devirtualization
5245 introduce new calls.
5247 2013-09-13 Jan Hubicka <jh@suse.cz>
5249 * ipa-inline-analysis.c (struct growth_data): Add node.
5250 (do_estimate_growth_1): Fix detection of recursion.
5252 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5254 PR tree-optimization/58392
5255 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5256 to avoid shadowing of outer loop variable. If
5257 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5258 replace_by_duplicate_decl simduid of loops that have it set and
5259 set dest_cfun->has_simduid_loops and/or
5260 dest_cfun->has_force_vect_loops.
5261 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5262 instead of maybe_lookup_decl.
5263 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5264 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5265 call. Copy over force_vect and copy and remap simduid. Set
5266 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5267 (copy_cfg_body): Remove blocks_to_copy argument. Use
5268 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5269 caller. Don't set cfun->has_simduid_loops and/or
5270 cfun->has_force_vect_loops here.
5271 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5273 (expand_call_inline, tree_function_versioning): Adjust copy_body
5276 2013-09-13 Martin Jambor <mjambor@suse.cz>
5279 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5280 the assert if the edge was a speculative one.
5282 2013-09-13 Richard Biener <rguenther@suse.de>
5284 * tree-data-ref.h (known_dependences_p): Move here ...
5285 * tree-loop-distribution.c (known_dependences_p): ... from here.
5286 (dump_rdg_component, debug_rdg_component): Remove.
5288 (generate_loops_for_partition): Use gimple_uid instead of
5289 relying on matching stmt visit order.
5290 (rdg_build_partitions): Take starting stmt vector.
5291 (ldist_gen): Merge into ...
5292 (distribute_loop): ... this function. Do not compute starting
5294 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5296 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5298 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5299 Set type attribute correctly. Set predicable_short_it attribute.
5300 (cmpsi_shiftsi): Remove %? from output template.
5302 2013-09-13 Richard Biener <rguenther@suse.de>
5304 * tree-loop-distribution.c (struct rdg_component,
5305 rdg_defs_used_in_other_loops_p, free_rdg_components,
5306 rdg_build_components): Remove.
5307 (stmts_from_loop): Do not record virtual PHIs.
5308 (generate_loops_for_partition): Skip virtual PHIs.
5309 (build_rdg_partition_for_component): Rename to ...
5310 (build_rdg_partition_for_vertex): ... this and adjust.
5311 (rdg_build_partitions): Take a vector of starting vertices
5312 instead of components. Remove unnecessary leftover handling.
5313 (ldist_gen): Do not build components or record other stores.
5314 (distribute_loop): Do not distribute loops containing stmts
5317 2013-09-13 Christian Bruel <christian.bruel@st.com>
5320 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5322 2013-09-13 Kai Tietz <ktietz@redhat.com>
5324 * config.gcc: Separate cases for mingw and cygwin targets,
5325 and add 64-bit cygwin target case.
5327 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5328 dll-export inline-functions.
5329 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5332 2013-09-13 Jeff Law <law@redhat.com>
5336 2013-09-06 Jeff Law <law@redhat.com>
5338 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5339 edge implied equivalences into successor phis.
5341 2013-09-12 DJ Delorie <dj@redhat.com>
5343 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5346 2013-09-12 Brooks Moses <bmoses@google.com>
5349 * Makefile.in: Do not install driver binaries in $(target)/bin.
5351 2013-09-12 DJ Delorie <dj@redhat.com>
5353 * config/rl78/rl78.opt (mrelax): New.
5354 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5355 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5357 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5359 (rl78_expand_epilogue): Likewise.
5361 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5364 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5365 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5366 pointer elimination with using remove_reg_equal_offset_note.
5368 2013-09-12 DJ Delorie <dj@redhat.com>
5370 * config/msp430/: New port.
5371 * config.gcc (msp430): Added.
5372 * doc/invoke.texi: Document MSP430 options.
5373 * doc/install.texi: Document msp430-elf
5374 * doc/md.texi: Document msp430-elf
5375 * doc/contrib.texi: Document msp430-elf
5377 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5380 2013-09-12 Martin Jambor <mjambor@suse.cz>
5383 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5384 reference descriptor is NULL.
5385 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5388 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5390 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5391 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5392 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5393 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5394 make_temp_ssa_name): move to tree-ssanames.h
5395 * tree-ssa-alias.h: Move prototype.
5396 * tree-ssa.h: Include tree-ssanames.h.
5397 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5398 * tree-ssanames.h: New. Move items from tree-flow*.h
5399 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5401 2013-09-12 Richard Biener <rguenther@suse.de>
5403 PR tree-optimization/58404
5404 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5405 propagate non-invariant addresses into dereferences wrapped
5406 in component references.
5408 2013-09-12 Richard Biener <rguenther@suse.de>
5410 PR tree-optimization/58402
5411 * passes.def: Move pass_late_warn_uninitialized later.
5413 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5415 * tree-ssa.h: New. Move content from tree-flow.h and
5417 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5418 Move prototypes belonging to tree-ssa.c.
5419 * tree-flow-inline.h (redirect_edge_var_map_def,
5420 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5422 * gimple.h: Adjust prototypes.
5423 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5425 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5426 * tree.h: Move prototype to tree-ssa.h.
5427 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5428 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5429 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5430 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5431 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5432 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5433 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5434 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5435 graphite-blocking.c, graphite-clast-to-gimple.c,
5436 graphite-dependences.c, graphite-interchange.c,
5437 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5438 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5439 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5440 ipa-reference.c, ipa-split.c, ipa-utils.c,
5441 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5442 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5443 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5444 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5445 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5446 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5447 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5448 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5449 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5450 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5451 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5452 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5453 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5454 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5455 value-prof.c, var-tracking.c,
5456 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5458 2013-09-12 Richard Biener <rguenther@suse.de>
5460 PR tree-optimization/58396
5461 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5462 (build_rdg): Take a loop-nest parameter, fix memleaks.
5463 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5465 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5467 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5470 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5472 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5473 The FMA instruction names should have a 'v' prefix.
5475 2013-09-12 Richard Biener <rguenther@suse.de>
5477 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5478 (dot_rdg): Use popen instead of system in optional code.
5479 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5480 (already_processed_vertex_p): Adjust.
5481 (has_anti_or_output_dependence, predecessor_has_mem_write,
5482 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5483 rdg_flag_uses): Remove.
5484 (rdg_flag_vertex): Simplify.
5485 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5487 (build_rdg_partition_for_component): Process the first vertex
5488 of a component only.
5489 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5491 2013-09-12 Alan Modra <amodra@gmail.com>
5493 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5495 2013-09-11 DJ Delorie <dj@redhat.com>
5496 Nick Clifton <nickc@redhat.com>
5498 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5499 (rl78_stack_based_mem): New.
5500 * config/rl78/constraints.md (Iv08): New.
5509 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5511 (movsi): Change from expand to insn-and-split.
5512 (ashrsi3): Clobber AX.
5516 * config/rl78/rl78.md (CC_REG): Fix.
5517 (addsi3): Allow memory and immediate operands.
5518 (addsi3_internal): Split into...
5519 (addsi3_internal_virt): ...new, and ...
5520 (addsi3_internal_real): ...new.
5522 (subsi3_internal_virt): New.
5523 (subsi3_internal_real): New.
5524 (mulsi3): Add memory operand.
5525 (mulsi3_rl78): Likewise.
5526 (mulsi3_g13): Likewise.
5527 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5528 (cbranchqi4_real): Add more constraint options.
5529 (cbranchhi4_real): Expand pattern.
5530 (cbranchhi4_real_signed): New.
5531 (cbranchhi4_real_inverted): New.
5532 (cbranchsi4_real_lt): New.
5533 (cbranchsi4_real_ge): New.
5534 (cbranchsi4_real_signed): New.
5535 (cbranchsi4_real): New.
5537 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5539 (lshrsi3_virt): Likewise.
5540 (ashlsi3_virt): Likewise.
5541 (cbranchqi4_virt_signed): New.
5542 (cbranchhi4_virt_signed): New.
5543 (cbranchsi4_virt): New.
5544 * config/rl78/rl78.c: Whitespace fixes throughout.
5545 (move_elim_pass): New.
5546 (pass_data_rl78_move_elim): New.
5547 (pass_rl78_move_elim): New.
5548 (make_pass_rl78_move_elim): New.
5549 (rl78_devirt_info): Run devirt earlier.
5550 (rl78_move_elim_info): New.
5551 (rl78_asm_file_start): Register it.
5552 (rl78_split_movsi): New.
5553 (rl78_as_legitimate_address): Allow virtual base registers when
5555 (rl78_addr_space_convert): Remove spurious debug stuff.
5556 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5557 (rl78_print_operand): More cases for not printing '#'.
5558 (rl78_expand_compare): Remove most of the logic.
5559 (content_memory): New.
5560 (clear_content_memory): New.
5561 (get_content_index): New.
5562 (get_content_name): New.
5563 (display_content_memory): New.
5564 (update_content): New.
5565 (record_content): New.
5566 (already_contains): New.
5567 (insn_ok_now): Re-recog insns with virtual registers.
5568 (add_postponed_content_update): New.
5569 (process_postponed_content_update): New.
5570 (gen_and_emit_move): New.
5571 (transcode_memory_rtx): Record new location content.
5572 Use gen_and_emit_move.
5573 (force_into_acc): New.
5574 (move_to_acc): Use gen_and_emit_move.
5575 (move_from_acc): Likewise.
5576 (move_acc_to_reg): Likewise.
5577 (move_to_x): Likewise.
5578 (move_to_hl): Likewise.
5579 (move_to_de): Likewise.
5580 (rl78_alloc_physical_registers_op1): Record location content.
5581 (has_constraint): New.
5582 (rl78_alloc_physical_registers_op2): Record location content.
5584 (rl78_alloc_physical_registers_ro1): Likewise.
5585 (rl78_alloc_physical_registers_cmp): Likewise.
5586 (rl78_alloc_physical_registers_umul): Likewise.
5587 (rl78_alloc_address_registers_macax): New.
5588 (rl78_alloc_physical_registers): Initialize and set location
5589 content memory as needed.
5590 (rl78_reorg): Make sure split2 is called.
5591 (rl78_rtx_costs): New.
5593 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5595 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5596 for (not (neg ...)) and (neg (not ...)) cases.
5598 2013-09-11 Richard Biener <rguenther@suse.de>
5601 * passes.def: Split critical edges before late uninit warning passes.
5602 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5604 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5606 PR tree-optimization/58385
5607 * fold-const.c (build_range_check): If both low and high are NULL,
5608 use omit_one_operand_loc to preserve exp side-effects.
5610 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5612 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5614 2013-09-11 Richard Biener <rguenther@suse.de>
5616 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5617 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5618 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5619 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5620 stmts_from_loop, known_dependences_p, build_empty_rdg,
5621 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5622 * tree-loop-distribution.c: ... here.
5623 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5624 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5625 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5626 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5627 * tree-loop-distribution.c: ... here.
5628 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5629 (struct partition_s): Add loops member.
5630 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5631 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5632 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5634 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5635 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5636 Sergey Lega <sergey.s.lega@intel.com>
5637 Anna Tikhonova <anna.tikhonova@intel.com>
5638 Ilya Tocar <ilya.tocar@intel.com>
5639 Andrey Turetskiy <andrey.turetskiy@intel.com>
5640 Ilya Verbin <ilya.verbin@intel.com>
5641 Kirill Yukhin <kirill.yukhin@intel.com>
5642 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5644 * config/i386/constraints.md (k): New.
5646 * config/i386/i386.c (const regclass_map): Add new mask registers.
5647 (dbx_register_map): Ditto.
5648 (dbx64_register_map): Ditto.
5649 (svr4_dbx_register_map): Ditto.
5650 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5652 (ix86_preferred_reload_class): Disable constants for mask registers.
5653 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5654 (ix86_hard_regno_mode_ok): Support new mask registers.
5655 (x86_order_regs_for_local_alloc): Ditto.
5656 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5657 (FIXED_REGISTERS): Add new mask registers.
5658 (CALL_USED_REGISTERS): Ditto.
5659 (REG_ALLOC_ORDER): Ditto.
5660 (VALID_MASK_REG_MODE): New.
5661 (FIRST_MASK_REG): Ditto.
5662 (LAST_MASK_REG): Ditto.
5663 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5664 (MAYBE_MASK_CLASS_P): New.
5665 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5666 (REG_CLASS_CONTENTS): Ditto.
5667 (MASK_REGNO_P): New.
5668 (ANY_MASK_REG_P): Ditto.
5669 (HI_REGISTER_NAMES): Add new mask registers.
5670 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5671 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5673 (attribute "type"): Add mskmov, msklog.
5674 (attribute "length_immediate"): Support them.
5675 (attribute "memory"): Ditto.
5676 (attribute "prefix_0f"): Ditto.
5677 (*movhi_internal): Support new mask registers.
5678 (*movqi_internal): Ditto.
5679 (define_split): Split out clobber pattern is a logic
5680 insn on mask registers.
5681 (*k<logic><mode>): New.
5682 (*andhi_1): Extend to support mask regs.
5683 (*andqi_1): Extend to support mask regs.
5685 (define_split): Split and-not to and and not if operands
5687 (*<code><mode>_1): Separate HI mode to new pattern...
5688 (*<code>hi_1): This.
5689 (*<code>qi_1): Extend to support mask regs.
5691 (kortestzhi): Ditto.
5692 (kortestchi): Ditto.
5694 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5695 (*one_cmplhi2_1): ...Here, now with mask registers support.
5696 (*one_cmplqi2_1): Support new mask registers.
5697 (HI/QImode arithmetics splitter): Don't split if mask registers
5699 (HI/QImode not splitter): Ditto.
5700 * config/i386/predicated.md (mask_reg_operand): New.
5701 (general_reg_operand): Ditto.
5703 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5705 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5706 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5708 2013-09-10 Jeff Law <law@redhat.com>
5710 PR tree-optimization/58380
5711 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5712 of threading through a buried loop header.
5714 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5715 return value for single successor case.
5717 2013-09-10 Jan Hubicka <jh@suse.cz>
5719 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5721 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5724 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5725 support conditional execution.
5726 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5728 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5730 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5732 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5733 Check destination too to check move insn.
5734 (undo_optional_reloads): Add check that the original peudo did not
5735 changed its allocation and the optional reload was inherited on last
5736 inheritance pass. Break loop after deciding to keep optional reload.
5737 (lra_undo_inheritance): Add check that inherited pseudo still in
5740 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5742 * config/aarch64/aarch64.md (generic_sched): New.
5743 * config/aarch64/aarch64-generic.md (load): Make conditional
5744 on generic_sched attribute.
5745 (nonload): Likewise.
5747 2013-09-10 Jan Hubicka <jh@suse.cz>
5749 * lto-cgraph.c: Include ipa-utils.h.
5750 (compute_ltrans_boundary): Also add possible targets into the boundary.
5752 2013-09-10 Jan Hubicka <jh@suse.cz>
5754 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5755 VAR_DECL of vtable rather than full expression.
5757 2013-09-10 Jan Hubicka <jh@suse.cz>
5758 Paolo Carlini <paolo.carlini@oracle.com>
5760 * cgraphunit.c (analyze_functions): Save input_location, set it
5761 to UNKNOWN_LOCATION and restore it at the end.
5763 2013-09-10 Martin Jambor <mjambor@suse.cz>
5765 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5766 represented by a thunk.
5768 2013-09-10 Jeff Law <law@redhat.com>
5770 PR tree-optimization/58343
5771 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5772 jump threading requests through loop headers buried in the middle
5773 of a jump threading path.
5775 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5776 in return value/type.
5778 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5780 PR rtl-optimization/58365
5781 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5782 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5785 2013-09-10 Richard Biener <rguenther@suse.de>
5787 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5788 * tree-data-ref.c (create_rdg_vertices): Collect all data
5789 references, signal failure to the caller, use data-ref API.
5790 (build_rdg): Compute data references only once. Maintain lifetime
5791 of data references and data dependences from within RDG.
5792 (free_rdg): Free dependence relations.
5793 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5794 inventing extra dependences.
5795 (distribute_loop): Update for RDG API changes.
5797 2013-09-10 Kai Tietz <ktietz@redhat.com>
5799 * doc/invoke.texi (fms-extensions): Document changed
5800 behavior for ms-abi targets.
5801 * config/i386/i386.c (ix86_option_override_internal):
5802 Set default value of option -fms-extension for ms-abi targets.
5804 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5806 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5808 2013-09-10 Alan Modra <amodra@gmail.com>
5811 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5813 2013-09-10 Alan Modra <amodra@gmail.com>
5815 * config/rs6000/predicates.md (add_cint_operand): New.
5816 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5817 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5818 using add_cint_operand.
5819 (largetoc_high_plus_aix): Likewise.
5821 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5823 PR tree-optimization/58364
5824 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5825 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5826 the current range can't be an unconditional true or false.
5828 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5830 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5832 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5834 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5836 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5839 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5841 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5843 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5844 NO_REGS for immediate that can't be moved directly into FP_REGS.
5846 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5848 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5849 comparison with negated operand.
5850 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5853 2013-09-09 Richard Biener <rguenther@suse.de>
5856 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5857 of a subloop record all its block as affecting loop-closed SSA form.
5859 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5861 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5862 of an rtx/bitpos pair.
5863 (store_fixed_bit_field): Update accordingly.
5865 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5867 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5869 * builtins.c (expand_errno_check): Likewise.
5870 * dwarf2cfi.c (init_return_column_size): Likewise.
5871 * except.c (sjlj_mark_call_sites): Likewise.
5872 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5873 * lra-constraints.c (emit_inc): Likewise.
5874 * ree.c (combine_set_extension): Likewise.
5875 * regmove.c (fixup_match_2): Likewise.
5876 * reload1.c (inc_for_reload): Likewise.
5878 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5880 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5881 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5882 Use gen_int_mode with the mode of the associated simplify_* call.
5883 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5884 * expmed.c (expand_shift_1): Likewise.
5885 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5886 * loop-iv.c (iv_number_of_iterations): Likewise.
5887 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5888 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5890 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5892 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5893 of the associated expand_* call.
5894 (asan_emit_stack_protection): Likewise.
5895 * builtins.c (round_trampoline_addr): Likewise.
5896 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5897 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5898 (emit_store_flag): Likewise.
5899 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5900 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5902 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5903 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5904 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5905 * modulo-sched.c (generate_prolog_epilog): Likewise.
5906 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5907 (expand_ctz, expand_ffs, expand_unop): Likewise.
5909 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5911 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5912 of the associated gen_rtx_* call.
5913 * caller-save.c (init_caller_save): Likewise.
5914 * combine.c (find_split_point, make_extraction): Likewise.
5915 (make_compound_operation): Likewise.
5916 * dwarf2out.c (mem_loc_descriptor): Likewise.
5917 * explow.c (plus_constant, probe_stack_range): Likewise.
5918 * expmed.c (expand_mult_const): Likewise.
5919 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5920 * reload1.c (init_reload): Likewise.
5921 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5922 * var-tracking.c (adjust_mems): Likewise.
5923 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5924 rather than gen_rtx_fmt_ee.
5926 2013-09-09 Jan Hubicka <jh@suse.cz>
5929 * value-prof.c (gimple_ic): Copy also abnormal edges.
5931 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5933 * asan.c (asan_shadow_cst): Use gen_int_mode.
5935 2013-09-08 Jan Hubicka <jh@suse.cz>
5937 * ipa-profile.c: Add toplevel comment.
5938 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5939 (contains_hot_call_p): New function.
5940 (ipa_propagate_frequency): Set frequencies based on counts when
5942 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5943 profile; do not tamper with profile after inlining if it is read.
5945 2013-09-08 Jan Hubicka <jh@suse.cz>
5947 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5950 2013-09-08 Jan Hubicka <jh@suse.cz>
5952 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5955 2013-09-08 Jeff Law <law@redhat.com>
5958 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5961 2013-09-08 Andi Kleen <ak@linux.intel.com>
5963 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5965 2013-09-08 Jan Hubicka <jh@suse.cz>
5967 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5969 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5971 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5972 for non-debug insns.
5973 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5974 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5975 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5978 2013-09-08 Jan Hubicka <jh@suse.cz>
5980 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5981 targets and devirtualize to BUILT_IN_UNREACHABLE.
5982 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5983 * ipa.c (walk_polymorphic_call_targets): New function.
5984 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5985 functions; use the new timevar.
5986 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5987 was removed from the program.
5988 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5989 not consider it in the walk when its vtable is dead.
5990 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5992 (devirt_variable_node_removal_hook): New function.
5993 (possible_polymorphic_call_targets): Also register
5994 devirt_variable_node_removal_hook.
5995 (ipa_devirt): Do not do non-speculative devirtualization.
5996 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5998 2013-09-08 Jan Hubicka <jh@suse.cz>
6000 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6001 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6002 varpool_remove_variable_insertion_hook): Declare.
6003 * varpool.c (varpool_node_hook_list): New structure.
6004 (first_varpool_node_removal_hook,
6005 first_varpool_variable_insertion_hook): New variables.
6006 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6007 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6008 varpool_remove_variable_insertion_hook,
6009 varpool_call_variable_insertion_hooks): New functions.
6010 (varpool_remove_node): Use it.
6012 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6015 * diagnostic.c (diagnostic_build_prefix): When s.file is
6016 "<built-in>" don't output line and column numbers.
6018 2013-09-06 Jan Hubicka <jh@suse.cz>
6020 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6021 * lto-streamer-out.c: Stream thunks, too.
6022 * lto-streamer-in.c (input_function): Pop cfun here
6023 (lto_read_body): Instead of here.
6025 2013-09-06 Caroline Tice <cmtice@google.com>
6027 * doc/install.texi: Add documentation for the --enable-vtable-verify
6028 and the --disable-libvtv configure options.
6030 2013-09-06 Jeff Law <law@redhat.com>
6032 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6033 edge implied equivalences into successor phis.
6035 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6037 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6039 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6041 * resource.c (mark_target_live_regs): Compute resources taking
6042 into account if a call is predicated or not.
6044 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6046 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6047 the compiler in the function names.
6049 2013-09-06 Jan Hubicka <jh@suse.cz>
6052 * ipa-inline.c (has_caller_p): New function.
6053 (want_inline_function_to_all_callers_p): Use it.
6054 (sum_callers, inline_to_all_callers): Break out from ...
6055 (ipa_inline): ... here.
6057 2013-09-06 Jan Hubicka <jh@suse.cz>
6059 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6060 only when AVX is enabled.
6062 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6064 * config/aarch64/aarch64.md
6065 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6066 is fpsimd_<load/store>2.
6067 (load_pair<mode>): Likewise.
6068 (store_pair<mode>): Likewise.
6070 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6072 * config/arm/types.md (type): Add "mrs" type.
6073 * config/aarch64/aarch64.md
6074 (aarch64_load_tp_hard): Make type "mrs".
6076 (load_tp_hard): Make type "mrs".
6077 * config/arm/cortex-a15.md: Update with new attributes.
6078 * config/arm/cortex-a5.md: Update with new attributes.
6079 * config/arm/cortex-a53.md: Update with new attributes.
6080 * config/arm/cortex-a7.md: Update with new attributes.
6081 * config/arm/cortex-a8.md: Update with new attributes.
6082 * config/arm/cortex-a9.md: Update with new attributes.
6083 * config/arm/cortex-m4.md: Update with new attributes.
6084 * config/arm/cortex-r4.md: Update with new attributes.
6085 * config/arm/fa526.md: Update with new attributes.
6086 * config/arm/fa606te.md: Update with new attributes.
6087 * config/arm/fa626te.md: Update with new attributes.
6088 * config/arm/fa726te.md: Update with new attributes.
6090 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6092 * config/aarch64/aarch64.md
6093 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6094 (*movtf_aarch64): Likewise.
6096 (thumb1_movdi_insn): Use "multiple" for type where more than one
6097 instruction is used for a move.
6098 (*arm32_movhf): Likewise.
6099 (*thumb_movdf_insn): Likewise.
6101 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6103 * config/arm/types.md (type): Rename fcpys to fmov.
6105 (*arm_movsi_vfp): Rename type fcpys as fmov.
6106 (*thumb2_movsi_vfp): Likewise
6107 (*movhf_vfp_neon): Likewise
6108 (*movhf_vfp): Likewise
6109 (*movsf_vfp): Likewise
6110 (*thumb2_movsf_vfp): Likewise
6111 (*movsfcc_vfp): Likewise
6112 (*thumb2_movsfcc_vfp): Likewise
6113 * config/aarch64/aarch64-simd.md
6114 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6115 * config/aarch64/aarch64.md
6116 (*movsi_aarch64): Replace type mov_reg with fmovs.
6117 (*movdi_aarch64): Likewise
6118 (*movsf_aarch64): Likewise
6119 (*movdf_aarch64): Likewise
6121 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6122 * config/arm/iwmmxt.md
6123 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6124 * config/arm/arm1020e.md: Update with new attributes.
6125 * config/arm/cortex-a15-neon.md: Update with new attributes.
6126 * config/arm/cortex-a5.md: Update with new attributes.
6127 * config/arm/cortex-a53.md: Update with new attributes.
6128 * config/arm/cortex-a7.md: Update with new attributes.
6129 * config/arm/cortex-a8-neon.md: Update with new attributes.
6130 * config/arm/cortex-a9.md: Update with new attributes.
6131 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6132 * config/arm/cortex-r4f.md: Update with new attributes.
6133 * config/arm/marvell-pj4.md: Update with new attributes.
6134 * config/arm/vfp11.md: Update with new attributes.
6136 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6138 * config/aarch64/aarch64.md
6139 (*madd<mode>): Fix type attribute.
6140 (*maddsi_uxtw): Likewise.
6141 (*msub<mode>): Likewise.
6142 (*msubsi_uxtw): Likewise.
6143 (<su_optab>maddsidi4): Likewise.
6144 (<su_optab>msubsidi4): Likewise.
6146 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6148 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6149 * config/arm/arm.md (core_cycles): Remove fdiv.
6150 * config/arm/vfp.md:
6151 (*sqrtsf2_vfp): Update for attribute changes.
6152 (*sqrtdf2_vfp): Likewise.
6153 * config/aarch64/aarch64.md:
6154 (sqrt<mode>2): Update for attribute changes.
6155 * config/arm/arm1020e.md: Update with new attributes.
6156 * config/arm/cortex-a15-neon.md: Update with new attributes.
6157 * config/arm/cortex-a5.md: Update with new attributes.
6158 * config/arm/cortex-a53.md: Update with new attributes.
6159 * config/arm/cortex-a7.md: Update with new attributes.
6160 * config/arm/cortex-a8-neon.md: Update with new attributes.
6161 * config/arm/cortex-a9.md: Update with new attributes.
6162 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6163 * config/arm/cortex-r4f.md: Update with new attributes.
6164 * config/arm/marvell-pj4.md: Update with new attributes.
6165 * config/arm/vfp11.md: Update with new attributes.
6167 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6169 * config/arm/types.md
6170 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6171 * config/aarch64/aarch64.md
6172 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6174 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6175 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6176 (float<GPI:mode><GPF:mode>2): Likewise.
6178 (*truncsisf2_vfp): Update with new attributes.
6179 (*truncsidf2_vfp): Likewise.
6180 (fixuns_truncsfsi2): Likewise.
6181 (fixuns_truncdfsi2): Likewise.
6182 (*floatsisf2_vfp): Likewise.
6183 (*floatsidf2_vfp): Likewise.
6184 (floatunssisf2): Likewise.
6185 (floatunssidf2): Likewise.
6186 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6187 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6188 * config/arm/arm1020e.md: Update with new attributes.
6189 * config/arm/cortex-a15-neon.md: Update with new attributes.
6190 * config/arm/cortex-a5.md: Update with new attributes.
6191 * config/arm/cortex-a53.md: Update with new attributes.
6192 * config/arm/cortex-a7.md: Update with new attributes.
6193 * config/arm/cortex-a8-neon.md: Update with new attributes.
6194 * config/arm/cortex-a9.md: Update with new attributes.
6195 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6196 * config/arm/cortex-r4f.md: Update with new attributes.
6197 * config/arm/marvell-pj4.md: Update with new attributes.
6198 * config/arm/vfp11.md: Update with new attributes.
6200 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6202 * config/aarch64/arm_neon.h
6203 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6204 (vqtbx<1,2,3,4><q>_s8): Likewise.
6206 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6208 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6209 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6210 (add<mode>3): Add type attribute.
6211 (add<mode>3): Likewise.
6212 (usadd<mode>3): Likewise.
6213 (ssadd<mode>3): Likewise.
6214 (sub<mode>3): Likewise.
6215 (sub<mode>3): Likewise.
6216 (ussub<mode>3): Likewise.
6217 (sssub<mode>3): Likewise.
6218 (ssmulsa3): Likewise.
6219 (usmulusa3): Likewise.
6220 (arm_usatsihi): Likewise.
6222 (*movdi_vfp): Add types for all instructions.
6223 (*movdi_vfp_cortexa8): Likewise.
6224 (*movhf_vfp_neon): Likewise.
6225 (*movhf_vfp): Likewise.
6226 (*movdf_vfp): Likewise.
6227 (*thumb2_movdf_vfp): Likewise.
6228 (*thumb2_movdfcc_vfp): Likewise.
6229 * config/arm/arm.md: Add type attribute to all insn patterns.
6230 (*thumb1_adddi3): Add type attribute.
6231 (*arm_adddi3): Likewise.
6232 (*adddi_sesidi_di): Likewise.
6233 (*adddi_zesidi_di): Likewise.
6234 (*thumb1_addsi3): Likewise.
6235 (addsi3_compare0): Likewise.
6236 (*addsi3_compare0_scratch): Likewise.
6237 (*compare_negsi_si): Likewise.
6238 (cmpsi2_addneg): Likewise.
6239 (*addsi3_carryin_<optab>): Likewise.
6240 (*addsi3_carryin_alt2_<optab>): Likewise.
6241 (*addsi3_carryin_clobercc_<optab>): Likewise.
6242 (*subsi3_carryin): Likewise.
6243 (*subsi3_carryin_const): Likewise.
6244 (*subsi3_carryin_compare): Likewise.
6245 (*subsi3_carryin_compare_const): Likewise.
6246 (*arm_subdi3): Likewise.
6247 (*thumb_subdi3): Likewise.
6248 (*subdi_di_zesidi): Likewise.
6249 (*subdi_di_sesidi): Likewise.
6250 (*subdi_zesidi_di): Likewise.
6251 (*subdi_sesidi_di): Likewise.
6252 (*subdi_zesidi_ze): Likewise.
6253 (thumb1_subsi3_insn): Likewise.
6254 (*arm_subsi3_insn): Likewise.
6255 (*anddi3_insn): Likewise.
6256 (*anddi_zesidi_di): Likewise.
6257 (*anddi_sesdi_di): Likewise.
6258 (*ne_zeroextracts): Likewise.
6259 (*ne_zeroextracts): Likewise.
6260 (*ite_ne_zeroextr): Likewise.
6261 (*ite_ne_zeroextr): Likewise.
6262 (*anddi_notdi_di): Likewise.
6263 (*anddi_notzesidi): Likewise.
6264 (*anddi_notsesidi): Likewise.
6265 (andsi_notsi_si): Likewise.
6266 (thumb1_bicsi3): Likewise.
6267 (*iordi3_insn): Likewise.
6268 (*iordi_zesidi_di): Likewise.
6269 (*iordi_sesidi_di): Likewise.
6270 (*thumb1_iorsi3_insn): Likewise.
6271 (*xordi3_insn): Likewise.
6272 (*xordi_zesidi_di): Likewise.
6273 (*xordi_sesidi_di): Likewise.
6274 (*arm_xorsi3): Likewise.
6275 (*andsi_iorsi3_no): Likewise.
6276 (*smax_0): Likewise.
6277 (*smax_m1): Likewise.
6278 (*arm_smax_insn): Likewise.
6279 (*smin_0): Likewise.
6280 (*arm_smin_insn): Likewise.
6281 (*arm_umaxsi3): Likewise.
6282 (*arm_uminsi3): Likewise.
6283 (*minmax_arithsi): Likewise.
6284 (*minmax_arithsi_): Likewise.
6285 (*satsi_<SAT:code>): Likewise.
6286 (arm_ashldi3_1bit): Likewise.
6287 (arm_ashrdi3_1bit): Likewise.
6288 (arm_lshrdi3_1bit): Likewise.
6289 (*arm_negdi2): Likewise.
6290 (*thumb1_negdi2): Likewise.
6291 (*arm_negsi2): Likewise.
6292 (*thumb1_negsi2): Likewise.
6293 (*negdi_extendsid): Likewise.
6294 (*negdi_zero_extend): Likewise.
6295 (*arm_abssi2): Likewise.
6296 (*thumb1_abssi2): Likewise.
6297 (*arm_neg_abssi2): Likewise.
6298 (*thumb1_neg_abss): Likewise.
6299 (one_cmpldi2): Likewise.
6300 (extend<mode>di2): Likewise.
6301 (*compareqi_eq0): Likewise.
6302 (*arm_extendhisi2addsi): Likewise.
6303 (*arm_movdi): Likewise.
6304 (*thumb1_movdi_insn): Likewise.
6305 (*arm_movt): Likewise.
6306 (*thumb1_movsi_insn): Likewise.
6307 (pic_add_dot_plus_four): Likewise.
6308 (pic_add_dot_plus_eight): Likewise.
6309 (tls_load_dot_plus_eight): Likewise.
6310 (*thumb1_movhi_insn): Likewise.
6311 (*thumb1_movsf_insn): Likewise.
6312 (*movdf_soft_insn): Likewise.
6313 (*thumb_movdf_insn): Likewise.
6314 (cbranchsi4_insn): Likewise.
6315 (cbranchsi4_scratch): Likewise.
6316 (*negated_cbranchsi4): Likewise.
6317 (*tbit_cbranch): Likewise.
6318 (*tlobits_cbranch): Likewise.
6319 (*tstsi3_cbranch): Likewise.
6320 (*cbranchne_decr1): Likewise.
6321 (*addsi3_cbranch): Likewise.
6322 (*addsi3_cbranch_scratch): Likewise.
6323 (*arm_cmpdi_insn): Likewise.
6324 (*arm_cmpdi_unsig): Likewise.
6325 (*arm_cmpdi_zero): Likewise.
6326 (*thumb_cmpdi_zero): Likewise.
6327 (*deleted_compare): Likewise.
6328 (*mov_scc): Likewise.
6329 (*mov_negscc): Likewise.
6330 (*mov_notscc): Likewise.
6331 (*cstoresi_eq0_thumb1_insn): Likewise.
6332 (cstoresi_nltu_thumb1): Likewise.
6333 (cstoresi_ltu_thu): Likewise.
6334 (thumb1_addsi3_addgeu): Likewise.
6335 (*arm_jump): Likewise.
6336 (*thumb_jump): Likewise.
6337 (*check_arch2): Likewise.
6338 (arm_casesi_internal): Likewise.
6339 (thumb1_casesi_dispatch): Likewise.
6340 (*arm_indirect_jump): Likewise.
6341 (*thumb1_indirect_jump): Likewise.
6343 (*and_scc): Likewise.
6344 (*ior_scc): Likewise.
6345 (*compare_scc): Likewise.
6346 (*cond_move): Likewise.
6347 (*cond_arith): Likewise.
6348 (*cond_sub): Likewise.
6349 (*cmp_ite0): Likewise.
6350 (*cmp_ite1): Likewise.
6351 (*cmp_and): Likewise.
6352 (*cmp_ior): Likewise.
6353 (*ior_scc_scc): Likewise.
6354 (*ior_scc_scc_cmp): Likewise.
6355 (*and_scc_scc): Likewise.
6356 (*and_scc_scc_cmp): Likewise.
6357 (*and_scc_scc_nod): Likewise.
6358 (*negscc): Likewise.
6359 (movcond_addsi): Likewise.
6360 (movcond): Likewise.
6361 (*ifcompare_plus_move): Likewise.
6362 (*if_plus_move): Likewise.
6363 (*ifcompare_move_plus): Likewise.
6364 (*if_move_plus): Likewise.
6365 (*ifcompare_arith_arith): Likewise.
6366 (*if_arith_arith): Likewise.
6367 (*ifcompare_arith_move): Likewise.
6368 (*if_arith_move): Likewise.
6369 (*ifcompare_move_arith): Likewise.
6370 (*if_move_arith): Likewise.
6371 (*ifcompare_move_not): Likewise.
6372 (*if_move_not): Likewise.
6373 (*ifcompare_not_move): Likewise.
6374 (*if_not_move): Likewise.
6375 (*ifcompare_shift_move): Likewise.
6376 (*if_shift_move): Likewise.
6377 (*ifcompare_move_shift): Likewise.
6378 (*if_move_shift): Likewise.
6379 (*ifcompare_shift_shift): Likewise.
6380 (*ifcompare_not_arith): Likewise.
6381 (*ifcompare_arith_not): Likewise.
6382 (*if_arith_not): Likewise.
6383 (*ifcompare_neg_move): Likewise.
6384 (*if_neg_move): Likewise.
6385 (*ifcompare_move_neg): Likewise.
6386 (*if_move_neg): Likewise.
6387 (prologue_thumb1_interwork): Likewise.
6388 (*cond_move_not): Likewise.
6389 (*sign_extract_onebit): Likewise.
6390 (*not_signextract_onebit): Likewise.
6391 (stack_tie): Likewise.
6392 (align_4): Likewise.
6393 (align_8): Likewise.
6394 (consttable_end): Likewise.
6395 (consttable_1): Likewise.
6396 (consttable_2): Likewise.
6397 (consttable_4): Likewise.
6398 (consttable_8): Likewise.
6399 (consttable_16): Likewise.
6400 (*thumb1_tablejump): Likewise.
6401 (prefetch): Likewise.
6402 (force_register_use): Likewise.
6403 (thumb_eh_return): Likewise.
6404 (load_tp_hard): Likewise.
6405 (load_tp_soft): Likewise.
6406 (tlscall): Likewise.
6407 (*arm_movtas_ze): Likewise.
6408 (*arm_rev): Likewise.
6409 (*arm_revsh): Likewise.
6410 (*arm_rev16): Likewise.
6411 * config/arm/thumb2.md
6412 (*thumb2_smaxsi3): Likewise.
6413 (*thumb2_sminsi3): Likewise.
6414 (*thumb32_umaxsi3): Likewise.
6415 (*thumb2_uminsi3): Likewise.
6416 (*thumb2_negdi2): Likewise.
6417 (*thumb2_abssi2): Likewise.
6418 (*thumb2_neg_abss): Likewise.
6419 (*thumb2_movsi_insn): Likewise.
6420 (tls_load_dot_plus_four): Likewise.
6421 (*thumb2_movhi_insn): Likewise.
6422 (*thumb2_mov_scc): Likewise.
6423 (*thumb2_mov_negs): Likewise.
6424 (*thumb2_mov_negs): Likewise.
6425 (*thumb2_mov_nots): Likewise.
6426 (*thumb2_mov_nots): Likewise.
6427 (*thumb2_movsicc_): Likewise.
6428 (*thumb2_movsfcc_soft_insn): Likewise.
6429 (*thumb2_indirect_jump): Likewise.
6430 (*thumb2_and_scc): Likewise.
6431 (*thumb2_ior_scc): Likewise.
6432 (*thumb2_ior_scc_strict_it): Likewise.
6433 (*thumb2_cond_move): Likewise.
6434 (*thumb2_cond_arith): Likewise.
6435 (*thumb2_cond_ari): Likewise.
6436 (*thumb2_cond_sub): Likewise.
6437 (*thumb2_negscc): Likewise.
6438 (*thumb2_movcond): Likewise.
6439 (thumb2_casesi_internal): Likewise.
6440 (thumb2_casesi_internal_pic): Likewise.
6441 (*thumb2_alusi3_short): Likewise.
6442 (*thumb2_mov<mode>_shortim): Likewise.
6443 (*thumb2_addsi_short): Likewise.
6444 (*thumb2_subsi_short): Likewise.
6445 (thumb2_addsi3_compare0): Likewise.
6446 (*thumb2_cbz): Likewise.
6447 (*thumb2_cbnz): Likewise.
6448 (*thumb2_one_cmplsi2_short): Likewise.
6449 (*thumb2_negsi2_short): Likewise.
6450 (*orsi_notsi_si): Likewise.
6451 * config/arm/arm1020e.md: Update with new attributes.
6452 * config/arm/arm1026ejs.md: Update with new attributes.
6453 * config/arm/arm1136jfs.md: Update with new attributes.
6454 * config/arm/arm926ejs.md: Update with new attributes.
6455 * config/arm/cortex-a15.md: Update with new attributes.
6456 * config/arm/cortex-a5.md: Update with new attributes.
6457 * config/arm/cortex-a53.md: Update with new attributes.
6458 * config/arm/cortex-a7.md: Update with new attributes.
6459 * config/arm/cortex-a8.md: Update with new attributes.
6460 * config/arm/cortex-a9.md: Update with new attributes.
6461 * config/arm/cortex-m4.md: Update with new attributes.
6462 * config/arm/cortex-r4.md: Update with new attributes.
6463 * config/arm/fa526.md: Update with new attributes.
6464 * config/arm/fa606te.md: Update with new attributes.
6465 * config/arm/fa626te.md: Update with new attributes.
6466 * config/arm/fa726te.md: Update with new attributes.
6468 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6470 * config/aarch64/aarch64-simd.md
6471 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6472 <vwx> iterator to ensure correct register choice.
6473 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6474 (aarch64_sqdmull_n<mode>): Likewise.
6475 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6476 * config/aarch64/arm_neon.h
6477 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6478 (vml<as><q>_n_<su>16): Likewise.
6479 (vml<as>l_high_lane<q>_<su>16): Likewise.
6480 (vml<as>l_high_n_<su>16): Likewise.
6481 (vml<as>l_lane<q>_<su>16): Likewise.
6482 (vml<as>l_n_<su>16): Likewise.
6483 (vmul<q>_lane<q>_<su>16): Likewise.
6484 (vmul<q>_n_<su>16): Likewise.
6485 (vmull_lane<q>_<su>16): Likewise.
6486 (vmull_n_<su>16): Likewise.
6487 (vmull_high_lane<q>_<su>16): Likewise.
6488 (vmull_high_n_<su>16): Likewise.
6489 (vqrdmulh<q>_n_s16): Likewise.
6491 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6493 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6494 have the correct lane parameter.
6496 2013-09-06 Richard Biener <rguenther@suse.de>
6498 * cfganal.c (control_dependences::~control_dependences):
6499 Properly free all of the vector.
6501 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6504 * config/i386/i386.c (ix86_conditional_register_usage):
6505 Proper initialize extended SSE registers.
6507 2013-09-06 Jan Hubicka <jh@suse.cz>
6509 PR tree-optimization/58311
6510 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6512 2013-09-06 Jan Hubicka <jh@suse.cz>
6514 * Makefile.in (tree-sra.o): Update dependencies.
6515 * tree-sra.c: Include ipa-utils.h
6516 (scan_function): Use recursive_call_p.
6517 (has_caller_p): New function.
6518 (cgraph_for_node_and_aliases): Count also callers of aliases.
6520 2013-09-06 Jan Hubicka <jh@suse.cz>
6523 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6524 * tree-tailcall.c: Include ipa-utils.h.
6525 (find_tail_calls): Use it.
6526 * ipa-pure-const.c (check_call): Likewise.
6527 * ipa-utils.c (recursive_call_p): New function.
6528 * ipa-utils.h (recursive_call_p): Dclare.
6529 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6530 (symtab_semantically_equivalent_p): New function.
6531 * Makefile.in (tree-tailcall.o): Update dependencies.
6533 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6535 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6538 2013-09-06 Richard Biener <rguenther@suse.de>
6540 * lto-streamer.h (lto_global_var_decls): Remove.
6541 * Makefile.in (OBJS): Remove lto-symtab.o.
6542 (lto-symtab.o): Remove.
6543 (GTFILES): Remove lto-symtab.c
6544 * lto-symtab.c: Move to lto/
6546 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6548 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6549 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6550 (UNSPEC_FPINT_RINT): New constant definitions.
6551 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6552 definition with 2 attributes.
6553 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6554 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6557 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6559 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6560 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6561 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6562 attribute to "z196_alone".
6563 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6566 2013-09-06 Richard Biener <rguenther@suse.de>
6568 * basic-block.h (class control_dependences): New.
6569 * tree-ssa-dce.c (control_dependence_map): Remove.
6571 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6572 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6573 find_pdom, find_control_dependence, find_all_control_dependences):
6575 (mark_control_dependent_edges_necessary,
6576 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6577 tree_dce_done, perform_tree_ssa_dce): Adjust.
6578 * cfganal.c (set_control_dependence_map_bit,
6579 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6580 find_all_control_dependences): Move from tree-ssa-dce.c and
6581 implement as methods of control_dependences class.
6582 (control_dependences::control_dependences): New.
6583 (control_dependences::~control_dependences): Likewise.
6584 (control_dependences::get_edges_dependent_on): Likewise.
6585 (control_dependences::get_edge): Likewise.
6587 2013-09-04 Jan Hubicka <jh@suse.cz>
6589 * tree.c (types_same_for_odr): Drop overactive check.
6590 * ipa-devirt.c (hash_type_name): Likewise.
6592 2013-09-04 Jan Hubicka <jh@suse.cz>
6594 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6595 (analyze_functions): ... here.
6597 2013-09-04 Jan Hubicka <jh@suse.cz>
6600 * cgraphunit.c (analyze_functions): Clear AUX fields
6601 after processing; initialize assembler name has.
6603 2013-09-05 Jeff Law <law@redhat.com>
6605 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6606 from thread_around_empty_block. Record threading path into PATH.
6607 Recurse if threading through the initial block is successful.
6608 (thread_across_edge): Corresponding changes to slightly simplify.
6610 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6612 * config/aarch64/aarch64.md
6613 (type): Remove frecpe, frecps, frecpx.
6614 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6615 fix to be a TARGET_SIMD instruction.
6616 (aarch64_frecps): Remove.
6617 * config/aarch64/aarch64-simd.md
6618 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6619 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6621 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6622 Sofiane Naci <sofiane.naci@arm.com>
6624 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6625 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6626 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6627 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6628 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6629 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6630 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6631 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6632 into "clz, "rbit". Rename "shift" to "shift_imm".
6633 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6634 changes. Update for attribute changes all occurrences of arlo_* and
6636 * config/arm/arm-fixed.md: Update for attribute changes
6637 all occurrences of arlo_* types.
6638 * config/arm/thumb2.md: Update for attribute changes all occurrences
6640 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6641 (cortexa7_older_only): Likewise.
6642 (cortexa7_younger): Likewise.
6643 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6644 (1020alu_shift_op): Likewise.
6645 (1020alu_shift_reg_op): Likewise.
6646 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6647 (alu_shift_op): Likewise.
6648 (alu_shift_reg_op): Likewise.
6649 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6650 (11_alu_shift_op): Likewise.
6651 (11_alu_shift_reg_op): Likewise.
6652 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6653 (9_alu_shift_reg_op): Likewise.
6654 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6656 (cortex_a15_alu_shift): Likewise.
6657 (cortex_a15_alu_shift_reg): Likewise.
6658 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6660 (cortex_a5_alu_shift): Likewise.
6661 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6663 (cortex_a53_alu_shift): Likewise.
6664 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6666 (cortex_a7_alu_reg): Likewise.
6667 (cortex_a7_alu_shift): Likewise.
6668 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6670 (cortex_a8_alu_shift): Likewise.
6671 (cortex_a8_alu_shift_reg): Likewise.
6672 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6673 (cortex_a9_dp_shift): Likewise.
6674 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6676 * config/arm/cortex-r4.md
6677 (cortex_r4_alu): Update for attribute changes.
6678 (cortex_r4_mov): Likewise.
6679 (cortex_r4_alu_shift_reg): Likewise.
6680 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6681 (526_alu_shift_op): Likewise.
6682 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6683 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6684 (626te_alu_shift_op): Likewise.
6685 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6686 (726te_alu_shift_op): Likewise.
6687 (726te_alu_shift_reg_op): Likewise.
6688 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6689 (mp626_alu_shift_op): Likewise.
6690 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6691 (pj4_alu_conds): Likewise.
6692 (pj4_shift): Likewise.
6693 (pj4_shift_conds): Likewise.
6694 (pj4_alu_shift): Likewise.
6695 (pj4_alu_shift_conds): Likewise.
6696 * config/aarch64/aarch64.md: Update for attribute change
6697 all occurrences of arlo_* and shift* types.
6699 2013-09-05 Mike Stump <mikestump@comcast.net>
6701 * tree.h: Move documentation for tree_function_decl to tree-core.h
6702 with the declaration.
6704 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6707 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6708 looking for widest mode.
6710 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6712 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6714 2013-09-05 Richard Biener <rguenther@suse.de>
6716 PR tree-optimization/58137
6717 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6718 Do not create vectors of pointers.
6719 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6720 types for the components of the vector initializer.
6721 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6722 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6724 2013-09-05 Martin Jambor <mjambor@suse.cz>
6726 * ipa-prop.c (remove_described_reference): Accept missing references,
6727 return false if that hppens, otherwise return true.
6728 (cgraph_node_for_jfunc): New function.
6729 (try_decrement_rdesc_refcount): Likewise.
6730 (try_make_edge_direct_simple_call): Use them.
6731 (ipa_edge_removal_hook): Remove references from rdescs.
6732 (ipa_edge_duplication_hook): Clone rdescs and their references
6733 when the new edge has the same caller as the old one.
6734 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6735 reference before removing any edges.
6737 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6739 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6741 * thumb2.md (thumb2_storewb_parisi): New pattern.
6743 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6745 * config/aarch64/aarch64-option-extensions.def: Add
6746 AARCH64_OPT_EXTENSION of 'crc'.
6747 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6748 (AARCH64_ISA_CRC): Ditto.
6749 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6750 description of the CRC extension.
6752 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6754 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6755 * config/rs6000/linux.h: Ditto.
6756 * alpha/linux.h: Ditto.
6757 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6758 no_c99_libc_has_function.
6759 * config/c6x/uclinux-elf.h: Ditto.
6760 * config/lm32/uclinux-elf.h: Ditto.
6761 * config/m68k/uclinux.h: Ditto.
6762 * config/moxie/uclinux.h: Ditto.
6763 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6764 (crisv32-*-linux*, cris-*-linux*): Ditto.
6765 * config/bfin/bfin.c: Include "tm_p.h".
6767 2013-09-05 Richard Biener <rguenther@suse.de>
6769 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6770 check for a definition without a basic-block.
6772 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6773 Sofiane Naci <sofiane.naci@arm.com>
6775 * config/aarch64/aarch64.md
6776 (*movti_aarch64): Rename r_2_f and f_2_r.
6777 (*movsf_aarch64): Likewise.
6778 (*movdf_aarch64): Likewise.
6779 (*movtf_aarch64): Likewise.
6780 (aarch64_movdi_<mode>low): Likewise.
6781 (aarch64_movdi_<mode>high): Likewise.
6782 (aarch64_mov<mode>high_di): Likewise.
6783 (aarch64_mov<mode>low_di): Likewise.
6784 (aarch64_movtilow_tilow): Likewise.
6785 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6786 values to config/arm/types.md
6787 (attribute "conds"): Update for attribute change.
6788 (anddi3_insn): Likewise.
6789 (iordi3_insn): Likewise.
6790 (xordi3_insn): Likewise.
6791 (one_cmpldi2): Likewise.
6792 * config/arm/types.md (type): Add Neon types.
6793 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6794 use "type" attribute.
6795 (movmisalign<mode>_neon_store): Likewise.
6796 (movmisalign<mode>_neon_load): Likewise.
6797 (vec_set<mode>_internal): Likewise.
6798 (vec_setv2di_internal): Likewise.
6799 (vec_extract<mode>): Likewise.
6800 (vec_extractv2di): Likewise.
6801 (add<mode>3_neon): Likewise.
6802 (adddi3_neon): Likewise.
6803 (sub<mode>3_neon): Likewise.
6804 (subdi3_neon): Likewise.
6805 (mul<mode>3_neon): Likewise.
6806 (mul<mode>3add<mode>_neon): Likewise.
6807 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6808 (fma<VCVTF:mode>4)): Likewise.
6809 (fma<VCVTF:mode>4_intrinsic): Likewise.
6810 (fmsub<VCVTF:mode>4)): Likewise.
6811 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6812 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6813 (ior<mode>3): Likewise.
6814 (and<mode>3): Likewise.
6815 (anddi3_neon): Likewise.
6816 (orn<mode>3_neon): Likewise.
6817 (orndi3_neon): Likewise.
6818 (bic<mode>3_neon): Likewise.
6819 (bicdi3_neon): Likewise.
6820 (xor<mode>3): Likewise.
6821 (one_cmpl<mode>2): Likewise.
6822 (abs<mode>2): Likewise.
6823 (neg<mode>2): Likewise.
6824 (umin<mode>3_neon): Likewise.
6825 (umax<mode>3_neon): Likewise.
6826 (smin<mode>3_neon): Likewise.
6827 (smax<mode>3_neon): Likewise.
6828 (vashl<mode>3): Likewise.
6829 (vashr<mode>3_imm): Likewise.
6830 (vlshr<mode>3_imm): Likewise.
6831 (ashl<mode>3_signed): Likewise.
6832 (ashl<mode>3_unsigned): Likewise.
6833 (neon_load_count): Likewise.
6834 (ashldi3_neon_noclobber): Likewise.
6835 (signed_shift_di3_neon): Likewise.
6836 (unsigned_shift_di3_neon): Likewise.
6837 (ashrdi3_neon_imm_noclobber): Likewise.
6838 (lshrdi3_neon_imm_noclobber): Likewise.
6839 (widen_ssum<mode>3): Likewise.
6840 (widen_usum<mode>3): Likewise.
6841 (quad_halves_<code>v4si): Likewise.
6842 (quad_halves_<code>v4sf): Likewise.
6843 (quad_halves_<code>v8hi): Likewise.
6844 (quad_halves_<code>v16qi): Likewise.
6845 (reduc_splus_v2di): Likewise.
6846 (neon_vpadd_internal<mode>): Likewise.
6847 (neon_vpsmin<mode>): Likewise.
6848 (neon_vpsmax<mode>): Likewise.
6849 (neon_vpumin<mode>): Likewise.
6850 (neon_vpumax<mode>): Likewise.
6851 (ss_add<mode>_neon): Likewise.
6852 (us_add<mode>_neon): Likewise.
6853 (ss_sub<mode>_neon): Likewise.
6854 (us_sub<mode>_neon): Likewise.
6855 (neon_vadd<mode>_unspec): Likewise.
6856 (neon_vaddl<mode>): Likewise.
6857 (neon_vaddw<mode>): Likewise.
6858 (neon_vhadd<mode>): Likewise.
6859 (neon_vqadd<mode>): Likewise.
6860 (neon_vaddhn<mode>): Likewise.
6861 (neon_vmul<mode>): Likewise.
6862 (neon_vmla<mode>): Likewise.
6863 (neon_vmlal<mode>): Likewise.
6864 (neon_vmls<mode>): Likewise.
6865 (neon_vmlsl<mode>): Likewise.
6866 (neon_vqdmulh<mode>): Likewise.
6867 (neon_vqdmlal<mode>): Likewise.
6868 (neon_vqdmlsl<mode>): Likewise.
6869 (neon_vmull<mode>): Likewise.
6870 (neon_vqdmull<mode>): Likewise.
6871 (neon_vsub<mode>_unspec): Likewise.
6872 (neon_vsubl<mode>): Likewise.
6873 (neon_vsubw<mode>): Likewise.
6874 (neon_vqsub<mode>): Likewise.
6875 (neon_vhsub<mode>): Likewise.
6876 (neon_vsubhn<mode>): Likewise.
6877 (neon_vceq<mode>): Likewise.
6878 (neon_vcge<mode>): Likewise.
6879 (neon_vcgeu<mode>): Likewise.
6880 (neon_vcgt<mode>): Likewise.
6881 (neon_vcgtu<mode>): Likewise.
6882 (neon_vcle<mode>): Likewise.
6883 (neon_vclt<mode>): Likewise.
6884 (neon_vcage<mode>): Likewise.
6885 (neon_vcagt<mode>): Likewise.
6886 (neon_vtst<mode>): Likewise.
6887 (neon_vabd<mode>): Likewise.
6888 (neon_vabdl<mode>): Likewise.
6889 (neon_vaba<mode>): Likewise.
6890 (neon_vabal<mode>): Likewise.
6891 (neon_vmax<mode>): Likewise.
6892 (neon_vmin<mode>): Likewise.
6893 (neon_vpaddl<mode>): Likewise.
6894 (neon_vpadal<mode>): Likewise.
6895 (neon_vpmax<mode>): Likewise.
6896 (neon_vpmin<mode>): Likewise.
6897 (neon_vrecps<mode>): Likewise.
6898 (neon_vrsqrts<mode>): Likewise.
6899 (neon_vqabs<mode>): Likewise.
6900 (neon_vqneg<mode>): Likewise.
6901 (neon_vcls<mode>): Likewise.
6902 (clz<mode>2): Likewise.
6903 (popcount<mode>2): Likewise.
6904 (neon_vrecpe): Likewise.
6905 (neon_vrsqrte): Likewise.
6906 (neon_vget_lane<mode>_sext_internal): Likewise.
6907 (neon_vget_lane<mode>_zext_internal): Likewise.
6908 (neon_vdup_n<mode>): Likewise.
6909 (neon_vdup_nv2di): Likewise.
6910 (neon_vdpu_lane<mode>_internal): Likewise.
6911 (neon_vswp<mode>): Likewise.
6912 (float<mode><V_cvtto>2): Likewise.
6913 (floatuns<mode><V_cvtto>2): Likewise.
6914 (fix_trunc<mode><V_cvtto>)2): Likewise
6915 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6916 (neon_vcvt<mode>): Likewise.
6917 (neon_vcvtv4sfv4hf): Likewise.
6918 (neon_vcvtv4hfv4sf): Likewise.
6919 (neon_vcvt_n<mode>): Likewise.
6920 (neon_vmovn<mode>): Likewise.
6921 (neon_vqmovn<mode>): Likewise.
6922 (neon_vqmovun<mode>): Likewise.
6923 (neon_vmovl<mode>): Likewise.
6924 (neon_vmul_lane<mode>): Likewise.
6925 (neon_vmull_lane<mode>): Likewise.
6926 (neon_vqdmull_lane<mode>): Likewise.
6927 (neon_vqdmulh_lane<mode>): Likewise.
6928 (neon_vmla_lane<mode>): Likewise.
6929 (neon_vmlal_lane<mode>): Likewise.
6930 (neon_vqdmlal_lane<mode>): Likewise.
6931 (neon_vmls_lane<mode>): Likewise.
6932 (neon_vmlsl_lane<mode>): Likewise.
6933 (neon_vqdmlsl_lane<mode>): Likewise.
6934 (neon_vext<mode>): Likewise.
6935 (neon_vrev64<mode>): Likewise.
6936 (neon_vrev32<mode>): Likewise.
6937 (neon_vrev16<mode>): Likewise.
6938 (neon_vbsl<mode>_internal): Likewise.
6939 (neon_vshl<mode>): Likewise.
6940 (neon_vqshl<mode>): Likewise.
6941 (neon_vshr_n<mode>): Likewise.
6942 (neon_vshrn_n<mode>): Likewise.
6943 (neon_vqshrn_n<mode>): Likewise.
6944 (neon_vqshrun_n<mode>): Likewise.
6945 (neon_vshl_n<mode>): Likewise.
6946 (neon_vqshl_n<mode>): Likewise.
6947 (neon_vqshlu_n<mode>): Likewise.
6948 (neon_vshll_n<mode>): Likewise.
6949 (neon_vsra_n<mode>): Likewise.
6950 (neon_vsri_n<mode>): Likewise.
6951 (neon_vsli_n<mode>): Likewise.
6952 (neon_vtbl1v8qi): Likewise.
6953 (neon_vtbl2v8qi): Likewise.
6954 (neon_vtbl3v8qi): Likewise.
6955 (neon_vtbl4v8qi): Likewise.
6956 (neon_vtbx1v8qi): Likewise.
6957 (neon_vtbx2v8qi): Likewise.
6958 (neon_vtbx3v8qi): Likewise.
6959 (neon_vtbx4v8qi): Likewise.
6960 (neon_vtrn<mode>_internal): Likewise.
6961 (neon_vzip<mode>_internal): Likewise.
6962 (neon_vuzp<mode>_internal): Likewise.
6963 (neon_vld1<mode>): Likewise.
6964 (neon_vld1_lane<mode>): Likewise.
6965 (neon_vld1_dup<mode>): Likewise.
6966 (neon_vld1_dupv2di): Likewise.
6967 (neon_vst1<mode>): Likewise.
6968 (neon_vst1_lane<mode>): Likewise.
6969 (neon_vld2<mode>): Likewise.
6970 (neon_vld2_lane<mode>): Likewise.
6971 (neon_vld2_dup<mode>): Likewise.
6972 (neon_vst2<mode>): Likewise.
6973 (neon_vst2_lane<mode>): Likewise.
6974 (neon_vld3<mode>): Likewise.
6975 (neon_vld3qa<mode>): Likewise.
6976 (neon_vld3qb<mode>): Likewise.
6977 (neon_vld3_lane<mode>): Likewise.
6978 (neon_vld3_dup<mode>): Likewise.
6979 (neon_vst3<mode>): Likewise.
6980 (neon_vst3qa<mode>): Likewise.
6981 (neon_vst3qb<mode>): Likewise.
6982 (neon_vst3_lane<mode>): Likewise.
6983 (neon_vld4<mode>): Likewise.
6984 (neon_vld4qa<mode>): Likewise.
6985 (neon_vld4qb<mode>): Likewise.
6986 (neon_vld4_lane<mode>): Likewise.
6987 (neon_vld4_dup<mode>): Likewise.
6988 (neon_vst4<mode>): Likewise.
6989 (neon_vst4qa<mode>): Likewise.
6990 (neon_vst4qb<mode>): Likewise.
6991 (neon_vst4_lane<mode>): Likewise.
6992 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6993 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6994 (neon_vec_<US>mult_lo_<mode>): Likewise.
6995 (neon_vec_<US>mult_hi_<mode>): Likewise.
6996 (neon_vec_<US>shiftl_<mode>): Likewise.
6997 (neon_unpack<US>_<mode>): Likewise.
6998 (neon_vec_<US>mult_<mode>): Likewise.
6999 (vec_pack_trunc_<mode>): Likewise.
7000 (neon_vec_pack_trunk_<mode>): Likewise.
7001 (neon_vabd<mode>_2): Likewise.
7002 (neon_vabd<mode>_3): Likewise.
7003 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7004 (thumb2_movsi_vfp): Likewise.
7005 (movdi_vfp): Likewise.
7006 (movdi_vfp_cortexa8): Likewise.
7007 (movhf_vfp_neon): Likewise.
7008 (movhf_vfp): Likewiwse.
7009 (movsf_vfp): Likewiwse.
7010 (thumb2_movsf_vfp): Likewiwse.
7011 (movdf_vfp): Likewise.
7012 (thumb2_movdf_vfp): Likewise.
7013 (movsfcc_vfp): Likewise.
7014 (thumb2_movsfcc_vfp): Likewise.
7015 (movdfcc_vfp): Likewise.
7016 (thumb2_movdfcc_vfp): Likewise.
7017 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7018 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7019 (v10_v2c): Likewise.
7020 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7022 (cortex_a15_neon_int_2): Likewise.
7023 (cortex_a15_neon_int_3): Likewise.
7024 (cortex_a15_neon_int_4): Likewise.
7025 (cortex_a15_neon_int_5): Likewise.
7026 (cortex_a15_neon_vqneg_vqabs): Likewise.
7027 (cortex_a15_neon_vmov): Likewise.
7028 (cortex_a15_neon_vaba): Likewise.
7029 (cortex_a15_neon_vaba_qqq): Likewise.
7030 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7031 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7032 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7033 scalar_64_32_long_scalar): Likewise.
7034 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7035 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7036 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7037 lotype_qdd_64_32_long): Likewise.
7038 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7039 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7040 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7041 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7042 (cortex_a15_neon_shift_1): Likewise.
7043 (cortex_a15_neon_shift_2): Likewise.
7044 (cortex_a15_neon_shift_3): Likewise.
7045 (cortex_a15_neon_vshl_ddd): Likewise.
7046 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7047 (cortex_a15_neon_vsra_vrsra): Likewise.
7048 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7049 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7050 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7051 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7052 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7053 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7054 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7055 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7056 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7057 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7058 (cortex_a15_neon_bp_simple): Likewise.
7059 (cortex_a15_neon_bp_2cycle): Likewise.
7060 (cortex_a15_neon_bp_3cycle): Likewise.
7061 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7062 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7063 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7064 (cortex_a15_neon_vld2_4_regs): Likewise.
7065 (cortex_a15_neon_vld3_vld4): Likewise.
7066 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7067 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7068 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7069 (cortex_a15_neon_vst3_vst4): Likewise.
7070 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7071 (cortex_a15_neon_vld3_vld4_lane" 10
7072 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7073 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7074 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7075 (cortex_a15_neon_ldm_2): Likewise.0
7076 (cortex_a15_neon_stm_2): Likewise.
7077 (cortex_a15_neon_mcr): Likewise.
7078 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7079 (cortex_a15_neon_mrc): Likewise.
7080 (cortex_a15_neon_mrrc): Likewise.
7081 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7083 (cortex_a15_alu_shift): Likewise.
7084 (cortex_a15_alu_shift_reg): Likewise.
7085 (cortex_a15_mult32): Likewise.
7086 (cortex_a15_mult64): Likewise.
7087 (cortex_a15_block): Likewise.
7088 (cortex_a15_branch): Likewise.
7089 (cortex_a15_load1): Likewise.
7090 (cortex_a15_load3): Likewise.
7091 (cortex_a15_store1): Likewise.
7092 (cortex_a15_store3): Likewise.
7093 (cortex_a15_call): Likewise.
7094 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7095 (cortex_a5_f2r): Likewise.
7096 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7098 (cortex_a53_f2r): Likewise.
7099 * config/arm/cortex-a7.md
7100 (cortex_a7_branch): Update for attribute change.
7101 (cortex_a7_call): Likewise.
7102 (cortex_a7_alu_imm): Likewise.
7103 (cortex_a7_alu_reg): Likewise.
7104 (cortex_a7_alu_shift): Likewise.
7105 (cortex_a7_mul): Likewise.
7106 (cortex_a7_load1): Likewise.
7107 (cortex_a7_store1): Likewise.
7108 (cortex_a7_load2): Likewise.
7109 (cortex_a7_store2): Likewise.
7110 (cortex_a7_load3): Likewise.
7111 (cortex_a7_store3): Likewise.
7112 (cortex_a7_load4): Likewise.
7113 (cortex_a7_store4): Likewise.
7114 (cortex_a7_fpalu): Likewise.
7115 (cortex_a7_fconst): Likewise.
7116 (cortex_a7_fpmuls): Likewise.
7117 (cortex_a7_neon_mul): Likewise.
7118 (cortex_a7_fpmacs): Likewise.
7119 (cortex_a7_neon_mla: Likewise.
7120 (cortex_a7_fpmuld: Likewise.
7121 (cortex_a7_fpmacd: Likewise.
7122 (cortex_a7_fpfmad: Likewise.
7123 (cortex_a7_fdivs: Likewise.
7124 (cortex_a7_fdivd: Likewise.
7125 (cortex_a7_r2f: Likewise.
7126 (cortex_a7_f2r: Likewise.
7127 (cortex_a7_f_flags: Likewise.
7128 (cortex_a7_f_loads: Likewise.
7129 (cortex_a7_f_loadd: Likewise.
7130 (cortex_a7_f_stores: Likewise.
7131 (cortex_a7_f_stored: Likewise.
7132 (cortex_a7_neon): Likewise.
7133 * config/arm/cortex-a8-neon.md
7134 (cortex_a8_neon_mrc): Update for attribute change.
7135 (cortex_a8_neon_mrrc): Likewise.
7136 (cortex_a8_neon_int_1): Likewise.
7137 (cortex_a8_neon_int_2): Likewise.
7138 (cortex_a8_neon_int_3): Likewise.
7139 (cortex_a8_neon_int_4): Likewise.
7140 (cortex_a8_neon_int_5): Likewise.
7141 (cortex_a8_neon_vqneg_vqabs): Likewise.
7142 (cortex_a8_neon_vmov): Likewise.
7143 (cortex_a8_neon_vaba): Likewise.
7144 (cortex_a8_neon_vaba_qqq): Likewise.
7145 (cortex_a8_neon_vsma): Likewise.
7146 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7147 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7148 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7149 long_scalar): Likewise.
7150 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7151 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7152 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7153 long_scalar_qdd_64_32_long): Likewise.
7154 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7155 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7156 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7157 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7158 (cortex_a8_neon_shift_1): Likewise.
7159 (cortex_a8_neon_shift_2): Likewise.
7160 (cortex_a8_neon_shift_3): Likewise.
7161 (cortex_a8_neon_vshl_ddd): Likewise.
7162 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7163 (cortex_a8_neon_vsra_vrsra): Likewise.
7164 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7165 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7166 (cortex_a8_neon_fp_vsum): Likewise.
7167 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7168 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7169 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7170 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7171 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7172 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7173 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7174 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7175 (cortex_a8_neon_bp_simple): Likewise.
7176 (cortex_a8_neon_bp_2cycle): Likewise.
7177 (cortex_a8_neon_bp_3cycle): Likewise.
7178 (cortex_a8_neon_ldr): Likewise.
7179 (cortex_a8_neon_str): Likewise.
7180 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7181 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7182 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7183 (cortex_a8_neon_vld2_4_regs): Likewise.
7184 (cortex_a8_neon_vld3_vld4): Likewise.
7185 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7186 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7187 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7188 (cortex_a8_neon_vst3_vst4): Likewise.
7189 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7190 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7191 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7192 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7193 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7194 (cortex_a8_neon_mcr): Likewise.
7195 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7196 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7197 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7199 (ca9_neon_mrrc): Likewise.
7200 (cortex_a9_neon_int_1): Likewise.
7201 (cortex_a9_neon_int_2): Likewise.
7202 (cortex_a9_neon_int_3): Likewise.
7203 (cortex_a9_neon_int_4): Likewise.
7204 (cortex_a9_neon_int_5): Likewise.
7205 (cortex_a9_neon_vqneg_vqabs): Likewise.
7206 (cortex_a9_neon_vmov): Likewise.
7207 (cortex_a9_neon_vaba): Likewise.
7208 (cortex_a9_neon_vaba_qqq): Likewise.
7209 (cortex_a9_neon_vsma): Likewise.
7210 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7211 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7212 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7213 long_scalar): Likewise.
7214 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7215 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7216 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7217 long_scalar_qdd_64_32_long): Likewise.
7218 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7219 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7220 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7221 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7222 (cortex_a9_neon_shift_1): Likewise.
7223 (cortex_a9_neon_shift_2): Likewise.
7224 (cortex_a9_neon_shift_3): Likewise.
7225 (cortex_a9_neon_vshl_ddd): Likewise.
7226 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7227 (cortex_a9_neon_vsra_vrsra): Likewise.
7228 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7229 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7230 (cortex_a9_neon_fp_vsum): Likewise.
7231 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7232 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7233 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7234 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7235 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7236 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7237 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7238 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7239 (cortex_a9_neon_bp_simple): Likewise.
7240 (cortex_a9_neon_bp_2cycle): Likewise.
7241 (cortex_a9_neon_bp_3cycle): Likewise.
7242 (cortex_a9_neon_ldr): Likewise.
7243 (cortex_a9_neon_str): Likewise.
7244 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7245 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7246 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7247 (cortex_a9_neon_vld2_4_regs): Likewise.
7248 (cortex_a9_neon_vld3_vld4): Likewise.
7249 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7250 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7251 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7252 (cortex_a9_neon_vst3_vst4): Likewise.
7253 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7254 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7255 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7256 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7257 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7258 (cortex_a9_neon_mcr): Likewise.
7259 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7260 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7261 (cortex_a9_fps): Likewise.
7262 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7264 (cortex_m4_fmuls): Likewise.
7265 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7267 (cortex_r4_mrc): Likewise.
7268 * config/arm/iterators.md: Update comment referring to neon_type.
7269 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7270 (iwmmxt_movsi_insn): Likewise.
7271 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7273 (pj4_core_to_vfp): Likewise.
7274 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7276 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7277 (vfp_fstore): Likewise.
7278 * doc/md.texi: Change references to neon_type to refer to type.
7280 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7282 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7284 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7286 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7289 2013-09-04 Easwaran Raman <eraman@google.com>
7292 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7293 (build_and_add_sum): Use it.
7294 (appears_later_in_bb): Simplify code.
7296 2013-09-04 Teresa Johnson <tejohnson@google.com>
7298 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7300 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7302 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7304 2013-09-04 Jan Hubicka <jh@suse.cz>
7306 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7307 * ipa-devirt.c: Include diganostic.h
7308 (odr_type_d): Add types and types_set.
7309 (hash_type_name): Work for types with vtables during LTO.
7310 (odr_hasher::remove): Fix comment; destroy types_set.
7311 (add_type_duplicate): New function,
7312 (get_odr_type): Use it.
7313 (dump_type_inheritance_graph): Dump type duplicates.
7314 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7316 * tree.c (types_same_for_odr): Give exact answers on types with
7319 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7321 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7322 explaining their differences.
7324 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7326 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7328 2013-09-03 Jeff Law <law@redhat.com>
7330 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7331 when not threading through a joiner block. Pass joiner/no joiner
7332 state to register_jump_thread.
7333 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7334 state from argument rather than implying on path length.
7335 Dump the entire jump thread path into debugging dump.
7336 * tree-flow.h (register_jump_thread): Update prototype.
7338 2013-08-29 Xinliang David Li <davidxl@google.com>
7340 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7341 Remove a trivial gcc_assert.
7343 2013-08-29 Xinliang David Li <davidxl@google.com>
7345 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7346 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7347 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7348 Delay base decl alignment adjustment.
7349 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7350 * tree-vectorizer.h: New data structure.
7351 (set_dr_misalignment): New function.
7352 (dr_misalignment): Ditto.
7353 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7354 (vectorizable_load): Ditto.
7355 (ensure_base_align): New function.
7356 (vectorize_loops): Add dbg_cnt support.
7357 (execute_vect_slp): Ditto.
7358 * dbgcnt.def: New debug counter.
7359 * Makefile: New dependency.
7361 2013-09-03 Meador Inge <meadori@codesourcery.com>
7365 2013-08-30 Meador Inge <meadori@codesourcery.com>
7367 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7369 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7371 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7372 function descriptor.
7374 2013-09-03 Richard Biener <rguenther@suse.de>
7376 * tree-affine.c (add_elt_to_tree): Fix association issue,
7377 avoid useless converts and make sure to always return a
7378 properly typed result.
7380 2013-09-03 Richard Biener <rguenther@suse.de>
7383 * fold-const.c (negate_expr_p): Fix division case.
7384 (negate_expr): Likewise.
7386 2013-09-03 Richard Biener <rguenther@suse.de>
7389 * tree-streamer-out.c: Include tm.h.
7390 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7392 2013-09-03 Jan Hubicka <jh@suse.cz>
7394 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7396 2013-09-03 Alan Modra <amodra@gmail.com>
7398 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7399 multilibs to big-endian target and vice versa.
7400 * config/rs6000/t-linux64: Use := assignment on all vars.
7401 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7402 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7403 * config/rs6000/t-linux64le: New file.
7404 * config/rs6000/t-linux64bele: New file.
7405 * config/rs6000/t-linux64lebe: New file.
7407 2013-09-02 Jan Hubicka <jh@suse.cz>
7409 * ipa-inline-transform.c (inline_transform): Do not
7410 optimize_inline_calls when not optimizing.
7412 2013-09-02 Jan Hubicka <jh@suse.cz>
7414 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7415 duplicated nodes for assembler names.
7416 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7417 hard registers from assembler name hash.
7419 2013-09-02 Jan Hubicka <jh@suse.cz>
7421 * ipa-split.c (execute_split_functions): Split externally visible
7422 functions called once.
7424 2013-09-02 Martin Jambor <mjambor@suse.cz>
7427 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7428 linked list. When finding the correct duplicate, also consider also
7429 the caller in additon to its inlined_to node.
7431 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7433 * config/aarch64/aarch64-simd-builtins.def
7434 (dup_lane_scalar): Remove.
7435 * config/aarch64/aarch64-simd.md
7436 (aarch64_simd_dup): Add 'w->w' alternative.
7437 (aarch64_dup_lane<mode>): Allow for VALL.
7438 (aarch64_dup_lane_scalar<mode>): Remove.
7439 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7440 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7441 (aarch64_get_lane_unsigned<mode>): Likewise.
7442 (aarch64_get_lane<mode>): Likewise.
7443 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7444 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7445 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7446 (VCON): Change container of V2SF.
7447 (vswap_width_name): Likewise.
7448 * config/aarch64/arm_neon.h
7449 (__aarch64_vdup_lane_any): New.
7450 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7451 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7452 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7454 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7457 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7458 the source or the destination is a single hard register.
7460 2013-09-02 Richard Biener <rguenther@suse.de>
7463 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7466 2013-09-02 Richard Biener <rguenther@suse.de>
7468 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7469 arithmetic to sizetype.
7471 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7473 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7474 Find auto-increment use both before and after candidate.
7476 2013-09-02 Marek Polacek <polacek@redhat.com>
7478 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7480 2013-09-01 Jan Hubicka <jh@suse.cz>
7482 * Makefile.in: Add ipa-profile.o
7483 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7484 * cgraph.c (struct cgraph_propagate_frequency_data,
7485 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7486 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7487 * cgraph.h (cgraph_propagate_frequency): Remove.
7488 * ipa-inline-analysis.c: Include ipa-utils.h;
7489 drop duplicated cfgloop.h.
7490 (inline_update_callee_summaries): Update.
7491 * ipa-profile.c: New file.
7492 * ipa-utils.h (ipa_propagate_frequency): Declare.
7493 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7494 data-streamer.h, value-prof.h.
7495 (symtab_remove_unreachable_nodes): Update profile.
7496 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7497 account_time_size, cmp_counts, dump_histogram,
7498 ipa_profile_generate_summary, ipa_profile_write_summary,
7499 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7500 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7501 Move to ipa-profile.c.
7503 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7505 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7507 2013-09-01 Jan Hubicka <jh@suse.cz>
7509 * common.opt (fdevirtualize-speculatively): New function.
7510 * invoke.texi (fdevirtualize-speculatively): Document.
7511 * ipa-devirt.c: Include ipa-inline.h
7512 (likely_target_p): New function.
7513 (ipa_devirt): New function.
7514 (gate_ipa_devirt): New function.
7515 (pass_data_ipa_devirt): New static var.
7516 (pass_ipa_devirt): Likewise.
7517 (make_pass_ipa_devirt): New function.
7518 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7519 (common_handle_option): Disable devirtualization when
7520 value range profiling is available.
7521 * passes.def (pass_ipa_devirt): Add.
7522 * timever.def (TV_IPA_DEVIRT): New timevar.
7523 * tree-pass.h (make_pass_ipa_devirt):
7525 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7527 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7528 include sanitize(undefined).
7530 2013-08-31 Diego Novillo <dnovillo@google.com>
7532 * Makefile.in (TREE_CORE_H): Define.
7534 (GTFILES): Add tree-core.h.
7535 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7537 * tree-core.h: New file.
7538 Move all data structures, enum, typedefs, global
7539 declarations and constants from ...
7542 2013-08-31 Jan Hubicka <jh@suse.cz>
7544 * bulitins.c (expand_builtin): Do not early exit for gcov
7545 instrumented functions.
7547 2013-08-31 Marek Polacek <polacek@redhat.com>
7549 * ubsan.c: Include tm_p.h.
7551 2013-08-31 Jan Hubicka <jh@suse.cz>
7553 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7556 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7557 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7558 * tree-cfg.c (verify_gimple_label): ... here.
7559 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7560 (ipa_merge_profiles): New function.
7561 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7562 (lto_input_function_body): Likewise.
7563 * ipa-utils.h (ipa_merge_profiles): Declare.
7564 * lto-streamer.h (lto_input_function_body): Update prototype.
7565 (emit_label_in_global_context_p): Remove.
7566 * lto-symtab.c: Include ipa-utils.h
7567 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7569 2013-08-31 Jan Hubicka <jh@suse.cz>
7571 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7573 2013-08-31 Jan Hubicka <jh@suse.cz>
7575 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7577 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7579 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7580 "cmp" RTX before signed_comparison_operator check to account
7583 2013-08-30 Jan Hubicka <jh@suse.cz>
7585 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7586 (detect_type_change_1): Rename to ...
7587 (detect_type_change): ... this one; early return on non-polymorphic
7589 (detect_type_change_ssa): Add comp_type parameter; update
7590 use of detect_type_change.
7591 (compute_complex_assign_jump_func): Add param_type parameter;
7592 update use of detect_type_change_ssa.
7593 (compute_complex_ancestor_jump_func): Likewise.
7594 (ipa_get_callee_param_type): New function.
7595 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7596 update calls to the jump function computation functions.
7598 2013-08-30 Teresa Johnson <tejohnson@google.com>
7599 Steven Bosscher <steven@gcc.gnu.org>
7601 * cfgrtl.c (fixup_new_cold_bb): New routine.
7602 (commit_edge_insertions): Invoke fixup_partitions.
7603 (find_partition_fixes): New routine.
7604 (fixup_partitions): Ditto.
7605 (verify_hot_cold_block_grouping): Update comments.
7606 (rtl_verify_edges): Invoke find_partition_fixes.
7607 (rtl_verify_bb_pointers): Update comments.
7608 (rtl_verify_bb_layout): Ditto.
7609 * basic-block.h (probably_never_executed_edge_p): Declare.
7610 (fixup_partitions): Ditto.
7611 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7612 * bb-reorder.c (sanitize_hot_paths): New function.
7613 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7615 * predict.c (probably_never_executed_edge_p): New routine.
7616 * cfg.c (check_bb_profile): Add partition insanity warnings.
7618 2013-08-30 Meador Inge <meadori@codesourcery.com>
7620 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7622 2013-08-30 Marek Polacek <polacek@redhat.com>
7624 * Makefile.in (ubsan.o): Add.
7625 (c-family/c-ubsan.o): Add.
7626 (builtins.o): Add ubsan.h dependency.
7627 * ubsan.h: New file.
7628 * ubsan.c: New file.
7629 * common.opt: Add -fsanitize=undefined option.
7630 (flag_sanitize): Add variable.
7631 (fsanitize=): Add option. Add Driver.
7632 (fsanitize=thread): Remove option.
7633 (fsanitize=address): Likewise.
7634 (static-libubsan): New option.
7635 * doc/invoke.texi: Document the new flag and -static-libubsan.
7636 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7637 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7638 * builtin-attrs.def (ATTR_COLD): Define.
7639 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7640 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7641 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7642 * flag-types.h (sanitize_code): New enum.
7643 * opts.c (common_handle_option): Parse command line arguments
7644 of -fsanitize=. Add -fsanitize=unreachable option.
7645 * varasm.c (get_variable_section): Adjust.
7646 (assemble_noswitch_variable): Likewise.
7647 (assemble_variable): Likewise.
7648 (output_constant_def_contents): Likewise.
7649 (categorize_decl_for_section): Likewise.
7650 (place_block_symbol): Likewise.
7651 (output_object_block): Likewise.
7652 * builtins.def: Likewise.
7653 * toplev.c (compile_file): Likewise.
7654 (process_options): Likewise.
7655 * cppbuiltin.c: Likewise.
7656 * tsan.c (tsan_pass): Likewise.
7657 (tsan_gate): Likewise.
7658 (tsan_gate_O0): Likewise.
7659 * cfgexpand.c (partition_stack_vars): Likewise.
7660 (expand_stack_vars): Likewise.
7661 (defer_stack_allocation): Likewise.
7662 (expand_used_vars): Likewise.
7663 * cfgcleanup.c (old_insns_match_p): Likewise.
7664 * asan.c (asan_finish_file): Likewise.
7665 (asan_instrument): Likewise.
7666 (gate_asan): Likewise.
7667 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7668 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7669 (asan_global_struct): Use pointer_sized_int_node instead
7670 calling build_nonstandard_integer_type.
7671 (initialize_sanitizer_builtins): Likewise.
7672 (asan_finish_file): Likewise.
7673 * gcc.c: Document %{%:function(args):X}.
7674 (static_spec_functions): Add sanitize.
7675 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7676 store funcval != NULL there.
7677 (do_spec_1): Adjust handle_spec_function caller.
7678 (handle_braces): Allow %:function(args) as condition.
7679 (sanitize_spec_function): New function.
7680 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7681 (LIBUBSAN_SPEC): Likewise.
7682 (LIBUBSAN_EARLY_SPEC): Likewise.
7683 (SANITIZER_SPEC): Handle libubsan.
7684 (SANITIZER_EARLY_SPEC): Likewise.
7685 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7686 instead of fsanitize=address.
7687 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7688 instead of fsanitize=address*.
7689 * builtins.c: Include ubsan.h.
7690 (fold_builtin_0): Instrument __builtin_unreachable.
7691 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7692 instead of flag_asan.
7693 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7694 (pointer_sized_int_node): Define.
7695 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7697 2013-08-30 Mike Stump <mikestump@comcast.net>
7699 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7702 2013-08-29 Jan Hubicka <jh@suse.cz>
7704 * cgraph.c (cgraph_function_body_availability): Handle weakref
7706 * passes.def: Remove pass_fixup_cfg.
7707 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7708 track when we need to remove functions.
7709 (gate_ipa_inline): Execute inlining always; add comment why.
7710 (pass_data_ipa_inline): Remove TODO_remove_functions.
7711 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7712 do not produce summaries.
7713 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7714 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7715 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7716 constant pool and vtable.
7718 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7720 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7721 New arm_neon.h's internal macros to specify 64-bit constants.
7722 Avoid using stdint.h's macros.
7724 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7726 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7727 before applying REG_P.
7729 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7731 PR tree-optimization/58277
7732 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7733 after seeing too many stmts with vdef in between dombb and current
7734 bb, invalidate everything.
7736 2013-08-30 Richard Biener <rguenther@suse.de>
7738 * fold-const.c (fold_single_bit_test): Fix overflow test.
7740 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7742 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7743 and which can live in a register, always retrieve the value on entry.
7744 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7745 passed by invisible reference specially.
7746 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7747 (vt_add_function_parameter): Correctly deal with a parameter passed by
7748 invisible reference.
7750 2013-08-30 Jan Hubicka <jh@suse.cz>
7752 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7754 2013-08-30 Richard Biener <rguenther@suse.de>
7756 PR tree-optimization/58228
7757 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7758 allow invariant loads in nested loop vectorization.
7760 2013-08-30 Richard Biener <rguenther@suse.de>
7762 PR tree-optimization/58223
7763 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7764 (has_anti_or_output_dependence): ... this and adjust to also
7765 look for output dependences.
7766 (mark_nodes_having_upstream_mem_writes): Adjust.
7767 (rdg_flag_uses): Likewise.
7769 2013-08-30 Richard Biener <rguenther@suse.de>
7771 PR tree-optimization/58010
7772 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7773 assert that we have a loop-closed PHI.
7775 2013-08-29 Jan Hubicka <jh@suse.cz>
7777 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7778 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7779 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7780 (lto_free_function_in_decl_state_for_node): New function.
7782 2013-08-29 Xinliang David Li <davidxl@google.com>
7784 * loop-unroll.c (report_unroll_peel): Minor message change.
7785 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7786 Emit alignment peeling message with default -fopt-info.
7787 (vect_loop_versioning): Emit loop version info message.
7788 * tree-vectorizer.c (vectorize_loops): Minor message change.
7789 (execute_vect_slp): Ditto.
7791 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7793 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7794 of the clone from the DECL_NAME of the original function.
7796 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7798 * passes.c (register_pass): Add overload.
7799 * tree-pass.h (register_pass): Forward declare it. Add comment.
7801 2013-08-29 Jan Hubicka <jh@suse.cz>
7803 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7804 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7805 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7806 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7807 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7808 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7809 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7810 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7812 2013-08-29 Teresa Johnson <tejohnson@google.com>
7814 * dumpfile.c (dump_loc): Output column number.
7815 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7816 * doc/invoke.texi: Document optall -fopt-info flag.
7817 * profile.c (read_profile_edge_counts): Use new dump framework.
7818 (compute_branch_probabilities): Ditto.
7819 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7820 when pass not in any opt group.
7821 * pass_manager.h (pass_manager::get_pass_profile): New method.
7822 * value-prof.c (check_counter): Use new dump framework.
7823 (check_ic_target): Ditto.
7824 * coverage.c (get_coverage_counts): Ditto.
7825 (coverage_init): Setup new dump framework.
7827 2013-08-29 Richard Biener <rguenther@suse.de>
7829 PR tree-optimization/58246
7830 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7831 handle the dominance check inside a basic-block.
7833 2013-08-29 Richard Biener <rguenther@suse.de>
7836 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7837 of default defs that appear in abnormal PHI nodes.
7839 2013-08-29 Richard Biener <rguenther@suse.de>
7841 PR tree-optimization/57685
7842 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7843 single-use operands to avoid exponential complexity.
7845 2013-08-28 Dehao Chen <dehao@google.com>
7847 * ipa-inline.c (edge_badness): Fix integer underflow.
7849 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7851 * gtm-builtins.def (_ITM_free): Declare leaf.
7853 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7856 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7857 (*tls_local_dynamic_base_64_largepic): Likewise.
7858 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7859 Remove predicate from call operand.
7860 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7861 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7863 2013-08-28 Jeff Law <law@redhat.com>
7865 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7866 checks for the number of predecessors and successors allowed.
7867 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7868 which require copying a joiner block if there is a request which
7869 is a subpath that requires no joiner block copying.
7871 2013-08-28 Jan Hubicka <jh@suse.cz>
7873 * lto-streamer-out.c (DFS_write_tree_body): Drop
7874 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7875 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7876 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7877 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7878 Do not read DECL_ERROR_ISSUED.
7879 (unpack_ts_decl_with_vis_value_fields): Do not read
7881 (lto_input_ts_binfo_tree_pointers): Do not read
7882 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7883 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7884 write DECL_ERROR_ISSUED..
7885 (pack_ts_decl_with_vis_value_fields): Do not write
7887 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7888 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7889 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7890 * tree.h (tree_decl_common): Update comment.
7891 (DECL_ERROR_ISSUED): Remove.
7893 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7896 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7898 2013-08-28 Jan Hubicka <jh@suse.cz>
7900 * builtins.def (free): Declare leaf.
7902 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7905 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7906 * configure: Regenerate.
7908 2013-08-27 Martin Jambor <mjambor@suse.cz>
7910 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7911 (ipa_ancestor_jf_data): Likewise.
7912 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7913 (ipa_get_jf_pass_through_type_preserved): New function.
7914 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7915 (ipa_get_jf_ancestor_type_preserved): New function.
7916 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7917 (ipa_get_jf_ancestor_result): Likewise.
7918 (propagate_vals_accross_pass_through): Use
7919 ipa_get_jf_pass_through_result to do all the value mappings.
7920 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7921 type_preserved flag.
7922 (ipa_set_jf_cst_copy): New function.
7923 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7924 (ipa_set_jf_arith_pass_through): Likewise.
7925 (ipa_set_ancestor_jf): Likewise.
7926 (compute_complex_assign_jump_func): Set type_preserved instead of
7928 (ipa_compute_jump_functions_for_edge): Likewise.
7929 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7930 (update_jump_functions_after_inlining): Update type_preserved.
7931 Explicitely create jump functions when combining one with pass_through.
7932 (ipa_write_jump_function): Stream the type_preserved flags.
7933 (ipa_read_jump_function): Likewise.
7935 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7936 Aldy Hernandez <aldyh@redhat.com>
7938 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7939 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7940 * function.h (struct function): Add has_force_vect_loops and
7942 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7943 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7945 * gimple.def: Update CLAUSES comments.
7946 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7947 (gimple_build_omp_for): Add argument to prototype.
7948 (gimple_omp_for_kind): New.
7949 (gimple_omp_for_set_kind): New.
7950 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7951 GOVD_DATA_SHARE_CLASS.
7952 (enum omp_region_type): Add ORT_SIMD.
7953 (gimple_add_tmp_var): Handle ORT_SIMD.
7954 (gimplify_var_or_parm_decl): Same.
7955 (is_gimple_stmt): Same.
7956 (omp_firstprivatize_variable): Same.
7957 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7958 (omp_notice_variable): Handle ORT_SIMD.
7959 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7960 (omp_check_private): Handle ORT_SIMD.
7961 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7963 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7964 Handle OMP_CLAUSE_LASTPRIVATE.
7965 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7967 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7968 (gimplify_expr): Handle OMP_SIMD.
7969 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7970 (expand_GOMP_SIMD_VF): New.
7971 (expand_GOMP_SIMD_LAST_LANE): New.
7972 * internal-fn.def (GOMP_SIMD_LANE): New.
7973 (GOMP_SIMD_VF): New.
7974 (GOMP_SIMD_LAST_LANE): New.
7975 * omp-low.c: Include target.h.
7976 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7978 (check_omp_nesting_restrictions): Same.
7980 (lower_rec_simd_input_clauses): New.
7981 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7983 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7984 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7985 (expand_omp_build_assign): New.
7986 (expand_omp_for_init_counts): New.
7987 (expand_omp_for_init_vars): New.
7988 (extract_omp_for_update_vars): New.
7989 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7990 and rewrite accordingly.
7991 (expand_omp_simd): New.
7992 (expand_omp_for): Use expand_omp_simd.
7993 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7994 (lower_omp_for): Do not lower the body.
7995 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7997 * tree-flow.h (find_omp_clause): Remove prototype.
7998 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7999 forcing vectorization of the loop, or if flag_tree_vectorize.
8000 (gate_tree_if_conversion): Similarly.
8001 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8002 gimple_build_omp_for.
8003 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8004 * tree-parloops (create_parallel_loop): Pass kind argument to
8005 gimple_build_omp_for.
8006 * tree-pretty-print.c (dump_omp_clause): Add cases for
8007 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8008 OMP_CLAUSE__SIMDUID_.
8009 (dump_generic_node): Handle OMP_SIMD.
8010 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8011 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8012 unroll OMP_SIMD loops here.
8013 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8014 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8016 (vect_analyze_data_refs): Handle simd loops.
8017 * tree-vect-loop.c (vectorizable_live_operation): Handle
8019 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8020 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8021 (vectorizable_load): Same.
8022 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8023 (struct simduid_to_vf): New.
8024 (simduid_to_vf::hash): New.
8025 (simduid_to-vf::equal): New.
8026 (struct simd_array_to_simduid): New.
8027 (simd_array_to_simduid::hash): New.
8028 (simd_array_to_simduid::equal): New.
8029 (adjust_simduid_builtins): New.
8030 (struct note_simd_array_uses_struct): New.
8031 (note_simd_array_uses_cb): New.
8032 (note_simd_array_uses): New.
8033 (vectorize_loops): Handle simd hints and adjust simd builtins
8035 * tree-vectorizer.h (struct _stmt_vec_info): Add
8036 simd_lane_access_p field.
8037 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8038 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8039 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8040 (omp_clause_code_name): Same.
8041 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8042 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8043 * tree.def (OMP_SIMD): New entry.
8044 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8045 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8046 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8047 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8048 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8049 (OMP_CLAUSE_LINEAR_STEP): New.
8050 (OMP_CLAUSE_SAFELEN_EXPR): New.
8051 (OMP_CLAUSE__SIMDUID__DECL): New.
8052 (find_omp_clause): New prototype.
8054 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8056 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8057 Haswell processor detection.
8059 2013-08-27 Christian Widmer <shadow@umbrox.de>
8062 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8063 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8064 AVX2 capable processors.
8066 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8068 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8069 of vget_low_* with implementations in terms of other intrinsics.
8071 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8074 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8075 values to -1, 0 and 1.
8077 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8079 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8080 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8081 (<optab><mode>3_insn): Remove.
8082 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8083 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8084 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8085 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8086 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8087 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8088 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8089 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8090 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8091 (ror<mode>3_insn): Likewise.
8092 * config/aarch64/predicates.md (aarch64_simd_register): New.
8094 2013-08-27 Richard Biener <rguenther@suse.de>
8096 PR tree-optimization/57521
8097 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8098 one edge is non-critical.
8099 (find_phi_replacement_condition): Make sure to use a non-critical
8100 edge. Cleanup and remove old bug workarounds.
8101 (bb_postdominates_preds): Remove.
8102 (if_convertible_loop_p_1): Do not compute post-dominators.
8103 (combine_blocks): Do not free post-dominators.
8104 (main_tree_if_conversion): Likewise.
8105 (pass_data_if_conversion): Add TODO_verify_ssa.
8107 2013-08-27 DJ Delorie <dj@redhat.com>
8109 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8111 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8113 * function.c (assign_parm_find_data_types): Set passed_mode and
8114 nominal_mode to the TYPE_MODE of nominal_type for the built
8115 pointer type in case of the struct-pass-by-reference.
8117 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8119 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8120 (UINT16_TYPE): Change default to "unsigned int".
8122 * config/avr/avr.opt (mfract-convert-truncate): New option.
8123 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8124 is set, round negative fractional integers according to n1169
8125 when converting to integer types.
8127 2013-08-26 Jan Hubicka <jh@suse.cz>
8129 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8130 methods can not be called indirectly when their address is not taken.
8132 2013-08-26 Jan Hubicka <jh@suse.cz>
8134 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8137 2013-08-26 Jan Hubicka <jh@suse.cz>
8139 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8142 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8144 * reload.c (find_valid_class): Allow classes that do not include
8145 FIRST_PSEUDO_REGISTER - 1.
8147 2013-08-26 Jan Hubicka <jh@suse.cz>
8149 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8150 fix edge count/frequency when speculation failed; fix type check
8151 for the direct call.
8153 2013-08-26 Jan Hubicka <jh@suse.cz>
8155 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8157 2013-08-26 Jan Hubicka <jh@suse.cz>
8159 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8160 to be changed by edge redirection.
8162 2013-08-26 Jan Hubicka <jh@suse.cz>
8164 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8165 formating; add sanity check.
8166 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8167 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8168 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8169 (resolve_noninline_speculation): Update callee keys, too.
8171 2013-08-26 Jan Hubicka <jh@suse.cz>
8173 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8174 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8176 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8178 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8181 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8183 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8184 (x86_64_elf_select_section): Ditto.
8185 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8187 (ix86_pass_by_reference): Ditto.
8188 (output_set_got): Ditto.
8189 (ix86_unary_operator_ok): Ditto.
8190 (ix86_expand_builtin): Ditto.
8192 2013-08-23 Jan Hubicka <jh@suse.cz>
8194 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8196 2013-08-23 Jan Hubicka <jh@suse.cz>
8198 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8199 (tree_decl_with_vis): Add FINAL field.
8201 2013-08-23 Jeff Law <law@redhat.com>
8203 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8204 the debugging dump when the expression is fully redundant.
8206 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8208 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8209 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8210 * pretty-print.c (pp_formatted_text_data): Likewise.
8211 (pp_write_text_to_stream): Likewise.
8212 (pp_write_text_as_dot_label_to_stream): Likewise.
8213 (pp_append_r): Likewise.
8214 (pp_format): Likewise.
8215 (pp_flush): Likewise.
8216 (pp_clear_output_area): Likewise.
8217 (pp_append_text): Likewise.
8218 (pp_formatted_text): Likewise.
8219 (pp_remaining_character_count_for_line): Likewise.
8220 (pp_newline): Likewise.
8221 (pp_character): Likewise.
8222 (output_buffer::~output_buffer): Define.
8223 (pretty_printer::~pretty_printer): Destruct output buffer.
8224 * pretty-print.h (output_buffer::~output_buffer): Declare.
8225 (pretty_printer::~pretty_printer): Declare virtual.
8227 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8230 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8231 HOST_WIDE_INT_M1U): New macros.
8232 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8233 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8234 unsigned -1 for lshift.
8235 * cse.c (cse_insn): Likewise.
8236 * double-int.c (rshift_double, lshift_double): Likewise.
8237 * builtins.c (fold_builtin_bitop): Likewise.
8238 * combine.c (force_to_mode): Likewise.
8239 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8240 * simplify-rtx.c (simplify_const_unary_operation,
8241 simplify_const_binary_operation): Likewise.
8242 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8243 check_va_list_escapes): Likewise.
8244 * rtlanal.c (nonzero_bits1): Likewise.
8245 * expmed.c (expand_smod_pow2): Likewise.
8246 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8248 2013-08-23 Jan Hubicka <jh@suse.cz>
8250 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8251 as having address taken.
8253 2013-08-23 Jan Hubicka <jh@suse.cz>
8255 * ipa-utils.h (method_class_type): Declare.
8256 * ipa-devirt.c (method_class_type): Export.
8258 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8259 do not walk base classes of anonymous types.
8261 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8263 PR rtl-optimization/58220
8265 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8266 handle SEQUENCE insns properly.
8268 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8270 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8272 (pp_newline_and_indent): Likewise.
8273 (pp_separate_with): Likewise.
8274 * pretty-print.c (pp_newline_and_flush): Define.
8275 (pp_newline_and_indent): Likewise.
8276 (pp_separate_with): Likewise.
8278 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8281 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8282 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8284 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8286 * config/i386/predicates.md (ext_sse_reg_operand): New.
8287 * config/i386/i386.md (*movti_internal): Use
8288 predicate to determine if EVEX is needed.
8289 (*movsi_internal): Ditto.
8290 (*movdf_internal): Ditto.
8291 (*movsf_internal): Ditto.
8292 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8294 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8296 PR tree-optimization/58209
8297 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8298 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8299 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8300 emit POINTER_PLUS_EXPR.
8301 (create_tailcall_accumulator): For pointer result type accumulate in
8304 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8306 * configure.ac: Add backslashes missing from the last change.
8307 * configure: Regenerate.
8309 2013-08-22 Jan Hubicka <jh@susue.cz>
8311 * ipa.c (function_and_variable_visibility): First remember function
8312 was global and then make it local.
8314 2013-08-22 Julian Brown <julian@codesourcery.com>
8316 * configure.ac: Add aarch64 to list of arches which use "nop" in
8318 * configure: Regenerate.
8320 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8322 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8323 gnu_libc_has_function.
8324 * config/s390/tpf.h: Likewise.
8326 2013-08-22 Jan Hubicka <jh@susue.cz>
8328 * timevar.c (validate_phases): Add cast.
8330 2013-08-22 Jan Hubicka <jh@susue.cz>
8332 * timevar.c (validate_phases): Use size_t for memory.
8333 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8335 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8337 * pretty-print.h (output_buffer::output_buffer): Declare.
8338 (pretty_printer::pretty_printer): Likewise.
8339 (pp_construct): Remove.
8340 * pretty-print.c (output_buffer::output_buffer): Define.
8341 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8342 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8343 (print_gimple_expr): Likewise.
8344 (print_gimple_seq): Likewise.
8345 (gimple_dump_bb): Likewise.
8346 * sched-vis.c (dump_value_slim): Likewise.
8347 (dump_insn_slim): Likewise.
8348 (dump_rtl_slim): Likewise.
8349 (str_pattern_slim): Likewise.
8350 * tree-mudflap.c (mf_varname_tree): Likewise.
8351 * graph.c (print_graph_cfg): Likewise.
8352 (start_graph_dump): Likewise.
8353 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8355 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8356 pretty printer initialization.
8357 * coretypes.h (diagnostic_context): Remove superflous type alias
8359 (pretty_printer): Likewise. Declare directly as a class.
8360 (pretty_print_info): Remove declaration as class.
8361 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8362 and pp_clear_output_area.
8363 (asan_add_global): Likewise.
8365 2013-08-22 Jan Hubicka <jh@suse.cz>
8367 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8368 * ipa-utils.h (update_type_inheritance_graph): Declare.
8369 (possible_polymorphic_call_target_p): Declare.
8370 (possible_polymorphic_call_target_p): New.
8371 * ipa-devirt.c: Update toplevel comments.
8372 (cached_polymorphic_call_targets): Move up.
8373 (odr_type_d): Move ID down.
8374 (polymorphic_type_binfo_p): Update comment.
8375 (odr_hasher::remove): Likewise;
8376 (get_odr_type): Set anonymous_namespace.
8377 (dump_odr_type): Dump it.
8378 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8379 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8380 (record_binfo): Add comment.
8381 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8383 (devirt_node_removal_hook): Do not iCE when cache is freed.
8384 (possible_polymorphic_call_target_p): New predicate.
8385 (update_type_inheritance_graph): New function.
8387 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8388 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8389 Sergey Lega <sergey.s.lega@intel.com>
8390 Anna Tikhonova <anna.tikhonova@intel.com>
8391 Ilya Tocar <ilya.tocar@intel.com>
8392 Andrey Turetskiy <andrey.turetskiy@intel.com>
8393 Ilya Verbin <ilya.verbin@intel.com>
8394 Kirill Yukhin <kirill.yukhin@intel.com>
8395 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8397 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8398 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8399 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8400 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8401 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8402 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8403 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8404 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8405 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8406 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8407 OPT_mavx512pf, OPT_mavx512er cases.
8408 * config/i386/constraints.md (v): New constraint.
8409 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8410 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8411 (bit_AVX512CD): New.
8412 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8413 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8414 * config/i386/i386-c.c (ix86_target_macros_internal):
8415 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8417 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8418 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8419 * config/i386/i386.c (regclass_map, dbx_register_map)
8420 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8421 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8422 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8424 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8425 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8426 -mavx512pf options. Fix formatting.
8427 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8428 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8429 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8430 -mavx512cd, -mavx512pf options.
8431 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8432 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8433 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8435 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8436 (ix86_set_reg_reg_cost): Ditto.
8437 (x86_order_regs_for_local_alloc): Ditto.
8438 (MAX_VECT_LEN): Extend to 64-byte.
8439 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8440 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8441 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8442 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8443 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8444 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8445 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8446 (SSE_REG_MODE_P): Support new modes.
8447 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8448 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8449 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8450 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8451 (REG_CLASS_CONTENTS): Add new registers.
8452 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8453 (EXT_REX_SSE_REGNO_P): New.
8454 (HI_REGISTER_NAMES): Add new registers.
8455 * config/i386/i386.md: Define constants for new registers.
8456 (mode): Add new 512-bit modes.
8457 (prefix): Support evex prefix.
8458 (isa): Support avx512f, noavx512f, fma_avx512f.
8459 (ssemodesuffix): Add new 512-bit modes.
8461 (*movxi_internal_avx512f): Ditto.
8462 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8464 (*movsi_internal): Likewise.
8465 (*movdf_internal): Likewise.
8466 (*movsf_internal): Likewise.
8467 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8468 (<code><mode>3): Likewise.
8469 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8471 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8472 with the new constraint "v".
8473 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8475 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8476 with the new constraint "v".
8477 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8478 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8479 (*<plusminus_insn><mode>3): Likewise.
8480 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8481 (*mul<mode>3): Likewise.
8482 (<sse>_vmmul<mode>3): Likewise.
8483 (<sse>_div<mode>3): Likewise.
8484 (<sse>_vmdiv<mode>3): Likewise.
8485 (<sse>_sqrt<mode>2): Likewise.
8486 (<sse>_vmsqrt<mode>2): Likewise.
8487 (*<code><mode>3_finite): Likewise.
8488 (*<code><mode>3) <smaxmin>: Likewise.
8489 (<sse>_vm<code><mode>3): Likewise.
8490 (*<code><mode>3) <any_logic>: Likewise.
8491 (*fma_fmadd_<mode>): Likewise.
8492 (*fma_fmsub_<mode>): Likewise.
8493 (*fma_fnmadd_<mode>): Likewise.
8494 (*fma_fnmsub_<mode>): Likewise.
8495 (*fma_fmaddsub_<mode>): Likewise.
8496 (*fma_fmsubadd_<mode>): Likewise.
8497 (*fmai_fmadd_<mode>): Likewise.
8498 (*fmai_fmsub_<mode>): Likewise.
8499 (*fmai_fnmadd_<mode>): Likewise.
8500 (*fmai_fnmsub_<mode>): Likewise.
8501 (sse_cvtsi2ss): Likewise.
8502 (sse_cvtsi2ssq): Likewise.
8503 (sse_cvtss2si): Likewise.
8504 (sse_cvtss2si_2): Likewise.
8505 (sse_cvtss2siq): Likewise.
8506 (sse_cvtss2siq_2): Likewise.
8507 (sse_cvttss2si): Likewise.
8508 (sse_cvtss2siq_2): Likewise.
8509 (float<sseintvecmodelower><mode>2): Likewise.
8510 (sse2_cvtsd2si_2): Likewise.
8511 (sse2_cvtsd2siq_2): Likewise.
8512 (*<plusminus_insn><mode>3): Likewise.
8513 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8514 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8515 (ashr<mode>3): Likewise.
8516 (<shift_insn><mode>3): Likewise.
8517 (avx2_<code><mode>3): Likewise.
8518 (*avx2_<code><mode>3): Likewise.
8519 (*andnot<mode>3): Likewise.
8520 (*<code><mode>3) <any_logic>: Likewise.
8521 (abs<mode>2): Likewise.
8522 (avx2_permvar<mode>): Likewise.
8523 (avx2_perm<mode>_1): Likewise.
8524 (*avx_vpermilp<mode>): Likewise.
8525 (avx_vpermilvar<mode>3): Likewise.
8526 (avx2_ashrv<mode>): Likewise.
8527 (avx2_<shift_insn>v<mode>): Likewise.
8528 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8530 * doc/rtl.texi: Document XImode.
8532 2013-08-21 Jeff Law <law@redhat.com>
8534 * tree-flow.h (register_jump_thread): Pass vector of edges
8535 instead of each important edge.
8536 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8537 thread path into a vector and pass that to register_jump_thread.
8538 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8539 passed in edge vector to the current 3-edge form.
8542 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8544 * dce.c (fini_dce): Call df_analyze again just in case
8545 delete_unmarked_insns removed anything.
8547 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8549 * reload.h (struct reg_equivs): Rename to ..
8550 (struct reg_equivs_s): .. this.
8552 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8554 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8556 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8558 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8560 2013-08-21 Jeff Law <law@redhat.com>
8562 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8563 simplify assignments too. If the RHS collapses to a singleton
8564 range, then return the value for the range.
8566 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8568 * config/i386/sse.md (V16): Rename to...
8570 (mov<mode>): Update iterator name.
8571 (*mov<mode>_internal): Ditto.
8572 (push<mode>1): Ditto.
8573 (movmisalign<mode>): Ditto.
8575 2013-08-20 Jan Hubicka <jh@suse.cz>
8578 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8579 entry for direct edges.
8580 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8582 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8584 Revert my last two changes, r201865 and r201864:
8587 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8589 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8590 instances can own GC refs.
8592 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8593 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8594 (gcc::context::gt_pch_nx): Likewise.
8595 (gcc::context::gt_pch_nx): Likewise.
8596 * ggc.h (gt_ggc_mx <T>): New.
8597 (gt_pch_nx_with_op <T>): New.
8598 (gt_pch_nx <T>): New.
8599 * passes.c (opt_pass::gt_ggc_mx): New.
8600 (opt_pass::gt_pch_nx): New.
8601 (opt_pass::gt_pch_nx_with_op): New.
8602 (pass_manager::gt_ggc_mx): New.
8603 (pass_manager::gt_pch_nx): New.
8604 (pass_manager::gt_pch_nx_with_op): New.
8605 (pass_manager::operator new): Use
8606 ggc_internal_cleared_alloc_stat rather than xcalloc.
8607 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8608 (pass_manager::gt_ggc_mx): New.
8609 (pass_manager::gt_pch_nx): New.
8610 (pass_manager::gt_pch_nx_with_op): New.
8611 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8612 (opt_pass::operator new): New.
8613 (opt_pass::gt_ggc_mx): New.
8614 (opt_pass::gt_pch_nx): New.
8615 (opt_pass::gt_pch_nx_with_op): New.
8618 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8620 * Makefile.in (GTFILES): Add context.h.
8621 * context.c (gcc::context::operator new): New.
8622 (gcc::context::gt_ggc_mx): New.
8623 (gcc::context::gt_pch_nx): New.
8624 (gcc::context::gt_pch_nx): New.
8625 * context.h (gcc::context): Add GTY((user)) marking.
8626 (gcc::context::operator new): New.
8627 (gcc::context::gt_ggc_mx): New.
8628 (gcc::context::gt_pch_nx): New.
8629 (gcc::context::gt_pch_nx): New.
8630 (g): Add GTY marking.
8631 (gt_ggc_mx (gcc::context *)): New.
8632 (gt_pch_nx (gcc::context *)): New.
8633 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8634 void *cookie)): New.
8635 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8637 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8639 * dce.c (fini_dce): Call df_analyze again just in case
8640 delete_unmarked_insns removed anything.
8642 2013-08-20 Teresa Johnson <tejohnson@google.com>
8644 PR rtl-optimizations/57451
8645 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8646 from crossing split section boundaries.
8648 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8650 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8651 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8652 with MULTLIB_DEFAULTS.
8654 2013-08-20 Nick Clifton <nickc@redhat.com>
8656 * target.def (narrow_volatile_bitfield): Note that the default
8657 value is false, not !TARGET_STRICT_ALIGN.
8658 * doc/tm.texi: Regenerate.
8660 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8662 Fix LIB_SPEC for systems without libpthread.
8664 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8665 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8667 * config/i386/linux-common.h: Likewise.
8668 * config/mips/linux-common.h: Likewise.
8670 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8672 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8675 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8677 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8678 instances can own GC refs.
8680 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8681 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8682 (gcc::context::gt_pch_nx): Likewise.
8683 (gcc::context::gt_pch_nx): Likewise.
8684 * ggc.h (gt_ggc_mx <T>): New.
8685 (gt_pch_nx_with_op <T>): New.
8686 (gt_pch_nx <T>): New.
8687 * passes.c (opt_pass::gt_ggc_mx): New.
8688 (opt_pass::gt_pch_nx): New.
8689 (opt_pass::gt_pch_nx_with_op): New.
8690 (pass_manager::gt_ggc_mx): New.
8691 (pass_manager::gt_pch_nx): New.
8692 (pass_manager::gt_pch_nx_with_op): New.
8693 (pass_manager::operator new): Use
8694 ggc_internal_cleared_alloc_stat rather than xcalloc.
8695 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8696 (pass_manager::gt_ggc_mx): New.
8697 (pass_manager::gt_pch_nx): New.
8698 (pass_manager::gt_pch_nx_with_op): New.
8699 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8700 (opt_pass::operator new): New.
8701 (opt_pass::gt_ggc_mx): New.
8702 (opt_pass::gt_pch_nx): New.
8703 (opt_pass::gt_pch_nx_with_op): New.
8705 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8707 * Makefile.in (GTFILES): Add context.h.
8708 * context.c (gcc::context::operator new): New.
8709 (gcc::context::gt_ggc_mx): New.
8710 (gcc::context::gt_pch_nx): New.
8711 (gcc::context::gt_pch_nx): New.
8712 * context.h (gcc::context): Add GTY((user)) marking.
8713 (gcc::context::operator new): New.
8714 (gcc::context::gt_ggc_mx): New.
8715 (gcc::context::gt_pch_nx): New.
8716 (gcc::context::gt_pch_nx): New.
8717 (g): Add GTY marking.
8718 (gt_ggc_mx (gcc::context *)): New.
8719 (gt_pch_nx (gcc::context *)): New.
8720 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8721 void *cookie)): New.
8722 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8724 2013-08-20 Alan Modra <amodra@gmail.com>
8727 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8728 (rs6000_emit_epilogue): Likewise.
8730 2013-08-19 Dehao Chen <dehao@google.com>
8732 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8734 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8735 Jakub Jelinek <jakub@redhat.com>
8737 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8738 (BUILT_IN_FABSD64): Likewise.
8739 (BUILT_IN_FABSD128): Likewise.
8740 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8741 (fold_builtin_1): Likewise.
8742 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8743 destination and source operands.
8744 (*abstd2_fpr): Likewise.
8745 (*nabstd2_fpr): Likewise.
8747 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8749 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8752 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8754 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8757 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8759 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8760 * builtins.c (default_libc_has_function): New.
8761 (gnu_libc_has_function): Ditto.
8762 (no_c99_libc_has_function): Ditto.
8763 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8764 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8765 (fold_builtin_sincos): Likewise.
8766 (fold_builtin_cexp): Likewise.
8767 * builtins.def (DEF_C94_BUILTIN): Likewise.
8768 (DEF_C99_BUILTIN): Likewise.
8769 (DEF_C99_C90RES_BUILTIN): Likewise.
8770 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8771 definitions to using this define.
8772 * config/darwin-protos.h (darwin_libc_has_function): New.
8773 * config/darwin.c (darwin_libc_has_function): Ditto.
8774 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8775 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8776 * config/darwin.h: Ditto.
8777 * config/elfos.h: Ditto.
8778 * config/freebsd.h: Ditto.
8779 * config/i386/cygming.h: Ditto.
8780 * config/i386/djgpp.h: Ditto.
8781 * config/i386/i386-interix.h: Ditto.
8782 * config/microblaze/microblaze.h: Ditto.
8783 * config/mmix/mmix.h: Ditto.
8784 * config/gnu-user.h: Ditto.
8785 * config/ia64/hpux.h: Ditto.
8786 * config/pa/pa-hpux.h: Ditto.
8787 * config/pdp11/pdp11.h: Ditto.
8788 * config/picochip/picochip.h: Ditto.
8789 * config/linux.h: Ditto.
8790 * config/netbsd.h: Ditto.
8791 * config/openbsd.h: Ditto.
8792 * config/rs6000/aix43.h: Ditto.
8793 * config/rs6000/aix51.h: Ditto.
8794 * config/rs6000/aix52.h: Ditto.
8795 * config/rs6000/aix53.h: Ditto.
8796 * config/rs6000/aix61.h: Ditto.
8797 * config/rs6000/darwin.h: Ditto.
8798 * config/rs6000/linux.h: Ditto.
8799 * config/rs6000/linux64.h: Ditto.
8800 * config/s390/tpf.h: Ditto.
8801 * config/sol2-10.h: Ditto.
8802 * config/sol2.h: Ditto.
8803 * config/vms/vms.h: Ditto.
8804 * config/vxworks.h: Ditto.
8805 * config/linux-android.c (linux_android_libc_has_function):
8806 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8807 * config/linux-protos.h (linux_android_libc_has_function):
8809 * config/i386/i386.c (ix86_libc_has_function): New.
8810 * config/i386/i386-protos.h
8811 (ix86_libc_has_function): New declaration.
8812 * config/i386/i386.md
8813 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8814 ("isinf<mode>2): Likewise.
8815 * convert.c (convert_to_integer): Using new target hook
8816 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8817 TARGET_C99_FUNCTIONS.
8818 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8819 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8820 * coretypes.h (function_class): New enum for different
8821 classes of functions.
8822 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8823 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8824 (TARGET_HAS_SINCOS): Likewise.
8825 (TARGET_LIBC_HAS_FUNCTION): New.
8826 * doc/tm.texi: Regenerated.
8827 * targhooks.h (default_libc_has_function): New declaration.
8828 (no_c99_libc_has_function): Ditto.
8829 (gnu_libc_has_function): Ditto.
8830 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8831 and TARGET_HAS_SINCOS.
8833 2013-08-18 Jan Hubicka <jh@suse.cz>
8835 * Makeifle-in (ipa-devirt.o): New.
8836 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8837 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8838 (analyze_functions): Look into possible targets of polymorphic call.
8839 * dumpfile.c (dump_files): Add type-inheritance dump.
8840 * dumpfile.h (TDI_inheritance): New.
8841 * ipa-devirt.c: New file.
8842 * ipa-utils.h (odr_type_d): Forward declare.
8843 (odr_type): New type.
8844 (build_type_inheritance_graph): Declare.
8845 (possible_polymorphic_call_targets): Declare and introduce inline
8846 variant when only edge is pased.
8847 (dump_possible_polymorphic_call_targets): Likewise.
8848 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8849 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8850 (types_same_for_odr): ... here.
8851 * tree.h (type_in_anonymous_namespace_p): Declare.
8853 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8855 PR tree-optimization/58006
8856 * tree-parloops.c (take_address_of): Don't ICE if get_name
8858 (eliminate_local_variables_stmt): Remove clobber stmts.
8860 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8862 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8863 error message is issued for an alias to undefined symbol.
8865 2013-08-18 Jan Hubicka <jh@suse.cz>
8867 * cgraph.c (cgraph_create_indirect_edge): Discover
8868 polymorphic calls and record basic info into indirect_info.
8869 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8870 devirtualization, ignore objc function calls.
8871 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8872 call with no parm index info.
8873 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8874 * tree.c (virtual_method_call_p): New function.
8875 * tree.h (virtual_method_call_p): Declare.
8877 2013-08-16 Jan Hubicka <jh@suse.cz>
8880 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8882 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8884 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8887 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8889 * gengtype.c (type_for_name): Add special-case support for
8890 locating types within the "gcc::" namespace.
8891 (open_base_files): Emit a "using namespace gcc" directive.
8893 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8896 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8897 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8899 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8900 array instead of each individual operand as a separate argument.
8901 (emit_fusion_gpr_load): Likewise.
8902 (expand_fusion_gpr_load): Add new function declaration.
8904 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8905 signature to have the operands passed as an array, instead of as
8906 separate arguments. Allow ZERO_EXTEND to be in the memory
8907 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8908 depend on the register live/dead flags when peepholes are run.
8909 (expand_fusion_gpr_load): New function to be called from the
8910 peephole2 pass, to change the register that addis sets to be the
8912 (emit_fusion_gpr_load): Change the calling signature to have the
8913 operands passed as an array, instead of as separate arguments.
8914 Allow ZERO_EXTEND to be in the memory address, and also
8915 SIGN_EXTEND if -mpower8-fusion-sign.
8917 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8919 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8920 adjust the register addis loads up in the peephole2 pass. Do not
8921 depend on the register live/dead state when the peephole pass is done.
8923 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8925 * gengtype.c (create_user_defined_type): Ensure that the kind
8926 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8927 declaration is seen before the GTY((user)) marking.
8929 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8932 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8934 2013-08-16 Jan Hubicka <jh@suse.cz>
8936 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8937 arugment expected_type.
8938 (gimple_fold_call): Use it.
8939 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8940 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8941 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8942 (try_make_edge_direct_virtual_call): Likewise.
8943 * tree.c (obj_type_ref_class): New.
8944 * tree.h (obj_type_ref_class): Use it.
8946 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8948 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8949 (rtl_slim_pp): Likewise.
8950 (init_rtl_slim_pretty_print): Likewise.
8951 (dump_value_slim): Don't call it. Use local pretty printer.
8952 (dump_insn_slim): Likewise.
8953 (dump_rtl_slim): Likewise.
8954 (str_pattern_slim): Likewise.
8955 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8958 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8960 PR tree-optimization/58164
8961 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8962 walk gimple_goto_dest of GIMPLE_GOTO.
8964 PR tree-optimization/58165
8965 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8966 bi_call must be the last stmt in a bb, don't split_block, instead
8967 use fallthru edge from it and give up if there is none.
8968 Release conds vector when returning early.
8970 2013-08-14 Xinliang David Li <davidxl@google.com>
8972 * config/i386/i386.c (ix86_option_override_internal):
8973 Remove unused variable and field.
8975 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8978 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8979 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8980 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8981 and Linux, correct BLKmode alignment when 128-bit alignment is
8982 required and compatibility flag is not set.
8983 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8984 for zero-size arguments when compatibility flag is not set.
8986 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8988 PR tree-optimization/58145
8989 * tree-sra.c (build_ref_for_offset): If prev_base has
8990 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8992 2013-08-14 Xinliang David Li <davidxl@google.com>
8994 * config/i386/i386.c (ix86_option_override_internal):
8995 Fix uninitialized variable error.
8997 2013-08-14 Xinliang David Li <davidxl@google.com>
8999 * config/i386/i386.opt: Define two new options.
9000 * config/i386/x86-tune.def: Add arch selector field in macros.
9001 * config/i386/i386.h: Adjust macro definition.
9002 * config/i386/i386.c (ix86_option_override_internal):
9004 (parse_mtune_ctrl_str): New function.
9005 (set_ix86_tune_features): New function.
9006 (ix86_function_specific_restore): Call the new helper function.
9008 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9010 PR rtl-optimization/57662
9011 * sel-sched.c (code_motion_process_successors): When the current insn
9012 is removed after the recursive traversal, break from the loop.
9013 Add comments and debug printouts.
9015 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9016 Alexandre Oliva <aoliva@redhat.com>
9019 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9020 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9021 there also UNSPEC_PLTOFF.
9023 2013-08-14 Marek Polacek <polacek@redhat.com>
9025 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9026 NUM_CONDITIONS bit positions.
9028 2013-08-13 Cary Coutant <ccoutant@google.com>
9030 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9031 (attr_checksum): Hash vector contents instead of pointer.
9032 (attr_checksum_ordered): Likewise.
9034 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9036 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9037 when Pmode != word_mode. Add length_address attribute.
9038 (sse3_monitor_<mode>): Merge from sse3_monitor and
9039 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9040 Pmode != word_mode. Update insn length attribute.
9041 * config/i386/i386.c (ix86_option_override_internal): Update
9042 ix86_gen_monitor selection for merged sse3_monitor insn.
9044 2013-08-13 Julian Brown <julian@codesourcery.com>
9046 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9047 perform invalid legitimization on greater-than-word-size modes for
9050 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9052 * ira.c (setup_class_translate_array): Use aclass instead of cl
9053 for classes not fully covered by allocno classes.
9055 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9057 PR tree-optimization/57661
9058 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9059 * tree-inline.c (tree_function_versioning): Initialize it.
9060 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9061 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9062 that is not being copied.
9065 * asan.c (instrument_strlen_call): Fix typo in comment.
9066 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9068 2013-08-13 Steve Ellcey <sellcey@mips.com>
9070 * config/mips/mips.md (prefetch): Use lw instead of ld on
9071 loongson in 32bit mode.
9073 2013-08-13 Nick Clifton <nickc@redhat.com>
9075 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9077 2013-08-13 Jan Hubicka <jh@suse.cz>
9079 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9080 introduced edge; fix typo in sanity check.
9081 (cgraph_resolve_speculation): Export; improve diagnostic.
9082 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9083 speculation at type mismatch.
9084 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9085 (cgraph_resolve_speculation): Declare.
9086 (symtab_can_be_discarded): New function.
9087 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9088 * ipa-inline-transform.c (speculation_removed): New global var.
9089 (clone_inlined_nodes): See if speculation can be removed.
9090 (inline_call): If speculations was removed, we growths may not match.
9091 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9092 (speculation_useful_p): New function.
9093 (resolve_noninline_speculation): New function.
9094 (inline_small_functions): Resolve useless speculations.
9095 * ipa-inline.h (speculation_useful_p): Declare
9096 * ipa.c (can_replace_by_local_alias): Simplify.
9097 (ipa_profile): Produce speculative calls in non-lto, too;
9098 add simple cost model; produce local aliases.
9100 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9102 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9105 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9107 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9109 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9111 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9112 iterator instead of X87MODEF.
9114 2013-08-12 Perez Read <netfirewall@gmail.com>
9117 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9118 operand 0 for intel asm alternative.
9119 (*movabs<mode>_2): Ditto for operand 1.
9121 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9123 * config/aarch64/arm_none.h
9124 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9126 2013-08-12 Nick Clifton <nickc@redhat.com>
9128 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9130 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9132 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9133 expand for QI/HImode operand to produce more effictive code for
9134 unsigned char(short) --> float(double) conversion.
9136 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9138 * doc/invoke.texi: Mention that -ftls-model does not force the final
9141 2013-08-12 Marek Polacek <polacek@redhat.com>
9142 Marc Glisse <marc.glisse@inria.fr>
9144 PR tree-optimization/57980
9145 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9146 when creating -1 constant.
9148 2013-08-10 Jan Hubicka <jh@suse.cz>
9150 Workaround binutils PR14342.
9151 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9152 (gimple_init_edge_profiler): Likewise.
9153 (gimple_gen_ic_func_profiler): Likewise.
9155 2013-08-09 Jan Hubicka <jh@suse.cz>
9157 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9159 2013-08-09 Xinliang David Li <davidxl@google.com>
9161 * config/i386/stringop.def: New file.
9162 * config/i386/stringop.opt: New file.
9163 * config/i386/i386-opts.h: Include stringopt.def.
9164 * config/i386/i386.opt: Include stringopt.opt.
9165 * config/i386/i386.c (ix86_option_override_internal):
9166 Override default size based stringop inline strategies with options.
9167 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9170 2013-08-09 Jan Hubicka <jh@suse.cz>
9172 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9174 2013-08-09 Jan Hubicka <jh@suse.cz>
9176 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9178 (dump_cgraph_node): Dump profile-id.
9179 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9180 and common_target_probability.
9181 * lto-cgraph.c (lto_output_edge): Stream common targets.
9182 (lto_output_node): Stream profile ids.
9183 (input_node): Stream profile ids.
9184 (input_edge): Stream common targets.
9185 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9186 * ipa.c: Include value-prof.h
9187 (ipa_profile_generate_summary): Turn indirect call statement histograms
9188 into common targets.
9189 (ipa_profile): Turn common targets into speculative edges.
9191 2013-08-09 Jan Hubicka <jh@suse.cz>
9193 * cgraph.h (cgraph_node): Add profile_id.
9194 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9195 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9196 (del_node_map): Update.
9197 (find_func_by_funcdef_no): Replace by ...
9198 (find_func_by_profile_id): ... this one.
9199 (gimple_ic_transform): Do not remove useful histograms when
9200 speculation is not done; dump info when indirect call removal
9202 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9203 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9204 (__gcov_indirect_call_profiler_v2): .. this one.
9205 * profile.h (init_node_map): Update.
9206 * coverage.c (coverage_compute_profile_id): New function.
9207 * coverage.h (coverage_compute_profile_id): Declare.
9208 * tree-profile.c (init_ic_make_global_vars): Make
9209 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9210 (gimple_init_edge_profiler): Update prototype of
9211 __gcov_indirect_call_profiler.
9212 (gimple_gen_ic_func_profiler): Simplify.
9213 (tree_profiling): Use init_node_map
9215 2013-08-09 Jan Hubicka <jh@suse.cz>
9217 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9218 non-speculative refs.
9219 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9220 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9221 (cgraph_set_call_stmt): Likewise.
9222 (cgraph_create_edge_1): Fix release checking compilatoin;
9224 (cgraph_free_edge): Free indirect info.
9225 (cgraph_turn_edge_to_speculative): New function.
9226 (cgraph_speculative_call_info): New function.
9227 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9228 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9229 (dump_cgraph_node): Dump speculation.
9230 (verify_edge_count_and_frequency): Accept speculative edges.
9231 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9232 (verify_cgraph_node): Handle speculation.
9233 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9234 (cgraph_set_call_stmt): Update prototype.
9235 (cgraph_make_edge_direct): Update prototype.
9236 (cgraph_speculative_call_info): Declare.
9237 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9238 to change; update call of ipa_find_references.
9239 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9240 lto_stmt_uid and speculative flags.
9241 (ipa_dump_references): Dump speculation.
9242 (ipa_clone_references): Clone speculative flag.
9243 (ipa_clone_referring): Likewise.
9244 (ipa_clone_ref): New function.
9245 (ipa_find_reference): Look into lto_stmt_uids
9246 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9247 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9248 (ipa_find_reference): Update declaration.
9249 (ipa_clone_ref): Declare.
9250 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9251 stream speculative flag.
9252 (lto_output_ref): Stream statements uids and speculation.
9253 (input_ref): Likewise.
9254 (input_edge): Stream speuclation.
9255 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9256 (cgraph_set_call_stmt_including_clones): Handle speculation.
9257 * ipa-inline.c (heap_edge_removal_hook): New function.
9258 (inline_small_functions): Register it.
9259 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9260 also initialize refs.
9261 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9263 (try_make_edge_direct_simple_call): Likewise.
9264 (try_make_edge_direct_simple_call): Likewise.
9265 (update_indirect_edges_after_inlining): Likewise.
9266 (remove_described_reference): Look proper lto_stmt_uid.
9267 (propagate_controlled_uses): Likewise.
9268 (propagate_controlled_uses): Liekwise.
9269 * tree-inline.c (copy_bb): Copy speculative edges.
9270 (redirect_all_calls): New function.
9271 (copy_cfg_body): Do redirection after loop info is updated.
9272 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9274 2013-08-09 Jan Hubicka <jh@suse.cz>
9276 * lto-streamer-out.c (output_function): Renumber PHIs.
9277 * lto-streamer-in.c (input_function): Likewise.
9279 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9281 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9282 (get_lane_unsigned): Likewise.
9283 (dup_lane_scalar): Likewise.
9284 (get_lane): enable for VALL.
9285 * config/aarch64/aarch64-simd.md
9286 (aarch64_dup_lane_scalar<mode>): Remove.
9287 (aarch64_get_lane_signed<mode>): Likewise.
9288 (aarch64_get_lane_unsigned<mode>): Likewise.
9289 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9290 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9291 (aarch64_get_lane<mode>): Enable for all vector modes.
9292 (aarch64_get_lanedi): Remove misleading constraints.
9293 * config/aarch64/arm_neon.h
9294 (__aarch64_vget_lane_any): Define.
9295 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9296 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9297 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9298 * config/aarch64/iterators.md (VDQQH): New.
9302 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9304 * configure.ac: Add GAS check for LEON instructions on SPARC.
9305 * configure: Regenerate.
9306 * config.in: Likewise.
9307 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9309 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9310 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9312 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9313 (AS_LEON_FLAG): New macro.
9314 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9315 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9316 Deal with LEON and LEON3 for the memory model.
9317 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9318 (atomic_compare_and_swap<mode>_1): Likewise.
9319 (*atomic_compare_and_swap<mode>_1): Likewise.
9321 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9323 * config/arm/neon.md (vcond): Fix floating-point vector
9324 comparisons against 0.
9326 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9328 * lra-constraints.c (emit_spill_move): Remove assert.
9329 (process_alt_operands): Add more debugging
9330 output. Increase reject for spilling into memory. Decrease
9331 reject for reloading scratch.
9332 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9334 2013-08-08 Steve Ellcey <sellcey@mips.com>
9336 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9337 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9338 micromips incompatible. Add nan2008.
9339 (MULTILIB_DIRNAMES): Add nan2008.
9340 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9341 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9342 and micromips incompatible. Add nan2008.
9343 (MULTILIB_DIRNAMES): Add nan2008.
9344 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9346 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9348 PR rtl-optimization/58079
9349 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9350 simplify_comparison has widened a comparison with an integer.
9352 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9354 * config/arm/neon.md (movmisalign<mode>): Disable when we
9355 don't allow unaligned accesses.
9356 (*movmisalign<mode>_neon_store): Likewise.
9357 (*movmisalign<mode>_neon_load): Likewise.
9358 (*movmisalign<mode>_neon_store): Likewise.
9359 (*movmisalign<mode>_neon_load): Likewise.
9361 2013-08-08 Jan Hubicka <jh@suse.cz>
9363 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9364 (make_pass_rebuild_cgraph_edges): Also clear references.
9365 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9366 * ipa-inline-transform.c (inline_transform): Remove all references
9368 * cgraphunit.c (expand_function): Remove all references after
9370 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9371 (ipa_find_reference): Rewrite to iterator.
9372 (remove_stmt_references): Likewise.
9373 (ipa_clear_stmts_in_references): New function.
9374 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9375 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9377 * ipa-split.c (split_function): Remove references in split function.
9379 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9382 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9383 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9385 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9388 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9389 suggested mode for the assignment isn't compatible with the
9392 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9395 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9397 2013-08-07 Xinliang David Li <davidxl@google.com>
9399 * config/i386/i386.opt: New option -mtune-ctrl=.
9400 * config/i386/x86-tune.def: New file.
9401 * config/i386/i386.h: include x86-tune.def.
9402 * config/i386/i386.c (ix86_option_override_internal):
9403 Parsing -mtune-ctrl= option and set tune features.
9405 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9408 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9409 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9412 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9414 * rtl.h (update_alignments): Declare.
9415 * final.c (grow_label_align): New function extracted from...
9416 (shorten_branches): ...here. Call it.
9417 (update_alignments): New function.
9418 * reorg.c (sibling_labels): New variable.
9419 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9420 the new label along with it onto the sibling_labels vector.
9421 (fill_simple_delay_slots): Adjust call to get_label_before.
9422 (fill_slots_from_thread): Likewise.
9423 (relax_delay_slots): Likewise.
9424 (make_return_insns): Likewise.
9425 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9427 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9429 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9430 document its semantics.
9431 (diagnostic_report_diagnostic): Adjust accordingly.
9433 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9435 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9436 (sparc_option_override): ...and port to new C++ pass API.
9437 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9439 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9441 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9443 2013-08-06 Caroline Tice <cmtice@google.com>
9445 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9446 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9447 * tree-pass.h: Add pass_vtable_verify.
9448 * varasm.c (assemble_variable): Add code to properly set the comdat
9449 section and name for the .vtable_map_vars section.
9450 (assemble_vtyv_preinit_initializer): New function.
9451 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9453 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9454 * vtable-verify.c: New file.
9455 * vtable-verify.h: New file.
9456 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9457 initialiation levels.
9458 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9459 * passes.def: Insert pass_vtable_verify.
9460 * aclocal.m4: Reorder includes.
9461 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9462 -fvtv-counts options.
9463 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9464 as appropriate, if -fvtable-verify=... is used.
9465 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9466 -fvtable-verify=... is used.
9467 * Makefile.in (OBJS): Add vtable-verify.o to list.
9468 (vtable-verify.o): Add new build rule.
9469 (GTFILES): Add vtable-verify.c to list.
9470 * common.opt (fvtable-verify=): New flag.
9471 (vtv_priority): Values for fvtable-verify= flag.
9472 (fvtv-counts): New flag.
9473 (fvtv-debug): New flag.
9474 * tree.h (save_vtable_map_decl): New extern function decl.
9476 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9478 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9479 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9480 (pass_data_rl78_devirt): ...new pass_data instance and...
9481 (make_pass_rl78_devirt): ...new function.
9482 (rl78_asm_file_start): Port pass registration to new C++ API.
9484 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9486 * coretypes.h (rtl_opt_pass): Add.
9487 (gcc::context): Add.
9488 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9489 (epiphany_init): Port to new C++ pass API.
9490 (epiphany_optimize_mode_switching): Likewise.
9491 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9492 (pass_manager::get_pass_mode_switching): New.
9493 (pass_manager::get_pass_peephole2): New.
9494 * mode-switching.c (pass_mode_switching): Add clone method.
9495 * recog.c (pass_peephole2): Add clone method.
9496 (pass_split_all_insns): Add clone method.
9498 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9500 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9501 (mips_option_override): ...here, porting to new C++ API for passes.
9503 2013-08-06 Jan Hubicka <jh@suse.cz>
9505 * cgraph.c (cgraph_get_body): New function based on lto.c
9507 * cgraph.h (cgraph_get_body): Declare.
9508 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9510 * cgraphunit.c (expand_function): Get body prior expanding.
9511 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9512 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9514 * passes.c (do_per_function_toporder): Get body.
9515 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9516 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9518 2013-08-06 Martin Jambor <mjambor@suse.cz>
9521 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9522 is not re-finalized. Rename second parameter to no_collect.
9524 2013-08-06 Martin Jambor <mjambor@suse.cz>
9527 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9528 MEM_REF has proper alignment information.
9530 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9533 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9535 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9536 rtx (*) (rtx, ...) with insn_gen_fn.
9537 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9538 insn_gen_fn::stored_funcptr. Add initializer braces.
9540 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9542 Rewrite how instances of passes are cloned to remove assumptions
9543 about their sizes (thus allowing pass subclasses to have
9544 additional data fields, albeit non-GC-managed ones at this point).
9546 * passes.c (make_pass_instance): Now that passes have clone
9547 methods, rewrite this function to eliminate XNEW and memcpy
9548 calls that used hardcoded sizes. Since this function no longer
9549 creates pass instances, rename it to...
9550 (add_pass_instance): ...this. Document the old way that passes were
9551 numbered and flagged, and rework this function to continue using it.
9552 (next_pass_1): Add an initial_pass argument for use by
9554 (position_pass): When adding multiple instances of a pass, use
9555 the pass's clone method, rather than relying on the XNEW/memcpy
9556 within the former make_pass_instance (now add_pass_instance).
9557 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9558 the initial instance of the current pass within the pass manager.
9560 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9562 This is the automated part of the conversion of passes from C
9563 structs to C++ classes.
9565 Patch autogenerated by refactor_passes.py from
9566 https://github.com/davidmalcolm/gcc-refactoring-scripts
9567 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9569 * asan.c (pass_asan): Convert from a global struct to a subclass of
9570 gimple_opt_pass along with...
9571 (pass_data_asan): ...new pass_data instance and...
9572 (make_pass_asan): ...new function.
9573 (pass_asan_O0): Convert from a global struct to a subclass of
9574 gimple_opt_pass along with...
9575 (pass_data_asan_O0): ...new pass_data instance and...
9576 (make_pass_asan_O0): ...new function.
9577 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9578 subclass of rtl_opt_pass along with...
9579 (pass_data_inc_dec): ...new pass_data instance and...
9580 (make_pass_inc_dec): ...new function.
9581 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9582 a subclass of rtl_opt_pass along with...
9583 (pass_data_reorder_blocks): ...new pass_data instance and...
9584 (make_pass_reorder_blocks): ...new function.
9585 (pass_duplicate_computed_gotos): Convert from a global struct to a
9586 subclass of rtl_opt_pass along with...
9587 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9588 (make_pass_duplicate_computed_gotos): ...new function.
9589 (pass_partition_blocks): Convert from a global struct to a subclass of
9590 rtl_opt_pass along with...
9591 (pass_data_partition_blocks): ...new pass_data instance and...
9592 (make_pass_partition_blocks): ...new function.
9593 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9594 struct to a subclass of rtl_opt_pass along with...
9595 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9597 (make_pass_branch_target_load_optimize1): ...new function.
9598 (pass_branch_target_load_optimize2): Convert from a global struct to a
9599 subclass of rtl_opt_pass along with...
9600 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9602 (make_pass_branch_target_load_optimize2): ...new function.
9603 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9604 of rtl_opt_pass along with...
9605 (pass_data_jump): ...new pass_data instance and...
9606 (make_pass_jump): ...new function.
9607 (pass_jump2): Convert from a global struct to a subclass of
9608 rtl_opt_pass along with...
9609 (pass_data_jump2): ...new pass_data instance and...
9610 (make_pass_jump2): ...new function.
9611 * cfgexpand.c (pass_expand): Convert from a global struct to a
9612 subclass of rtl_opt_pass along with...
9613 (pass_data_expand): ...new pass_data instance and...
9614 (make_pass_expand): ...new function.
9615 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9616 of rtl_opt_pass along with...
9617 (pass_data_free_cfg): ...new pass_data instance and...
9618 (make_pass_free_cfg): ...new function.
9619 (pass_into_cfg_layout_mode): Convert from a global struct to a
9620 subclass of rtl_opt_pass along with...
9621 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9622 (make_pass_into_cfg_layout_mode): ...new function.
9623 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9624 subclass of rtl_opt_pass along with...
9625 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9626 (make_pass_outof_cfg_layout_mode): ...new function.
9627 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9628 struct to a subclass of gimple_opt_pass along with...
9629 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9630 (make_pass_build_cgraph_edges): ...new function.
9631 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9632 subclass of gimple_opt_pass along with...
9633 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9634 (make_pass_rebuild_cgraph_edges): ...new function.
9635 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9636 subclass of gimple_opt_pass along with...
9637 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9639 (make_pass_remove_cgraph_callee_edges): ...new function.
9640 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9641 struct to a subclass of rtl_opt_pass along with...
9642 (pass_data_stack_adjustments): ...new pass_data instance and...
9643 (make_pass_stack_adjustments): ...new function.
9644 * combine.c (pass_combine): Convert from a global struct to a subclass
9645 of rtl_opt_pass along with...
9646 (pass_data_combine): ...new pass_data instance and...
9647 (make_pass_combine): ...new function.
9648 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9649 global struct to a subclass of rtl_opt_pass along with...
9650 (pass_data_compare_elim_after_reload): ...new pass_data instance
9652 (make_pass_compare_elim_after_reload): ...new function.
9653 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9654 of rtl_opt_pass along with...
9655 (pass_data_rtl_cprop): ...new pass_data instance and...
9656 (make_pass_rtl_cprop): ...new function.
9657 * cse.c (pass_cse): Convert from a global struct to a subclass of
9658 rtl_opt_pass along with...
9659 (pass_data_cse): ...new pass_data instance and...
9660 (make_pass_cse): ...new function.
9661 (pass_cse2): Convert from a global struct to a subclass of
9662 rtl_opt_pass along with...
9663 (pass_data_cse2): ...new pass_data instance and...
9664 (make_pass_cse2): ...new function.
9665 (pass_cse_after_global_opts): Convert from a global struct to a
9666 subclass of rtl_opt_pass along with...
9667 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9668 (make_pass_cse_after_global_opts): ...new function.
9669 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9670 of rtl_opt_pass along with...
9671 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9672 (make_pass_ud_rtl_dce): ...new function.
9673 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9674 rtl_opt_pass along with...
9675 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9676 (make_pass_fast_rtl_dce): ...new function.
9677 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9678 a subclass of rtl_opt_pass along with...
9679 (pass_data_df_initialize_opt): ...new pass_data instance and...
9680 (make_pass_df_initialize_opt): ...new function.
9681 (pass_df_initialize_no_opt): Convert from a global struct to a
9682 subclass of rtl_opt_pass along with...
9683 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9684 (make_pass_df_initialize_no_opt): ...new function.
9685 (pass_df_finish): Convert from a global struct to a subclass of
9686 rtl_opt_pass along with...
9687 (pass_data_df_finish): ...new pass_data instance and...
9688 (make_pass_df_finish): ...new function.
9689 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9690 rtl_opt_pass along with...
9691 (pass_data_rtl_dse1): ...new pass_data instance and...
9692 (make_pass_rtl_dse1): ...new function.
9693 (pass_rtl_dse2): Convert from a global struct to a subclass of
9694 rtl_opt_pass along with...
9695 (pass_data_rtl_dse2): ...new pass_data instance and...
9696 (make_pass_rtl_dse2): ...new function.
9697 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9698 subclass of rtl_opt_pass along with...
9699 (pass_data_dwarf2_frame): ...new pass_data instance and...
9700 (make_pass_dwarf2_frame): ...new function.
9701 * except.c (pass_set_nothrow_function_flags): Convert from a global
9702 struct to a subclass of rtl_opt_pass along with...
9703 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9705 (make_pass_set_nothrow_function_flags): ...new function.
9706 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9707 subclass of rtl_opt_pass along with...
9708 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9710 (make_pass_convert_to_eh_region_ranges): ...new function.
9711 * final.c (pass_compute_alignments): Convert from a global struct to a
9712 subclass of rtl_opt_pass along with...
9713 (pass_data_compute_alignments): ...new pass_data instance and...
9714 (make_pass_compute_alignments): ...new function.
9715 (pass_final): Convert from a global struct to a subclass of
9716 rtl_opt_pass along with...
9717 (pass_data_final): ...new pass_data instance and...
9718 (make_pass_final): ...new function.
9719 (pass_shorten_branches): Convert from a global struct to a subclass of
9720 rtl_opt_pass along with...
9721 (pass_data_shorten_branches): ...new pass_data instance and...
9722 (make_pass_shorten_branches): ...new function.
9723 (pass_clean_state): Convert from a global struct to a subclass of
9724 rtl_opt_pass along with...
9725 (pass_data_clean_state): ...new pass_data instance and...
9726 (make_pass_clean_state): ...new function.
9727 * function.c (pass_instantiate_virtual_regs): Convert from a global
9728 struct to a subclass of rtl_opt_pass along with...
9729 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9730 (make_pass_instantiate_virtual_regs): ...new function.
9731 (pass_leaf_regs): Convert from a global struct to a subclass of
9732 rtl_opt_pass along with...
9733 (pass_data_leaf_regs): ...new pass_data instance and...
9734 (make_pass_leaf_regs): ...new function.
9735 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9736 subclass of rtl_opt_pass along with...
9737 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9739 (make_pass_thread_prologue_and_epilogue): ...new function.
9740 (pass_match_asm_constraints): Convert from a global struct to a
9741 subclass of rtl_opt_pass along with...
9742 (pass_data_match_asm_constraints): ...new pass_data instance and...
9743 (make_pass_match_asm_constraints): ...new function.
9744 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9745 subclass of rtl_opt_pass along with...
9746 (pass_data_rtl_fwprop): ...new pass_data instance and...
9747 (make_pass_rtl_fwprop): ...new function.
9748 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9749 rtl_opt_pass along with...
9750 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9751 (make_pass_rtl_fwprop_addr): ...new function.
9752 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9753 rtl_opt_pass along with...
9754 (pass_data_rtl_pre): ...new pass_data instance and...
9755 (make_pass_rtl_pre): ...new function.
9756 (pass_rtl_hoist): Convert from a global struct to a subclass of
9757 rtl_opt_pass along with...
9758 (pass_data_rtl_hoist): ...new pass_data instance and...
9759 (make_pass_rtl_hoist): ...new function.
9760 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9761 subclass of gimple_opt_pass along with...
9762 (pass_data_lower_cf): ...new pass_data instance and...
9763 (make_pass_lower_cf): ...new function.
9764 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9765 from a global struct to a subclass of gimple_opt_pass along with...
9766 (pass_data_strength_reduction): ...new pass_data instance and...
9767 (make_pass_strength_reduction): ...new function.
9768 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9769 of rtl_opt_pass along with...
9770 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9771 (make_pass_rtl_ifcvt): ...new function.
9772 (pass_if_after_combine): Convert from a global struct to a subclass of
9773 rtl_opt_pass along with...
9774 (pass_data_if_after_combine): ...new pass_data instance and...
9775 (make_pass_if_after_combine): ...new function.
9776 (pass_if_after_reload): Convert from a global struct to a subclass of
9777 rtl_opt_pass along with...
9778 (pass_data_if_after_reload): ...new pass_data instance and...
9779 (make_pass_if_after_reload): ...new function.
9780 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9781 a subclass of rtl_opt_pass along with...
9782 (pass_data_initialize_regs): ...new pass_data instance and...
9783 (make_pass_initialize_regs): ...new function.
9784 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9785 of ipa_opt_pass_d along with...
9786 (pass_data_ipa_cp): ...new pass_data instance and...
9787 (make_pass_ipa_cp): ...new function.
9788 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9789 global struct to a subclass of gimple_opt_pass along with...
9790 (pass_data_inline_parameters): ...new pass_data instance and...
9791 (make_pass_inline_parameters): ...new function.
9792 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9793 subclass of gimple_opt_pass along with...
9794 (pass_data_early_inline): ...new pass_data instance and...
9795 (make_pass_early_inline): ...new function.
9796 (pass_ipa_inline): Convert from a global struct to a subclass of
9797 ipa_opt_pass_d along with...
9798 (pass_data_ipa_inline): ...new pass_data instance and...
9799 (make_pass_ipa_inline): ...new function.
9800 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9801 struct to a subclass of gimple_opt_pass along with...
9802 (pass_data_local_pure_const): ...new pass_data instance and...
9803 (make_pass_local_pure_const): ...new function.
9804 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9805 ipa_opt_pass_d along with...
9806 (pass_data_ipa_pure_const): ...new pass_data instance and...
9807 (make_pass_ipa_pure_const): ...new function.
9808 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9809 to a subclass of ipa_opt_pass_d along with...
9810 (pass_data_ipa_reference): ...new pass_data instance and...
9811 (make_pass_ipa_reference): ...new function.
9812 * ipa-split.c (pass_split_functions): Convert from a global struct to
9813 a subclass of gimple_opt_pass along with...
9814 (pass_data_split_functions): ...new pass_data instance and...
9815 (make_pass_split_functions): ...new function.
9816 (pass_feedback_split_functions): Convert from a global struct to a
9817 subclass of gimple_opt_pass along with...
9818 (pass_data_feedback_split_functions): ...new pass_data instance and...
9819 (make_pass_feedback_split_functions): ...new function.
9820 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9821 global struct to a subclass of simple_ipa_opt_pass along with...
9822 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9824 (make_pass_ipa_function_and_variable_visibility): ...new function.
9825 (pass_ipa_free_inline_summary): Convert from a global struct to a
9826 subclass of simple_ipa_opt_pass along with...
9827 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9828 (make_pass_ipa_free_inline_summary): ...new function.
9829 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9830 subclass of ipa_opt_pass_d along with...
9831 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9833 (make_pass_ipa_whole_program_visibility): ...new function.
9834 (pass_ipa_profile): Convert from a global struct to a subclass of
9835 ipa_opt_pass_d along with...
9836 (pass_data_ipa_profile): ...new pass_data instance and...
9837 (make_pass_ipa_profile): ...new function.
9838 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9839 ipa_opt_pass_d along with...
9840 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9841 (make_pass_ipa_cdtor_merge): ...new function.
9842 * ira.c (pass_ira): Convert from a global struct to a subclass of
9843 rtl_opt_pass along with...
9844 (pass_data_ira): ...new pass_data instance and...
9845 (make_pass_ira): ...new function.
9846 (pass_reload): Convert from a global struct to a subclass of
9847 rtl_opt_pass along with...
9848 (pass_data_reload): ...new pass_data instance and...
9849 (make_pass_reload): ...new function.
9850 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9851 subclass of rtl_opt_pass along with...
9852 (pass_data_cleanup_barriers): ...new pass_data instance and...
9853 (make_pass_cleanup_barriers): ...new function.
9854 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9855 of rtl_opt_pass along with...
9856 (pass_data_loop2): ...new pass_data instance and...
9857 (make_pass_loop2): ...new function.
9858 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9859 rtl_opt_pass along with...
9860 (pass_data_rtl_loop_init): ...new pass_data instance and...
9861 (make_pass_rtl_loop_init): ...new function.
9862 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9863 rtl_opt_pass along with...
9864 (pass_data_rtl_loop_done): ...new pass_data instance and...
9865 (make_pass_rtl_loop_done): ...new function.
9866 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9867 subclass of rtl_opt_pass along with...
9868 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9869 (make_pass_rtl_move_loop_invariants): ...new function.
9870 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9871 rtl_opt_pass along with...
9872 (pass_data_rtl_unswitch): ...new pass_data instance and...
9873 (make_pass_rtl_unswitch): ...new function.
9874 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9875 subclass of rtl_opt_pass along with...
9876 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9878 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9879 (pass_rtl_doloop): Convert from a global struct to a subclass of
9880 rtl_opt_pass along with...
9881 (pass_data_rtl_doloop): ...new pass_data instance and...
9882 (make_pass_rtl_doloop): ...new function.
9883 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9884 a subclass of rtl_opt_pass along with...
9885 (pass_data_lower_subreg): ...new pass_data instance and...
9886 (make_pass_lower_subreg): ...new function.
9887 (pass_lower_subreg2): Convert from a global struct to a subclass of
9888 rtl_opt_pass along with...
9889 (pass_data_lower_subreg2): ...new pass_data instance and...
9890 (make_pass_lower_subreg2): ...new function.
9891 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9892 struct to a subclass of ipa_opt_pass_d along with...
9893 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9894 (make_pass_ipa_lto_gimple_out): ...new function.
9895 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9896 of ipa_opt_pass_d along with...
9897 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9898 (make_pass_ipa_lto_finish_out): ...new function.
9899 * mode-switching.c (pass_mode_switching): Convert from a global struct
9900 to a subclass of rtl_opt_pass along with...
9901 (pass_data_mode_switching): ...new pass_data instance and...
9902 (make_pass_mode_switching): ...new function.
9903 * modulo-sched.c (pass_sms): Convert from a global struct to a
9904 subclass of rtl_opt_pass along with...
9905 (pass_data_sms): ...new pass_data instance and...
9906 (make_pass_sms): ...new function.
9907 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9908 subclass of gimple_opt_pass along with...
9909 (pass_data_expand_omp): ...new pass_data instance and...
9910 (make_pass_expand_omp): ...new function.
9911 (pass_lower_omp): Convert from a global struct to a subclass of
9912 gimple_opt_pass along with...
9913 (pass_data_lower_omp): ...new pass_data instance and...
9914 (make_pass_lower_omp): ...new function.
9915 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9916 of gimple_opt_pass along with...
9917 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9918 (make_pass_diagnose_omp_blocks): ...new function.
9919 * passes.c (pass_early_local_passes): Convert from a global struct to
9920 a subclass of simple_ipa_opt_pass along with...
9921 (pass_data_early_local_passes): ...new pass_data instance and...
9922 (make_pass_early_local_passes): ...new function.
9923 (pass_all_early_optimizations): Convert from a global struct to a
9924 subclass of gimple_opt_pass along with...
9925 (pass_data_all_early_optimizations): ...new pass_data instance and...
9926 (make_pass_all_early_optimizations): ...new function.
9927 (pass_all_optimizations): Convert from a global struct to a subclass
9928 of gimple_opt_pass along with...
9929 (pass_data_all_optimizations): ...new pass_data instance and...
9930 (make_pass_all_optimizations): ...new function.
9931 (pass_all_optimizations_g): Convert from a global struct to a subclass
9932 of gimple_opt_pass along with...
9933 (pass_data_all_optimizations_g): ...new pass_data instance and...
9934 (make_pass_all_optimizations_g): ...new function.
9935 (pass_rest_of_compilation): Convert from a global struct to a subclass
9936 of rtl_opt_pass along with...
9937 (pass_data_rest_of_compilation): ...new pass_data instance and...
9938 (make_pass_rest_of_compilation): ...new function.
9939 (pass_postreload): Convert from a global struct to a subclass of
9940 rtl_opt_pass along with...
9941 (pass_data_postreload): ...new pass_data instance and...
9942 (make_pass_postreload): ...new function.
9943 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9944 subclass of rtl_opt_pass along with...
9945 (pass_data_gcse2): ...new pass_data instance and...
9946 (make_pass_gcse2): ...new function.
9947 * postreload.c (pass_postreload_cse): Convert from a global struct to
9948 a subclass of rtl_opt_pass along with...
9949 (pass_data_postreload_cse): ...new pass_data instance and...
9950 (make_pass_postreload_cse): ...new function.
9951 * predict.c (pass_profile): Convert from a global struct to a subclass
9952 of gimple_opt_pass along with...
9953 (pass_data_profile): ...new pass_data instance and...
9954 (make_pass_profile): ...new function.
9955 (pass_strip_predict_hints): Convert from a global struct to a subclass
9956 of gimple_opt_pass along with...
9957 (pass_data_strip_predict_hints): ...new pass_data instance and...
9958 (make_pass_strip_predict_hints): ...new function.
9959 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9960 of rtl_opt_pass along with...
9961 (pass_data_peephole2): ...new pass_data instance and...
9962 (make_pass_peephole2): ...new function.
9963 (pass_split_all_insns): Convert from a global struct to a subclass of
9964 rtl_opt_pass along with...
9965 (pass_data_split_all_insns): ...new pass_data instance and...
9966 (make_pass_split_all_insns): ...new function.
9967 (pass_split_after_reload): Convert from a global struct to a subclass
9968 of rtl_opt_pass along with...
9969 (pass_data_split_after_reload): ...new pass_data instance and...
9970 (make_pass_split_after_reload): ...new function.
9971 (pass_split_before_regstack): Convert from a global struct to a
9972 subclass of rtl_opt_pass along with...
9973 (pass_data_split_before_regstack): ...new pass_data instance and...
9974 (make_pass_split_before_regstack): ...new function.
9975 (pass_split_before_sched2): Convert from a global struct to a subclass
9976 of rtl_opt_pass along with...
9977 (pass_data_split_before_sched2): ...new pass_data instance and...
9978 (make_pass_split_before_sched2): ...new function.
9979 (pass_split_for_shorten_branches): Convert from a global struct to a
9980 subclass of rtl_opt_pass along with...
9981 (pass_data_split_for_shorten_branches): ...new pass_data instance
9983 (make_pass_split_for_shorten_branches): ...new function.
9984 * ree.c (pass_ree): Convert from a global struct to a subclass of
9985 rtl_opt_pass along with...
9986 (pass_data_ree): ...new pass_data instance and...
9987 (make_pass_ree): ...new function.
9988 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9989 subclass of rtl_opt_pass along with...
9990 (pass_data_stack_regs): ...new pass_data instance and...
9991 (make_pass_stack_regs): ...new function.
9992 (pass_stack_regs_run): Convert from a global struct to a subclass of
9993 rtl_opt_pass along with...
9994 (pass_data_stack_regs_run): ...new pass_data instance and...
9995 (make_pass_stack_regs_run): ...new function.
9996 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9997 subclass of rtl_opt_pass along with...
9998 (pass_data_cprop_hardreg): ...new pass_data instance and...
9999 (make_pass_cprop_hardreg): ...new function.
10000 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10001 subclass of rtl_opt_pass along with...
10002 (pass_data_reginfo_init): ...new pass_data instance and...
10003 (make_pass_reginfo_init): ...new function.
10004 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10005 of rtl_opt_pass along with...
10006 (pass_data_regmove): ...new pass_data instance and...
10007 (make_pass_regmove): ...new function.
10008 * regrename.c (pass_regrename): Convert from a global struct to a
10009 subclass of rtl_opt_pass along with...
10010 (pass_data_regrename): ...new pass_data instance and...
10011 (make_pass_regrename): ...new function.
10012 * reorg.c (pass_delay_slots): Convert from a global struct to a
10013 subclass of rtl_opt_pass along with...
10014 (pass_data_delay_slots): ...new pass_data instance and...
10015 (make_pass_delay_slots): ...new function.
10016 (pass_machine_reorg): Convert from a global struct to a subclass of
10017 rtl_opt_pass along with...
10018 (pass_data_machine_reorg): ...new pass_data instance and...
10019 (make_pass_machine_reorg): ...new function.
10020 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10021 of rtl_opt_pass along with...
10022 (pass_data_sched): ...new pass_data instance and...
10023 (make_pass_sched): ...new function.
10024 (pass_sched2): Convert from a global struct to a subclass of
10025 rtl_opt_pass along with...
10026 (pass_data_sched2): ...new pass_data instance and...
10027 (make_pass_sched2): ...new function.
10028 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10029 to a subclass of rtl_opt_pass along with...
10030 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10031 (make_pass_stack_ptr_mod): ...new function.
10032 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10033 to a subclass of rtl_opt_pass along with...
10034 (pass_data_rtl_store_motion): ...new pass_data instance and...
10035 (make_pass_rtl_store_motion): ...new function.
10036 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10037 of gimple_opt_pass along with...
10038 (pass_data_tracer): ...new pass_data instance and...
10039 (make_pass_tracer): ...new function.
10040 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10041 to a subclass of gimple_opt_pass along with...
10042 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10043 (make_pass_diagnose_tm_blocks): ...new function.
10044 (pass_lower_tm): Convert from a global struct to a subclass of
10045 gimple_opt_pass along with...
10046 (pass_data_lower_tm): ...new pass_data instance and...
10047 (make_pass_lower_tm): ...new function.
10048 (pass_tm_init): Convert from a global struct to a subclass of
10049 gimple_opt_pass along with...
10050 (pass_data_tm_init): ...new pass_data instance and...
10051 (make_pass_tm_init): ...new function.
10052 (pass_tm_mark): Convert from a global struct to a subclass of
10053 gimple_opt_pass along with...
10054 (pass_data_tm_mark): ...new pass_data instance and...
10055 (make_pass_tm_mark): ...new function.
10056 (pass_tm_edges): Convert from a global struct to a subclass of
10057 gimple_opt_pass along with...
10058 (pass_data_tm_edges): ...new pass_data instance and...
10059 (make_pass_tm_edges): ...new function.
10060 (pass_tm_memopt): Convert from a global struct to a subclass of
10061 gimple_opt_pass along with...
10062 (pass_data_tm_memopt): ...new pass_data instance and...
10063 (make_pass_tm_memopt): ...new function.
10064 (pass_ipa_tm): Convert from a global struct to a subclass of
10065 simple_ipa_opt_pass along with...
10066 (pass_data_ipa_tm): ...new pass_data instance and...
10067 (make_pass_ipa_tm): ...new function.
10068 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10069 subclass of gimple_opt_pass along with...
10070 (pass_data_call_cdce): ...new pass_data instance and...
10071 (make_pass_call_cdce): ...new function.
10072 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10073 subclass of gimple_opt_pass along with...
10074 (pass_data_build_cfg): ...new pass_data instance and...
10075 (make_pass_build_cfg): ...new function.
10076 (pass_split_crit_edges): Convert from a global struct to a subclass of
10077 gimple_opt_pass along with...
10078 (pass_data_split_crit_edges): ...new pass_data instance and...
10079 (make_pass_split_crit_edges): ...new function.
10080 (pass_warn_function_return): Convert from a global struct to a
10081 subclass of gimple_opt_pass along with...
10082 (pass_data_warn_function_return): ...new pass_data instance and...
10083 (make_pass_warn_function_return): ...new function.
10084 (pass_warn_function_noreturn): Convert from a global struct to a
10085 subclass of gimple_opt_pass along with...
10086 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10087 (make_pass_warn_function_noreturn): ...new function.
10088 (pass_warn_unused_result): Convert from a global struct to a subclass
10089 of gimple_opt_pass along with...
10090 (pass_data_warn_unused_result): ...new pass_data instance and...
10091 (make_pass_warn_unused_result): ...new function.
10092 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10093 a subclass of gimple_opt_pass along with...
10094 (pass_data_merge_phi): ...new pass_data instance and...
10095 (make_pass_merge_phi): ...new function.
10096 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10097 a subclass of gimple_opt_pass along with...
10098 (pass_data_lower_complex): ...new pass_data instance and...
10099 (make_pass_lower_complex): ...new function.
10100 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10101 gimple_opt_pass along with...
10102 (pass_data_lower_complex_O0): ...new pass_data instance and...
10103 (make_pass_lower_complex_O0): ...new function.
10104 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10105 subclass of gimple_opt_pass along with...
10106 (pass_data_lower_eh): ...new pass_data instance and...
10107 (make_pass_lower_eh): ...new function.
10108 (pass_refactor_eh): Convert from a global struct to a subclass of
10109 gimple_opt_pass along with...
10110 (pass_data_refactor_eh): ...new pass_data instance and...
10111 (make_pass_refactor_eh): ...new function.
10112 (pass_lower_resx): Convert from a global struct to a subclass of
10113 gimple_opt_pass along with...
10114 (pass_data_lower_resx): ...new pass_data instance and...
10115 (make_pass_lower_resx): ...new function.
10116 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10117 of gimple_opt_pass along with...
10118 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10119 (make_pass_lower_eh_dispatch): ...new function.
10120 (pass_cleanup_eh): Convert from a global struct to a subclass of
10121 gimple_opt_pass along with...
10122 (pass_data_cleanup_eh): ...new pass_data instance and...
10123 (make_pass_cleanup_eh): ...new function.
10124 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10125 to a subclass of simple_ipa_opt_pass along with...
10126 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10127 (make_pass_ipa_lower_emutls): ...new function.
10128 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10129 a subclass of gimple_opt_pass along with...
10130 (pass_data_if_conversion): ...new pass_data instance and...
10131 (make_pass_if_conversion): ...new function.
10132 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10133 subclass of gimple_opt_pass along with...
10134 (pass_data_build_ssa): ...new pass_data instance and...
10135 (make_pass_build_ssa): ...new function.
10136 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10137 global struct to a subclass of gimple_opt_pass along with...
10138 (pass_data_loop_distribution): ...new pass_data instance and...
10139 (make_pass_loop_distribution): ...new function.
10140 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10141 subclass of gimple_opt_pass along with...
10142 (pass_data_mudflap_1): ...new pass_data instance and...
10143 (make_pass_mudflap_1): ...new function.
10144 (pass_mudflap_2): Convert from a global struct to a subclass of
10145 gimple_opt_pass along with...
10146 (pass_data_mudflap_2): ...new pass_data instance and...
10147 (make_pass_mudflap_2): ...new function.
10148 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10149 subclass of gimple_opt_pass along with...
10150 (pass_data_mudflap_1): ...new pass_data instance and...
10151 (make_pass_mudflap_1): ...new function.
10152 (pass_mudflap_2): Convert from a global struct to a subclass of
10153 gimple_opt_pass along with...
10154 (pass_data_mudflap_2): ...new pass_data instance and...
10155 (make_pass_mudflap_2): ...new function.
10156 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10157 gimple_opt_pass along with...
10158 (pass_data_nrv): ...new pass_data instance and...
10159 (make_pass_nrv): ...new function.
10160 (pass_return_slot): Convert from a global struct to a subclass of
10161 gimple_opt_pass along with...
10162 (pass_data_return_slot): ...new pass_data instance and...
10163 (make_pass_return_slot): ...new function.
10164 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10165 to a subclass of gimple_opt_pass along with...
10166 (pass_data_object_sizes): ...new pass_data instance and...
10167 (make_pass_object_sizes): ...new function.
10168 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10169 global struct to a subclass of gimple_opt_pass along with...
10170 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10172 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10173 (pass_fixup_cfg): Convert from a global struct to a subclass of
10174 gimple_opt_pass along with...
10175 (pass_data_fixup_cfg): ...new pass_data instance and...
10176 (make_pass_fixup_cfg): ...new function.
10177 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10178 (make_pass_mudflap_1): ...new function.
10179 (pass_mudflap_2): Replace declaration with that of...
10180 (make_pass_mudflap_2): ...new function.
10181 (pass_asan): Replace declaration with that of...
10182 (make_pass_asan): ...new function.
10183 (pass_asan_O0): Replace declaration with that of...
10184 (make_pass_asan_O0): ...new function.
10185 (pass_tsan): Replace declaration with that of...
10186 (make_pass_tsan): ...new function.
10187 (pass_tsan_O0): Replace declaration with that of...
10188 (make_pass_tsan_O0): ...new function.
10189 (pass_lower_cf): Replace declaration with that of...
10190 (make_pass_lower_cf): ...new function.
10191 (pass_refactor_eh): Replace declaration with that of...
10192 (make_pass_refactor_eh): ...new function.
10193 (pass_lower_eh): Replace declaration with that of...
10194 (make_pass_lower_eh): ...new function.
10195 (pass_lower_eh_dispatch): Replace declaration with that of...
10196 (make_pass_lower_eh_dispatch): ...new function.
10197 (pass_lower_resx): Replace declaration with that of...
10198 (make_pass_lower_resx): ...new function.
10199 (pass_build_cfg): Replace declaration with that of...
10200 (make_pass_build_cfg): ...new function.
10201 (pass_early_tree_profile): Replace declaration with that of...
10202 (make_pass_early_tree_profile): ...new function.
10203 (pass_cleanup_eh): Replace declaration with that of...
10204 (make_pass_cleanup_eh): ...new function.
10205 (pass_sra): Replace declaration with that of...
10206 (make_pass_sra): ...new function.
10207 (pass_sra_early): Replace declaration with that of...
10208 (make_pass_sra_early): ...new function.
10209 (pass_early_ipa_sra): Replace declaration with that of...
10210 (make_pass_early_ipa_sra): ...new function.
10211 (pass_tail_recursion): Replace declaration with that of...
10212 (make_pass_tail_recursion): ...new function.
10213 (pass_tail_calls): Replace declaration with that of...
10214 (make_pass_tail_calls): ...new function.
10215 (pass_tree_loop): Replace declaration with that of...
10216 (make_pass_tree_loop): ...new function.
10217 (pass_tree_loop_init): Replace declaration with that of...
10218 (make_pass_tree_loop_init): ...new function.
10219 (pass_lim): Replace declaration with that of...
10220 (make_pass_lim): ...new function.
10221 (pass_tree_unswitch): Replace declaration with that of...
10222 (make_pass_tree_unswitch): ...new function.
10223 (pass_predcom): Replace declaration with that of...
10224 (make_pass_predcom): ...new function.
10225 (pass_iv_canon): Replace declaration with that of...
10226 (make_pass_iv_canon): ...new function.
10227 (pass_scev_cprop): Replace declaration with that of...
10228 (make_pass_scev_cprop): ...new function.
10229 (pass_empty_loop): Replace declaration with that of...
10230 (make_pass_empty_loop): ...new function.
10231 (pass_record_bounds): Replace declaration with that of...
10232 (make_pass_record_bounds): ...new function.
10233 (pass_graphite): Replace declaration with that of...
10234 (make_pass_graphite): ...new function.
10235 (pass_graphite_transforms): Replace declaration with that of...
10236 (make_pass_graphite_transforms): ...new function.
10237 (pass_if_conversion): Replace declaration with that of...
10238 (make_pass_if_conversion): ...new function.
10239 (pass_loop_distribution): Replace declaration with that of...
10240 (make_pass_loop_distribution): ...new function.
10241 (pass_vectorize): Replace declaration with that of...
10242 (make_pass_vectorize): ...new function.
10243 (pass_slp_vectorize): Replace declaration with that of...
10244 (make_pass_slp_vectorize): ...new function.
10245 (pass_complete_unroll): Replace declaration with that of...
10246 (make_pass_complete_unroll): ...new function.
10247 (pass_complete_unrolli): Replace declaration with that of...
10248 (make_pass_complete_unrolli): ...new function.
10249 (pass_parallelize_loops): Replace declaration with that of...
10250 (make_pass_parallelize_loops): ...new function.
10251 (pass_loop_prefetch): Replace declaration with that of...
10252 (make_pass_loop_prefetch): ...new function.
10253 (pass_iv_optimize): Replace declaration with that of...
10254 (make_pass_iv_optimize): ...new function.
10255 (pass_tree_loop_done): Replace declaration with that of...
10256 (make_pass_tree_loop_done): ...new function.
10257 (pass_ch): Replace declaration with that of...
10258 (make_pass_ch): ...new function.
10259 (pass_ccp): Replace declaration with that of...
10260 (make_pass_ccp): ...new function.
10261 (pass_phi_only_cprop): Replace declaration with that of...
10262 (make_pass_phi_only_cprop): ...new function.
10263 (pass_build_ssa): Replace declaration with that of...
10264 (make_pass_build_ssa): ...new function.
10265 (pass_build_alias): Replace declaration with that of...
10266 (make_pass_build_alias): ...new function.
10267 (pass_build_ealias): Replace declaration with that of...
10268 (make_pass_build_ealias): ...new function.
10269 (pass_dominator): Replace declaration with that of...
10270 (make_pass_dominator): ...new function.
10271 (pass_dce): Replace declaration with that of...
10272 (make_pass_dce): ...new function.
10273 (pass_dce_loop): Replace declaration with that of...
10274 (make_pass_dce_loop): ...new function.
10275 (pass_cd_dce): Replace declaration with that of...
10276 (make_pass_cd_dce): ...new function.
10277 (pass_call_cdce): Replace declaration with that of...
10278 (make_pass_call_cdce): ...new function.
10279 (pass_merge_phi): Replace declaration with that of...
10280 (make_pass_merge_phi): ...new function.
10281 (pass_split_crit_edges): Replace declaration with that of...
10282 (make_pass_split_crit_edges): ...new function.
10283 (pass_pre): Replace declaration with that of...
10284 (make_pass_pre): ...new function.
10285 (pass_profile): Replace declaration with that of...
10286 (make_pass_profile): ...new function.
10287 (pass_strip_predict_hints): Replace declaration with that of...
10288 (make_pass_strip_predict_hints): ...new function.
10289 (pass_lower_complex_O0): Replace declaration with that of...
10290 (make_pass_lower_complex_O0): ...new function.
10291 (pass_lower_complex): Replace declaration with that of...
10292 (make_pass_lower_complex): ...new function.
10293 (pass_lower_vector): Replace declaration with that of...
10294 (make_pass_lower_vector): ...new function.
10295 (pass_lower_vector_ssa): Replace declaration with that of...
10296 (make_pass_lower_vector_ssa): ...new function.
10297 (pass_lower_omp): Replace declaration with that of...
10298 (make_pass_lower_omp): ...new function.
10299 (pass_diagnose_omp_blocks): Replace declaration with that of...
10300 (make_pass_diagnose_omp_blocks): ...new function.
10301 (pass_expand_omp): Replace declaration with that of...
10302 (make_pass_expand_omp): ...new function.
10303 (pass_expand_omp_ssa): Replace declaration with that of...
10304 (make_pass_expand_omp_ssa): ...new function.
10305 (pass_object_sizes): Replace declaration with that of...
10306 (make_pass_object_sizes): ...new function.
10307 (pass_strlen): Replace declaration with that of...
10308 (make_pass_strlen): ...new function.
10309 (pass_fold_builtins): Replace declaration with that of...
10310 (make_pass_fold_builtins): ...new function.
10311 (pass_stdarg): Replace declaration with that of...
10312 (make_pass_stdarg): ...new function.
10313 (pass_early_warn_uninitialized): Replace declaration with that of...
10314 (make_pass_early_warn_uninitialized): ...new function.
10315 (pass_late_warn_uninitialized): Replace declaration with that of...
10316 (make_pass_late_warn_uninitialized): ...new function.
10317 (pass_cse_reciprocals): Replace declaration with that of...
10318 (make_pass_cse_reciprocals): ...new function.
10319 (pass_cse_sincos): Replace declaration with that of...
10320 (make_pass_cse_sincos): ...new function.
10321 (pass_optimize_bswap): Replace declaration with that of...
10322 (make_pass_optimize_bswap): ...new function.
10323 (pass_optimize_widening_mul): Replace declaration with that of...
10324 (make_pass_optimize_widening_mul): ...new function.
10325 (pass_warn_function_return): Replace declaration with that of...
10326 (make_pass_warn_function_return): ...new function.
10327 (pass_warn_function_noreturn): Replace declaration with that of...
10328 (make_pass_warn_function_noreturn): ...new function.
10329 (pass_cselim): Replace declaration with that of...
10330 (make_pass_cselim): ...new function.
10331 (pass_phiopt): Replace declaration with that of...
10332 (make_pass_phiopt): ...new function.
10333 (pass_forwprop): Replace declaration with that of...
10334 (make_pass_forwprop): ...new function.
10335 (pass_phiprop): Replace declaration with that of...
10336 (make_pass_phiprop): ...new function.
10337 (pass_tree_ifcombine): Replace declaration with that of...
10338 (make_pass_tree_ifcombine): ...new function.
10339 (pass_dse): Replace declaration with that of...
10340 (make_pass_dse): ...new function.
10341 (pass_nrv): Replace declaration with that of...
10342 (make_pass_nrv): ...new function.
10343 (pass_rename_ssa_copies): Replace declaration with that of...
10344 (make_pass_rename_ssa_copies): ...new function.
10345 (pass_sink_code): Replace declaration with that of...
10346 (make_pass_sink_code): ...new function.
10347 (pass_fre): Replace declaration with that of...
10348 (make_pass_fre): ...new function.
10349 (pass_check_data_deps): Replace declaration with that of...
10350 (make_pass_check_data_deps): ...new function.
10351 (pass_copy_prop): Replace declaration with that of...
10352 (make_pass_copy_prop): ...new function.
10353 (pass_vrp): Replace declaration with that of...
10354 (make_pass_vrp): ...new function.
10355 (pass_uncprop): Replace declaration with that of...
10356 (make_pass_uncprop): ...new function.
10357 (pass_return_slot): Replace declaration with that of...
10358 (make_pass_return_slot): ...new function.
10359 (pass_reassoc): Replace declaration with that of...
10360 (make_pass_reassoc): ...new function.
10361 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10362 (make_pass_rebuild_cgraph_edges): ...new function.
10363 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10364 (make_pass_remove_cgraph_callee_edges): ...new function.
10365 (pass_build_cgraph_edges): Replace declaration with that of...
10366 (make_pass_build_cgraph_edges): ...new function.
10367 (pass_local_pure_const): Replace declaration with that of...
10368 (make_pass_local_pure_const): ...new function.
10369 (pass_tracer): Replace declaration with that of...
10370 (make_pass_tracer): ...new function.
10371 (pass_warn_unused_result): Replace declaration with that of...
10372 (make_pass_warn_unused_result): ...new function.
10373 (pass_diagnose_tm_blocks): Replace declaration with that of...
10374 (make_pass_diagnose_tm_blocks): ...new function.
10375 (pass_lower_tm): Replace declaration with that of...
10376 (make_pass_lower_tm): ...new function.
10377 (pass_tm_init): Replace declaration with that of...
10378 (make_pass_tm_init): ...new function.
10379 (pass_tm_mark): Replace declaration with that of...
10380 (make_pass_tm_mark): ...new function.
10381 (pass_tm_memopt): Replace declaration with that of...
10382 (make_pass_tm_memopt): ...new function.
10383 (pass_tm_edges): Replace declaration with that of...
10384 (make_pass_tm_edges): ...new function.
10385 (pass_split_functions): Replace declaration with that of...
10386 (make_pass_split_functions): ...new function.
10387 (pass_feedback_split_functions): Replace declaration with that of...
10388 (make_pass_feedback_split_functions): ...new function.
10389 (pass_strength_reduction): Replace declaration with that of...
10390 (make_pass_strength_reduction): ...new function.
10391 (pass_ipa_lower_emutls): Replace declaration with that of...
10392 (make_pass_ipa_lower_emutls): ...new function.
10393 (pass_ipa_function_and_variable_visibility): Replace declaration with
10395 (make_pass_ipa_function_and_variable_visibility): ...new function.
10396 (pass_ipa_tree_profile): Replace declaration with that of...
10397 (make_pass_ipa_tree_profile): ...new function.
10398 (pass_early_local_passes): Replace declaration with that of...
10399 (make_pass_early_local_passes): ...new function.
10400 (pass_ipa_whole_program_visibility): Replace declaration with that
10402 (make_pass_ipa_whole_program_visibility): ...new function.
10403 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10404 (make_pass_ipa_lto_gimple_out): ...new function.
10405 (pass_ipa_increase_alignment): Replace declaration with that of...
10406 (make_pass_ipa_increase_alignment): ...new function.
10407 (pass_ipa_inline): Replace declaration with that of...
10408 (make_pass_ipa_inline): ...new function.
10409 (pass_ipa_free_lang_data): Replace declaration with that of...
10410 (make_pass_ipa_free_lang_data): ...new function.
10411 (pass_ipa_free_inline_summary): Replace declaration with that of...
10412 (make_pass_ipa_free_inline_summary): ...new function.
10413 (pass_ipa_cp): Replace declaration with that of...
10414 (make_pass_ipa_cp): ...new function.
10415 (pass_ipa_reference): Replace declaration with that of...
10416 (make_pass_ipa_reference): ...new function.
10417 (pass_ipa_pure_const): Replace declaration with that of...
10418 (make_pass_ipa_pure_const): ...new function.
10419 (pass_ipa_pta): Replace declaration with that of...
10420 (make_pass_ipa_pta): ...new function.
10421 (pass_ipa_lto_finish_out): Replace declaration with that of...
10422 (make_pass_ipa_lto_finish_out): ...new function.
10423 (pass_ipa_tm): Replace declaration with that of...
10424 (make_pass_ipa_tm): ...new function.
10425 (pass_ipa_profile): Replace declaration with that of...
10426 (make_pass_ipa_profile): ...new function.
10427 (pass_ipa_cdtor_merge): Replace declaration with that of...
10428 (make_pass_ipa_cdtor_merge): ...new function.
10429 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10431 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10432 (pass_init_datastructures): Replace declaration with that of...
10433 (make_pass_init_datastructures): ...new function.
10434 (pass_fixup_cfg): Replace declaration with that of...
10435 (make_pass_fixup_cfg): ...new function.
10436 (pass_expand): Replace declaration with that of...
10437 (make_pass_expand): ...new function.
10438 (pass_instantiate_virtual_regs): Replace declaration with that of...
10439 (make_pass_instantiate_virtual_regs): ...new function.
10440 (pass_rtl_fwprop): Replace declaration with that of...
10441 (make_pass_rtl_fwprop): ...new function.
10442 (pass_rtl_fwprop_addr): Replace declaration with that of...
10443 (make_pass_rtl_fwprop_addr): ...new function.
10444 (pass_jump): Replace declaration with that of...
10445 (make_pass_jump): ...new function.
10446 (pass_jump2): Replace declaration with that of...
10447 (make_pass_jump2): ...new function.
10448 (pass_lower_subreg): Replace declaration with that of...
10449 (make_pass_lower_subreg): ...new function.
10450 (pass_cse): Replace declaration with that of...
10451 (make_pass_cse): ...new function.
10452 (pass_fast_rtl_dce): Replace declaration with that of...
10453 (make_pass_fast_rtl_dce): ...new function.
10454 (pass_ud_rtl_dce): Replace declaration with that of...
10455 (make_pass_ud_rtl_dce): ...new function.
10456 (pass_rtl_dce): Replace declaration with that of...
10457 (make_pass_rtl_dce): ...new function.
10458 (pass_rtl_dse1): Replace declaration with that of...
10459 (make_pass_rtl_dse1): ...new function.
10460 (pass_rtl_dse2): Replace declaration with that of...
10461 (make_pass_rtl_dse2): ...new function.
10462 (pass_rtl_dse3): Replace declaration with that of...
10463 (make_pass_rtl_dse3): ...new function.
10464 (pass_rtl_cprop): Replace declaration with that of...
10465 (make_pass_rtl_cprop): ...new function.
10466 (pass_rtl_pre): Replace declaration with that of...
10467 (make_pass_rtl_pre): ...new function.
10468 (pass_rtl_hoist): Replace declaration with that of...
10469 (make_pass_rtl_hoist): ...new function.
10470 (pass_rtl_store_motion): Replace declaration with that of...
10471 (make_pass_rtl_store_motion): ...new function.
10472 (pass_cse_after_global_opts): Replace declaration with that of...
10473 (make_pass_cse_after_global_opts): ...new function.
10474 (pass_rtl_ifcvt): Replace declaration with that of...
10475 (make_pass_rtl_ifcvt): ...new function.
10476 (pass_into_cfg_layout_mode): Replace declaration with that of...
10477 (make_pass_into_cfg_layout_mode): ...new function.
10478 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10479 (make_pass_outof_cfg_layout_mode): ...new function.
10480 (pass_loop2): Replace declaration with that of...
10481 (make_pass_loop2): ...new function.
10482 (pass_rtl_loop_init): Replace declaration with that of...
10483 (make_pass_rtl_loop_init): ...new function.
10484 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10485 (make_pass_rtl_move_loop_invariants): ...new function.
10486 (pass_rtl_unswitch): Replace declaration with that of...
10487 (make_pass_rtl_unswitch): ...new function.
10488 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10489 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10490 (pass_rtl_doloop): Replace declaration with that of...
10491 (make_pass_rtl_doloop): ...new function.
10492 (pass_rtl_loop_done): Replace declaration with that of...
10493 (make_pass_rtl_loop_done): ...new function.
10494 (pass_web): Replace declaration with that of...
10495 (make_pass_web): ...new function.
10496 (pass_cse2): Replace declaration with that of...
10497 (make_pass_cse2): ...new function.
10498 (pass_df_initialize_opt): Replace declaration with that of...
10499 (make_pass_df_initialize_opt): ...new function.
10500 (pass_df_initialize_no_opt): Replace declaration with that of...
10501 (make_pass_df_initialize_no_opt): ...new function.
10502 (pass_reginfo_init): Replace declaration with that of...
10503 (make_pass_reginfo_init): ...new function.
10504 (pass_inc_dec): Replace declaration with that of...
10505 (make_pass_inc_dec): ...new function.
10506 (pass_stack_ptr_mod): Replace declaration with that of...
10507 (make_pass_stack_ptr_mod): ...new function.
10508 (pass_initialize_regs): Replace declaration with that of...
10509 (make_pass_initialize_regs): ...new function.
10510 (pass_combine): Replace declaration with that of...
10511 (make_pass_combine): ...new function.
10512 (pass_if_after_combine): Replace declaration with that of...
10513 (make_pass_if_after_combine): ...new function.
10514 (pass_ree): Replace declaration with that of...
10515 (make_pass_ree): ...new function.
10516 (pass_partition_blocks): Replace declaration with that of...
10517 (make_pass_partition_blocks): ...new function.
10518 (pass_match_asm_constraints): Replace declaration with that of...
10519 (make_pass_match_asm_constraints): ...new function.
10520 (pass_regmove): Replace declaration with that of...
10521 (make_pass_regmove): ...new function.
10522 (pass_split_all_insns): Replace declaration with that of...
10523 (make_pass_split_all_insns): ...new function.
10524 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10525 (make_pass_fast_rtl_byte_dce): ...new function.
10526 (pass_lower_subreg2): Replace declaration with that of...
10527 (make_pass_lower_subreg2): ...new function.
10528 (pass_mode_switching): Replace declaration with that of...
10529 (make_pass_mode_switching): ...new function.
10530 (pass_sms): Replace declaration with that of...
10531 (make_pass_sms): ...new function.
10532 (pass_sched): Replace declaration with that of...
10533 (make_pass_sched): ...new function.
10534 (pass_ira): Replace declaration with that of...
10535 (make_pass_ira): ...new function.
10536 (pass_reload): Replace declaration with that of...
10537 (make_pass_reload): ...new function.
10538 (pass_clean_state): Replace declaration with that of...
10539 (make_pass_clean_state): ...new function.
10540 (pass_branch_prob): Replace declaration with that of...
10541 (make_pass_branch_prob): ...new function.
10542 (pass_value_profile_transformations): Replace declaration with that
10544 (make_pass_value_profile_transformations): ...new function.
10545 (pass_postreload_cse): Replace declaration with that of...
10546 (make_pass_postreload_cse): ...new function.
10547 (pass_gcse2): Replace declaration with that of...
10548 (make_pass_gcse2): ...new function.
10549 (pass_split_after_reload): Replace declaration with that of...
10550 (make_pass_split_after_reload): ...new function.
10551 (pass_branch_target_load_optimize1): Replace declaration with that
10553 (make_pass_branch_target_load_optimize1): ...new function.
10554 (pass_thread_prologue_and_epilogue): Replace declaration with that
10556 (make_pass_thread_prologue_and_epilogue): ...new function.
10557 (pass_stack_adjustments): Replace declaration with that of...
10558 (make_pass_stack_adjustments): ...new function.
10559 (pass_peephole2): Replace declaration with that of...
10560 (make_pass_peephole2): ...new function.
10561 (pass_if_after_reload): Replace declaration with that of...
10562 (make_pass_if_after_reload): ...new function.
10563 (pass_regrename): Replace declaration with that of...
10564 (make_pass_regrename): ...new function.
10565 (pass_cprop_hardreg): Replace declaration with that of...
10566 (make_pass_cprop_hardreg): ...new function.
10567 (pass_reorder_blocks): Replace declaration with that of...
10568 (make_pass_reorder_blocks): ...new function.
10569 (pass_branch_target_load_optimize2): Replace declaration with that
10571 (make_pass_branch_target_load_optimize2): ...new function.
10572 (pass_leaf_regs): Replace declaration with that of...
10573 (make_pass_leaf_regs): ...new function.
10574 (pass_split_before_sched2): Replace declaration with that of...
10575 (make_pass_split_before_sched2): ...new function.
10576 (pass_compare_elim_after_reload): Replace declaration with that of...
10577 (make_pass_compare_elim_after_reload): ...new function.
10578 (pass_sched2): Replace declaration with that of...
10579 (make_pass_sched2): ...new function.
10580 (pass_stack_regs): Replace declaration with that of...
10581 (make_pass_stack_regs): ...new function.
10582 (pass_stack_regs_run): Replace declaration with that of...
10583 (make_pass_stack_regs_run): ...new function.
10584 (pass_df_finish): Replace declaration with that of...
10585 (make_pass_df_finish): ...new function.
10586 (pass_compute_alignments): Replace declaration with that of...
10587 (make_pass_compute_alignments): ...new function.
10588 (pass_duplicate_computed_gotos): Replace declaration with that of...
10589 (make_pass_duplicate_computed_gotos): ...new function.
10590 (pass_variable_tracking): Replace declaration with that of...
10591 (make_pass_variable_tracking): ...new function.
10592 (pass_free_cfg): Replace declaration with that of...
10593 (make_pass_free_cfg): ...new function.
10594 (pass_machine_reorg): Replace declaration with that of...
10595 (make_pass_machine_reorg): ...new function.
10596 (pass_cleanup_barriers): Replace declaration with that of...
10597 (make_pass_cleanup_barriers): ...new function.
10598 (pass_delay_slots): Replace declaration with that of...
10599 (make_pass_delay_slots): ...new function.
10600 (pass_split_for_shorten_branches): Replace declaration with that of...
10601 (make_pass_split_for_shorten_branches): ...new function.
10602 (pass_split_before_regstack): Replace declaration with that of...
10603 (make_pass_split_before_regstack): ...new function.
10604 (pass_convert_to_eh_region_ranges): Replace declaration with that
10606 (make_pass_convert_to_eh_region_ranges): ...new function.
10607 (pass_shorten_branches): Replace declaration with that of...
10608 (make_pass_shorten_branches): ...new function.
10609 (pass_set_nothrow_function_flags): Replace declaration with that of...
10610 (make_pass_set_nothrow_function_flags): ...new function.
10611 (pass_dwarf2_frame): Replace declaration with that of...
10612 (make_pass_dwarf2_frame): ...new function.
10613 (pass_final): Replace declaration with that of...
10614 (make_pass_final): ...new function.
10615 (pass_rtl_seqabstr): Replace declaration with that of...
10616 (make_pass_rtl_seqabstr): ...new function.
10617 (pass_release_ssa_names): Replace declaration with that of...
10618 (make_pass_release_ssa_names): ...new function.
10619 (pass_early_inline): Replace declaration with that of...
10620 (make_pass_early_inline): ...new function.
10621 (pass_inline_parameters): Replace declaration with that of...
10622 (make_pass_inline_parameters): ...new function.
10623 (pass_update_address_taken): Replace declaration with that of...
10624 (make_pass_update_address_taken): ...new function.
10625 (pass_convert_switch): Replace declaration with that of...
10626 (make_pass_convert_switch): ...new function.
10627 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10628 to a subclass of simple_ipa_opt_pass along with...
10629 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10630 (make_pass_ipa_tree_profile): ...new function.
10631 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10632 subclass of gimple_opt_pass along with...
10633 (pass_data_sra_early): ...new pass_data instance and...
10634 (make_pass_sra_early): ...new function.
10635 (pass_sra): Convert from a global struct to a subclass of
10636 gimple_opt_pass along with...
10637 (pass_data_sra): ...new pass_data instance and...
10638 (make_pass_sra): ...new function.
10639 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10640 gimple_opt_pass along with...
10641 (pass_data_early_ipa_sra): ...new pass_data instance and...
10642 (make_pass_early_ipa_sra): ...new function.
10643 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10644 subclass of gimple_opt_pass along with...
10645 (pass_data_ccp): ...new pass_data instance and...
10646 (make_pass_ccp): ...new function.
10647 (pass_fold_builtins): Convert from a global struct to a subclass of
10648 gimple_opt_pass along with...
10649 (pass_data_fold_builtins): ...new pass_data instance and...
10650 (make_pass_fold_builtins): ...new function.
10651 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10652 subclass of gimple_opt_pass along with...
10653 (pass_data_copy_prop): ...new pass_data instance and...
10654 (make_pass_copy_prop): ...new function.
10655 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10656 global struct to a subclass of gimple_opt_pass along with...
10657 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10658 (make_pass_rename_ssa_copies): ...new function.
10659 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10660 subclass of gimple_opt_pass along with...
10661 (pass_data_dce): ...new pass_data instance and...
10662 (make_pass_dce): ...new function.
10663 (pass_dce_loop): Convert from a global struct to a subclass of
10664 gimple_opt_pass along with...
10665 (pass_data_dce_loop): ...new pass_data instance and...
10666 (make_pass_dce_loop): ...new function.
10667 (pass_cd_dce): Convert from a global struct to a subclass of
10668 gimple_opt_pass along with...
10669 (pass_data_cd_dce): ...new pass_data instance and...
10670 (make_pass_cd_dce): ...new function.
10671 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10672 subclass of gimple_opt_pass along with...
10673 (pass_data_dominator): ...new pass_data instance and...
10674 (make_pass_dominator): ...new function.
10675 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10676 gimple_opt_pass along with...
10677 (pass_data_phi_only_cprop): ...new pass_data instance and...
10678 (make_pass_phi_only_cprop): ...new function.
10679 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10680 subclass of gimple_opt_pass along with...
10681 (pass_data_dse): ...new pass_data instance and...
10682 (make_pass_dse): ...new function.
10683 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10684 a subclass of gimple_opt_pass along with...
10685 (pass_data_forwprop): ...new pass_data instance and...
10686 (make_pass_forwprop): ...new function.
10687 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10688 struct to a subclass of gimple_opt_pass along with...
10689 (pass_data_tree_ifcombine): ...new pass_data instance and...
10690 (make_pass_tree_ifcombine): ...new function.
10691 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10692 subclass of gimple_opt_pass along with...
10693 (pass_data_ch): ...new pass_data instance and...
10694 (make_pass_ch): ...new function.
10695 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10696 subclass of gimple_opt_pass along with...
10697 (pass_data_tree_loop): ...new pass_data instance and...
10698 (make_pass_tree_loop): ...new function.
10699 (pass_tree_loop_init): Convert from a global struct to a subclass of
10700 gimple_opt_pass along with...
10701 (pass_data_tree_loop_init): ...new pass_data instance and...
10702 (make_pass_tree_loop_init): ...new function.
10703 (pass_lim): Convert from a global struct to a subclass of
10704 gimple_opt_pass along with...
10705 (pass_data_lim): ...new pass_data instance and...
10706 (make_pass_lim): ...new function.
10707 (pass_tree_unswitch): Convert from a global struct to a subclass of
10708 gimple_opt_pass along with...
10709 (pass_data_tree_unswitch): ...new pass_data instance and...
10710 (make_pass_tree_unswitch): ...new function.
10711 (pass_predcom): Convert from a global struct to a subclass of
10712 gimple_opt_pass along with...
10713 (pass_data_predcom): ...new pass_data instance and...
10714 (make_pass_predcom): ...new function.
10715 (pass_vectorize): Convert from a global struct to a subclass of
10716 gimple_opt_pass along with...
10717 (pass_data_vectorize): ...new pass_data instance and...
10718 (make_pass_vectorize): ...new function.
10719 (pass_graphite): Convert from a global struct to a subclass of
10720 gimple_opt_pass along with...
10721 (pass_data_graphite): ...new pass_data instance and...
10722 (make_pass_graphite): ...new function.
10723 (pass_graphite_transforms): Convert from a global struct to a subclass
10724 of gimple_opt_pass along with...
10725 (pass_data_graphite_transforms): ...new pass_data instance and...
10726 (make_pass_graphite_transforms): ...new function.
10727 (pass_check_data_deps): Convert from a global struct to a subclass of
10728 gimple_opt_pass along with...
10729 (pass_data_check_data_deps): ...new pass_data instance and...
10730 (make_pass_check_data_deps): ...new function.
10731 (pass_iv_canon): Convert from a global struct to a subclass of
10732 gimple_opt_pass along with...
10733 (pass_data_iv_canon): ...new pass_data instance and...
10734 (make_pass_iv_canon): ...new function.
10735 (pass_scev_cprop): Convert from a global struct to a subclass of
10736 gimple_opt_pass along with...
10737 (pass_data_scev_cprop): ...new pass_data instance and...
10738 (make_pass_scev_cprop): ...new function.
10739 (pass_record_bounds): Convert from a global struct to a subclass of
10740 gimple_opt_pass along with...
10741 (pass_data_record_bounds): ...new pass_data instance and...
10742 (make_pass_record_bounds): ...new function.
10743 (pass_complete_unroll): Convert from a global struct to a subclass of
10744 gimple_opt_pass along with...
10745 (pass_data_complete_unroll): ...new pass_data instance and...
10746 (make_pass_complete_unroll): ...new function.
10747 (pass_complete_unrolli): Convert from a global struct to a subclass of
10748 gimple_opt_pass along with...
10749 (pass_data_complete_unrolli): ...new pass_data instance and...
10750 (make_pass_complete_unrolli): ...new function.
10751 (pass_parallelize_loops): Convert from a global struct to a subclass
10752 of gimple_opt_pass along with...
10753 (pass_data_parallelize_loops): ...new pass_data instance and...
10754 (make_pass_parallelize_loops): ...new function.
10755 (pass_loop_prefetch): Convert from a global struct to a subclass of
10756 gimple_opt_pass along with...
10757 (pass_data_loop_prefetch): ...new pass_data instance and...
10758 (make_pass_loop_prefetch): ...new function.
10759 (pass_iv_optimize): Convert from a global struct to a subclass of
10760 gimple_opt_pass along with...
10761 (pass_data_iv_optimize): ...new pass_data instance and...
10762 (make_pass_iv_optimize): ...new function.
10763 (pass_tree_loop_done): Convert from a global struct to a subclass of
10764 gimple_opt_pass along with...
10765 (pass_data_tree_loop_done): ...new pass_data instance and...
10766 (make_pass_tree_loop_done): ...new function.
10767 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10768 struct to a subclass of gimple_opt_pass along with...
10769 (pass_data_cse_reciprocals): ...new pass_data instance and...
10770 (make_pass_cse_reciprocals): ...new function.
10771 (pass_cse_sincos): Convert from a global struct to a subclass of
10772 gimple_opt_pass along with...
10773 (pass_data_cse_sincos): ...new pass_data instance and...
10774 (make_pass_cse_sincos): ...new function.
10775 (pass_optimize_bswap): Convert from a global struct to a subclass of
10776 gimple_opt_pass along with...
10777 (pass_data_optimize_bswap): ...new pass_data instance and...
10778 (make_pass_optimize_bswap): ...new function.
10779 (pass_optimize_widening_mul): Convert from a global struct to a
10780 subclass of gimple_opt_pass along with...
10781 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10782 (make_pass_optimize_widening_mul): ...new function.
10783 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10784 subclass of gimple_opt_pass along with...
10785 (pass_data_phiopt): ...new pass_data instance and...
10786 (make_pass_phiopt): ...new function.
10787 (pass_cselim): Convert from a global struct to a subclass of
10788 gimple_opt_pass along with...
10789 (pass_data_cselim): ...new pass_data instance and...
10790 (make_pass_cselim): ...new function.
10791 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10792 subclass of gimple_opt_pass along with...
10793 (pass_data_phiprop): ...new pass_data instance and...
10794 (make_pass_phiprop): ...new function.
10795 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10796 subclass of gimple_opt_pass along with...
10797 (pass_data_pre): ...new pass_data instance and...
10798 (make_pass_pre): ...new function.
10799 (pass_fre): Convert from a global struct to a subclass of
10800 gimple_opt_pass along with...
10801 (pass_data_fre): ...new pass_data instance and...
10802 (make_pass_fre): ...new function.
10803 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10804 subclass of gimple_opt_pass along with...
10805 (pass_data_reassoc): ...new pass_data instance and...
10806 (make_pass_reassoc): ...new function.
10807 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10808 subclass of gimple_opt_pass along with...
10809 (pass_data_sink_code): ...new pass_data instance and...
10810 (make_pass_sink_code): ...new function.
10811 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10812 subclass of gimple_opt_pass along with...
10813 (pass_data_strlen): ...new pass_data instance and...
10814 (make_pass_strlen): ...new function.
10815 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10816 struct to a subclass of gimple_opt_pass along with...
10817 (pass_data_build_alias): ...new pass_data instance and...
10818 (make_pass_build_alias): ...new function.
10819 (pass_build_ealias): Convert from a global struct to a subclass of
10820 gimple_opt_pass along with...
10821 (pass_data_build_ealias): ...new pass_data instance and...
10822 (make_pass_build_ealias): ...new function.
10823 (pass_ipa_pta): Convert from a global struct to a subclass of
10824 simple_ipa_opt_pass along with...
10825 (pass_data_ipa_pta): ...new pass_data instance and...
10826 (make_pass_ipa_pta): ...new function.
10827 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10828 subclass of gimple_opt_pass along with...
10829 (pass_data_uncprop): ...new pass_data instance and...
10830 (make_pass_uncprop): ...new function.
10831 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10832 global struct to a subclass of gimple_opt_pass along with...
10833 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10834 (make_pass_late_warn_uninitialized): ...new function.
10835 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10836 to a subclass of gimple_opt_pass along with...
10837 (pass_data_init_datastructures): ...new pass_data instance and...
10838 (make_pass_init_datastructures): ...new function.
10839 (pass_early_warn_uninitialized): Convert from a global struct to a
10840 subclass of gimple_opt_pass along with...
10841 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10842 (make_pass_early_warn_uninitialized): ...new function.
10843 (pass_update_address_taken): Convert from a global struct to a
10844 subclass of gimple_opt_pass along with...
10845 (pass_data_update_address_taken): ...new pass_data instance and...
10846 (make_pass_update_address_taken): ...new function.
10847 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10848 struct to a subclass of gimple_opt_pass along with...
10849 (pass_data_release_ssa_names): ...new pass_data instance and...
10850 (make_pass_release_ssa_names): ...new function.
10851 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10852 subclass of gimple_opt_pass along with...
10853 (pass_data_stdarg): ...new pass_data instance and...
10854 (make_pass_stdarg): ...new function.
10855 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10856 global struct to a subclass of gimple_opt_pass along with...
10857 (pass_data_convert_switch): ...new pass_data instance and...
10858 (make_pass_convert_switch): ...new function.
10859 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10860 to a subclass of gimple_opt_pass along with...
10861 (pass_data_tail_recursion): ...new pass_data instance and...
10862 (make_pass_tail_recursion): ...new function.
10863 (pass_tail_calls): Convert from a global struct to a subclass of
10864 gimple_opt_pass along with...
10865 (pass_data_tail_calls): ...new pass_data instance and...
10866 (make_pass_tail_calls): ...new function.
10867 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10868 struct to a subclass of gimple_opt_pass along with...
10869 (pass_data_lower_vector): ...new pass_data instance and...
10870 (make_pass_lower_vector): ...new function.
10871 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10872 gimple_opt_pass along with...
10873 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10874 (make_pass_lower_vector_ssa): ...new function.
10875 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10876 to a subclass of gimple_opt_pass along with...
10877 (pass_data_slp_vectorize): ...new pass_data instance and...
10878 (make_pass_slp_vectorize): ...new function.
10879 (pass_ipa_increase_alignment): Convert from a global struct to a
10880 subclass of simple_ipa_opt_pass along with...
10881 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10882 (make_pass_ipa_increase_alignment): ...new function.
10883 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10884 gimple_opt_pass along with...
10885 (pass_data_vrp): ...new pass_data instance and...
10886 (make_pass_vrp): ...new function.
10887 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10888 subclass of simple_ipa_opt_pass along with...
10889 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10890 (make_pass_ipa_free_lang_data): ...new function.
10891 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10892 gimple_opt_pass along with...
10893 (pass_data_tsan): ...new pass_data instance and...
10894 (make_pass_tsan): ...new function.
10895 (pass_tsan_O0): Convert from a global struct to a subclass of
10896 gimple_opt_pass along with...
10897 (pass_data_tsan_O0): ...new pass_data instance and...
10898 (make_pass_tsan_O0): ...new function.
10899 * var-tracking.c (pass_variable_tracking): Convert from a global
10900 struct to a subclass of rtl_opt_pass along with...
10901 (pass_data_variable_tracking): ...new pass_data instance and...
10902 (make_pass_variable_tracking): ...new function.
10903 * web.c (pass_web): Convert from a global struct to a subclass of
10904 rtl_opt_pass along with...
10905 (pass_data_web): ...new pass_data instance and...
10906 (make_pass_web): ...new function.
10907 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10908 declaration with that of...
10909 (make_pass_mode_switch_use): ...new function.
10910 (pass_resolve_sw_modes): Replace declaration with that of...
10911 (make_pass_resolve_sw_modes): ...new function.
10912 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10913 from a global struct to a subclass of rtl_opt_pass along with...
10914 (pass_data_mode_switch_use): ...new pass_data instance and...
10915 (make_pass_mode_switch_use): ...new function.
10916 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10917 from a global struct to a subclass of rtl_opt_pass along with...
10918 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10919 (make_pass_resolve_sw_modes): ...new function.
10920 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10921 struct to a subclass of rtl_opt_pass along with...
10922 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10923 (make_pass_insert_vzeroupper): ...new function.
10924 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10925 global struct to a subclass of rtl_opt_pass along with...
10926 (pass_data_work_around_errata): ...new pass_data instance and...
10927 (make_pass_work_around_errata): ...new function.
10928 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10929 struct to a subclass of rtl_opt_pass along with...
10930 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10931 (make_pass_mips_machine_reorg2): ...new function.
10933 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10935 * passes.c (pass_manager::operator new): New.
10937 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10939 Handwritten part of conversion of passes to C++ classes.
10941 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10942 (toplev.o): Add dep on PASS_MANAGER_H.
10943 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10944 of early local pases to reflect this moving from a global to a
10945 member of gcc::pass_manager.
10946 (cgraph_add_new_function): Likewise.
10947 * lto-cgraph.c (lto_output_node): Update for conversion of
10948 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10949 * passes.c (opt_pass::clone): New.
10950 (opt_pass::gate): New.
10951 (opt_pass::execute): New.
10952 (opt_pass::opt_pass): New.
10953 (pass_manager::execute_early_local_passes): New.
10954 (pass_manager::execute_pass_mode_switching): new.
10955 (finish_optimization_passes): Convert to...
10956 (pass_manager::finish_optimization_passes): ...this.
10957 (finish_optimization_passes): Update for conversion of passes to
10959 (register_dump_files_1): Use has_gate since we cannot portably
10960 check a vtable entry against NULL.
10961 (dump_one_pass): Likewise.
10962 (ipa_write_summaries_2): Likewise.
10963 (ipa_write_optimization_summaries_1): Likewise.
10964 (ipa_read_summaries_1): Likewise.
10965 (ipa_read_optimization_summaries_1): Likewise.
10966 (execute_ipa_stmt_fixups): Likewise.
10967 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10968 pass-creation functions rather than wiring up globals, and
10969 storing the results in fields of pass_manager generated using
10970 pass-instances.def.
10971 (pass_manager::dump_profile_report): Update for conversion of
10972 passes to C++ classes.
10973 (pass_manager::execute_ipa_summary_passes): Likewise.
10974 (execute_one_ipa_transform_pass): Likewise.
10975 (execute_one_pass): Use has_gate and has_execute since we cannot
10976 portably check a vtable entry against NULL.
10977 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10978 (pass_manager): Use pass-instances.def to add fields for the
10979 various pass instances.
10980 * toplev.c (finalize): Update for move of
10981 finish_optimization_passes to a method of gcc::pass_manager.
10982 * toplev.h (finish_optimization_passes): Move to method of class
10984 * tree-pass.h (struct pass_data): New.
10985 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10986 (opt_pass::gate): Convert to virtual function.
10987 (opt_pass::~opt_pass): New.
10988 (opt_pass::clone): New.
10989 (opt_pass::execute): Convert to virtual function.
10990 (opt_pass::opt_pass): New.
10991 (opt_pass::ctxt_): new.
10992 (gimple_opt_pass): Convert to subclass of opt_pass.
10993 (gimple_opt_pass::gimple_opt_pass): New.
10994 (rtl_opt_pass): Convert to subclass of opt_pass.
10995 (rtl_opt_pass::rtl_opt_pass): New.
10996 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10997 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10998 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10999 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11000 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11001 invocation of pass_mode_switching to reflect this moving from a
11002 global to a member of gcc::pass_manager.
11003 (ix86_option_override): Rework how pass_insert_vzeroupper is
11004 added to the pass_manager to reflect autogenerated changes.
11005 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11007 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11009 PR rtl-optimization/57708
11010 * recog.c (peep2_find_free_register): Validate all regs in a
11013 2013-08-05 Jan Hubicka <jh@suse.cz>
11016 * cgraph.c (verify_cgraph_node): Accept local flags from other
11018 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11019 (function_and_variable_visibility): Likewise.
11020 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11022 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11024 * graph.c (init_graph_slim_pretty_print): Remove.
11025 (print_graph_cfg): Do not call it. Use local pretty printer.
11026 (start_graph_dump): Likewise.
11028 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11030 * gimple-pretty-print.c (buffer): Remove.
11031 (initialized): Likewise.
11032 (maybe_init_pretty_print): Likewise.
11033 (print_gimple_stmt): Do not call it. Use non-static local
11034 pretty_printer variable.
11035 (print_gimple_expr): Likewise.
11036 (print_gimple_seq): Likewise.
11037 (gimple_dump_bb): Likewise.
11039 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11041 * asan.c (asan_pp): Remove.
11042 (asan_pp_initialized): Likewise.
11043 (asan_pp_initialize): Likewise.
11044 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11045 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11046 (asan_add_global): Likewise.
11048 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11050 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11051 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11052 * pretty-print.c (pp_formatted_text_data): Likewise.
11053 (pp_indent): Rename from pp_base_indent.
11054 (pp_format): Rename from pp_base_format.
11055 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11056 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11057 (pp_flush): Rename from pp_base_flush.
11058 (pp_set_line_maximum_length): Rename from
11059 pp_base_set_line_maximum_length.
11060 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11061 (pp_set_prefix): Rename from pp_base_set_prefix.
11062 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11063 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11064 (pp_append_text): Rename from pp_base_append_text.
11065 (pp_formatted_text): Rename from pp_base_formatted_text.
11066 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11067 (pp_remaining_character_count_for_line): Rename from
11068 pp_base_remaining_character_count_for_line.
11069 (pp_newline): Rename from pp_base_newline.
11070 (pp_character): Rename from pp_base_character.
11071 (pp_string): Rename from pp_base_string.
11072 (pp_maybe_space): Rename from pp_base_maybe_space.
11073 * asan.c (asan_pp_string): Adjust.
11074 (asan_emit_stack_protection): Likewise.
11075 (asan_add_global): Likewise.
11076 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11077 * tree-mudflap.c (mf_varname_tree): Likewise.
11078 * tree-pretty-print.c (pp_tree_identifier): Rename from
11079 pp_base_tree_identifier.
11080 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11081 Declare as function.
11083 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11085 * pretty-print.h (pp_bar_bar): New.
11086 (pp_ampersand_ampersand): Likewise.
11087 (pp_less_equal): Likewise.
11088 (pp_greater_equal): Likewise.
11089 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11090 printer functions instead of pp_string or operators and punctuators.
11091 (dump_gimple_call): Likewise.
11092 (dump_gimple_omp_for): Likewise.
11093 (dump_gimple_transaction): Likewise.
11094 (dump_gimple_phi): Likewise.
11095 (pp_gimple_stmt_1): Likewise.
11096 * sched-vis.c (print_insn): Likewise.
11097 * tree-mudflap.c (mf_varname_tree): Likewise.
11098 * tree-pretty-print.c (dump_block_node): Likewise.
11099 (dump_generic_node): Likewise.
11101 2013-08-02 Jan Hubicka <jh@suse.cz>
11103 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11105 * lto-streamer-out.c (tree_is_indexable): Results decls and
11106 parm decls are not indexable.
11107 (DFS_write_tree_body): Do not follow args and results.
11108 (hash_tree): Likewise.
11109 (output_functions): Rearrange so struct function is needed
11110 only when real body is output; be able to also ouptut abstract
11111 functions; output DECL_ARGUMENTS and DECL_RESULT.
11112 (lto_output): When not in WPA, ale store abstract functions.
11113 (write_symbol): Do not care about RESULT_DECL.
11114 (output_symbol_p): Handle correctly sbtract decls.
11115 * lto-streamer-in.c (input_function): Rearrange so struct
11116 function can be NULL at entry; allow streaming of
11117 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11118 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11119 sanity check during LTO.
11120 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11121 RESULT_DECl and DECL_ARGUMENTS.
11122 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11125 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11127 * pretty-print.h (pp_underscore): New.
11129 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11130 printer functions instead of pp_character.
11131 (dump_binary_rhs): Likewise.
11132 (dump_ternary_rhs): Likewise.
11133 (dump_gimple_call_args): Likewise.
11134 (pp_points_to_solution): Likewise.
11135 (dump_gimple_call): Likewise.
11136 (dump_gimple_switch): Likewise.
11137 (dump_gimple_cond): Likewise.
11138 (dump_gimple_bind): Likewise.
11139 (dump_gimple_try): Likewise.
11140 (dump_gimple_omp_for): Likewise.
11141 (dump_gimple_omp_continue): Likewise.
11142 (dump_gimple_omp_single): Likewise.
11143 (dump_gimple_omp_sections): Likewise.
11144 (dump_gimple_omp_block): Likewise.
11145 (dump_gimple_omp_critical): Likewise.
11146 (dump_gimple_transaction): Likewise.
11147 (dump_gimple_asm): Likewise.
11148 (dump_gimple_phi): Likewise.
11149 (dump_gimple_omp_parallel): Likewise.
11150 (dump_gimple_omp_task): Likewise.
11151 (dump_gimple_omp_atomic_load): Likewise.
11152 (dump_gimple_omp_atomic_store): Likewise.
11153 (dump_gimple_mem_ops): Likewise.
11154 (pp_gimple_stmt_1): Likewise.
11155 (pp_cfg_jump): Likewise.
11156 (dump_implicit_edges): Likewise.
11157 (gimple_dump_bb_for_graph): Likewise.
11158 * graph.c (draw_cfg_node): Likewise.
11159 * langhooks.c (lhd_print_error_function): Likewise.
11160 * sched-vis.c (print_exp): Likewise.
11161 (print_value): Likewise.
11162 (print_pattern): Likewise.
11163 (print_insn): Likewise.
11164 (rtl_dump_bb_for_graph): Likewise.
11165 * tree-pretty-print.c (dump_function_declaration): Likewise.
11166 (dump_array_domain): Likewise.
11167 (dump_omp_clause): Likewise.
11168 (dump_location): Likewise.
11169 (dump_generic_node): Likewise.
11170 (print_struct_decl): Likewise.
11171 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11173 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11175 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11176 candidate table when replacing a candidate statement.
11177 (replace_rhs_if_not_dup): Likewise.
11178 (replace_one_candidate): Likewise.
11180 2013-08-02 Jan Hubicka <jh@suse.cz>
11181 Martin Liska <marxin.liska@gmail.com>
11183 * cgraphunit.c (add_new_function): Fix logic when adding from
11185 (assemble_thunk): Rename to ...
11186 (expand_thunk); .. this one; export; get it working with
11187 general functions; make produced gimple valid.
11188 * cgraph.h (expand_thunk): Declare.
11190 2013-08-02 Jan Hubicka <jh@suse.cz>
11192 * ipa-cp.c (gather_context_independent_values): Use
11193 ipa_get_param_move_cost.
11194 (get_replacement_map): Remove PARAM; move parameter folding
11196 (create_specialized_node): Update.
11197 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11198 assert that we have gimple body; update move_cost.
11199 (count_formal_params): Assert that we have gimple body.
11200 (ipa_dump_param): New function.
11201 (ipa_alloc_node_params): Break out from ...
11202 (ipa_initialize_node_params): ... here.
11203 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11204 (ipa_write_node_info): Stream move costs.
11205 (ipa_read_node_info): Read move costs.
11206 (ipa_update_after_lto_read): Do not recompute node params.
11207 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11208 (ipa_get_param): Check we are not in WPA.
11209 (ipa_get_param_move_cost): New.
11210 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11211 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11212 parm numbers to be present.
11214 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11216 PR rtl-optimization/58048
11217 * lra-constraints.c (process_alt_operands): Don't check asm
11218 operand on register.
11220 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11222 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11223 the implied StoreLoad barrier for atomic operations if before.
11225 2013-08-02 Jan Hubicka <jh@suse.cz>
11226 Martin Liska <marxin.liska@gmail.com>
11228 * cgraph.c (cgraph_function_body_availability): Do not check
11230 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11231 symtab_node_availability): Declare.
11232 * ipa.c (can_replace_by_local_alias): New.
11233 (function_and_variable_visibility): Use it.
11234 * symtab.c (symtab_for_node_and_aliases,
11235 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11237 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11239 PR rtl-optimization/57963
11240 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11241 (lra_constraints): Use them.
11243 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11245 * config/arm/types.md (define_attr "type"): Add "load_acq"
11247 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11249 (cortex_a53_store1): Likewise.
11251 2013-08-01 Jan Hubicka <jh@suse.cz>
11253 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11254 partitions are not needed.
11256 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11258 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11259 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11260 MAYBE_NON_Q_CLASS_P where appropriate.
11262 2013-08-01 Jan Hubicka <jh@suse.cz>
11264 * cgraph.h (release_function_body): Declare.
11265 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11266 of unused delcarations.
11268 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11270 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11271 RTL form when subtracting a constant.
11273 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11275 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11276 Generate canonical plus rtx with negated immediate instead of minus
11278 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11280 2013-08-01 Jan Hubicka <jh@suse.cz>
11282 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11283 (cgraph_release_function_body): Likewise.
11284 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11285 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11286 to used_as_abstract_origin.
11287 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11288 symbols used as abstract origins.
11289 * cgraphunit.c (analyze_functions): Update.
11290 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11291 used_as_abstract_origin.
11292 * tree-inline.c (tree_function_versioning): Update
11293 used_as_abstract_origin; be ready for DECL_RESULT and
11294 DECL_ARGUMENTS to be NULL.
11296 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11297 for abstract functions.
11298 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11301 2013-08-01 Jan Hubicka <jh@suse.cz>
11303 * profile.c (compute_value_histograms): Fix thinko.
11305 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11307 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11308 aarch-common-protos.h to extra_headers.
11309 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11310 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11311 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11313 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11315 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11316 Include "../arm/types.md". Define "type" attribute for all patterns.
11317 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11320 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11322 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11323 to support power8 load fusion.
11324 (fusion_gpr_mem_load): Likewise.
11326 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11328 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11329 declarations for power8 load fusion.
11330 (emit_fusion_gpr_load): Likewise.
11332 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11333 tuning for power8, turn on fusion mode by default. Turn on sign
11334 extending fusion mode if normal fusion mode is on, and we are at
11336 (fusion_gpr_load_p): New function, return true if we can fuse an
11337 addis instruction with a dependent load to a GPR.
11338 (emit_fusion_gpr_load): Emit the instructions for power8 load
11341 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11342 (VSX load fusion peepholes): New peepholes to fuse together an
11343 addi instruction with a VSX load instruction.
11345 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11346 peepholes to fuse an addis instruction with a load to a GPR base
11347 register. If we are supporting sign extending fusions, convert
11348 sign extending loads to zero extending loads and add an explicit
11351 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11353 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11354 aarch-common-protos.h to extra_headers.
11355 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11356 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11357 (arm_early_store_addr_dep): Likewise.
11358 (arm_no_early_alu_shift_dep): Likewise.
11359 (arm_no_early_alu_shift_value_dep): Likewise.
11360 (arm_no_early_mul_dep): Likewise.
11361 (arm_no_early_store_addr_dep): Likewise.
11362 (arm_mac_accumulator_is_mul_result): Likewise.
11363 (arm_mac_accumulator_is_result): Likewise.
11364 * config/arm/aarch-common.c: ... here. New file.
11365 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11367 (arm_early_store_addr_dep): Likewise.
11368 (arm_no_early_alu_shift_dep): Likewise.
11369 (arm_no_early_alu_shift_value_dep): Likewise.
11370 (arm_no_early_mul_dep): Likewise.
11371 (arm_no_early_store_addr_dep): Likewise.
11372 (arm_mac_accumulator_is_mul_result): Likewise.
11373 (arm_mac_accumulator_is_result): Likewise.
11374 * config/arm/aarch-common-protos.h: ... here. New file.
11375 * config/arm/t-arm (aarch-common.o): Define.
11377 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11379 * config/arm/arm.md: Include new file "types.md".
11380 (define_attr "type"): Move from here to ...
11381 (define_attr "mul32"): Likewise.
11382 (define_attr "mul64"): Likewise.
11383 * config/arm/types.md: ... here. New file.
11385 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11387 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11388 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11390 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11392 * gen-pass-instances.awk: Fix offset of substr().
11394 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11396 * Makefile.in (pass-instances.def): New.
11397 (passes.o): Replace dependency on passes.def with one on
11400 * gen-pass-instances.awk: New.
11402 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11403 rather than passes.def, updating local definition of NEXT_PASS
11404 macro to add an extra NUM parameter (currently unused).
11406 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11408 * Makefile.in (PASS_MANAGER_H): New.
11409 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11410 (passes.o): Likewise.
11411 (statistics.o): Likewise.
11412 (cgraphunit.o): Likewise.
11413 (context.o): Depend on PASS_MANAGER_H.
11415 * pass_manager.h: New.
11417 * cgraphunit.c (cgraph_add_new_function): Update for moves
11418 of globals to fields of pass_manager.
11419 (analyze_function): Likewise.
11420 (expand_function): Likewise.
11421 (ipa_passes): Likewise.
11422 (compile): Likewise.
11424 * context.c (context::context): New.
11425 * context.h (context::context): New.
11426 (context::get_passes): New.
11427 (context::passes_): New.
11429 * lto-cgraph.c (input_node): Update for moves of globals to
11430 fields of pass_manager.
11432 * passes.c (all_passes): Remove, in favor of a field of the
11433 same name within the new class pass_manager.
11434 (all_small_ipa_passes): Likewise.
11435 (all_lowering_passes): Likewise.
11436 (all_regular_ipa_passes): Likewise.
11437 (all_late_ipa_passes): Likewise.
11438 (all_lto_gen_passes): Likewise.
11439 (passes_by_id): Likewise.
11440 (passes_by_id_size): Likewise.
11441 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11442 the new class pass_manager.
11443 (set_pass_for_id): Convert to...
11444 (pass_manager::set_pass_for_id): ...method.
11445 (get_pass_for_id): Convert to...
11446 (pass_manager::get_pass_for_id): ...method.
11447 (register_one_dump_file): Move body of implementation into...
11448 (pass_manager::register_one_dump_file): ...here.
11449 (register_dump_files_1): Convert to...
11450 (pass_manager::register_dump_files_1): ...method.
11451 (register_dump_files): Convert to...
11452 (pass_manager::register_dump_files): ...method.
11453 (create_pass_tab): Update for moves of globals to fields of
11455 (dump_passes): Move body of implementation into...
11456 (pass_manager::dump_passes): ...here.
11457 (register_pass): Move body of implementation into...
11458 (pass_manager::register_pass): ...here.
11459 (init_optimization_passes): Convert into...
11460 (pass_manager::pass_manager): ...constructor for new
11461 pass_manager class, and initialize the pass_lists array.
11462 (check_profile_consistency): Update for moves of globals to
11463 fields of pass_manager.
11464 (dump_profile_report): Move body of implementation into...
11465 (pass_manager::dump_profile_report): ...here.
11466 (ipa_write_summaries_1): Update for moves of pass lists from
11467 being globals to fields of pass_manager.
11468 (ipa_write_optimization_summaries): Likewise.
11469 (ipa_read_summaries): Likewise.
11470 (ipa_read_optimization_summaries): Likewise.
11471 (execute_all_ipa_stmt_fixups): Likewise.
11473 * statistics.c (statistics_fini): Update for moves of globals to
11474 fields of pass_manager.
11476 * toplev.c (general_init): Replace call to
11477 init_optimization_passes with construction of the pass_manager
11480 * tree-pass.h (all_passes): Remove, in favor of a field of the
11481 same name within the new class pass_manager.
11482 (all_small_ipa_passes): Likewise.
11483 (all_lowering_passes): Likewise.
11484 (all_regular_ipa_passes): Likewise.
11485 (all_lto_gen_passes): Likewise.
11486 (all_late_ipa_passes): Likewise.
11487 (passes_by_id): Likewise.
11488 (passes_by_id_size): Likewise.
11489 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11490 the new class pass_manager.
11491 (get_pass_for_id): Remove.
11493 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11495 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11498 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11500 * arm.md (mulhi3): New expand pattern.
11502 2013-07-30 Jan Hubicka <jh@suse.cz>
11503 Martin Liska <marxin.liska@gmail.com>
11505 * profile.c (compute_value_histograms): Do not ICE when
11506 there is mismatch only on some counters.
11508 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11510 PR rtl-optimization/57637
11511 * function.c (move_insn_for_shrink_wrap): Also check the
11512 GEN set of the LIVE problem for the liveness analysis
11513 if it exists, otherwise give up.
11515 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11517 PR tree-optimization/57993
11518 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11519 replaced statement in the candidate table.
11520 (phi_add_costs): Return infinite cost when the hidden basis does
11521 not dominate all phis on which the candidate is dependent.
11522 (replace_one_candidate): Record replaced statement in the
11525 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11527 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11528 (ashlv2si3): New expander.
11529 (*ashlv2si3_i): New define_insn_and_split.
11530 * predicates.md (float_operation): Allow patterns with three
11531 basic sub-patterns.
11533 PR rtl-optimization/58021
11534 * mode-switching.c (create_pre_exit): Always split off preceding
11535 insns if we are not at the basic block head.
11537 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11539 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11540 (UCLIBC_DYNAMIC_LINKER): New macro.
11541 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11543 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11544 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11546 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11547 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11548 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11549 for SF and DF modes. Use ieee_quad_format for TF mode.
11550 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11551 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11552 (mips_option_override): Handle `-mnan=legacy'.
11553 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11554 `-mabs=2008' and `-mnan=2008'.
11555 (OPTION_DEFAULT_SPECS): Add "nan" default.
11556 (ASM_SPEC): Handle `-mnan='.
11557 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11558 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11559 comment accordingly.
11560 (neg<mode>2): Likewise.
11561 * config/mips/mips.opt (mabs, mnan): New options.
11562 * doc/install.texi (Configuration): Document `--with-nan=' option.
11563 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11565 (MIPS Options): Document them.
11566 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11567 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11568 * configure: Regenerate.
11569 * config.in: Regenerate.
11571 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11573 * config/i386/i386.md (float post-reload splitters): Do not check
11574 for subregs of SSE registers.
11576 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11577 H.J. Lu <hongjiu.lu@intel.com>
11581 * config/i386/i386.md (post-reload splitter
11582 to avoid partial SSE reg dependency stalls): New pattern.
11584 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11586 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11588 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11590 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11591 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11593 2013-07-26 Cary Coutant <ccoutant@google.com>
11595 * dwarf2out.c (die_checksum_ordered): Don't include template
11596 instantiations in signature.
11597 (is_template_parameter): New function.
11598 (is_template_instantiation): New function.
11599 (generate_skeleton_bottom_up): Don't include template instantiations
11601 (generate_skeleton): Likewise.
11602 (break_out_comdat_types): Move recursive call to break out nested
11604 (prune_unused_types_mark_generic_parms_dies): Call
11605 is_template_parameter.
11607 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11609 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11610 uses vector registers.
11611 * config/aarch64/iterators.md: Add attributes rtn and vas.
11613 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11614 Richard Earnshaw <richard.earnshaw@arm.com>
11616 * combine.c (simplify_comparison): Re-canonicalize operands
11618 * config/arm/arm.md (movcond_addsi): New splitter.
11620 2013-07-25 Sterling Augustine <saugustine@google.com>
11622 * dwarf2out.c (size_of_pubnames): Move code to...
11623 (include_pubname_in_output): ...here. New.
11624 (want_pubnames): Rearrange.
11625 (output_pubnames): Call include_pubname_in_output. Move assertion.
11627 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11629 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11631 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11634 * doc/extend.texi: Remove obsolete builtins. Fix
11635 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11637 2013-07-25 Jan Hubicka <jh@suse.cz>
11639 * cgraph.c (release_function_body): Break out from ...
11640 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11641 and DECL_ARGUMENTS.
11642 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11643 old_tree in the map.
11644 (create_specialized_node): Update.
11645 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11647 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11648 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11649 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11650 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11653 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11655 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11656 addsi3_carryin_alt2_<optab>): Correct output template.
11658 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11660 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11661 Adjust for arm_restrict_it.
11662 Remove trailing whitespace.
11664 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11666 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11667 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11669 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11671 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11673 PR rtl-optimization/57960
11674 * lra-constraints.c (process_alt_operands): Use the right mode
11675 when checking strict_low.
11677 2013-07-25 Jan Hubicka <jh@suse.cz>
11679 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11680 * cgraph.c (cgraph_remove_node): Do not release function body
11681 when in cgraph streaming.
11682 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11683 in other partitions are not considered reachable; fix handling of
11686 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11688 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11690 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11695 * config/arm/arm.md ("*sibcall_insn): Replace use of
11696 Ss with US. Adjust output for v5 and v4t.
11697 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11699 * config/arm/constraints.md ("Ss"): Rename to US.
11701 2013-07-25 Terry Guo <terry.guo@arm.com>
11703 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11704 shift_add/shift_sub0/shift_sub1 RTXs.
11706 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11707 Anton Blanchard <anton@au1.ibm.com>
11709 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11710 (altivec_vpks<VI_char>ss): Likewise.
11711 (altivec_vpks<VI_char>us): Likewise.
11712 (altivec_vpku<VI_char>us): Likewise.
11713 (altivec_vpku<VI_char>um): Likewise.
11715 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11717 Introduce context class.
11719 * Makefile.in (CONTEXT_H): New.
11720 (OBJS): Add context.o.
11721 (toplev.o): Add CONTEXT_H to dependencies.
11724 * toplev.c (general_init): Create the singleton gcc::context instance.
11730 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11732 PR rtl-optimization/57968
11733 * mode-switching.c (create_pre_exit): Allow instructions that
11734 don't set a return register to need a non-exit mode.
11736 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11737 Anton Blanchard <anton@au1.ibm.com>
11739 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11740 operands to vperm for little endian.
11741 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11742 of lvsl to create the control mask for a vperm for little endian.
11744 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11745 Anton Blanchard <anton@au1.ibm.com>
11747 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11748 two operands for little-endian.
11750 2013-07-23 Steve Ellcey <sellcey@mips.com>
11752 * config/mips/mips.c (mips_case_values_threshold): New.
11753 (TARGET_CASE_VALUES_THRESHOLD): Define.
11755 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11756 Anton Blanchard <anton@au1.ibm.com>
11758 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11759 selection of field for vector splat in little endian mode.
11761 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11763 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11764 expanders to rs6000.md.
11765 (ior<mode>3): Likewise.
11766 (and<mode>3): Likewise.
11767 (one_cmpl<mode>2): Likewise.
11768 (nor<mode>3): Likewise.
11769 (andc<mode>3): Likewise.
11770 (eqv<mode>3): Likewise.
11771 (nand<mode>3): Likewise.
11772 (orc<mode>3): Likewise.
11774 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11777 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11778 to split multi-word logical operations.
11779 (rs6000_split_logical_di): Likewise.
11780 (rs6000_split_logical): Likewise.
11782 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11783 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11784 and allow TImode operations in 32-bit.
11785 (vsx_and<mode>3_64bit): Likewise.
11786 (vsx_ior<mode>3_32bit): Likewise.
11787 (vsx_ior<mode>3_64bit): Likewise.
11788 (vsx_xor<mode>3_32bit): Likewise.
11789 (vsx_xor<mode>3_64bit): Likewise.
11790 (vsx_one_cmpl<mode>2_32bit): Likewise.
11791 (vsx_one_cmpl<mode>2_64bit): Likewise.
11792 (vsx_nor<mode>3_32bit): Likewise.
11793 (vsx_nor<mode>3_64bit): Likewise.
11794 (vsx_andc<mode>3_32bit): Likewise.
11795 (vsx_andc<mode>3_64bit): Likewise.
11796 (vsx_eqv<mode>3_32bit): Likewise.
11797 (vsx_eqv<mode>3_64bit): Likewise.
11798 (vsx_nand<mode>3_32bit): Likewise.
11799 (vsx_nand<mode>3_64bit): Likewise.
11800 (vsx_orc<mode>3_32bit): Likewise.
11801 (vsx_orc<mode>3_64bit): Likewise.
11803 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11804 logical types in GPRs.
11806 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11807 logical insns to rs6000.md, and allow TImode operations in
11809 (altivec_ior<mode>3): Likewise.
11810 (altivec_xor<mode>3): Likewise.
11811 (altivec_one_cmpl<mode>2): Likewise.
11812 (altivec_nor<mode>3): Likewise.
11813 (altivec_andc<mode>3): Likewise.
11815 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11816 attributes for moving the 128-bit logical operations into
11818 (BOOL_REGS_OUTPUT): Likewise.
11819 (BOOL_REGS_OP1): Likewise.
11820 (BOOL_REGS_OP2): Likewise.
11821 (BOOL_REGS_UNARY): Likewise.
11822 (BOOL_REGS_AND_CR0): Likewise.
11823 (one_cmpl<mode>2): Add support for DI logical operations on
11824 32-bit, splitting the operations to 32-bit.
11825 (anddi3): Likewise.
11826 (iordi3): Likewise.
11827 (xordi3): Likewise.
11828 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11829 changes to combine the 32/64-bit code, allow logical operations on
11830 TI mode in 32-bit, and to use similar match_operator patterns like
11831 scalar mode uses. Combine the Altivec and VSX code for logical
11832 operations, and move it here.
11833 (ior<mode>3, 128-bit types): Likewise.
11834 (xor<mode>3, 128-bit types): Likewise.
11835 (one_cmpl<mode>3, 128-bit types): Likewise.
11836 (nor<mode>3, 128-bit types): Likewise.
11837 (andc<mode>3, 128-bit types): Likewise.
11838 (eqv<mode>3, 128-bit types): Likewise.
11839 (nand<mode>3, 128-bit types): Likewise.
11840 (orc<mode>3, 128-bit types): Likewise.
11841 (and<mode>3_internal): Likewise.
11842 (bool<mode>3_internal): Likewise.
11843 (boolc<mode>3_internal1): Likewise.
11844 (boolc<mode>3_internal2): Likewise.
11845 (boolcc<mode>3_internal1): Likewise.
11846 (boolcc<mode>3_internal2): Likewise.
11847 (eqv<mode>3_internal1): Likewise.
11848 (eqv<mode>3_internal2): Likewise.
11849 (one_cmpl1<mode>3_internal): Likewise.
11851 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11853 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11854 Rename flag_stack_usage to flag_stack_usage_info.
11856 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11858 * config/microblaze/sync.md: New file.
11859 * config/microblaze/microblaze.md: Include sync.md
11860 * config/microblaze/microblaze.c: Add print_operand 'y'.
11861 * config/microblaze/constraints.md: Add memory_contraint
11862 'Q' which is a single register.
11864 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11866 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11868 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11870 * reload.c (find_reloads): Exit loop once we find this operand
11871 cannot be reloaded somehow for this alternative.
11873 * reload.c (find_reloads): Exit loop once we find a hard register.
11875 * rtlanal.c (computed_jump_p): Exit loop once we find label
11878 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11880 * cfgloopmanip.c (remove_path): Exit loop after setting
11883 * gensupport.c (subst_dup): Avoid loop if code is not
11884 MATCH_DUP nor MATCH_OP_DUP.
11886 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11888 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11890 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11892 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11893 true for SP_REGNUM if mode == ptr_mode.
11894 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11895 with value R0_REGNUM + 31.
11897 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11899 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11900 pad pointer-typed argument downward.
11902 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11904 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11905 and __ILP32__ when the ILP32 model is in use.
11907 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11909 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11910 (aarch64_load_symref_appropriately): In the case of
11911 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11912 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11913 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11914 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11915 if 'mode' doesn't equal to 'ptr_mode'.
11916 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11917 'vcall_offset'; change to call aarch64_emit_move differently depending
11918 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11919 to calculate the upper bound of 'vcall_offset'.
11920 (aarch64_cannot_force_const_mem): Change to also return true if
11922 (aarch64_legitimize_reload_address): In the case of large
11923 displacements, add new local variable 'xmode' and an assertion
11924 based on it; change to use 'xmode' to generate the new rtx and
11926 (aarch64_asm_trampoline_template): Change to generate the template
11927 differently depending on TARGET_ILP32 or not; change to use
11928 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11929 (aarch64_trampoline_size): Removed.
11930 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11931 and replace immediate literals with it. Change to use 'ptr_mode'
11932 instead of 'DImode' and call convert_memory_address if the mode
11933 of 'fnaddr' doesn't equal to 'ptr_mode'.
11934 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11936 (aarch64_elf_asm_destructor): Likewise.
11937 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11938 on TARGET_ILP32 instead of aarch64_trampoline_size.
11939 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11940 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11941 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11942 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11943 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11944 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11945 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11946 depending on the value of 'mode'.
11947 (add_losym_<mode>): New.
11948 (ldr_got_small_<mode>): New, based on ldr_got_small.
11949 (ldr_got_small): Remove.
11950 (ldr_got_small_sidi): New.
11951 * config/aarch64/iterators.md (P): New.
11952 (PTR): Change to 'ptr_mode' in the condition.
11954 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11956 * config.gcc (aarch64*-*-*): Support --with-abi.
11957 (aarch64*-*-elf): Support --with-multilib-list.
11958 (aarch64*-*-linux*): Likewise.
11959 (supported_defaults): Add abi to aarch64*-*-*.
11960 * configure.ac: Mention AArch64 for --with-multilib-list.
11961 * configure: Re-generated.
11962 * config/aarch64/biarchilp32.h: New file.
11963 * config/aarch64/biarchlp64.h: New file.
11964 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11966 (MULTILIB_DEFAULTS): Ditto.
11967 (DRIVER_SELF_SPECS): Ditto.
11968 (ASM_SPEC): Update to also substitute -mabi.
11969 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11970 file whose name depends on -mabi= and -mbig-endian.
11971 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11973 (POINTER_SIZE): New define.
11974 (POINTERS_EXTEND_UNSIGNED): Ditto.
11975 (enum aarch64_abi_type): New enumeration tag.
11976 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11977 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11978 (TARGET_ILP32): New define.
11979 * config/aarch64/aarch64.opt (mabi): New.
11980 (aarch64_abi): New.
11981 (ilp32, lp64): New values for -mabi.
11982 * config/aarch64/t-aarch64 (comma): New define.
11983 (MULTILIB_OPTIONS): Ditto.
11984 (MULTILIB_DIRNAMES): Ditto.
11985 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11986 * doc/invoke.texi: Document -mabi for AArch64.
11988 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11990 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11992 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11993 Anton Blanchard <anton@au1.ibm.com>
11995 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11996 endianness when selecting field to splat.
11998 2013-07-22 Eric Christopher <echristo@gmail.com>
12000 * dwarf2out.c (die_odr_checksum): New function to use
12001 CHECKSUM_ macros and ULEB128 for DIE tag.
12002 (generate_type_signature): Use.
12004 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12006 * config.gcc (sparc*-*-*): Accept leon3 processor.
12007 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12008 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12009 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12010 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12011 (mfix-ut699): Adjust comment.
12012 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12013 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12014 (CPP_CPU_SPEC): Likewise.
12015 (ASM_CPU_SPEC): Likewise.
12016 * config/sparc/sparc.c (leon3_cost): New constant.
12017 (sparc_option_override): Add leon3 support.
12018 (mem_ref): New function.
12019 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12020 (sparc_do_work_around_errata): Look into the instruction in the delay
12021 slot and adjust accordingly. Add fix for the data cache nullify issues
12022 of the UT699. Change insertion position for the NOP.
12023 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12024 (leon3_load): New reservation.
12025 (leon_store): Bump latency to 2.
12026 (grfpu): New automaton.
12027 (grfpu_alu): New unit.
12028 (grfpu_ds): Likewise.
12029 (leon_fp_alu): Adjust.
12030 (leon_fp_mult): Delete.
12031 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12032 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12033 * config/sparc/sparc.md (cpu): Add leon3.
12034 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12035 (swapsi): Likewise.
12036 (atomic_test_and_set): Likewise.
12037 (ldstub): Likewise.
12039 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12041 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12042 default for R5900 targets.
12043 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12044 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12045 * config/mips/mips.c (mips_option_override): Report an error for
12046 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12047 for -march=r5900 -mhard-float.
12049 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12051 * df-problems.c (can_move_insns_across): Exit loop once we
12052 find a non-fixed, non-global register.
12054 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12057 * omega.c (omega_eliminate_red): Break after setting red_found.
12058 (omega_problem_has_red_equations): Similarly after setting found.
12059 (omega_query_variable): Similarly after setting coupled.
12061 2013-07-22 Marek Polacek <polacek@redhat.com>
12063 * gimplify.c: Don't include gimple.h twice.
12065 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12067 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12068 instead of TARGET_THUMB1.
12069 (Pz): New constraint.
12070 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12072 (compare_negsi_si): Likewise.
12073 (compare_addsi2_op0): Likewise.
12074 (compare_addsi2_op1): Likewise.
12075 (addsi3_carryin_<optab>): Likewise.
12076 (addsi3_carryin_alt2_<optab>): Likewise.
12077 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12078 for arm_restrict_it.
12079 (subsi3_carryin): Likewise.
12080 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12081 (minmax_arithsi): Disable for arm_restrict_it.
12082 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12083 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12084 (satsi_<SAT:code>_shift): Likewise.
12085 (arm_shiftsi3): Add alternative for 16-bit encoding.
12086 (arm32_movhf): Disable for arm_restrict_it.
12087 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12088 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12090 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12092 * config/arm/arm.md (attribute "insn"): Delete.
12093 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12094 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12095 (not_shiftsi): Update for attribute change.
12096 (not_shiftsi_compare0): Likewise.
12097 (not_shiftsi_compare0_scratch): Likewise.
12098 (arm_one_cmplsi2): Likewise.
12099 (thumb1_one_cmplsi2): Likewise.
12100 (notsi_compare0): Likewise.
12101 (notsi_compare0_scratch): Likewise.
12102 (thumb1_movdi_insn): Likewise.
12103 (arm_movsi_insn): Likewise.
12104 (movhi_insn_arch4): Likewise.
12105 (movhi_bytes): Likewise.
12106 (arm_movqi_insn): Likewise.
12107 (thumb1_movqi_insn): Likewise.
12108 (arm32_movhf): Likewise.
12109 (thumb1_movhf): Likewise.
12110 (arm_movsf_soft_insn): Likewise.
12111 (thumb1_movsf_insn): Likewise.
12112 (thumb_movdf_insn): Likewise.
12113 (movsicc_insn): Likewise.
12114 (movsfcc_soft_insn): Likewise.
12115 (and_scc): Likewise.
12116 (cond_move): Likewise.
12117 (if_move_not): Likewise.
12118 (if_not_move): Likewise.
12119 (if_shift_move): Likewise.
12120 (if_move_shift): Likewise.
12121 (if_shift_shift): Likewise.
12122 (if_not_arith): Likewise.
12123 (if_arith_not): Likewise.
12124 (cond_move_not): Likewise.
12125 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12126 (neon_mov<mode>): Likewise.
12127 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12128 (thumb2_movsi_vfp): Likewise.
12129 (movsf_vfp): Likewise.
12130 (thumb2_movsf_vfp): Likewise.
12131 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12133 (cortexa7_older_only): Likewise.
12134 (cortexa7_younger): Likewise.
12135 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12136 (1020alu_shift_op): Likewise.
12137 (1020alu_shift_reg_op): Likewise.
12138 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12139 (alu_shift_op): Likewise.
12140 (alu_shift_reg_op): Likewise.
12141 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12142 (11_alu_shift_op): Likewise.
12143 (11_alu_shift_reg_op): Likewise.
12144 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12145 (9_alu_shift_reg_op): Likewise.
12146 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12148 (cortex_a15_alu_shift): Likewise.
12149 (cortex_a15_alu_shift_reg): Likewise.
12150 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12151 (cortex_a5_alu_shift): Likewise.
12152 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12154 (cortex_a53_alu_shift): Likewise.
12155 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12157 (cortex_a7_alu_reg): Likewise.
12158 (cortex_a7_alu_shift): Likewise.
12159 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12160 (cortex_a8_alu_shift): Likewise.
12161 (cortex_a8_alu_shift_reg): Likewise.
12162 (cortex_a8_mov): Likewise.
12163 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12164 (cortex_a9_dp_shift): Likewise.
12165 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12166 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12167 (cortex_r4_mov): Likewise.
12168 (cortex_r4_alu_shift): Likewise.
12169 (cortex_r4_alu_shift_reg): Likewise.
12170 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12171 (526_alu_shift_op): Likewise.
12172 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12173 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12174 (626te_alu_shift_op): Likewise.
12175 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12176 (726te_alu_op): Likewise.
12177 (726te_alu_shift_op): Likewise.
12178 (726te_alu_shift_reg_op): Likewise.
12179 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12180 (mp626_alu_shift_op): Likewise.
12181 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12182 (pj4_alu_e1_conds): Likewise.
12183 (pj4_alu): Likewise.
12184 (pj4_alu_conds): Likewise.
12185 (pj4_shift): Likewise.
12186 (pj4_shift_conds): Likewise.
12187 (pj4_alu_shift): Likewise.
12188 (pj4_alu_shift_conds): Likewise.
12190 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12192 * config/arm/predicates.md (shiftable_operator_strict_it):
12194 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12195 Disable cond_exec version for arm_restrict_it.
12196 (thumb2_smaxsi3): Convert to generate cond_exec.
12197 (thumb2_sminsi3): Likewise.
12198 (thumb32_umaxsi3): Likewise.
12199 (thumb2_uminsi3): Likewise.
12200 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12201 (thumb2_neg_abssi2): Likewise.
12202 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12203 (thumb2_movsicc_insn): Adjust alternatives.
12204 (thumb2_mov_negscc): Disable for arm_restrict_it.
12205 (thumb2_mov_negscc_strict_it): New pattern.
12206 (thumb2_mov_notscc_strict_it): New pattern.
12207 (thumb2_mov_notscc): Disable for arm_restrict_it.
12208 (thumb2_ior_scc): Likewise.
12209 (thumb2_ior_scc_strict_it): New pattern.
12210 (thumb2_cond_move): Adjust for arm_restrict_it.
12211 (thumb2_cond_arith): Disable for arm_restrict_it.
12212 (thumb2_cond_arith_strict_it): New pattern.
12213 (thumb2_cond_sub): Adjust for arm_restrict_it.
12214 (thumb2_movcond): Likewise.
12215 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12216 (thumb2_zero_extendhisi2_v6): Likewise.
12217 (thumb2_zero_extendqisi2_v6): Likewise.
12218 (orsi_notsi_si): Likewise.
12219 (orsi_not_shiftsi_si): Likewise.
12221 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12223 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12224 instruction sequence is 1 byte shorter.
12226 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12228 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12229 it is not needed after split.
12231 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12234 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12235 second label for nonlocal goto receivers. Don't output pic base labels
12236 unless we're producing PIC; mark that action unreachable().
12237 (ix86_save_reg): If the function contains a nonlocal label, save the
12239 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12240 * config/darwin.c (emitted_pic_label_num): New GTY.
12241 (update_pic_label_number_if_needed): New.
12242 (machopic_output_function_base_name): Adjust for nonlocal receiver
12244 (machopic_should_output_picbase_label): New.
12245 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12246 (nonlocal_goto_receiver): New insn and split.
12248 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12250 * config/aarch64/aarch64-builtins.c
12251 (aarch64_fold_builtin): Fold abs in all modes.
12252 * config/aarch64/aarch64-simd-builtins.def
12253 (abs): Enable for all modes.
12254 * config/aarch64/arm_neon.h
12255 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12256 (vabs_f64): Add missing intrinsic.
12258 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12260 * config/aarch64/arm_neon.h (vabs_s64): New function
12262 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12265 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12266 * config/avr/avr.md (adjust_len): Add `round'.
12267 * config/avr/avr-protos.h (avr_out_round): New prototype.
12268 (avr_out_plus): Add `out_label' argument.
12269 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12270 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12271 Handle the case where `insn' is just a pattern.
12272 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12273 (avr_out_round): New function.
12274 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12276 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12278 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12279 Add check for flag_stack_usage to handle -fstack-usage support
12281 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12283 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12284 interaction for new Power8 flags and VSX.
12286 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12288 PR middle-end/57698
12289 * tree-inline.c (expand_call_inline): Emit errors during
12290 early_inlining only if optimization is not turned on.
12292 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12296 * passes.c (init_optimization_passes): Move the construction of
12297 the pass hierarchy into a new passes.def file.
12299 * Makefile.in (passes.o): Add dependency on passes.def.
12301 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12303 * passes.c (init_optimization_passes): Introduce macros for
12304 constructing the tree of passes (INSERT_PASSES_AFTER,
12305 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12306 TERMINATE_PASS_LIST).
12308 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12309 Wei Mi <wmi@google.com>
12311 PR rtl-optimization/57878
12312 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12314 (reload_pseudo_compare_func): Check nregs first for reload
12317 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12319 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12321 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12323 * read-rtl.c (validate_const_int): Once an invalid character is
12324 seen, quit the loop.
12326 * gengtype.c (write_roots): Similarly once we find the "deletable"
12327 or "if_marked" option.
12329 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12331 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12332 "xtab" and "sat". Move value "clz" from here to ...
12333 (attriubte "type"): ... here.
12334 (satsi_<SAT:code>): Delete "insn" attribute.
12335 (satsi_<SAT:code>_shift): Likewise.
12336 (arm_zero_extendqisi2addsi): Likewise.
12337 (arm_extendqisi2addsi): Likewise.
12338 (clzsi2): Update for attribute changes.
12339 (rbitsi2): Likewise.
12340 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12342 (arm_usatsihi): Likewise.
12343 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12345 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12347 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12348 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12349 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12350 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12351 in alphabetical order.
12352 (attribute "core_cycles"): Update for attribute changes.
12353 (arm_addsi3): Likewise.
12354 (addsi3_compare0): Likewise.
12355 (addsi3_compare0_scratch): Likewise.
12356 (addsi3_compare_op1): Likewise.
12357 (addsi3_compare_op2): Likewise.
12358 (compare_addsi2_op0): Likewise.
12359 (compare_addsi2_op1): Likewise.
12360 (addsi3_carryin_shift_<optab>): Likewise.
12361 (subsi3_carryin_shift): Likewise.
12362 (rsbsi3_carryin_shift): Likewise.
12363 (arm_subsi3_insn): Likewise.
12364 (subsi3_compare0): Likewise.
12365 (subsi3_compare): Likewise.
12366 (arm_andsi3_insn): Likewise.
12367 (thumb1_andsi3_insn): Likewise.
12368 (andsi3_compare0): Likewise.
12369 (andsi3_compare0_scratch): Likewise.
12370 (zeroextractsi_compare0_scratch
12371 (andsi_not_shiftsi_si): Likewise.
12372 (iorsi3_insn): Likewise.
12373 (iorsi3_compare0): Likewise.
12374 (iorsi3_compare0_scratch): Likewise.
12375 (arm_xorsi3): Likewise.
12376 (thumb1_xorsi3_insn): Likewise.
12377 (xorsi3_compare0): Likewise.
12378 (xorsi3_compare0_scratch): Likewise.
12379 (satsi_<SAT:code>_shift): Likewise.
12381 (arm_shiftsi3): Likewise.
12382 (shiftsi3_compare0): Likewise.
12383 (not_shiftsi): Likewise.
12384 (not_shiftsi_compare0): Likewise.
12385 (not_shiftsi_compare0_scratch): Likewise.
12386 (arm_one_cmplsi2): Likewise.
12387 (thumb_one_complsi2): Likewise.
12388 (notsi_compare0): Likewise.
12389 (notsi_compare0_scratch): Likewise.
12390 (thumb1_zero_extendhisi2): Likewise.
12391 (arm_zero_extendhisi2): Likewise.
12392 (arm_zero_extendhisi2_v6): Likewise.
12393 (arm_zero_extendhisi2addsi): Likewise.
12394 (thumb1_zero_extendqisi2): Likewise.
12395 (thumb1_zero_extendqisi2_v6): Likewise.
12396 (arm_zero_extendqisi2): Likewise.
12397 (arm_zero_extendqisi2_v6): Likewise.
12398 (arm_zero_extendqisi2addsi): Likewise.
12399 (thumb1_extendhisi2): Likewise.
12400 (arm_extendhisi2): Likewise.
12401 (arm_extendhisi2_v6): Likewise.
12402 (arm_extendqisi): Likewise.
12403 (arm_extendqisi_v6): Likewise.
12404 (arm_extendqisi2addsi): Likewise.
12405 (thumb1_extendqisi2): Likewise.
12406 (thumb1_movdi_insn): Likewise.
12407 (arm_movsi_insn): Likewise.
12408 (movsi_compare0): Likewise.
12409 (movhi_insn_arch4): Likewise.
12410 (movhi_bytes): Likewise.
12411 (arm_movqi_insn): Likewise.
12412 (thumb1_movqi_insn): Likewise.
12413 (arm32_movhf): Likewise.
12414 (thumb1_movhf): Likewise.
12415 (arm_movsf_soft_insn): Likewise.
12416 (thumb1_movsf_insn): Likewise.
12417 (movdf_soft_insn): Likewise.
12418 (thumb_movdf_insn): Likewise.
12419 (arm_cmpsi_insn): Likewise.
12420 (cmpsi_shiftsi): Likewise.
12421 (cmpsi_shiftsi_swp): Likewise.
12422 (arm_cmpsi_negshiftsi_si): Likewise.
12423 (movsicc_insn): Likewise.
12424 (movsfcc_soft_insn): Likewise.
12425 (arith_shiftsi): Likewise.
12426 (arith_shiftsi_compare0
12427 (arith_shiftsi_compare0_scratch
12428 (sub_shiftsi): Likewise.
12429 (sub_shiftsi_compare0
12430 (sub_shiftsi_compare0_scratch
12431 (and_scc): Likewise.
12432 (cond_move): Likewise.
12433 (if_plus_move): Likewise.
12434 (if_move_plus): Likewise.
12435 (if_move_not): Likewise.
12436 (if_not_move): Likewise.
12437 (if_shift_move): Likewise.
12438 (if_move_shift): Likewise.
12439 (if_shift_shift): Likewise.
12440 (if_not_arith): Likewise.
12441 (if_arith_not): Likewise.
12442 (cond_move_not): Likewise.
12443 (thumb1_ashlsi3): Set type attribute.
12444 (thumb1_ashrsi3): Likewise.
12445 (thumb1_lshrsi3): Likewise.
12446 (thumb1_rotrsi3): Likewise.
12447 (shiftsi3_compare0_scratch): Likewise.
12448 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12449 (neon_mov<mode>): Likewise.
12450 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12452 (thumb2_movsi_insn): Likewise.
12453 (thumb2_cmpsi_neg_shiftsi): Likewise.
12454 (thumb2_extendqisi_v6): Likewise.
12455 (thumb2_zero_extendhisi2_v6): Likewise.
12456 (thumb2_zero_extendqisi2_v6): Likewise.
12457 (thumb2_shiftsi3_short): Likewise.
12458 (thumb2_addsi3_compare0_scratch): Likewise.
12459 (orsi_not_shiftsi_si): Likewise.
12460 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12461 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12463 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12464 (1020alu_shift_op): Likewise.
12465 (1020alu_shift_reg_op): Likewise.
12466 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12467 (alu_shift_op): Likewise.
12468 (alu_shift_reg_op): Likewise.
12469 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12470 (11_alu_shift_op): Likewise.
12471 (11_alu_shift_reg_op): Likewise.
12472 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12473 (9_alu_shift_reg_op): Likewise.
12474 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12476 (cortex_a15_alu_shift): Likewise.
12477 (cortex_a15_alu_shift_reg): Likewise.
12478 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12480 (cortex_a5_alu_shift): Likewise.
12481 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12483 (cortex_a53_alu_shift): Likewise.
12484 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12486 (cortex_a7_alu_reg): Likewise.
12487 (cortex_a7_alu_shift): Likewise.
12488 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12490 (cortex_a8_alu_shift): Likewise.
12491 (cortex_a8_alu_shift_reg): Likewise.
12492 (cortex_a8_mov): Likewise.
12493 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12494 (cortex_a9_dp_shift): Likewise.
12495 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12497 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12499 (cortex_r4_mov): Likewise.
12500 (cortex_r4_alu_shift): Likewise.
12501 (cortex_r4_alu_shift_reg): Likewise.
12502 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12503 (526_alu_shift_op): Likewise.
12504 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12505 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12506 (626te_alu_shift_op): Likewise.
12507 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12508 (726te_alu_op): Likewise.
12509 (726te_alu_shift_op): Likewise.
12510 (726te_alu_shift_reg_op): Likewise.
12511 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12512 (mp626_alu_shift_op): Likewise.
12513 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12514 (pj4_alu_e1_conds): Likewise.
12515 (pj4_alu): Likewise.
12516 (pj4_alu_conds): Likewise.
12517 (pj4_shift): Likewise.
12518 (pj4_shift_conds): Likewise.
12519 (pj4_alu_shift): Likewise.
12520 (pj4_alu_shift_conds): Likewise.
12521 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12523 (cortexa7_older_only): Likewise.
12524 (cortexa7_younger): Likewise.
12526 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12528 * ipa-pure-const.c (generate_summary): Rename to...
12529 (pure_const_generate_summary): ... this.
12531 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12533 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12535 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12538 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12541 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12543 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12544 enabled without -march=zEC12.
12545 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12548 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12550 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12551 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12552 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12553 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12554 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12555 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12556 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12558 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12559 and ISA_HAS_NMADD3_NMSUB3.
12560 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12561 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12562 (nmsub4<mode>, nmsub3<mode>): Likewise.
12563 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12565 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12567 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12568 TARGET_MIPS5400 checking.
12570 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12571 Peter Bergner <bergner@vnet.ibm.com>
12573 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12574 registers in the comment.
12575 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12576 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12577 rather than FIRST_PSEUDO_REGISTERS.
12579 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12581 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12582 enable extra ISA flags with TARGET_HTM.
12584 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12586 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12589 2013-07-15 Cong Hou <congh@google.com>
12591 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12592 in compare function for sorting.
12594 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12596 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12597 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12598 * config/rs6000/rs6000.opt: Add -mhtm option.
12599 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12600 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12601 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12602 __HTM__ if the HTM instructions are available.
12603 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12604 htm_spr_reg_operand): New define_predicates.
12605 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12606 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12608 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12609 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12610 HTM builtin functions.
12611 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12612 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12613 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12614 (rs6000_builtin_mask_calculate): Likewise.
12615 (rs6000_option_override_internal): Likewise.
12616 (bdesc_htm): Add new HTM builtin support.
12617 (htm_spr_num): New function.
12618 (htm_spr_regno): Likewise.
12619 (rs6000_htm_spr_icode): Likewise.
12620 (htm_expand_builtin): Likewise.
12621 (htm_init_builtins): Likewise.
12622 (rs6000_expand_builtin): Add support for HTM builtin functions.
12623 (rs6000_init_builtins): Likewise.
12624 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12626 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12627 (TARGET_HTM, MASK_HTM): Define macros.
12628 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12629 (FIXED_REGISTERS): Likewise.
12630 (CALL_USED_REGISTERS): Likewise.
12631 (CALL_REALLY_USED_REGISTERS): Likewise.
12632 (REG_ALLOC_ORDER): Likewise.
12633 (enum reg_class): Likewise.
12634 (REG_CLASS_NAMES): Likewise.
12635 (REG_CLASS_CONTENTS): Likewise.
12636 (REGISTER_NAMES): Likewise.
12637 (ADDITIONAL_REGISTER_NAMES): Likewise.
12638 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12639 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12640 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12641 * config/rs6000/htm.md: New file.
12642 * config/rs6000/htmintrin.h: New file.
12643 * config/rs6000/htmxlintrin.h: New file.
12645 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12647 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12648 Define SYMBOL_TINY_GOT, update comment.
12649 * config/aarch64/aarch64.c
12650 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12651 (aarch64_expand_mov_immediate): Likewise.
12652 (aarch64_print_operand): Likewise.
12653 (aarch64_classify_symbol): Likewise.
12654 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12655 (ldr_got_tiny): Define.
12657 2013-07-13 Tobias Grosser <tobias@grosser.es>
12659 PR tree-optimization/54094
12660 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12661 scheduling dimension for the parallelism check from the polyhedral
12662 information in the AST.
12663 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12666 2013-07-13 Jason Merrill <jason@redhat.com>
12668 * print-tree.c (debug_vec_tree): Use debug_raw.
12669 (debug_raw (vec<tree, va_gc> &)): New.
12670 (debug_raw (vec<tree, va_gc> *)): New.
12671 * tree.h: Declare them.
12673 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12675 * ifcvt.c (ifcvt_after_combine): New static variable.
12676 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12678 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12679 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12680 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12682 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12684 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12686 2013-07-12 Michael Matz <matz@suse.de>
12688 PR middle-end/55771
12689 * convert.c (convert_to_real): Reject non-float inner types.
12691 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12693 * config/aarch64/aarch64-protos.h
12694 (aarch64_simd_immediate_valid_for_move): Remove.
12695 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12696 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12698 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12700 2013-07-11 Steve Ellcey <sellcey@mips.com>
12702 * config/mips/mips.c (mips_conditional_register_usage): Do not
12703 use t[0-7] registers in MIPS16 mode when optimizing for size.
12705 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12707 * config/i386/i386.c (dispatch_function_versions): Fix array
12708 indexing of function_version_info to match actual_versions.
12710 2013-07-11 Teresa Johnson <tejohnson@google.com>
12712 * vec.h (struct va_gc): Move release out-of-line.
12713 (va_gc::release): Call ggc_free on released vec.
12715 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12717 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12718 Require GOT register as additional operand in UNSPEC.
12719 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12720 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12721 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12722 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12723 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12724 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12725 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12727 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12730 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12731 name seen by assembler/linker rather if available.
12733 2013-07-11 Andreas Schwab <schwab@suse.de>
12735 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12737 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12739 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12741 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12743 * doc/tm.texi.in: Move hook documentation to ....
12744 * target.def: ... here.
12746 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12747 text on @hook line.
12748 * doc/tm.texi: Regenerate.
12750 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12753 * doc/invoke.texi: Document -Wconditionally-supported.
12755 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12758 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12761 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12764 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12765 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12766 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12767 Remove duplicate devices.
12768 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12769 * config/avr/t-multilib: Regenerate.
12770 * config/avr/avr-tables.opt: Regenerate.
12771 * doc/avr-mmcu.texi: Regenerate.
12773 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12776 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12778 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12780 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12781 the cost of MULT when optimizing for size.
12783 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12785 * config/cr16/cr16-protos.h: Don't include target.h.
12787 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12789 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12790 adjust register size for TDmode and TFmode for VSX registers.
12792 2013-07-08 Kai Tietz <ktietz@redhat.com>
12795 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12796 hook_bool_const_tree_true.
12798 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12800 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12801 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12802 * config/s390/s390.md: Define FPR*_REGNUM constants.
12803 Fix FPR2_REGNUM constant (18 -> 17).
12804 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12805 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12806 ("trunc<BFP:mode><DFP_ALL:mode>2")
12807 ("trunc<DFP_ALL:mode><BFP:mode>2")
12808 ("*extend<BFP:mode><DFP_ALL:mode>2")
12809 ("*extend<DFP_ALL:mode><BFP:mode>2")
12810 ("extend<BFP:mode><DFP_ALL:mode>2")
12811 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12814 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12816 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12818 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12820 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12821 and cfun_fpr_bit_p to cfun_fpr_save_p.
12822 (s390_frame_area, s390_register_info, s390_frame_info)
12823 (s390_emit_prologue, s390_emit_epilogue)
12824 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12826 * config/s390/s390.h: Define *_REGNUM macros for floating point
12829 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12831 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12833 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12835 PR rtl-optimization/57786
12836 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12837 and break out of the loop when it is set to false.
12839 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12842 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12843 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12844 (const_int 63)) 0)).
12845 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12846 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12847 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12849 PR rtl-optimization/57829
12850 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12851 mask bits outside of mode are just sign-extension from mode to HWI.
12853 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12855 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12856 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12857 adjust_address instead of change_address to keep info about alignment.
12858 (emit_strmov): Remove.
12859 (emit_memmov): New function.
12860 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12861 (expand_movmem_epilogue): Likewise and return updated rtx for
12863 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12864 destination and source.
12865 (decide_alignment): Refactor, handle vector_loop.
12866 (ix86_expand_movmem): Likewise.
12867 (ix86_expand_setmem): Likewise.
12868 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12870 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12872 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12873 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12875 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12877 * config/i386/sse.md (sse_movlhps): Change alternative 3
12878 of operand 2 to "m".
12880 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12883 * config/i386/sse.md (iptr): New mode attribute.
12884 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12885 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12886 (<sse>_vmmul<mode>3): Ditto.
12887 (<sse>_vmdiv<mode>3): Ditto.
12888 (sse_vmrcpv4sf2): Ditto.
12889 (<sse>_vmsqrt<mode>2): Ditto.
12890 (sse_vmrsqrtv4sf2): Ditto.
12891 (<sse>_vm<code><mode>3): Ditto.
12892 (avx_vmcmp<mode>3): Ditto.
12893 (<sse>_vmmaskcmp<mode>3): Ditto.
12894 (<sse>_comi): Ditto.
12895 (<sse>_ucomi): Ditto.
12896 (*xop_vmfrcz_<mode>): Ditto.
12897 (*fmai_fmadd_<mode>): Ditto.
12898 (*fmai_fmsub_<mode>): Ditto.
12899 (*fmai_fnmadd_<mode>): Ditto.
12900 (*fmai_fnmsub_<mode>): Ditto.
12901 (*fma4i_vmfmadd_<mode>): Ditto.
12902 (*fma4i_vmfmsub_<mode>): Ditto.
12903 (*fma4i_vmfnmadd_<mode>): Ditto.
12904 (*fma4i_vmfnmsub_<mode>): Ditto.
12905 (*xop_vmfrcz_<mode>): Ditto.
12906 (sse_cvtps2pi): Ditto.
12907 (sse_cvttps2pi): Ditto.
12908 (sse_cvtss2si): Ditto.
12909 (sse_cvtss2si_2): Ditto.
12910 (sse_cvtss2siq_2): Ditto.
12911 (sse_cvttss2si): Ditto.
12912 (sse_cvttss2siq): Ditto.
12913 (sse_cvtsd2si): Ditto.
12914 (sse_cvtsd2si_2): Ditto.
12915 (sse_cvtsd2siq_2): Ditto.
12916 (sse_cvttsd2si): Ditto.
12917 (sse_cvttsd2siq): Ditto.
12918 (sse_cvtsd2ss): Ditto.
12919 (sse_cvtss2sd): Ditto.
12920 (avx2_pbroadcast<mode>): Ditto.
12921 (avx2_pbroadcast<mode>_1): Ditto.
12922 (*avx_vperm_broadcast_v4sf): Ditto.
12924 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12925 (sse_movlhps): Ditto.
12926 (sse_storehps): Ditto.
12927 (sse_loadhps): Ditto.
12928 (sse_storelps): Ditto.
12929 (sse_loadlps): Ditto.
12930 (*vec_concatv4sf): Ditto.
12931 (*vec_interleave_highv2df): Ditto.
12932 (*vec_interleave_lowv2df): Ditto.
12933 (*vec_extractv2df_1_sse): Ditto.
12934 (*vec_extractv2df_0_sse): Ditto.
12935 (sse2_storelpd): Ditto.
12936 (sse2_loadlpd): Ditto.
12937 (sse2_movsd): Ditto.
12938 (*vec_concatv4si): Ditto.
12939 (vec_concatv2di): Ditto.
12941 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12942 for Intel asm dialect.
12943 (mmx_punpcklwd): Ditto.
12944 (mmx_punpckldq): Ditto.
12946 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12947 for intel assembler dialect.
12949 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12952 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12953 for BUILT_IN_C{LZ,LRSB}*.
12954 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12955 * tree-vrp.c (extract_range_basic): Handle
12956 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12957 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12958 fall thru to code calling set_value*.
12959 * builtins.c (expand_builtin): Remove *IMAX cases.
12960 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12961 if width is bigger than 2*HWI.
12963 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12965 PR rtl-optimization/55342
12966 * lra-int.h (lra_subreg_reload_pseudos): New.
12967 * lra.c: Add undoing optional reloads to the block diagram.
12968 (lra_subreg_reload_pseudos): New.
12969 (lra_optional_reload_pseudos): Change comments.
12970 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12971 lra_optional_reload_pseudos after undo transformations.
12972 * lra-assigns.c (pseudo_prefix_title): New.
12973 (lra_setup_reg_renumber): Use it.
12974 (spill_for): Ditto. Check subreg reload pseudos too.
12975 (assign_by_spills): Consider subreg reload pseudos too.
12976 * lra-constraints.c (simplify_operand_subreg): Use
12977 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12978 (curr_insn_transform): Recognize and do optional reloads.
12979 (undo_optional_reloads): New.
12980 (lra_undo_inheritance): Call undo_optional_reloads.
12982 2013-07-05 Thomas Quinot <quinot@adacore.com>
12984 * tree-complex.c (expand_complex_operations_1): Fix typo.
12986 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12988 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12989 (tune_params): New member 'const vec_costs'.
12990 * config/aarch64/aarch64.c (generic_vector_cost): New.
12991 (generic_tunings): New member 'generic_vector_cost'.
12992 (aarch64_builtin_vectorization_cost): New.
12993 (aarch64_add_stmt_cost): New.
12994 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12995 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12997 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13000 * config/i386/predicates.md (vsib_address_operand): Disallow
13001 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13003 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13005 PR middle-end/55030
13006 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13007 expand_builtin_setjmp_receiver.
13008 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13009 with NULL for the label parameter.
13010 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13011 the frame-pointer. Adjust comments.
13012 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13013 only if LABEL is non-NULL.
13015 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13017 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13018 (ARM_ABI_AAPCS64): Ditto.
13020 (ARM_DEFAULT_ABI): Ditto.
13022 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13024 * config/aarch64/aarch64-builtins.c
13025 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13026 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13028 * config/aarch64/aarch64-simd.md
13029 (aarch64_ld1<VALL:mode>): New.
13030 (aarch64_st1<VALL:mode>): Likewise.
13031 * config/aarch64/arm_neon.h
13032 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13034 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13036 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13038 (ix86_option_override_internal):Turn on all -mavx target flags by
13039 default as they are dependent on AVX anyway.
13041 2013-07-02 Cary Coutant <ccoutant@google.com>
13043 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13044 deterministic hash.
13045 (loc_checksum_ordered): Likewise.
13046 (hash_loc_operands): Remove inline keyword.
13048 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13050 PR tree-optimization/57741
13051 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13052 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13053 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13054 Allow REAL_CST step_exprs if flag_associative_math.
13055 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13057 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13059 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13061 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13063 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13065 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13067 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13069 (iorsi3_insn): Likewise.
13070 (arm_xorsi3): Likewise.
13072 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13074 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13076 (attribute "type"): ... here, and prefix with "wmmx_".
13077 (attribute "core_cycles"): Update for attribute changes.
13078 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13079 (tbcstv4hi): Likewise.
13080 (tbcstv2si): Likewise.
13081 (iwmmxt_iordi3): Likewise.
13082 (iwmmxt_xordi3): Likewise.
13083 (iwmmxt_anddi3): Likewise.
13084 (iwmmxt_nanddi3): Likewise.
13085 (iwmmxt_arm_movdi): Likewise.
13086 (iwmmxt_movsi_insn): Likewise.
13087 (mov<mode>_internal): Likewise.
13088 (and<mode>3_iwmmxt): Likewise.
13089 (ior<mode>3_iwmmxt): Likewise.
13090 (xor<mode>3_iwmmxt): Likewise.
13091 (add<mode>3_iwmmxt): Likewise.
13092 (ssaddv8qi3): Likewise.
13093 (ssaddv4hi3): Likewise.
13094 (ssaddv2si3): Likewise.
13095 (usaddv8qi3): Likewise.
13096 (usaddv4hi3): Likewise.
13097 (usaddv2si3): Likewise.
13098 (sub<mode>3_iwmmxt): Likewise.
13099 (sssubv8qi3): Likewise.
13100 (sssubv4hi3): Likewise.
13101 (sssubv2si3): Likewise.
13102 (ussubv8qi3): Likewise.
13103 (ussubv4hi3): Likewise.
13104 (ussubv2si3): Likewise.
13105 (mulv4hi3_iwmmxt): Likewise.
13106 (smulv4hi3_highpart): Likewise.
13107 (umulv4hi3_highpart): Likewise.
13108 (iwmmxt_wmacs): Likewise.
13109 (iwmmxt_wmacsz): Likewise.
13110 (iwmmxt_wmacu): Likewise.
13111 (iwmmxt_wmacuz): Likewise.
13112 (iwmmxt_clrdi): Likewise.
13113 (iwmmxt_clrv8qi): Likewise.
13114 (iwmmxt_clr4hi): Likewise.
13115 (iwmmxt_clr2si): Likewise.
13116 (iwmmxt_uavgrndv8qi3): Likewise.
13117 (iwmmxt_uavgrndv4hi3): Likewise.
13118 (iwmmxt_uavgv8qi3): Likewise.
13119 (iwmmxt_uavgv4hi3): Likewise.
13120 (iwmmxt_tinsrb): Likewise.
13121 (iwmmxt_tinsrh): Likewise.
13122 (iwmmxt_tinsrw): Likewise.
13123 (iwmmxt_textrmub): Likewise.
13124 (iwmmxt_textrmsb): Likewise.
13125 (iwmmxt_textrmuh): Likewise.
13126 (iwmmxt_textrmsh): Likewise.
13127 (iwmmxt_textrmw): Likewise.
13128 (iwmxxt_wshufh): Likewise.
13129 (eqv8qi3): Likewise.
13130 (eqv4hi3): Likewise.
13131 (eqv2si3): Likewise.
13132 (gtuv8qi3): Likewise.
13133 (gtuv4hi3): Likewise.
13134 (gtuv2si3): Likewise.
13135 (gtv8qi3): Likewise.
13136 (gtv4hi3): Likewise.
13137 (gtv2si3): Likewise.
13138 (smax<mode>3_iwmmxt): Likewise.
13139 (umax<mode>3_iwmmxt): Likewise.
13140 (smin<mode>3_iwmmxt): Likewise.
13141 (umin<mode>3_iwmmxt): Likewise.
13142 (iwmmxt_wpackhss): Likewise.
13143 (iwmmxt_wpackwss): Likewise.
13144 (iwmmxt_wpackdss): Likewise.
13145 (iwmmxt_wpackhus): Likewise.
13146 (iwmmxt_wpackwus): Likewise.
13147 (iwmmxt_wpackdus): Likewise.
13148 (iwmmxt_wunpckihb): Likewise.
13149 (iwmmxt_wunpckihh): Likewise.
13150 (iwmmxt_wunpckihw): Likewise.
13151 (iwmmxt_wunpckilb): Likewise.
13152 (iwmmxt_wunpckilh): Likewise.
13153 (iwmmxt_wunpckilw): Likewise.
13154 (iwmmxt_wunpckehub): Likewise.
13155 (iwmmxt_wunpckehuh): Likewise.
13156 (iwmmxt_wunpckehuw): Likewise.
13157 (iwmmxt_wunpckehsb): Likewise.
13158 (iwmmxt_wunpckehsh): Likewise.
13159 (iwmmxt_wunpckehsw): Likewise.
13160 (iwmmxt_wunpckelub): Likewise.
13161 (iwmmxt_wunpckeluh): Likewise.
13162 (iwmmxt_wunpckeluw): Likewise.
13163 (iwmmxt_wunpckelsb): Likewise.
13164 (iwmmxt_wunpckelsh): Likewise.
13165 (iwmmxt_wunpckelsw): Likewise.
13166 (ror<mode>3): Likewise.
13167 (ashr<mode>3_iwmmxt): Likewise.
13168 (lshr<mode>3_iwmmxt): Likewise.
13169 (ashl<mode>3_iwmmxt): Likewise.
13170 (ror<mode>3_di): Likewise.
13171 (ashr<mode>3_di): Likewise.
13172 (lshr<mode>3_di): Likewise.
13173 (ashl<mode>3_di): Likewise.
13174 (iwmmxt_wmadds): Likewise.
13175 (iwmmxt_wmaddu): Likewise.
13176 (iwmmxt_tmia): Likewise.
13177 (iwmmxt_tmiaph): Likewise.
13178 (iwmmxt_tmiabb): Likewise.
13179 (iwmmxt_tmiatb): Likewise.
13180 (iwmmxt_tmiabt): Likewise.
13181 (iwmmxt_tmiatt): Likewise.
13182 (iwmmxt_tmovmskb): Likewise.
13183 (iwmmxt_tmovmskh): Likewise.
13184 (iwmmxt_tmovmskw): Likewise.
13185 (iwmmxt_waccb): Likewise.
13186 (iwmmxt_wacch): Likewise.
13187 (iwmmxt_waccw): Likewise.
13188 (iwmmxt_waligni): Likewise.
13189 (iwmmxt_walignr): Likewise.
13190 (iwmmxt_walignr0): Likewise.
13191 (iwmmxt_walignr1): Likewise.
13192 (iwmmxt_walignr2): Likewise.
13193 (iwmmxt_walignr3): Likewise.
13194 (iwmmxt_wsadb): Likewise.
13195 (iwmmxt_wsadh): Likewise.
13196 (iwmmxt_wsadbz): Likewise.
13197 (iwmmxt_wsadhz): Likewise.
13198 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13199 (iwmmxt_wabsdiffb): Likewise.
13200 (iwmmxt_wabsdiffh): Likewise.
13201 (iwmmxt_wabsdiffw): Likewise.
13202 (iwmmxt_waddsubhx): Likewise
13203 (iwmmxt_wsubaddhx): Likewise.
13204 (addc<mode>3): Likewise.
13205 (iwmmxt_avg4): Likewise.
13206 (iwmmxt_avg4r): Likewise.
13207 (iwmmxt_wmaddsx): Likewise.
13208 (iwmmxt_wmaddux): Likewise.
13209 (iwmmxt_wmaddsn): Likewise.
13210 (iwmmxt_wmaddun): Likewise.
13211 (iwmmxt_wmulwsm): Likewise.
13212 (iwmmxt_wmulwum): Likewise.
13213 (iwmmxt_wmulsmr): Likewise.
13214 (iwmmxt_wmulumr): Likewise.
13215 (iwmmxt_wmulwsmr): Likewise.
13216 (iwmmxt_wmulwumr): Likewise.
13217 (iwmmxt_wmulwl): Likewise.
13218 (iwmmxt_wqmulm): Likewise.
13219 (iwmmxt_wqmulwm): Likewise.
13220 (iwmmxt_wqmulmr): Likewise.
13221 (iwmmxt_wqmulwmr): Likewise.
13222 (iwmmxt_waddbhusm): Likewise.
13223 (iwmmxt_waddbhusl): Likewise.
13224 (iwmmxt_wqmiabb): Likewise.
13225 (iwmmxt_wqmiabt): Likewise.
13226 (iwmmxt_wqmiatb): Likewise.
13227 (iwmmxt_wqmiatt): Likewise.
13228 (iwmmxt_wqmiabbn): Likewise.
13229 (iwmmxt_wqmiabtn): Likewise.
13230 (iwmmxt_wqmiatbn): Likewise.
13231 (iwmmxt_wqmiattn): Likewise.
13232 (iwmmxt_wmiabb): Likewise.
13233 (iwmmxt_wmiabt): Likewise.
13234 (iwmmxt_wmiatb): Likewise.
13235 (iwmmxt_wmiatt): Likewise.
13236 (iwmmxt_wmiabbn): Likewise.
13237 (iwmmxt_wmiabtn): Likewise.
13238 (iwmmxt_wmiatbn): Likewise.
13239 (iwmmxt_wmiattn): Likewise.
13240 (iwmmxt_wmiawbb): Likewise.
13241 (iwmmxt_wmiawbt): Likewise.
13242 (iwmmxt_wmiawtb): Likewise.
13243 (iwmmxt_wmiawtt): Likewise.
13244 (iwmmxt_wmiawbbn): Likewise.
13245 (iwmmxt_wmiawbtn): Likewise.
13246 (iwmmxt_wmiawtbn): Likewise.
13247 (iwmmxt_wmiawttn): Likewise.
13248 (iwmmxt_wmerge): Likewise.
13249 (iwmmxt_tandc<mode>3): Likewise.
13250 (iwmmxt_torc<mode>3): Likewise.
13251 (iwmmxt_torvsc<mode>3): Likewise.
13252 (iwmmxt_textrc<mode>3): Likewise.
13253 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13254 (wmmxt_pack): Likewise.
13255 (wmmxt_mult_c1): Likewise.
13256 (wmmxt_mult_c2): Likewise.
13257 (wmmxt_alu_c1): Likewise.
13258 (wmmxt_alu_c2): Likewise.
13259 (wmmxt_alu_c3): Likewise.
13260 (wmmxt_transfer_c1): Likewise.
13261 (wmmxt_transfer_c2): Likewise.
13262 (wmmxt_transfer_c3): Likewise.
13263 (marvell_f_iwmmxt_wstr): Likewise.
13264 (marvell_f_iwmmxt_wldr): Likewise.
13266 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13268 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13270 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13273 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13274 * lra-constraints.c (need_for_split_p): Check call used hard regs
13275 living through calls.
13277 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13278 call used regs for call insn.
13280 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13283 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13284 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13286 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13288 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13290 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13292 * lra-constraints.c (need_for_split_p): Check call used hard regs
13293 living through calls.
13295 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13298 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13299 to tie with any other modes. Eliminate Altivec vector mode tests,
13300 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13301 code, to return 0 if testing MODE2 for a condition, if we've
13302 already tested MODE1 for the same condition.
13304 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13306 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13309 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13311 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13312 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13314 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13316 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13318 (aarch64_symbolic_constant_p): Remove.
13319 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13320 static. Fix line length and white space.
13321 (aarch64_symbolic_constant_p): Remove.
13322 * config/aarch64/predicates.md (aarch64_valid_symref):
13323 Use aarch64_classify_symbol_expression.
13325 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13327 * config/arm/constraints.md (Ts): New constraint.
13328 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13330 (compare_scc): Use "Ts" constraint for operand 0.
13331 (ior_scc_scc): Likewise.
13332 (and_scc_scc): Likewise.
13333 (and_scc_scc_nodom): Likewise.
13334 (ior_scc_scc_cmp): Likewise for operand 7.
13335 (and_scc_scc_cmp): Likewise.
13336 * config/arm/thumb2.md (thumb2_movsi_insn):
13337 Add alternatives for 16-bit encodings.
13338 (thumb2_movhi_insn): Likewise.
13339 (thumb2_movsicc_insn): Likewise.
13340 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13341 (thumb2_negscc): Use "Ts" constraint.
13342 Move mvn instruction outside cond_exec block.
13343 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13344 for 16-bit encodings.
13346 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13348 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13350 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13351 (mulsi3subsi): Likewise.
13352 (mulsidi3adddi): Likewise.
13353 (mulsidi3_v6): Likewise.
13354 (umulsidi3_v6): Likewise.
13355 (umulsidi3adddi_v6): Likewise.
13356 (smulsi3_highpart_v6): Likewise.
13357 (umulsi3_highpart_v6): Likewise.
13358 (mulhisi3tb): Likewise.
13359 (mulhisi3bt): Likewise.
13360 (mulhisi3tt): Likewise.
13361 (maddhisi4): Likewise.
13362 (maddhisi4tb): Likewise.
13363 (maddhisi4tt): Likewise.
13364 (maddhidi4): Likewise.
13365 (maddhidi4tb): Likewise.
13366 (maddhidi4tt): Likewise.
13367 (zeroextractsi_compare0_scratch): Likewise.
13368 (insv_zero): Likewise.
13369 (insv_t2): Likewise.
13370 (anddi_notzesidi_di): Likewise.
13371 (anddi_notsesidi_di): Likewise.
13372 (andsi_notsi_si): Likewise.
13373 (iordi_zesidi_di): Likewise.
13374 (xordi_zesidi_di): Likewise.
13375 (andsi_iorsi3_notsi): Likewise.
13376 (smax_0): Likewise.
13377 (smax_m1): Likewise.
13378 (smin_0): Likewise.
13379 (not_shiftsi): Likewise.
13380 (unaligned_loadsi): Likewise.
13381 (unaligned_loadhis): Likewise.
13382 (unaligned_loadhiu): Likewise.
13383 (unaligned_storesi): Likewise.
13384 (unaligned_storehi): Likewise.
13385 (extv_reg): Likewise.
13386 (extzv_t2): Likewise.
13387 (divsi3): Likewise.
13388 (udivsi3): Likewise.
13389 (arm_zero_extendhisi2addsi): Likewise.
13390 (arm_zero_extendqisi2addsi): Likewise.
13391 (compareqi_eq0): Likewise.
13392 (arm_extendhisi2_v6): Likewise.
13393 (arm_extendqisi2addsi): Likewise.
13394 (arm_movt): Likewise.
13395 (thumb2_ldrd): Likewise.
13396 (thumb2_ldrd_base): Likewise.
13397 (thumb2_ldrd_base_neg): Likewise.
13398 (thumb2_strd): Likewise.
13399 (thumb2_strd_base): Likewise.
13400 (thumb2_strd_base_neg): Likewise.
13401 (arm_negsi2): Add alternative for 16-bit encoding.
13402 (arm_one_cmplsi2): Likewise.
13404 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13406 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13407 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13408 (movdfcc): Likewise.
13409 * config/arm/vfp.md (*thumb2_movsf_vfp):
13410 Disable predication for arm_restrict_it.
13411 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13412 (*thumb2_movdfcc_vfp): Likewise.
13413 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13414 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13415 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13416 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13417 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13418 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13419 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13420 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13421 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13422 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13423 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13424 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13425 Disable predication for arm_restrict_it.
13427 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13429 * config/i386/bmiintrin.h (_bextr_u32): New.
13430 (_bextr_u64): Ditto.
13432 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13434 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13435 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13436 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13437 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13438 * config/mips/n32-elf.h: ...this new file.
13440 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13443 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13444 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13446 2013-06-27 Catherine Moore <clm@codesourcery.com>
13448 * config/mips/mips-tables.opt: Regenerate.
13449 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13450 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13451 * doc/invoke.texi: Add -m14kc.
13453 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13456 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13457 constraints of operand 1 and 2.
13460 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13461 to match RTL canonicalization. Swap predicates and
13462 constraints of operand 1 and 2.
13464 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13466 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13467 Process OP_INOUT regs for splitting too.
13469 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13471 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13472 decl before the loop, initialize to NULL.
13473 (vectorizable_load): Initialize ptr_incr to NULL.
13475 2013-06-27 Martin Jambor <mjambor@suse.cz>
13478 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13479 * ipa-ref.c (ipa_maybe_record_reference): New function.
13480 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13481 * ipa-cp.c (create_specialized_node): Record potential references from
13483 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13485 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13487 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13488 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13489 'mode' to force_reg.
13490 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13491 (aarch64_expand_mov_immediate): Likewise.
13493 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13495 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13496 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13498 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13500 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13501 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13502 (struct machine_function): Add tbegin_p.
13503 (s390_canonicalize_comparison): Fold CC mode compares to
13504 conditional jump if possible.
13505 (s390_emit_jump): Return the emitted jump.
13506 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13507 Handle CCRAWmode compares.
13508 (s390_option_override): Default to -mhtm if available.
13509 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13510 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13511 FPRs instead of df_regs_ever_live_p.
13512 (s390_optimize_nonescaping_tx): New function.
13513 (s390_init_frame_layout): Extend clobbered_regs array to cover
13515 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13516 (s390_expand_tbegin): New function.
13517 (enum s390_builtin): New enum definition.
13518 (code_for_builtin): New array definition.
13519 (s390_init_builtins): New function.
13520 (s390_expand_builtin): New function.
13521 (TARGET_INIT_BUILTINS): Define.
13522 (TARGET_EXPAND_BUILTIN): Define.
13523 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13524 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13525 (s390_alc_comparison): Likewise.
13526 * config/s390/s390-modes.def: Add CCRAWmode.
13527 * config/s390/s390.h (processor_flags): Add PF_TX.
13528 (TARGET_CPU_HTM): Define macro.
13529 (TARGET_HTM): Define macro.
13530 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13531 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13532 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13533 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13534 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13535 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13536 ("*cc_to_int"): Move up.
13537 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13538 constants other than 0.
13539 ("*ccraw_to_int"): New insn and splitter definition.
13540 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13541 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13542 ("tx_assist"): New expander.
13543 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13544 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13545 * config/s390/s390.opt: Add -mhtm option.
13546 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13547 * config/s390/htmxlintrin.h: New file.
13548 * config/s390/htmintrin.h: New file.
13549 * config/s390/s390intrin.h: New file.
13550 * doc/extend.texi: Document htm builtins.
13551 * config.gcc: Add the new header files to extra_headers.
13553 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13555 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13556 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13558 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13559 Pat Haugen <pthaugen@us.ibm.com>
13560 Peter Bergner <bergner@vnet.ibm.com>
13562 * config/rs6000/power8.md: New.
13563 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13564 setting for power8 entry.
13565 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13566 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13567 test for Power4/Power5 only.
13568 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13570 (force_new_group): Adjust comment.
13571 * config/rs6000/rs6000.md: Include power8.md.
13573 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13575 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13576 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13578 (tune_params): Update comment.
13579 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13580 (arm_max_conditional_execute): New function.
13581 (thumb2_final_prescan_insn): Use max_insn_skipped and
13582 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13584 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13586 PR tree-optimization/57705
13587 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13588 SSA_NAME step, provided that it is not defined inside the loop.
13589 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13590 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13592 2013-06-25 Martin Jambor <mjambor@suse.cz>
13594 PR middle-end/57670
13595 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13596 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13598 (ipa_note_param_call): Initialize member_ptr flag.
13599 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13600 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13601 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13602 (ipa_read_indirect_edge_info): Likewise.
13604 2013-06-25 Richard Biener <rguenther@suse.de>
13606 PR middle-end/56977
13607 * passes.c (init_optimization_passes): Move pass_fold_builtins
13608 and pass_dce earlier with -Og.
13610 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13612 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13613 <BIT_FIELD_REF>: Remove trailing TAB.
13614 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13617 2013-06-24 Martin Jambor <mjambor@suse.cz>
13619 PR tree-optimization/57358
13620 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13621 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13622 (ipa_analyze_params_uses): Generate pessimistic info when true.
13624 2013-06-24 Martin Jambor <mjambor@suse.cz>
13626 PR tree-optimization/57539
13627 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13628 global.inlined_to of the new node to it. All callers changed.
13629 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13630 inlining_into, pass it to cgraph_clone_node.
13631 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13632 ipa_free_edge_args_substructures.
13633 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13634 rdesc linked list. Do not assert rdesc edges have inlined caller.
13635 Assert we have found an rdesc in the rdesc list.
13637 2013-06-24 Richard Biener <rguenther@suse.de>
13639 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13640 (pointer_set_lookup): Declare.
13641 (class pointer_map): New template class implementing a
13642 generic pointer to T map.
13643 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13644 pointer_map<T>::contains, pointer_map<T>::insert,
13645 pointer_map<T>::traverse): New functions.
13646 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13647 (pointer_set_lookup): New function.
13648 (pointer_set_contains): Use pointer_set_lookup.
13649 (pointer_set_insert): Likewise.
13650 (insert_aux): Remove.
13651 (struct pointer_map_t): Embed a pointer_set_t.
13652 (pointer_map_create): Adjust.
13653 (pointer_map_destroy): Likewise.
13654 (pointer_map_contains): Likewise.
13655 (pointer_map_insert): Likewise.
13656 (pointer_map_traverse): Likewise.
13657 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13658 pointer_map<unsigned> instead of a pointer_map_t.
13659 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13660 (streamer_tree_cache_lookup): Likewise.
13661 (streamer_tree_cache_create): Likewise.
13662 (streamer_tree_cache_delete): Likewise.
13663 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13664 pointer_map<unsigned> instead of a pointer_map_t.
13665 (lto_init_tree_ref_encoder): Adjust.
13666 (lto_destroy_tree_ref_encoder): Likewise.
13667 * lto-section-out.c (lto_output_decl_index): Likewise.
13668 (lto_record_function_out_decl_state): Likewise.
13669 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13671 2013-06-24 Richard Biener <rguenther@suse.de>
13673 PR tree-optimization/57488
13674 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13676 2013-06-24 Alan Modra <amodra@gmail.com>
13678 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13679 (gen_easy_altivec_constant): Likewise.
13680 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13681 easy_vector_constant_msb): Likewise.
13683 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13686 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13687 add missing return true.
13689 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13692 * config/sh/predicates.md (general_extend_operand): Invoke
13693 general_movsrc_operand for memory operands.
13694 (general_movsrc_operand): Allow reg+reg addressing, do not use
13695 general_operand for memory operands.
13697 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13699 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13700 when current target options does not apply.
13701 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13702 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13703 * config/i386/bmiintrin.h: Pass appropriate target
13704 attributes to header.
13705 * config/i386/mmintrin.h: Ditto.
13706 * config/i386/nmmintrin.h: Ditto.
13707 * config/i386/avx2intrin.h: Ditto.
13708 * config/i386/fxsrintrin.h: Ditto.
13709 * config/i386/tbmintrin.h: Ditto.
13710 * config/i386/xsaveintrin.h: Ditto.
13711 * config/i386/f16cintrin.h: Ditto.
13712 * config/i386/xtestintrin.h: Ditto.
13713 * config/i386/xsaveoptintrin.h: Ditto.
13714 * config/i386/bmi2intrin.h: Ditto.
13715 * config/i386/lzcntintrin.h: Ditto.
13716 * config/i386/smmintrin.h: Ditto.
13717 * config/i386/wmmintrin.h: Ditto.
13718 * config/i386/x86intrin.h: Remove all header include guards.
13719 * config/i386/prfchwintrin.h: Ditto.
13720 * config/i386/pmmintrin.h: Ditto.
13721 * config/i386/tmmintrin.h: Ditto.
13722 * config/i386/xmmintrin.h: Ditto.
13723 * config/i386/popcntintrin.h: Ditto.
13724 * config/i386/rdseedintrin.h: Ditto.
13725 * config/i386/ammintrin.h: Ditto.
13726 * config/i386/emmintrin.h: Ditto.
13727 * config/i386/immintrin.h: Remove all header include guards.
13728 * config/i386/fma4intrin.h: Ditto.
13729 * config/i386/lwpintrin.h: Ditto.
13730 * config/i386/xopintrin.h: Ditto.
13731 * config/i386/ia32intrin.h: Ditto.
13732 * config/i386/avxintrin.h: Ditto.
13733 * config/i386/rtmintrin.h: Ditto.
13734 * config/i386/fmaintrin.h: Ditto.
13735 * config/i386/mm3dnow.h: Ditto.
13737 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13739 * common/config/i386/i386-common.c: Handle LZCNT.
13741 2013-06-22 Andi Kleen <ak@linux.intel.com>
13743 * doc/extend.texi: Use __atomic_store_n instead of
13744 __atomic_store in HLE example.
13746 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13748 * config/sh/sh.c: Remove <cstdlib> workaround.
13750 2013-06-21 Andi Kleen <ak@linux.intel.com>
13752 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13754 2013-06-21 Andi Kleen <ak@linux.intel.com>
13756 * doc/extend.texi: Document that __atomic_clear and
13757 __atomic_test_and_set should only be used with bool.
13759 2013-06-20 Jan Hubicka <jh@suse.cz>
13761 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13762 types_same_for_odr.
13763 * tree.c (decls_same_for_odr): New function.
13764 (same_for_edr): New function.
13765 (types_same_for_odr): New function.
13766 (get_binfo_at_offset): Use it.
13767 * tree.h (types_same_for_odr): Declare.
13769 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13770 Jason Merrill <jason@redhat.com>
13772 * system.h: Include <cstdlib> as well as <stdlib.h>.
13774 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13777 * config/i386/i386.c (construct_container): Report error if
13778 long double is used with disabled x87 float returns.
13780 2013-06-20 Jan Hubicka <jh@suse.cz>
13782 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13784 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13786 PR rtl-optimization/57425
13787 PR rtl-optimization/57569
13788 * alias.c (write_dependence_p): Remove parameters mem_mode and
13789 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13790 Changed all callers.
13791 (canon_anti_dependence): Get comments and semantics in sync.
13792 Add parameter mem_canonicalized. Changed all callers.
13793 * rtl.h (canon_anti_dependence): Update prototype.
13795 2013-06-20 Richard Biener <rguenther@suse.de>
13797 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13798 byte case, inline streamer_read_uchar and defer section
13801 2013-06-20 Richard Biener <rguenther@suse.de>
13803 PR tree-optimization/57584
13804 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13805 SSA names into the expanded expression that take part in
13806 abnormal coalescing.
13808 2013-06-19 Sharad Singhai <singhai@google.com>
13810 * gcov.c (print_usage): Handle new option.
13811 (process_args): Ditto.
13812 (get_gcov_intermediate_filename): New function.
13813 (output_intermediate_file): New function.
13814 (output_gcov_file): New function
13815 (generate_results): Handle new option.
13816 (release_function): Relase demangled name.
13817 (read_graph_file): Handle demangled name.
13818 (output_lines): Ditto.
13819 * doc/gcov.texi: Document gcov intermediate format.
13821 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13824 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13825 (lra_emit_add): Use the functions. Add comment about Y as an
13828 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13831 * collect2.c (collect_atexit): New.
13832 (collect_exit): Delete.
13833 (main): Register collect_atexit with atexit.
13834 (collect_wait): Change collect_exit to exit.
13836 * collect2.h (collect_exit): Delete.
13837 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13839 2013-06-19 Wei Mi <wmi@google.com>
13841 PR rtl-optimization/57518
13842 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13843 if regno is used in paradoxical subreg.
13844 (update_equiv_regs): Check pdx_subregs[regno] before
13845 set a reg to be equivalent with a mem.
13847 2013-06-19 Matthias Klose <doko@ubuntu.com>
13850 * file-find.h (find_a_file): Add a mode parameter.
13851 * file-find.c (find_a_file): Likewise.
13852 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13853 with X_OK for the executables.
13854 * collect2.c (main): Call find_a_file with X_OK.
13856 2013-06-19 Steve Ellcey <sellcey@mips.com>
13859 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13860 Use NEXT_INSN instead of next_real_insn.
13862 2013-06-19 Jan Hubicka <jh@suse.cz>
13864 * cgraph.h (const_value_known_p): Replace by ...
13865 (ctor_for_folding): .. this one.
13866 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13867 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13868 * expr.c (expand_expr_real_1): Likewise.
13869 (string_constant): Likewise.
13870 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13871 * ipa.c (process_references): Likewise.
13872 (symtab_remove_unreachable_nodes): Likewise.
13873 * ipa-inline-analysis.c (param_change_prob): Likewise.
13874 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13875 (get_base_constructor): Likwise.
13876 * varpool.c (varpool_remove_node): Likewise.
13877 (varpool_remove_initializer): LIkewise.
13878 (dump_varpool_node): LIkwise.
13879 (const_value_known_p): Rewrite to ...
13880 (ctor_for_folding): ... this one.
13882 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13885 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13886 PERSONALITY in $PATH derived prefixes.
13888 2013-06-19 Jeff Law <law@redhat.com>
13890 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13893 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13894 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13897 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13899 * config/arm/vfp.md: Move VFP instruction classification documentation
13901 * config/arm/arm.md: ... here. Update instruction classification
13904 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13906 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13907 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13908 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13910 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13913 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13914 with NEXT_INSN. Use tablejump_p to check for jump table data
13917 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13920 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13921 that now in C++ the value is correct per the C++ standards.
13923 2013-06-19 Richard Biener <rguenther@suse.de>
13925 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13926 for global context.
13928 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13931 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13934 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13935 with next_active_insn.
13937 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13939 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13940 functions are inlined during failures to flag an error.
13941 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13942 in early inline pass.
13944 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13946 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13949 2013-06-18 Julian Brown <julian@codesourcery.com>
13951 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13952 Permit virtual register pre-reload if !strict.
13953 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13955 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13957 * config/arm/neon.md (movmisalign<mode>): Use
13958 neon_perm_struct_or_reg_operand instead of
13959 neon_struct_or_register_operand.
13960 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13961 neon_permissive_struct_operand instead of neon_struct_operand.
13962 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13963 neon_vector_mem_operand.
13964 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13965 neon_vector_mem_operand.
13966 (neon_permissive_struct_operand): New.
13967 (neon_struct_or_register_operand): Rename to...
13968 (neon_perm_struct_or_reg_operand): This. Adjust call to
13969 neon_vector_mem_operand.
13971 2013-06-18 Richard Biener <rguenther@suse.de>
13973 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13974 * lto-streamer.h: Include pointer-set.h.
13975 (struct lto_decl_slot): Remove.
13976 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13977 Remove next_index entry.
13978 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13979 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13980 (lto_init_tree_ref_encoder): Adjust.
13981 (lto_destroy_tree_ref_encoder): Likewise.
13982 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13983 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13984 (lto_output_decl_index): Adjust.
13985 (lto_new_out_decl_state): Likewise.
13986 (lto_record_function_out_decl_state): Likewise.
13987 * lto-streamer-out.c (copy_function): Likewise.
13989 2013-06-18 Richard Biener <rguenther@suse.de>
13991 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13992 * cgraphunit.c: Include cfgloop.h.
13993 (init_lowered_empty_function): Initialize the loop tree.
13994 (assemble_thunk): Insert new BBs into loops.
13996 2013-06-18 Richard Biener <rguenther@suse.de>
13998 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13999 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14000 the map from cache entry to cache index optional.
14001 (streamer_tree_cache_replace_tree): Adjust accordingly.
14002 (streamer_tree_cache_append): Likewise.
14003 (streamer_tree_cache_delete): Likewise.
14004 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14005 streamer cache map from cache entry to cache index.
14006 * lto-streamer-out.c (create_output_block): Adjust.
14008 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14010 * config/arm/arm.md (attribute "insn"): Move multiplication and
14011 division attributes to...
14012 (attribute "type"): ... here. Remove mult.
14013 (attribute "mul32"): New attribute.
14014 (attribute "mul64"): Add umaal.
14015 (*arm_mulsi3): Update attributes.
14016 (*arm_mulsi3_v6): Likewise.
14017 (*thumb_mulsi3): Likewise.
14018 (*thumb_mulsi3_v6): Likewise.
14019 (*mulsi3_compare0): Likewise.
14020 (*mulsi3_compare0_v6): Likewise.
14021 (*mulsi_compare0_scratch): Likewise.
14022 (*mulsi_compare0_scratch_v6): Likewise.
14023 (*mulsi3addsi): Likewise.
14024 (*mulsi3addsi_v6): Likewise.
14025 (*mulsi3addsi_compare0): Likewise.
14026 (*mulsi3addsi_compare0_v6): Likewise.
14027 (*mulsi3addsi_compare0_scratch): Likewise.
14028 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14029 (*mulsi3subsi): Likewise.
14030 (*mulsidi3adddi): Likewise.
14031 (*mulsi3addsi_v6): Likewise.
14032 (*mulsidi3adddi_v6): Likewise.
14033 (*mulsidi3_nov6): Likewise.
14034 (*mulsidi3_v6): Likewise.
14035 (*umulsidi3_nov6): Likewise.
14036 (*umulsidi3_v6): Likewise.
14037 (*umulsidi3adddi): Likewise.
14038 (*umulsidi3adddi_v6): Likewise.
14039 (*smulsi3_highpart_nov6): Likewise.
14040 (*smulsi3_highpart_v6): Likewise.
14041 (*umulsi3_highpart_nov6): Likewise.
14042 (*umulsi3_highpart_v6): Likewise.
14043 (mulhisi3): Likewise.
14044 (*mulhisi3tb): Likewise.
14045 (*mulhisi3bt): Likewise.
14046 (*mulhisi3tt): Likewise.
14047 (maddhisi4): Likewise.
14048 (*maddhisi4tb): Likewise.
14049 (*maddhisi4tt): Likewise.
14050 (maddhidi4): Likewise.
14051 (*maddhidi4tb): Likewise.
14052 (*maddhidi4tt): Likewise.
14053 (divsi3): Likewise.
14054 (udivsi3): Likewise.
14055 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14056 (thumb2_mulsi_short_compare0): Likewise.
14057 (thumb2_mulsi_short_compare0_scratch): Likewise.
14058 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14059 (1020mult2): Likewise.
14060 (1020mult3): Likewise.
14061 (1020mult4): Likewise.
14062 (1020mult5): Likewise.
14063 (1020mult6): Likewise.
14064 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14066 (cortex_a15_mult64): Likewise.
14067 (cortex_a15_sdiv): Likewise.
14068 (cortex_a15_udiv): Likewise.
14069 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14075 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14076 (pj4_ir_div): Likewise.
14077 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14078 (11_mult2): Likewise.
14079 (11_mult3): Likewise.
14080 (11_mult4): Likewise.
14081 (11_mult5): Likewise.
14082 (11_mult6): Likewise.
14083 (11_mult7): Likewise.
14084 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14085 (cortex_a8_mla): Likewise.
14086 (cortex_a8_mull): Likewise.
14087 (cortex_a8_smulwy): Likewise.
14088 (cortex_a8_smlald): Likewise.
14089 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14090 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14091 (cortex_r4_mul_3): Likewise.
14092 (cortex_r4_mla_4): Likewise.
14093 (cortex_r4_mla_3): Likewise.
14094 (cortex_r4_smlald): Likewise.
14095 (cortex_r4_mull): Likewise.
14096 (cortex_r4_sdiv): Likewise.
14097 (cortex_r4_udiv): Likewise.
14098 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14099 (cortex_a7_idiv): Likewise.
14100 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14101 (9_mult2): Likewise.
14102 (9_mult3): Likewise.
14103 (9_mult4): Likewise.
14104 (9_mult5): Likewise.
14105 (9_mult6): Likewise.
14106 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14107 (cortex_a53_sdiv): Likewise.
14108 (cortex_a53_udiv): Likewise.
14109 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14110 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14111 (mp626_mult2): Likewise.
14112 (mp626_mult3): Likewise.
14113 (mp626_mult4): Likewise.
14114 * config/arm/fa526.md (526_mult1): Update attribute change.
14115 (526_mult2): Likewise.
14116 * config/arm/arm-generic.md (mult): Update attribute change.
14117 (mult_ldsched_strongarm): Likewise.
14118 (mult_ldsched): Likewise.
14119 (multi_cycle): Likewise.
14120 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14121 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14122 (606te_mult2): Likewise.
14123 (606te_mult3): Likewise.
14124 (606te_mult4): Likewise.
14125 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14126 (cortex_a9_mac16): Likewise.
14127 (cortex_a9_multiply): Likewise.
14128 (cortex_a9_mac): Likewise.
14129 (cortex_a9_multiply_long): Likewise.
14130 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14131 (626te_mult2): Likewise.
14132 (626te_mult3): Likewise.
14133 (626te_mult4): Likewise.
14135 2013-06-18 Richard Biener <rguenther@suse.de>
14138 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14140 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14143 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14144 with next_active_insn.
14146 2013-06-18 Alan Modra <amodra@gmail.com>
14148 * config/rs6000/rs6000.h (enum data_align): New.
14149 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14150 (DATA_ABI_ALIGNMENT): Define.
14151 (CONSTANT_ALIGNMENT): Correct comment.
14152 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14153 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14155 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14157 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14158 ATTRIBUTE_UNUSED marking.
14160 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14162 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14163 alternative and update.
14164 (aarch64_dup_lanedi): Delete.
14165 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14166 * config/aarch64/aarch64-simd-builtins.def: Update.
14168 2013-06-17 Richard Biener <rguenther@suse.de>
14170 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14171 (lto_input_scc): Declare.
14172 (lto_input_tree_1): Likewise.
14173 (struct lto_stats_d): Add num_tree_bodies_output and
14174 num_pickle_refs_output.
14175 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14176 (lto_read_tree_1): Split out from ...
14177 (lto_read_tree): ... this.
14178 (lto_input_scc): New function.
14179 (lto_input_tree_1): Split out from ...
14180 (lto_input_tree): ... this. Handle LTO_tree_scc.
14181 (lto_data_in_create): Create the streamer cache without hashes.
14182 * lto-streamer-out.c (create_output_block): Create the streamer
14183 cache with hashes when not doing WPA.
14184 (lto_write_tree_1): Split out from ...
14185 (lto_write_tree): ... this.
14186 (get_symbol_initial_value): New function.
14187 (lto_output_tree_1): Split out from ...
14188 (lto_output_tree): ... this. Write trees as series of SCCs
14189 using a DFS walk via DFS_write_tree.
14190 (struct sccs, struct scc_entry): New types.
14191 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14192 (DFS_write_tree_body): New function.
14193 (DFS_write_tree): Likewise.
14194 (hash_tree): Likewise.
14195 (scc_entry_compare): Likewise.
14196 (hash_scc): Likewise.
14197 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14198 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14199 TREE_CHAIN as regular reference.
14200 (streamer_read_integer_cst): Remove.
14201 (streamer_get_pickled_tree): Adjust.
14202 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14203 of DECL_EXTERNALs in BLOCK_VARS for now.
14204 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14206 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14207 Add hash value argument and record that if hashes are recorded
14209 (streamer_tree_cache_insert_1): Adjust.
14210 (streamer_tree_cache_insert): Likewise.
14211 (streamer_tree_cache_insert_at): Rename to ...
14212 (streamer_tree_cache_replace_tree): ... this and adjust.
14213 (streamer_tree_cache_append): Adjust.
14214 (record_common_node): Likewise.
14215 (streamer_tree_cache_create): Add argument whether to
14216 record hash values together with trees.
14217 (streamer_tree_cache_delete): Adjust.
14218 * tree-streamer.h (struct streamer_tree_cache_d): Add
14220 (streamer_read_integer_cst): Remove.
14221 (streamer_tree_cache_insert): Adjust.
14222 (streamer_tree_cache_append): Likewise.
14223 (streamer_tree_cache_insert_at): Rename to ...
14224 (streamer_tree_cache_replace_tree): ... this and adjust.
14225 (streamer_tree_cache_create): Add argument whether to record hashes.
14226 (streamer_tree_cache_get): Rename to ...
14227 (streamer_tree_cache_get_tree): ... this.
14228 (streamer_tree_cache_get_hash): New function.
14229 * tree.c (cache_integer_cst): New function.
14230 * tree.h (cache_integer_cst): Declare.
14231 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14232 * lto-symtab.c (lto_varpool_replace_node): Only release
14233 DECL_INITIAL of non-prevailing decls.
14234 * varpool.c (varpool_remove_initializer): Do not release
14235 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14237 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14239 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14240 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14241 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14242 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14243 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14244 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14245 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14246 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14247 instead of TARGET_64BIT.
14248 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14249 Require ISA_HAS_<D>DIV.
14251 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14253 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14254 (mips*-*-linux*): Move default with_llsc setting to where other
14256 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14258 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14259 Likewise. Remove default with_tune setting. Move default float
14260 setting to its own block. Handle with_llsc in the same block as above.
14262 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14264 PR rtl-optimization/57425
14265 PR rtl-optimization/57569
14266 * alias.c (write_dependence_p): Add new parameters mem_mode,
14267 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14268 Changed all callers.
14269 (canon_anti_dependence): New function.
14270 * cse.c (check_dependence): Use canon_anti_dependence.
14271 * cselib.c (cselib_invalidate_mem): Likewise.
14272 * rtl.h (canon_anti_dependence): Declare.
14274 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14276 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14277 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14278 ".set mips3" for 64-bit targets.
14280 2013-06-15 Dehao Chen <dehao@google.com>
14282 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14283 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14284 (gimple_check_call_args): Likewise.
14285 * value-prof.c (check_ic_target): Likewise.
14286 * ipa-inline.c (early_inliner): Likewise.
14287 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14288 * cgraph.c (cgraph_create_edge_1): Likewise.
14289 (cgraph_make_edge_direct): Likewise.
14291 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14294 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14295 rs6000_output_move_128bit to handle emitting quad memory
14296 operations. Set attribute length to 8 bytes.
14298 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14300 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14302 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14303 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14304 (aarch64_<su>mlsl<mode>): Likewise.
14306 2013-06-14 Mike Stump <mikestump@comcast.net>
14308 * Makefile.in (TARGET_H): Add insn-codes.h.
14310 2013-06-14 Alan Modra <amodra@gmail.com>
14312 PR middle-end/57134
14313 PR middle-end/57586
14314 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14315 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14316 bitfield expansion when EXPAND_MEMORY.
14317 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14319 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14321 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14322 test for clearing quad memory on 32-bit later.
14324 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14326 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14327 (fold_negate_expr): Likewise.
14328 (fold_real_zero_addition_p): Handle vectors.
14329 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14331 2013-06-14 Alan Modra <amodra@gmail.com>
14333 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14335 2013-06-13 Jan Hubicka <jh@suse.cz>
14337 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14338 Local comdats are not externally visible.
14339 * symtab.c (dump_symtab_base): Dump externally visible.
14340 (verify_symtab_base): Verify back links in the symtab hash.
14342 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14344 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14345 CONVERT_EXPR as equal nodes.
14347 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14349 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14351 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14353 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14354 Generalize to complex and vector.
14355 * tree.c (build_all_ones_cst): New function.
14356 * tree.h (build_all_ones_cst): Declare it.
14358 2013-06-13 Alan Modra <amodra@gmail.com>
14360 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14361 * config/rs6000/rs6000.md (signbittf2): New insn.
14362 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14363 (abstf2_internal, cmptf_internal2): Likewise.
14364 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14366 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14367 Pat Haugen <pthaugen@us.ibm.com>
14368 Peter Bergner <bergner@vnet.ibm.com>
14370 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14371 power8 byte, half-word, and quad-word atomic instructions.
14372 (emit_store_conditional): Likewise.
14373 (rs6000_expand_atomic_compare_and_swap): Likewise.
14374 (rs6000_expand_atomic_op): Likewise.
14376 * config/rs6000/sync.md (larx): Add new modes for power8.
14378 (AINT): New mode iterator to include TImode as well as normal
14379 integer modes on power8.
14380 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14381 that VSX registers are not considered. Use AINT mode iterator
14382 instead of INT1 to allow inclusion of quad word atomic operations
14384 (load_locked<mode>): Likewise.
14385 (store_conditional<mode>): Likewise.
14386 (atomic_compare_and_swap<mode>): Likewise.
14387 (atomic_exchange<mode>): Likewise.
14388 (atomic_nand<mode>): Likewise.
14389 (atomic_fetch_<fetchop_name><mode>): Likewise.
14390 (atomic_nand_fetch<mode>): Likewise.
14391 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14393 (ATOMIC): On power8, add QImode, HImode modes.
14394 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14395 modes that promote to SImode.
14396 (load_lockedti): Convert TImode arguments to PTImode, so that we
14397 get a guaranteed even/odd register pair.
14398 (load_lockedpti): Likewise.
14399 (store_conditionalti): Likewise.
14400 (store_conditionalpti): Likewise.
14402 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14403 atomic load/store instructions.
14406 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14408 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14410 (insn_count): New attribute, with most cases extracted from...
14411 (length): ...here. Redefine most cases in terms of insn_count.
14412 (single_insn): Delete.
14413 (can_delay): Use insn_count to check for single instructions.
14414 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14415 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14416 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14417 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14418 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14419 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14420 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14421 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14422 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14423 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14424 rather than "length".
14425 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14426 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14427 Use "insn_count" rather than "length".
14428 * config/mips/mips-dsp.md
14429 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14430 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14433 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14435 PR tree-optimization/57361
14436 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14438 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14440 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14442 (aarch64_simd_combine<mode>): New instruction expansion.
14443 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14444 function prototype.
14445 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14446 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14448 2013-06-12 Jan Hubicka <jh@suse.cz>
14450 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14451 decl has when in streaming stage.
14452 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14453 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14455 2013-06-12 Roland Stigge <stigge@antcom.de>
14458 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14460 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14462 PR tree-optimization/57537
14463 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14464 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14466 2013-06-12 Richard Biener <rguenther@suse.de>
14468 * data-streamer.h (streamer_write_char_stream): CSE
14469 obs->current_pointer.
14470 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14471 streamer_write_char_stream manually and optimize the resulting loop.
14472 (streamer_write_hwi_stream): Likewise.
14474 2013-06-12 Jan Hubicka <jh@suse.cz>
14476 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14477 * cgraph.h (varpool_create_empty_node): Declare.
14478 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14480 * symtab.c (symtab_unregister_node): Be lax about missin entries
14482 (symtab_get_node): Update comment.
14483 * varpool.c (varpool_create_empty_node): Break out from ...
14484 (varpool_node_for_decl): ... here.
14485 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14487 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14489 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14490 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14491 part. Use straight-line flow at the end.
14492 <COMPONENT_REF>: Remove superfluous else.
14493 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14495 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14498 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14499 target hook even for !TREE_PUBLIC decls. If no resolution info
14500 is available, return false for common and external decls.
14502 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14504 * config/rl78/constraints.md (U): New constraint.
14505 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14508 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14511 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14512 to allow returning address to AT_PLATFORM name.
14514 2013-06-11 Jan Hubicka <jh@suse.cz>
14516 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14517 * cgraph.h (symtab_node_base): Add weakref flag.
14518 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14519 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14520 (output_weakrefs): Use weakref flag.
14521 * fold-const.c (simple_operand_p): Handle WEAK.
14522 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14523 * ipa.c (varpool_externally_visible_p): Drop weakref.
14524 (function_and_variable_visibility): Update comment; fix weakref
14525 sanity checks; do not clear DECL_WEAK on them.
14526 * lto-cgraph.c (lto_output_node): update.
14527 (lto_output_varpool_node): Update.
14528 (input_overwrite_node): Update.
14529 (input_node): Update.
14530 (input_varpool_node): Update.
14531 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14532 (lto_symtab_merge_symbols): Add sanity check.
14533 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14534 * passes.c (rest_of_decl_compilation): Set static flag, too.
14535 * symtab.c (dump_symtab_base): Dump weakref.
14536 (verify_symtab_base): Sanity check weakrefs.
14537 (symtab_make_decl_local): Remove duplicated code.
14538 (symtab_alias_ultimate_target): Simplify.
14539 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14541 2013-06-11 Tom de Vries <tom@codesourcery.com>
14543 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14544 sequence_vect == NULL.
14546 2013-06-11 DJ Delorie <dj@redhat.com>
14548 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14549 (rl78_unwind_word_mode): New.
14551 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14553 * final.c (debug_prefix_maps): Make static.
14555 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14557 * function.c (initial_trampoline): Remove stray copy.
14559 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14561 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14563 2013-06-11 Martin Jambor <mjambor@suse.cz>
14565 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14566 within bounds at the beginning of the function.
14568 2013-06-11 Alan Modra <amodra@gmail.com>
14570 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14572 (get_named_section): Don't NULL !DECL_P decl.
14574 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14576 * doc/invoke.texi (core-avx2): Document.
14578 (atom): Updated with MOVBE.
14580 2013-06-11 Richard Biener <rguenther@suse.de>
14582 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14584 2013-06-11 Anton Blanchard <anton@samba.org>
14586 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14587 correct shift value in little-endian mode.
14589 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14592 * varasm.c (get_variable_align): Move #endif to the right place.
14594 2013-06-10 Cary Coutant <ccoutant@google.com>
14596 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14597 for hash so that hash table traversal order is deterministic.
14599 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14600 Pat Haugen <pthaugen@us.ibm.com>
14601 Peter Bergner <bergner@vnet.ibm.com>
14603 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14604 of vectors in GPRS if they are direct moves or quad word load or
14607 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14609 (direct_move_p): Likewise.
14610 (quad_load_store_p): Likewise.
14612 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14613 classes into bins based on the physical register type.
14614 (reg_class_to_reg_type): Likewise.
14615 (IS_STD_REG_TYPE): Likewise.
14616 (IS_FP_VECT_REG_TYPE): Likewise.
14617 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14618 use direct move instructions.
14619 (reload_gpr_vsx): Likewise.
14620 (reload_vsx_gpr): Likewise.
14621 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14622 information that is a simplification of register classes. Also
14623 precalculate direct move reload helpers.
14624 (direct_move_p): New function to return true if the operation can
14625 be done as a direct move instruciton.
14626 (quad_load_store_p): New function to return true if the operation
14627 is a quad memory operation.
14628 (rs6000_legitimize_address): If quad memory, only allow register
14629 indirect for TImode addresses.
14630 (rs6000_legitimate_address_p): Likewise.
14631 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14632 (rs6000_reload_register_type): Likewise.
14633 (register_to_reg_type): Return register type.
14634 (rs6000_secondary_reload_simple_move): New helper function for
14635 secondary reload and secondary memory needed to identify anything
14636 that is a simple move, and does not need reloading.
14637 (rs6000_secondary_reload_direct_move): New helper function for
14638 secondary reload to identify cases that can be done with several
14639 instructions via the direct move instructions.
14640 (rs6000_secondary_reload_move): New helper function for secondary
14641 reload to identify moves between register types that can be done.
14642 (rs6000_secondary_reload): Add support for quad memory operations
14643 and for direct move.
14644 (rs6000_secondary_memory_needed): Likewise.
14645 (rs6000_debug_secondary_memory_needed): Change argument names.
14646 (rs6000_output_move_128bit): New function to return the move to
14647 use for 128-bit moves, including knowing about the various
14648 limitations of quad memory operations.
14650 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14651 memory operations. call rs6000_output_move_128bit for the actual
14652 instruciton(s) to generate.
14653 (vsx_movti_64bit): Likewise.
14655 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14656 (UNSPEC_P8V_MTVSRWZ): Likewise.
14657 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14658 (UNSPEC_P8V_MTVSRD): Likewise.
14659 (UNSPEC_P8V_XXPERMDI): Likewise.
14660 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14661 (UNSPEC_FUSION_GPR): Likewise.
14662 (FMOVE128_GPR): New iterator for direct move.
14663 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14664 (f32_sv): Likewise.
14665 (f32_dm): Likewise.
14666 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14667 loads and direct move instructions.
14668 (zero_extendsidi2_lfiwzx): Likewise.
14669 (extendsidi2_lfiwax): Likewise.
14670 (extendsidi2_nocell): Likewise.
14671 (floatsi<mode>2_lfiwax): Likewise.
14672 (lfiwax): Likewise.
14673 (floatunssi<mode>2_lfiwzx): Likewise.
14674 (lfiwzx): Likewise.
14675 (fix_trunc<mode>_stfiwx): Likewise.
14676 (fixuns_trunc<mode>_stfiwx): Likewise.
14677 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14678 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14679 (parity<mode>2_cmpb): Set length/type attr.
14680 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14681 for 'mr.' to fast_compare.
14682 (bpermd_<mode>): Change type attr to popcnt.
14683 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14684 (p8_mtvsrwz_1): Likewise.
14685 (p8_mtvsrwz_2): Likewise.
14686 (reload_fpr_from_gpr<mode>): Likewise.
14687 (p8_mtvsrd_1): Likewise.
14688 (p8_mtvsrd_2): Likewise.
14689 (p8_xxpermdi_<mode>): Likewise.
14690 (reload_vsx_from_gpr<mode>): Likewise.
14691 (reload_vsx_from_gprsf): Likewise.
14692 (p8_mfvsrd_3_<mode>): LIkewise.
14693 (reload_gpr_from_vsx<mode>): Likewise.
14694 (reload_gpr_from_vsxsf): Likewise.
14695 (p8_mfvsrd_4_disf): Likewise.
14696 (multi-word GPR splits): Do not split direct moves or quad memory
14699 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14701 * tree-into-ssa.c (interesting_blocks): Make static.
14703 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14706 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14707 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14708 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14709 (get_variable_align): New function.
14710 (get_variable_section, emit_bss, emit_common,
14711 assemble_variable_contents, place_block_symbol): Use
14712 get_variable_align instead of DECL_ALIGN.
14713 (assemble_noswitch_variable): Add align argument, use it
14714 instead of DECL_ALIGN.
14715 (assemble_variable): Adjust caller. Use get_variable_align
14716 instead of DECL_ALIGN.
14717 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14719 (DATA_ABI_ALIGNMENT): Define.
14720 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14721 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14722 opt is false, only return the psABI mandated alignment increase.
14723 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14724 (DATA_ABI_ALIGNMENT): ... this.
14725 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14726 (DATA_ABI_ALIGNMENT): ... this.
14727 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14728 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14729 (DATA_ABI_ALIGNMENT): ... this.
14730 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14731 * doc/tm.texi: Regenerated.
14733 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14735 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14736 cmp_code to construct REG_EQUAL note.
14738 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14741 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14742 that operands[2] doesn't overlap with operands[0].
14744 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14745 Jan Hubicka <jh@suse.cz>
14747 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14748 hack to mark symbols as used.
14750 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14752 PR rtl-optimization/57559
14753 * lra-constraints.c (process_alt_operands): Don't discourage
14754 memory with known offset for offsetable memory constraint.
14755 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14757 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14759 * varasm.c (struct oc_local_state): Reorder fields.
14760 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14761 and adjust accordingly.
14762 (output_constructor): Reorder initialization code and adjust call to
14763 output_constructor_bitfield.
14765 2013-06-07 Jan Hubicka <jh@suse.cz>
14767 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14769 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14771 * tree-object-size.c (unknown): Make const.
14773 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14775 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14776 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14777 for last alternative in the cpu_facility attribute.
14779 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14782 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14783 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14784 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14785 * config/arm/constraints.md (Dg): New constraint.
14786 * config/arm/neon.md (xordi3_neon): Remove.
14787 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14788 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14790 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14792 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14793 Clean up alternatives.
14795 2013-06-07 Alan Modra <amodra@gmail.com>
14797 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14800 2013-06-07 Alan Modra <amodra@gmail.com>
14802 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14804 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14806 * config/arm/constraints.md (Df): New constraint.
14807 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14808 Correct length attribute for last two alternatives.
14810 2013-06-07 Alan Modra <amodra@gmail.com>
14812 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14813 override user -mfp-in-toc.
14814 (offsettable_ok_by_alignment): Consider just the current access
14815 rather than the whole object, unless BLKmode. Handle
14816 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14817 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14818 for -mcmodel=medium.
14819 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14820 override user -mfp-in-toc or -msum-in-toc. Default to
14821 -mno-fp-in-toc for -mcmodel=medium.
14823 2013-06-06 DJ Delorie <dj@redhat.com>
14825 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14826 TARGET_VALID_POINTER_MODE.
14828 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14829 Pat Haugen <pthaugen@us.ibm.com>
14830 Peter Bergner <bergner@vnet.ibm.com>
14832 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14833 Document new power8 builtins.
14835 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14836 condition code register, to allow 128-bit logical operations to be
14837 done in the VSX or GPR registers.
14838 (nor<mode>3): Use the canonical form for nor.
14839 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14840 vclz*, and vpopcnt* vector instructions.
14841 (nand<mode>3): Likewise.
14842 (orc<mode>3): Likewise.
14843 (clz<mode>2): LIkewise.
14844 (popcount<mode>2): Likewise.
14846 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14847 that only the GPRs are recognized.
14849 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14850 support for new power8 builtins.
14852 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14854 (xscvdpspn): Likewise.
14860 (vpopcnt): Likewise.
14861 (vpopcntb): Likewise.
14862 (vpopcnth): Likewise.
14863 (vpopcntw): Likewise.
14864 (vpopcntd): Likewise.
14866 (vmrgew): Likewise.
14867 (vmrgow): Likewise.
14869 (eqv_v16qi3): Likewise.
14870 (eqv_v8hi3): Likewise.
14871 (eqv_v4si3): Likewise.
14872 (eqv_v2di3): Likewise.
14873 (eqv_v4sf3): Likewise.
14874 (eqv_v2df3): Likewise.
14876 (nand_v16qi3): Likewise.
14877 (nand_v8hi3): Likewise.
14878 (nand_v4si3): Likewise.
14879 (nand_v2di3): Likewise.
14880 (nand_v4sf3): Likewise.
14881 (nand_v2df3): Likewise.
14883 (orc_v16qi3): Likewise.
14884 (orc_v8hi3): Likewise.
14885 (orc_v4si3): Likewise.
14886 (orc_v2di3): Likewise.
14887 (orc_v4sf3): Likewise.
14888 (orc_v2df3): Likewise.
14890 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14891 allow power8 quad mode in 64-bit.
14892 (rs6000_builtin_vectorized_function): Add support to vectorize
14893 ISA 2.07 count leading zeros, population count builtins.
14894 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14895 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14896 (builtin_function_type): Add vgbbd builtin function which takes an
14898 (altivec_expand_vec_perm_const): Add support for new power8 merge
14901 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14902 that does not include TImdoe for use with 32-bit.
14903 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14905 (UNSPEC_VSX_CVDPSPN): Likewise.
14906 (vsx_xscvdpspn): Likewise.
14907 (vsx_xscvspdpn): Likewise.
14908 (vsx_xscvdpspn_scalar): Likewise.
14909 (vsx_xscvspdpn_directmove): Likewise.
14910 (vsx_and<mode>3): Split logical operations into 32-bit and
14911 64-bit. Add support to do logical operations on TImode as well as
14912 VSX vector types. Allow logical operations to be done in either
14913 VSX registers or in general purpose registers in 64-bit mode. Add
14914 splitters if GPRs were used. For AND, add clobber of CCmode to
14915 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14917 (vsx_and<mode>3_32bit): Likewise.
14918 (vsx_and<mode>3_64bit): Likewise.
14919 (vsx_ior<mode>3): Likewise.
14920 (vsx_ior<mode>3_32bit): Likewise.
14921 (vsx_ior<mode>3_64bit): Likewise.
14922 (vsx_xor<mode>3): Likewise.
14923 (vsx_xor<mode>3_32bit): Likewise.
14924 (vsx_xor<mode>3_64bit): Likewise.
14925 (vsx_one_cmpl<mode>2): Likewise.
14926 (vsx_one_cmpl<mode>2_32bit): Likewise.
14927 (vsx_one_cmpl<mode>2_64bit): Likewise.
14928 (vsx_nor<mode>3): Likewise.
14929 (vsx_nor<mode>3_32bit): Likewise.
14930 (vsx_nor<mode>3_64bit): Likewise.
14931 (vsx_andc<mode>3): Likewise.
14932 (vsx_andc<mode>3_32bit): Likewise.
14933 (vsx_andc<mode>3_64bit): Likewise.
14934 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14935 and xxlorc instructions.
14936 (vsx_eqv<mode>3_64bit): Likewise.
14937 (vsx_nand<mode>3_32bit): Likewise.
14938 (vsx_nand<mode>3_64bit): Likewise.
14939 (vsx_orc<mode>3_32bit): Likewise.
14940 (vsx_orc<mode>3_64bit): Likewise.
14942 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14944 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14946 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14947 (p8_vmrgow): Likewise.
14948 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14949 GPRs to be split under VSX.
14950 (p8v_clz<mode>2): Add power8 count leading zero support.
14951 (p8v_popcount<mode>2): Add power8 population count support.
14952 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14955 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14958 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14960 (vec_nand): Likewise.
14961 (vec_vclz): Likewise.
14962 (vec_vclzb): Likewise.
14963 (vec_vclzd): Likewise.
14964 (vec_vclzh): Likewise.
14965 (vec_vclzw): Likewise.
14966 (vec_vgbbd): Likewise.
14967 (vec_vmrgew): Likewise.
14968 (vec_vmrgow): Likewise.
14969 (vec_vpopcnt): Likewise.
14970 (vec_vpopcntb): Likewise.
14971 (vec_vpopcntd): Likewise.
14972 (vec_vpopcnth): Likewise.
14973 (vec_vpopcntw): Likewise.
14975 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14977 PR rtl-optimization/57468
14978 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14981 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14983 PR rtl-optimization/57459
14984 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14985 type when setting live regs.
14987 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14989 * config/s390/s390.opt (mlra): New option.
14990 * config/s390/s390.c (s390_decompose_address): Check displacement
14991 for all registers for LRA.
14992 (s390_secondary_reload): Don't used secondary reloads for LRA.
14993 (s390_lra_p): New function.
14994 (TARGET_LRA_P): Define.
14995 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14996 of attribute cpu_facility to zarch for the last alternative.
14997 (*cmpmem_short): Ditto.
14999 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15001 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15002 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15003 (arm_expand_prologue): Likewise.
15005 2013-06-06 Teresa Johnson <tejohnson@google.com>
15008 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15009 as this is now done by redirect_edge_and_branch_force.
15010 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15011 barriers, and fix interaction with splitting.
15012 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15013 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15014 reflect changes made in the routine.
15015 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15016 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15017 since this is called in cfglayout mode, and replace partition fixup
15018 with assert as that is now done by force_nonfallthru_and_redirect.
15019 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15020 already be marked with region crossing note.
15021 (insert_section_boundary_note): Make non-static, gate on flag
15022 has_bb_partition, rewrite to also check for multiple partitions.
15023 (rest_of_handle_reorder_blocks): Remove call to
15024 insert_section_boundary_note, now done later during free_cfg.
15025 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15026 * bb-reorder.h (insert_section_boundary_note): Declare.
15027 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15028 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15029 invoke insert_section_boundary_note.
15030 (try_redirect_by_replacing_jump): Remove unnecessary
15031 check for region crossing note.
15032 (fixup_partition_crossing): New function.
15033 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15034 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15035 in non-cfglayout mode.
15036 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15037 remove old code that tried to do this. Emit barrier correctly
15038 when we are in cfglayout mode.
15039 (last_bb_in_partition): New function.
15040 (rtl_split_edge): Correctly fixup partition boundaries.
15041 (commit_one_edge_insertion): Remove old code that tried to
15042 fixup region crossing edge since this is now handled in
15043 split_block, and set up insertion point correctly since
15044 block may now end in a jump.
15045 (verify_hot_cold_block_grouping): Guard against checking when not in
15046 linearized RTL mode.
15047 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15049 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15050 rtl_verify_flow_info, so not called in cfglayout mode.
15051 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15052 (fixup_reorder_chain): Remove old code that attempted to fixup region
15053 crossing note as this is now handled in force_nonfallthru_and_redirect.
15054 (duplicate_insn_chain): Don't duplicate switch section notes.
15055 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15057 * basic-block.h (emit_barrier_after_bb): Declare.
15059 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15061 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15062 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15063 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15065 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15067 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15069 * config/arm/ldmstm.md: Regenerate.
15071 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15073 * config/arm/sync.md (atomic_loaddi_1):
15074 Disable predication for arm_restrict_it.
15075 (arm_load_exclusive<mode>): Likewise.
15076 (arm_load_exclusivesi): Likewise.
15077 (arm_load_exclusivedi): Likewise.
15078 (arm_load_acquire_exclusive<mode>): Likewise.
15079 (arm_load_acquire_exclusivesi): Likewise.
15080 (arm_load_acquire_exclusivedi): Likewise.
15081 (arm_store_exclusive<mode>): Likewise.
15082 (arm_store_exclusive<mode>): Likewise.
15083 (arm_store_release_exclusivedi): Likewise.
15084 (arm_store_release_exclusive<mode>): Likewise.
15086 2013-06-06 Richard Biener <rguenther@suse.de>
15088 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15090 (lto_tag_is_tree_code_p): Adjust.
15091 (lto_tag_is_gimple_code_p): Likewise.
15092 (lto_gimple_code_to_tag): Likewise.
15093 (lto_tag_to_gimple_code): Likewise.
15094 (lto_tree_code_to_tag): Likewise.
15095 (lto_tag_to_tree_code): Likewise.
15096 * data-streamer.h (streamer_write_hwi_in_range): Use
15097 uhwi streaming to stream the normalized range.
15098 (streamer_read_hwi_in_range): Likewise.
15100 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15102 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15103 (predicable_short_it): Likewise.
15104 (predicated): Likewise.
15105 (enabled): Handle above.
15106 (define_cond_exec): Set predicated attribute to yes.
15108 2013-06-05 Mike Stump <mikestump@comcast.net>
15110 * gdbinit.in (__FUNCTION__): Add.
15112 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15114 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15115 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15117 2013-06-05 Jan Hubicka <jh@suse.cz>
15119 * varasm.c (mark_decl_referenced): Revert the removal until targets
15122 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15124 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15125 instead of mark_decl_referenced.
15127 2013-06-05 Jan Hubicka <jh@suse.cz>
15129 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15130 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15131 and symtab_used_from_object_file_p.
15132 (cgraph_make_node_local_1): Clear forced_by_abi.
15133 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15134 * cgraph.h (symtab_node_base): Add forced_by_abi.
15135 (decide_is_variable_needed): Remove.
15136 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15137 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15138 (decide_is_symbol_needed): ... this one; handle symbols in general;
15139 always analyze virtuals; honnor forced_by_abi.
15140 (cgraph_finalize_function): Update.
15141 (varpool_finalize_decl): Update.
15142 (symbol_defined_and_needed): Remove.
15143 (analyze_functions): Update.
15144 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15145 output_refs, input_overwrite_node): Handle forced_by_abi.
15146 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15147 (address_taken_from_non_vtable_p): ... this one.
15148 (comdat_can_be_unshared_p_1): New function.
15149 (cgraph_comdat_can_be_unshared_p): Rename to ...
15150 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15151 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15152 (function_and_variable_visibility): Clear forced_by_abi as needed.
15153 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15154 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15155 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15156 * varpool.c (decide_is_variable_needed): Remove.
15158 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15160 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15161 (arm_option_override): Override arm_restrict_it where appropriate.
15162 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15163 * config/arm/arm.opt (mrestrict-it): New command-line option.
15164 * doc/invoke.texi: Document -mrestrict-it.
15166 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15168 * tsan.c (tsan_atomic_table): Make const.
15170 2013-06-05 Richard Biener <rguenther@suse.de>
15172 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15173 index associated with the tree we are supposed to replace.
15174 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15175 TREE_ASM_WRITTEN as zero for everything but SSA names.
15177 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15179 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15181 2013-06-04 Jan Hubicka <jh@suse.cz>
15183 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15184 (input_node, input_varpool_node): Handle correctly external same
15186 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15187 nodes at ltrans stage.
15189 2013-06-04 Jan Hubicka <jh@suse.cz>
15191 * ipa-inline.c (update_caller_keys): Fix availability test.
15192 (update_callee_keys): Likewise.
15193 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15194 to follow ELF standard.
15196 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15198 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15199 (mips64r5900el-*-elf*): New configurations.
15200 * config/mips/mips-cpus.def (r5900): New processor.
15201 * config/mips/mips-tables.opt: Regenerate.
15202 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15203 (mips_issue_rate): Handle PROCESSOR_R5900.
15204 (mips_reorg_process_insns): Force reorder mode for the R5900.
15205 * config/mips/mips.h (TARGET_MIPS5900): Define.
15206 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15208 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15210 * config/mips/mips.md (processor): Add r5900.
15211 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15213 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15215 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15216 into function to generate MOVI instruction.
15217 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15218 (aarch64_preferred_simd_mode): Turn into wrapper.
15219 (aarch64_output_scalar_simd_mov_immediate): New function.
15220 * config/aarch64/aarch64-protos.h: Add prototype for above.
15222 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15224 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15225 element_char member.
15226 (sizetochar): Return signed char.
15227 (aarch64_simd_valid_immediate): Remove elchar and other
15228 unnecessary variables.
15229 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15230 Calculate element_char as required.
15231 * config/aarch64/aarch64-protos.h: Update and move prototype
15232 for aarch64_output_simd_mov_immediate.
15233 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15236 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15238 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15239 information completed by aarch64_simd_valid_immediate.
15240 (aarch64_legitimate_constant_p): Update arguments.
15241 (aarch64_simd_valid_immediate): Work with struct rather than many
15243 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15244 (aarch64_simd_make_constant): Update arguments.
15245 (aarch64_output_simd_mov_immediate): Work with struct rather than
15246 many pointers. Output immediate directly rather than as operand.
15247 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15249 * config/aarch64/constraints.md (Dn): Update arguments.
15251 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15253 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15255 (aarch64_simd_immediate_valid_for_move): Remove.
15256 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15257 (aarch64_simd_make_constant): Update call.
15258 (aarch64_output_simd_mov_immediate): Update call.
15259 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15261 * config/aarch64/constraints.md (Dn): Update call.
15263 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15265 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15266 return type to bool for prototype.
15267 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15268 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15269 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15271 2013-06-04 Catherine Moore <clm@codesourcery.com>
15273 * config/mips/mips.opt (meva): New.
15274 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15275 (ASM_SPEC): Handle -meva.
15276 * doc/invoke.texi (meva): Document.
15278 2013-06-04 Alan Modra <amodra@gmail.com>
15280 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15283 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15285 * rtl.def: Add extra fourth optional field to define_cond_exec.
15286 * gensupport.c (process_one_cond_exec): Process attributes from
15288 * doc/md.texi: Document fourth field in define_cond_exec.
15290 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15292 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15293 out the processing order as in store_bit_field_1.
15295 2013-06-04 Jan Hubicka <jh@suse.cz>
15297 PR middle-end/57500
15298 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15299 non-VAR_DECL node if it does not exist yet.
15301 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15303 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15304 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15305 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15306 target_cpu_default setting.
15308 2013-06-03 Teresa Johnson <tejohnson@google.com>
15310 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15311 default to -fopt-info=optimized instead of all.
15312 * doc/invoke.texi: Ditto.
15313 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15314 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15315 (execute_vect_slp): Emit BB vectorization success under
15316 MSG_OPTIMIZED_LOCATIONS.
15317 * tree-vect-slp.c (vect_slp_transform_bb): Change
15318 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15319 * tree-vect-loop.c (vect_transform_loop): Ditto.
15321 2013-06-03 Jason Merrill <jason@redhat.com>
15324 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15325 Use TARGET_EXPR for C++.
15327 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15329 PR rtl-optimization/57268
15330 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15331 if DEBUG_INSN_P (insn).
15334 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15336 PR rtl-optimization/57268
15337 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15338 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15340 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15342 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15343 (ix86_avoid_lea_for_addr): Likewise.
15344 (exact_dependency_1): Likewise.
15345 (ix86_adjust_cost): Likewise.
15346 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15348 (do_reorder_for_imul): Fix typo, formatting and
15349 !reload_completed check removed.
15350 (ix86_sched_reorder): Fix typo and formatting.
15351 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15354 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15356 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15358 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15360 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15361 <STRING_CST>: Likewise.
15362 <VECTOR_CST>: Likewise.
15364 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15365 Mikael Morin <mikael@gcc.gnu.org>
15367 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15368 * config.in: Regenerated.
15369 * configure: Regenerated.
15371 2013-06-01 Jan Hubicka <jh@suse.cz>
15373 PR middle-end/57366
15374 * cgraphunit.c (compile): When weakref is not supported,
15375 set up transparent aliases before final output pass.
15376 * varasm.c (assemble_alias): Do not try to do it here.
15378 2013-06-01 Jan Hubicka <jh@suse.cz>
15380 PR middle-end/57467
15381 * passes.c (for_per_function): Skip unanalyzed functions.
15383 2013-06-01 Jan Hubicka <jh@suse.cz>
15385 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15386 (lto_symtab_merge_symbols_1): ... this one.
15387 (lto_symtab_merge_cgraph_nodes): Rename to ...
15388 (lto_symtab_merge_symbols): ... this one; simplify.
15389 * cgraph.c (same_body_aliases_done): Rename to ...
15390 (cpp_implicit_aliases_done): ... this one.
15391 (cgraph_create_function_alias): Update.
15392 (cgraph_same_body_alias): Update.
15393 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15394 (verify_edge_corresponds_to_fndecl): Simplify.
15395 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15396 (cgraph_node): Remove same_body_alias.
15397 (varpool_node): Remove alias_of and extra_name_alias.
15398 (same_body_aliases_done): Rename to ..
15399 (cpp_implicit_aliases_done): ... this one.
15400 (symtab_alias_ultimate_target): Add default parameter.
15401 (symtab_resolve_alias): New function.
15402 (fixup_same_cpp_alias_visibility): Declare.
15403 (cgraph_function_node): Add default parameter.
15404 (cgraph_node_asm_name): Likewise.
15405 (cgraph_function_or_thunk_node): Add default parameter; do
15406 not ICE when it is NULL.
15407 (varpool_variable_node): Likewise.
15408 * tree-emutls.c (create_emultls_var): Update.
15409 (ipa_lower_emutls): Update.
15410 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15411 (cgraph_reset_node): Reset alias info.
15412 (cgraph_finalize_function): Update.
15413 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15414 (analyze_function): Simplify.
15415 (cgraph_process_same_body_aliases): Simplify.
15416 (analyze_functions): Fixup same body aliases.
15417 (handle_alias_pairs): Simplify.
15418 (assemble_thunk): Update.
15419 (assemble_thunks_and_aliases): Update.
15420 (output_weakrefs): Rewrite.
15421 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15422 (lto_output_varpool_node): Likewise.
15423 (compute_ltrans_boundary): Remve assert.
15424 (get_alias_symbol): New functoin.
15425 (input_node): Rewrite alias handling.
15426 (input_varpool_node): Likewise.
15427 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15428 * ipa.c (process_references): Handle weakrefs correctly.
15429 (symtab_remove_unreachable_nodes): Likewise.
15430 * trans-mem.c (get_cg_data): Update.
15431 (ipa_tm_create_version_alias): Update.
15432 (ipa_tm_execute): Update.
15433 * symtab.c (dump_symtab_base): Dump aliases.
15434 (verify_symtab_base): Verify aliases.
15435 (symtab_node_availability): New function.
15436 (symtab_alias_ultimate_target): Simplify.
15437 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15438 handle all the fixup cases.
15439 (symtab_resolve_alias): New function.
15440 * passes.c (ipa_write_summaries): Handle weakrefs.
15441 * varpool.c (varpool_analyze_node): Simplify.
15442 (assemble_aliases): Update.
15443 (varpool_create_variable_alias): Simplify.
15444 (varpool_extra_name_alias): Simplify.
15445 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15446 (lto_symtab_merge_symbols): ... this one.
15448 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15451 PR rtl-optimization/57268
15452 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15453 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15455 2013-06-01 Tobias Burnus <burnus@net-b.de>
15457 Partially reverted:
15458 2013-05-31 Tobias Burnus <burnus@net-b.de>
15460 PR middle-end/57073
15461 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15464 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15466 PR rtl-optimization/57268
15467 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15468 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15470 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15472 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15473 unordered comparison operators when -fno-trapping-math is in effect
15475 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15476 and implement unordered comparison operators properly on the e500.
15478 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15480 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15481 for constant scalar integers.
15482 (simplify_relational_operation_1): Likewise.
15484 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15486 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15487 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15490 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15491 Igor Zamyatin <igor.zamyatin@intel.com>
15493 Silvermont (SLM) architecture performance tuning.
15494 * config/i386/i386.h (enum ix86_tune_indices): Add
15495 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15496 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15498 * config/i386/i386.c (initial_ix86_tune_features)
15499 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15500 (ix86_lea_outperforms): Handle Silvermont tuning.
15501 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15503 (ix86_use_lea_for_mov): Likewise.
15504 (ix86_avoid_lea_for_addr): Likewise.
15505 (ix86_lea_for_add_ok): Likewise.
15506 (exact_dependency_1): New function.
15507 (exact_store_load_dependency): Likewise.
15508 (ix86_adjust_cost): Handle Silvermont tuning.
15509 (do_reoder_for_imul): Likewise.
15510 (swap_top_of_ready_list): New function.
15511 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15513 * config/i386/i386.md (peepholes that split memory operand in fp
15516 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15518 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15519 Remove un-necessary braces.
15521 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15523 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15524 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15526 2013-05-31 Tobias Burnus <burnus@net-b.de>
15528 PR middle-end/57073
15529 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15531 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15534 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15535 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15536 * config/arm/neon.md (iordi3_neon): Remove.
15537 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15538 * config/arm/predicates.md (imm_for_neon_logic_operand):
15539 Move to earlier in the file.
15540 (neon_logic_op2): Likewise.
15541 (arm_iordi_operand_neon): New predicate.
15543 2013-05-31 Richard Biener <rguenther@suse.de>
15545 PR tree-optimization/57478
15546 PR tree-optimization/57453
15547 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15550 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15552 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15553 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15555 2013-05-30 Tobias Burnus <burnus@net-b.de>
15556 Thomas Koenig <tkoenig@gcc.gnu.org>
15558 PR middle-end/57073
15559 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15560 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15562 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15564 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15566 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15568 * target.def (register_usage_leveling_p): New hook.
15569 * targhooks.c (default_register_usage_leveling_p): New.
15570 * targhooks.h (default_register_usage_leveling_p): New prototype.
15571 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15572 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15573 * doc/tm.texi: Update.
15574 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15576 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15578 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15579 (*insv_reg<mode>): New define_insn.
15581 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15583 PR rtl-optimization/57439
15584 * postreload.c (move2add_valid_value_p): Check that we have
15585 a zero subreg_regno_offset when accessing the register in
15586 the requested mode.
15588 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15589 Igor Zamyatin <igor.zamyatin@intel.com>
15591 Silvermont (SLM) architecture pipeline model, tuning and
15593 * config.gcc: Add slm config options and target.
15595 * config/i386/slm.md: New.
15597 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15599 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15601 (ix86_target_macros_internal): Likewise.
15603 * config/i386/i386.c (slm_cost): New cost.
15604 (m_SLM): New macro flag.
15605 (initial_ix86_tune_features): Set m_SLM.
15606 (x86_accumulate_outgoing_args): Likewise.
15607 (x86_arch_always_fancy_math_387): Likewise.
15608 (processor_target_table): Add slm cost.
15609 (cpu_names): Add slm cpu name.
15610 (x86_option_override_internal): Set SLM ISA.
15611 (ix86_issue_rate): New case PROCESSOR_SLM.
15612 (ia32_multipass_dfa_lookahead): Likewise.
15613 (fold_builtin_cpu): Add slm.
15615 * config/i386/i386.h (TARGET_SLM): New target macro.
15616 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15617 (processor_type): Add PROCESSOR_SLM.
15619 * config/i386/i386.md (cpu): Add new value "slm".
15620 (slm.md): Include slm.md.
15622 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15623 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15625 * config/arm/arm-protos.h: Add and update function protos.
15626 * config/arm/arm.c (use_simple_return_p): New added.
15627 (thumb2_expand_return): Check simple_return flag.
15628 * config/arm/arm.md: Add simple_return and conditional simple_return.
15629 * config/arm/iterators.md: Add iterator for return and simple_return.
15631 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15633 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15634 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15635 (arm_emit_vfp_multi_reg_pop): Likewise.
15636 (thumb2_emit_ldrd_pop): Likewise.
15637 (arm_expand_epilogue): Add misc REG_CFA notes.
15638 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15640 2013-05-29 Lawrence Crowl <crowl@google.com>
15642 * config/arm/t-arm: Update for below.
15644 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15645 Change type to hash_table. Update dependent calls and types.
15647 * config/i386/t-cygming: Update for below.
15649 * config/i386/t-interix: Update for below.
15651 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15652 Change type to hash_table. Update dependent calls and types.
15653 (i386_find_on_wrapper_list::wrappers): Likewise.
15655 * config/ia64/t-ia64: Update for below.
15657 * config/ia64/ia64.c (bundle_state_table):
15658 Change type to hash_table. Update dependent calls and types.
15660 * config/mips/mips.c (mips_reorg_process_insns::htab):
15661 Change type to hash_table. Update dependent calls and types.
15663 * config/sol2.c (solaris_comdat_htab):
15664 Change type to hash_table. Update dependent calls and types.
15666 * config/t-sol2: Update for above.
15668 2013-05-29 Teresa Johnson <tejohnson@google.com>
15670 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15671 functions are not yet marked as defined.
15673 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15674 Pat Haugen <pthaugen@us.ibm.com>
15675 Peter Bergner <bergner@vnet.ibm.com>
15677 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15681 (vrotl<mode>3): Likewise.
15682 (vashl<mode>3): Likewise.
15683 (vlshr<mode>3): Likewise.
15684 (vashr<mode>3): Likewise.
15686 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15687 support for power8 V2DI builtins.
15689 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15690 power8 V2DI builtins.
15691 (vupkhsw): Likewise.
15692 (vupklsw): Likewise.
15693 (vaddudm): Likewise.
15694 (vminsd): Likewise.
15695 (vmaxsd): Likewise.
15696 (vminud): Likewise.
15697 (vmaxud): Likewise.
15698 (vpkudum): Likewise.
15699 (vpksdss): Likewise.
15700 (vpkudus): Likewise.
15701 (vpksdus): Likewise.
15706 (vsubudm): Likewise.
15707 (vcmpequd): Likewise.
15708 (vcmpgtsd): Likewise.
15709 (vcmpgtud): Likewise.
15710 (vcmpequd_p): Likewise.
15711 (vcmpgtsd_p): Likewise.
15712 (vcmpgtud_p): Likewise.
15713 (vupkhsw): Likewise.
15714 (vupklsw): Likewise.
15715 (vaddudm): Likewise.
15716 (vmaxsd): Likewise.
15717 (vmaxud): Likewise.
15718 (vminsd): Likewise.
15719 (vminud): Likewise.
15720 (vpksdss): Likewise.
15721 (vpksdus): Likewise.
15722 (vpkudum): Likewise.
15723 (vpkudus): Likewise.
15728 (vsubudm): Likewise.
15730 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15731 support for power8 V2DI instructions.
15733 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15734 power8 V2DI instructions. Combine pack and unpack insns to use an
15735 iterator for each mode. Check whether a particular mode supports
15736 Altivec instructions instead of just checking TARGET_ALTIVEC.
15737 (UNSPEC_VPKUWUM): Likewise.
15738 (UNSPEC_VPKSHSS): Likewise.
15739 (UNSPEC_VPKSWSS): Likewise.
15740 (UNSPEC_VPKUHUS): Likewise.
15741 (UNSPEC_VPKSHUS): Likewise.
15742 (UNSPEC_VPKUWUS): Likewise.
15743 (UNSPEC_VPKSWUS): Likewise.
15744 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15745 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15746 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15747 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15748 (UNSPEC_VUPKHSB): Likewise.
15749 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15750 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15751 (UNSPEC_VUPKHSH): Likewise.
15752 (UNSPEC_VUPKLSB): Likewise.
15753 (UNSPEC_VUPKLSH): Likewise.
15755 (VI_char): Likewise.
15756 (VI_scalar): Likewise.
15757 (VI_unit): Likewise.
15759 (VP_small): Likewise.
15760 (VP_small_lc): Likewise.
15761 (VU_char): Likewise.
15762 (add<mode>3): Likewise.
15763 (altivec_vaddcuw): Likewise.
15764 (altivec_vaddu<VI_char>s): Likewise.
15765 (altivec_vadds<VI_char>s): Likewise.
15766 (sub<mode>3): Likewise.
15767 (altivec_vsubcuw): Likewise.
15768 (altivec_vsubu<VI_char>s): Likewise.
15769 (altivec_vsubs<VI_char>s): Likewise.
15770 (altivec_vavgs<VI_char>): Likewise.
15771 (altivec_vcmpbfp): Likewise.
15772 (altivec_eq<mode>): Likewise.
15773 (altivec_gt<mode>): Likewise.
15774 (altivec_gtu<mode>): Likewise.
15775 (umax<mode>3): Likewise.
15776 (smax<mode>3): Likewise.
15777 (umin<mode>3): Likewise.
15778 (smin<mode>3): Likewise.
15779 (altivec_vpkuhum): Likewise.
15780 (altivec_vpkuwum): Likewise.
15781 (altivec_vpkshss): Likewise.
15782 (altivec_vpkswss): Likewise.
15783 (altivec_vpkuhus): Likewise.
15784 (altivec_vpkshus): Likewise.
15785 (altivec_vpkuwus): Likewise.
15786 (altivec_vpkswus): Likewise.
15787 (altivec_vpks<VI_char>ss): Likewise.
15788 (altivec_vpks<VI_char>us): Likewise.
15789 (altivec_vpku<VI_char>us): Likewise.
15790 (altivec_vpku<VI_char>um): Likewise.
15791 (altivec_vrl<VI_char>): Likewise.
15792 (altivec_vsl<VI_char>): Likewise.
15793 (altivec_vsr<VI_char>): Likewise.
15794 (altivec_vsra<VI_char>): Likewise.
15795 (altivec_vsldoi_<mode>): Likewise.
15796 (altivec_vupkhsb): Likewise.
15797 (altivec_vupkhs<VU_char>): Likewise.
15798 (altivec_vupkls<VU_char>): Likewise.
15799 (altivec_vupkhsh): Likewise.
15800 (altivec_vupklsb): Likewise.
15801 (altivec_vupklsh): Likewise.
15802 (altivec_vcmpequ<VI_char>_p): Likewise.
15803 (altivec_vcmpgts<VI_char>_p): Likewise.
15804 (altivec_vcmpgtu<VI_char>_p): Likewise.
15805 (abs<mode>2): Likewise.
15806 (vec_unpacks_hi_v16qi): Likewise.
15807 (vec_unpacks_hi_v8hi): Likewise.
15808 (vec_unpacks_lo_v16qi): Likewise.
15809 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15810 (vec_unpacks_lo_v8hi): Likewise.
15811 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15812 (vec_pack_trunc_v8h): Likewise.
15813 (vec_pack_trunc_v4si): Likewise.
15814 (vec_pack_trunc_<mode>): Likewise.
15816 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15818 (vec_vmaxsd): Likewise.
15819 (vec_vmaxud): Likewise.
15820 (vec_vminsd): Likewise.
15821 (vec_vminud): Likewise.
15822 (vec_vpksdss): Likewise.
15823 (vec_vpksdus): Likewise.
15824 (vec_vpkudum): Likewise.
15825 (vec_vpkudus): Likewise.
15826 (vec_vrld): Likewise.
15827 (vec_vsld): Likewise.
15828 (vec_vsrad): Likewise.
15829 (vec_vsrd): Likewise.
15830 (vec_vsubudm): Likewise.
15831 (vec_vupkhsw): Likewise.
15832 (vec_vupklsw): Likewise.
15834 2013-05-29 Jan Hubicka <jh@suse.cz>
15836 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15837 flags; reorder rest of fields in more consistent way.
15838 (varpool_node): Remove analyzed, finalized and alias.
15839 (cgraph_ndoe): Likewise.
15840 (symtab_alias_ultimate_target): New function.
15841 (cgraph_function_node): Move offline.
15842 (cgraph_reset_node): Declare.
15843 (cgraph_comdat_can_be_unshared_p): Remove.
15844 (varpool_remove_initializer): Declare.
15845 (varpool_first_defined_variable, varpool_next_defined_variable
15846 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15847 (cgraph_function_with_gimple_body_p): Update.
15848 (varpool_all_refs_explicit_p): Update.
15849 (symtab_alias_target): New function.
15850 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15851 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15852 (cgraph_function_or_thunk_node): Simplify using
15853 symtab_alias_ultimate_target.
15854 (varpool_variable_node): Likewise.
15855 * cgraph.c (cgraph_create_function_alias): Update.
15856 (cgraph_add_thunk): Update.
15857 (cgraph_remove_node): Update.
15858 (dump_cgraph_node): Do not dump removed flags.
15859 (cgraph_function_body_availability): Update.
15860 (cgraph_propagate_frequency): Update.
15861 (verify_cgraph_node): Check sanity of local flag.
15862 (cgraph_function_node): Move here from cgraph.h; revamp for
15863 cgraph_function_or_thunk_node.
15864 * lto-symtab.c (lto_varpool_replace_node): Update.
15865 (lto_symtab_resolve_can_prevail_p): Update.
15866 (lto_symtab_merge_cgraph_nodes): Update.
15867 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15868 propagate_constants_accross_call, devirtualization_time_bonus,
15869 ipcp_propagate_stage): Update.
15870 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15871 * ipa-inline-transform.c (clone_inlined_nodes,
15872 preserve_function_body_p): Update.
15873 * ipa-reference.c (propagate): Update.
15874 (write_node_summary_p): Update.
15875 * toplev.c (wrapup_global_declaration_2): Update.
15876 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15877 (analyze_function) ... this one.
15878 (cgraph_process_new_functions): Update.
15879 (cgraph_reset_node): Export.
15880 (cgraph_finalize_function): Update.
15881 (cgraph_add_new_function): Update.
15882 (process_function_and_variable_attributes): Update.
15883 (varpool_finalize_decl): Update.
15884 (symbol_finalized): Remove.
15885 (symbol_finalized_and_needed): Rename to ...
15886 (symbol_defined_and_needed): ... update.
15887 (cgraph_analyze_functions): Update.
15888 (handle_alias_pairs): Update.
15889 (mark_functions_to_output): Update.
15890 (assemble_thunk): Update.
15891 (output_in_order): Update.
15892 (output_weakrefs): Update.
15893 (finalize_compilation_unit): Update.
15894 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15895 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15896 input_node, input_varpool_node): Update.
15897 * dbxout.c (dbxout_expand_expr): Update.
15898 * cgraphclones.c (cgraph_clone_node): Update.
15899 (cgraph_copy_node_for_versioning): Update.
15900 (cgraph_materialize_clone): Update.
15901 (cgraph_materialize_all_clones): Update.
15902 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15903 propagate_pure_const, propagate_nothrow): Update.
15904 * lto-streamer-out.c (lto_output, write_symbol): Update.
15905 * ipa-utils.c (ipa_reverse_postorder): Update.
15906 * ipa-inline.c (can_inline_edge_p): Update.
15907 (update_caller_keys, ipa_inline): Update.
15908 * dwarf2out.c (reference_to_unused,
15909 premark_types_used_by_global_vars_helper): Update.
15910 * tree-eh.c (tree_could_trap_p): Update.
15911 * ipa-split.c (consider_split, execute_split_functions): Update.
15912 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15913 has_addr_references_p): Update; move ahead in file for better
15915 (process_references): Simplify.
15916 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15917 bodies are removed.
15918 (cgraph_comdat_can_be_unshared_p): Make static.
15919 (cgraph_externally_visible_p): Update.
15920 (varpool_externally_visible_p): Update.
15921 (function_and_variable_visibility): Update.
15922 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15923 ipa_tm_mark_force_output_node): Update.
15924 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15925 estimate_edge_devirt_benefit, inline_generate_summary,
15926 inline_write_summary): Update.
15927 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15928 * ipa-prop.c (ipa_compute_jump_functions): Update.
15929 (ipa_print_node_params, ipa_prop_read_section,
15930 ipa_update_after_lto_read, read_replacements_section): Update.
15931 * varasm.c (mark_decl_referenced): Update.
15932 (assemble_alias, dump_tm_clone_pairs): Update.
15933 * tree-inline.c (copy_bb): Update.
15934 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15936 * symtab.c (dump_symtab_base): Print new flags.
15937 (verify_symtab_base): Verify new flags.
15938 (symtab_alias_ultimate_target): New function.
15939 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15940 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15942 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15944 * i386.c (ix86_get_function_versions_dispatcher,
15945 ix86_generate_version_dispatcher_body): Update.
15946 (fold_builtin_cpu): Use varpool_add_new_variable.
15947 * varpool.c (varpool_remove_initializer): Break out from ...
15948 (varpool_remove_node): ... this one.
15949 (dump_varpool_node, varpool_node_for_asm,
15950 cgraph_variable_initializer_availability, varpool_analyze_node,
15951 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15952 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15955 2013-05-29 Jan Hubicka <jh@suse.cz>
15957 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15959 2013-05-29 Easwaran Raman <eraman@google.com>
15961 PR tree-optimization/57442
15962 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15963 when control exits the main loop.
15965 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15967 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15969 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15970 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15971 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15974 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15976 PR tree-optimization/57441
15977 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15978 Don't limit size of incr_vec to number of candidates.
15980 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15982 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15983 and mips16 directories.
15984 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15985 (MULTILIB_DIRNAMES): Ditto.
15986 (MULTILIB_EXCEPTIONS): Add new exceptions.
15987 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15988 (MULTILIB_DIRNAMES): Ditto.
15989 (MULTILIB_EXCEPTIONS): Add new exceptions.
15991 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15992 Marcus Shawcroft <marcus.shawcroft@arm.com>
15994 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15995 SYMBOL_TINY_ABSOLUTE.
15996 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15997 SYMBOL_TINY_ABSOLUTE.
15998 (aarch64_expand_mov_immediate): Likewise.
15999 (aarch64_classify_symbol): Likewise.
16000 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16001 Permit SYMBOL_TINY_ABSOLUTE.
16002 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16004 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16005 Marcus Shawcroft <marcus.shawcroft@arm.com>
16007 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16008 Refactor if/switch. Replace gcc_assert with if.
16010 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16012 * config/i386/i386.c (initial_ix86_tune_features): Enable
16013 FP Reassociation for AMD bdver1 and bdver2.
16015 2013-05-29 Martin Jambor <mjambor@suse.cz>
16017 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16018 and IMAGPART_EXPR do not occur within other handled_components.
16020 2013-05-29 Richard Biener <rguenther@suse.de>
16022 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16023 access on whether the use is in the BB we currently try to
16025 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16026 vectorize to vect_bb_slp_scalar_cost.
16028 2013-05-29 Richard Biener <rguenther@suse.de>
16030 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16031 computing scalar cost offsetted by stmts that are kept live
16033 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16034 for computation of scalar cost.
16036 2013-05-28 Steve Ellcey <sellcey@mips.com>
16038 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16040 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16042 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16043 array notation built-in reduction functions.
16044 * doc/passes.texi (Passes): Added documentation about changes done
16046 * doc/invoke.texi (C Dialect Options): Added documentation about
16047 the -fcilkplus flag.
16048 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16049 (BUILTINS_DEF): Depend on cilkplus.def.
16050 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16051 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16052 * cilkplus.def: New file.
16054 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16056 PR rtl-optimization/57439
16057 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16059 2013-05-28 Easwaran Raman <eraman@google.com>
16061 PR tree-optimization/57337
16062 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16063 (find_insert_point): Correctly identify the insertion point
16064 when two statements with the same UID is compared.
16066 2013-05-28 Richard Biener <rguenther@suse.de>
16068 PR tree-optimization/56787
16069 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16070 from the list of data references.
16071 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16073 (vect_analyze_loop_operations): Likewise.
16074 (vect_transform_loop): Remove clobbers.
16076 2013-05-28 Martin Jambor <mjambor@suse.cz>
16078 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16079 and REALPART_EXPRs have scalar type.
16081 2013-05-28 Richard Biener <rguenther@suse.de>
16083 PR tree-optimization/57411
16084 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16086 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16087 virtual operand propagation.
16089 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16091 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16092 destination register for bmasksi_vis.
16093 (vector_init_bshuffle): Likewise.
16094 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16096 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16098 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16099 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16100 mode if the instruction isn't available in the original mode.
16101 * config/sparc/sparc.opt (mfix-ut699): New option.
16102 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16103 (divdf3): Turn into expander.
16104 (divdf3_nofix): New insn.
16105 (divdf3_fix): Likewise.
16106 (divsf3): Disable if -mfix-ut699.
16107 (sqrtdf2): Turn into expander.
16108 (sqrtdf2_nofix): New insn.
16109 (sqrtdf2_fix): Likewise.
16110 (sqrtsf2): Disable if -mfix-ut699.
16112 2013-05-27 Richard Biener <rguenther@suse.de>
16114 PR middle-end/57412
16115 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16116 block for the new loop.
16118 2013-05-27 Richard Biener <rguenther@suse.de>
16120 PR tree-optimization/57343
16121 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16122 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16123 (number_of_iterations_cond): Do not build the folded tree.
16125 2013-05-27 Richard Biener <rguenther@suse.de>
16128 PR middle-end/57381
16129 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16130 OEP_CONSTANT_ADDRESS_OF retained.
16132 PR tree-optimization/57417
16133 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16134 for unchanged base.
16135 (set_ssa_val_to): Compare addresses using
16136 get_addr_base_and_unit_offset.
16138 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16140 PR rtl-optimization/56833
16141 * postreload.c (move2add_record_mode): New function.
16142 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16143 (move2add_use_add2_insn): Use move2add_record_sym_value.
16144 (move2add_use_add3_insn): Likewise.
16145 (reload_cse_move2add): Use move2add_valid_value_p and
16146 move2add_record_mode. Invalidate call-clobbered and REG_INC
16147 affected regs by setting reg_mode to VOIDmode.
16148 (move2add_note_store): Don't pretend the inside of a SUBREG is
16149 the actual destination. Invalidate single/leading registers by
16150 setting reg_mode to VOIDmode.
16151 Use move2add_record_sym_value, move2add_valid_value_p and
16152 move2add_record_mode.
16154 2013-05-27 Richard Biener <rguenther@suse.de>
16156 PR tree-optimization/57396
16157 * tree-affine.c (double_int_constant_multiple_p): Properly
16158 return false for val == 0 and div != 0.
16160 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16162 * config/mips/mips.h: Use #elif in preprocessor conditions.
16164 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16167 * config/mips/constraints.md (kl): New constraint.
16168 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16169 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16170 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16171 from LO for MIPS16.
16172 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16174 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16177 * config/mips/mips.c (mips_can_inline_p): New function.
16178 (TARGET_CAN_INLINE_P): Define.
16180 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16182 * sched-int.h (ds_t, dw_t): Make unsigned int.
16183 Fix documentation that describes how all the ds_t bits are used.
16184 Reserve the last bit for delayed-branch scheduling.
16185 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16186 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16187 (gen_dep_weak_1): Remove prototype.
16188 * sched-deps.c (get_dep_weak_1): Make static.
16189 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16190 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16191 * doc/tm.texi: Regenerate.
16192 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16194 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16197 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16199 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16200 Sandra Loosemore <sandra@codesourcery.com>
16202 * config.gcc (powerpc-*): Allow native for with-cpu.
16204 2013-05-24 Jeff Law <law@redhat.com>
16206 PR tree-optimization/57124
16207 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16208 conversion feeding a condition if the range has an overflow
16209 if -fstrict-overflow. Add warnings for when we do make the
16212 2013-05-24 Dehao Chen <dehao@google.com>
16214 * tree-cfg.c (locus_discrim_map): Fix the typo.
16215 (locus_discrim_hasher): Likewise.
16216 (locus_discrim_hasher::hash): Likewise.
16217 (locus_discrim_hasher::equal): Likewise.
16219 2013-05-24 Martin Jambor <mjambor@suse.cz>
16221 PR tree-optimization/57294
16222 * cgraph.h (ipa_record_stmt_references): Declare.
16223 * cgraphbuild.c (ipa_record_stmt_references): New function.
16224 (build_cgraph_edges): Use ipa_record_stmt_references.
16225 (rebuild_cgraph_edges): Likewise.
16226 (cgraph_rebuild_references): Likewise.
16227 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16228 associated with the old statement and build references from the
16229 newly built statements.
16230 * ipa-ref.c (ipa_remove_stmt_references): New function.
16231 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16233 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16235 * lra-constraints.c (emit_spill_move): Use smaller mode for
16237 (check_and_process_move): Consider mem-reg moves for secondary
16239 (curr_insn_transform): Don't lose insns emitted before for
16240 secondary memory moves.
16241 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16242 reg set up in the current insn.
16244 2013-05-24 Dehao Chen <dehao@google.com>
16246 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16248 (locus_descrim_hasher::equal): Likewise.
16249 (build_gimple_cfg): New discriminator assignment algorithm.
16250 (make_edges): Likewise.
16251 (next_discriminator_for_locus): Likewise.
16252 (same_line_p): Likewise.
16253 (assign_discriminators): Likewise.
16254 (make_cond_expr_edges): Likewise.
16255 (make_gimple_switch_edges): Likewise.
16256 (make_goto_expr_edges): Likewise.
16257 (make_gimple_asm_edges): Likewise.
16259 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16261 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16262 X format specifier to only display bottom 16 bits.
16263 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16264 immediate to match for operand 2, since it will be masked.
16266 2013-05-24 Richard Biener <rguenther@suse.de>
16268 PR tree-optimization/57287
16269 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16270 all SSA names that occur in abnormal PHIs.
16272 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16275 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16276 that index is not negative.
16278 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16280 PR rtl-optimization/55177
16281 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16282 (simplify_byte_swapping_operation): New.
16283 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16284 (simplify_relational_operation_1): Deal with BSWAP.
16286 2013-05-23 Richard Henderson <rth@redhat.com>
16289 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16290 (ix86_reorg): Call it.
16292 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16295 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16296 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16297 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16299 2013-05-23 Christian Bruel <christian.bruel@st.com>
16302 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16304 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16305 Marcus Shawcroft <marcus.shawcroft@arm.com>
16307 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16308 * config/aarch64/constraints.md (Usa): Remove.
16309 * doc/md.texi (AArch64 Usa): Remove.
16311 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16312 Marcus Shawcroft <marcus.shawcroft@arm.com>
16314 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16315 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16316 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16317 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16319 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16321 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16322 instruction (AdvSIMD).
16323 * config/aarch64/aarch64-builtins.c
16324 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16325 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16327 2013-05-23 Martin Jambor <mjambor@suse.cz>
16329 PR middle-end/57347
16330 * tree.h (contains_bitfld_component_ref_p): Declare.
16331 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16332 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16334 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16335 not access a bit-field. Assert all final offsets are byte-aligned.
16337 2013-05-23 Richard Biener <rguenther@suse.de>
16339 PR tree-optimization/57380
16340 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16341 least one invariant or re-used load.
16342 * passes.c (init_optimization_passes): Move pass_phiprop before
16345 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16347 * config/aarch64/aarch64-simd.md
16348 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16350 2013-05-23 Richard Biener <rguenther@suse.de>
16352 PR middle-end/57381
16353 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16354 OEP_CONSTANT_ADDRESS_OF retained.
16356 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16358 PR middle-end/57344
16359 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16360 don't lower unit. Handle unit not being always BITS_PER_WORD.
16362 2013-05-23 Richard Biener <rguenther@suse.de>
16364 PR rtl-optimization/57341
16365 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16366 instead of true_dependence.
16368 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16370 * bb-reorder.c (branch_threshold): Make const.
16371 (exec_threshold): Ditto.
16373 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16374 Pat Haugen <pthaugen@us.ibm.com>
16375 Peter Bergner <bergner@vnet.ibm.com>
16377 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16378 documentation for the power8 crypto builtins.
16380 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16382 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16383 macros for defining power8 builtin functions.
16384 (BU_P8V_AV_2): Likewise.
16385 (BU_P8V_AV_P): Likewise.
16386 (BU_P8V_VSX_1): Likewise.
16387 (BU_P8V_OVERLOAD_1): Likewise.
16388 (BU_P8V_OVERLOAD_2): Likewise.
16389 (BU_CRYPTO_1): Likewise.
16390 (BU_CRYPTO_2): Likewise.
16391 (BU_CRYPTO_3): Likewise.
16392 (BU_CRYPTO_OVERLOAD_1): Likewise.
16393 (BU_CRYPTO_OVERLOAD_2): Likewise.
16394 (XSCVSPDP): Fix typo, point to the correct instruction.
16395 (VCIPHER): Add power8 crypto builtins.
16396 (VCIPHERLAST): Likewise.
16397 (VNCIPHER): Likewise.
16398 (VNCIPHERLAST): Likewise.
16399 (VPMSUMB): Likewise.
16400 (VPMSUMH): Likewise.
16401 (VPMSUMW): Likewise.
16402 (VPERMXOR_V2DI): Likewise.
16403 (VPERMXOR_V4SI: Likewise.
16404 (VPERMXOR_V8HI: Likewise.
16405 (VPERMXOR_V16QI: Likewise.
16406 (VSHASIGMAW): Likewise.
16407 (VSHASIGMAD): Likewise.
16408 (VPMSUM): Likewise.
16409 (VPERMXOR): Likewise.
16410 (VSHASIGMA): Likewise.
16412 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16413 __CRYPTO__ if the crypto instructions are available.
16414 (altivec_overloaded_builtins): Add support for overloaded power8
16417 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16418 support for power8 crypto builtins.
16419 (builtin_function_type): Likewise.
16420 (altivec_init_builtins): Add support for builtins that take vector
16421 long long (V2DI) arguments.
16423 * config/rs6000/crypto.md: New file, define power8 crypto
16426 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16427 Pat Haugen <pthaugen@us.ibm.com>
16428 Peter Bergner <bergner@vnet.ibm.com>
16430 * doc/invoke.texi (Option Summary): Add power8 options.
16431 (RS/6000 and PowerPC Options): Likewise.
16433 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16434 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16435 wm, wn, wr documentation.
16437 * config/rs6000/constraints.md (wm): New constraint for VSX
16438 registers if direct move instructions are enabled.
16439 (wn): New constraint for no registers.
16440 (wq): New constraint for quad word even GPR registers.
16441 (wr): New constraint if 64-bit instructions are enabled.
16442 (wv): New constraint if power8 vector instructions are enabled.
16443 (wQ): New constraint for quad word memory locations.
16445 * config/rs6000/predicates.md (const_0_to_15_operand): New
16446 constraint for 0..15 for crypto instructions.
16447 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16448 as GPR and floating point registers.
16449 (int_reg_operand): New predicate to match only GPR registers.
16450 (base_reg_operand): New predicate to match base registers.
16451 (quad_int_reg_operand): New predicate to match even GPR registers
16452 for quad memory operations.
16453 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16454 operations in both GPR and VSX registers.
16455 (quad_memory_operand): New predicate for quad memory operations.
16456 (reg_or_indexed_operand): New predicate for direct move support.
16458 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16459 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16460 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16461 (POWERPC_MASKS): Add power8 options.
16462 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16465 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16466 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16468 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16469 (-mpower8-fusion): New power8 options.
16470 (-mpower8-fusion-sign): Likewise.
16471 (-mpower8-vector): Likewise.
16472 (-mcrypto): Likewise.
16473 (-mdirect-move): Likewise.
16474 (-mquad-memory): Likewise.
16476 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16477 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16479 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16480 (rs6000_debug_vector_unit): Add p8_vector.
16481 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16482 definitions. Also print fusion state.
16483 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16484 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16485 (rs6000_option_override_internal): Add support for power8.
16486 (rs6000_common_init_builtins): Add debugging for skipped builtins
16487 if -mdebug=builtin.
16488 (rs6000_adjust_cost): Add power8 support.
16489 (rs6000_issue_rate): Likewise.
16490 (insn_must_be_first_in_group): Likewise.
16491 (insn_must_be_last_in_group): Likewise.
16492 (force_new_group): Likewise.
16493 (rs6000_register_move_cost): Likewise.
16494 (rs6000_opt_masks): Likewise.
16496 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16497 power8 capable assembler, default to power7 options.
16498 (TARGET_DIRECT_MOVE): Likewise.
16499 (TARGET_CRYPTO): Likewise.
16500 (TARGET_P8_VECTOR): Likewise.
16501 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16502 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16503 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16504 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16505 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16506 (TARGET_XSCVDPSPN): Likewise.
16507 (TARGET_XSCVSPDPN): Likewsie.
16508 (TARGET_SYNC_HI_QI): Likewise.
16509 (TARGET_SYNC_TI): Likewise.
16510 (MASK_CRYPTO): Likewise.
16511 (MASK_DIRECT_MOVE): Likewise.
16512 (MASK_P8_FUSION): Likewise.
16513 (MASK_P8_VECTOR): Likewise.
16514 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16515 temporary used by some of the direct move instructions to get two FP
16516 temporary registers does not force creation of a stack frame.
16517 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16518 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16519 that any VSX registers are tieable, even if they are also an
16520 Altivec vector mode.
16521 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16522 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16523 (RS6000_BTM_CRYPTO): Likewise.
16524 (RS6000_BTM_COMMON): Likewise.
16526 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16527 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16528 (enum rs6000_vector): Add power8 vector support.
16530 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16534 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16535 (any_sibcall_could_use_r3): this and handle indirect calls.
16536 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16538 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16540 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16542 2013-05-22 Richard Biener <rguenther@suse.de>
16544 PR middle-end/57349
16545 * profile.c (branch_prob): Do not split blocks that are
16546 abnormally receiving from ECF_RETURNS_TWICE functions.
16548 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16550 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16551 address mode. Move pointer mode initialization to the same place.
16553 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16555 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16556 while it has any effect.
16558 2013-05-21 Easwaran Raman <eraman@google.com>
16560 PR tree-optimization/57322
16561 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16562 UID of the statement added to the BB to be 1.
16564 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16566 PR tree-optimization/57331
16567 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16568 of conversion from pointer type to integral type with integer.
16570 2013-05-21 Martin Jambor <mjambor@suse.cz>
16573 * ipa-prop.c (ipa_read_node_info): Process param_used and
16574 controlled_uses in the same order as when writing.
16576 2013-05-21 Magnus Granberg <baldrick@free.fr>
16579 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16581 2013-05-21 Richard Biener <rguenther@suse.de>
16583 PR tree-optimization/57318
16584 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16585 estimate stmts with side-effects as likely eliminated.
16587 2013-05-21 Richard Biener <rguenther@suse.de>
16589 PR tree-optimization/57330
16590 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16591 preserve the call stmts fntype.
16593 2013-05-21 Richard Biener <rguenther@suse.de>
16595 PR tree-optimization/57303
16596 * tree-ssa-sink.c (statement_sink_location): Improve killing
16597 stmt detection and properly handle self-assignments.
16599 2013-05-21 Christian Bruel <christian.bruel@st.com>
16601 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16602 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16603 registers. Set register size out of the PARALLEL loop.
16605 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16608 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16609 (*fmasf4, *fmasf4_media): New insns.
16611 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16613 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16614 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16615 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16616 (mips_idiv_insns): Update the comments to say that the returned
16617 instruction counts are in units of BASE_INSN_LENGTH.
16618 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16619 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16620 using 2 rather than 4 as the length of indirect MIPS16 and
16621 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16622 length of a NOP. Don't divide MIPS16 lengths by 2.
16623 (mips16_split_long_branches): Assume a branch is long if the
16624 length is greater than 4 rather than 8.
16625 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16626 rather than multiplying them by 2. Multiply instruction counts
16627 by BASE_INSN_LENGTH rather than 4.
16628 (*jump_mips16, tls_get_tp_mips16_<mode>)
16629 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16631 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16633 * config/mips/mips.md (extended_mips16): Remove branch case.
16634 (length): Remove duplicated extended_mips16 test.
16636 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16638 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16640 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16642 * recog.h (Recog_data): Rename to...
16643 (recog_data_d): ...this.
16644 (recog_data): Update accordingly.
16645 * recog.c (recog_data): Likewise.
16646 * reload.c (save_recog_data): Likewise.
16647 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16648 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16650 2013-05-17 Julian Brown <julian@codesourcery.com>
16652 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16653 found in a REG_EQUAL note, invalidate it.
16655 2013-05-17 Easwaran Raman <eraman@google.com>
16657 * tree-ssa-reassoc.c (find_insert_point): New function.
16658 (insert_stmt_after): Likewise.
16659 (get_def_stmt): Likewise.
16660 (ensure_ops_are_available): Likewise.
16661 (not_dominated_by): Likewise.
16662 (rewrite_expr_tree): Do not move statements beyond what is
16663 necessary. Remove call to swap_ops_for_binary_stmt...
16664 (reassociate_bb): ... and move it here.
16665 (build_and_add_sum): Assign UIDs for new statements.
16666 (linearize_expr): Likewise.
16667 (do_reassoc): Renumber gimple statement UIDs.
16669 2013-05-17 Jan Hubicka <jh@suse.cz>
16671 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16673 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16674 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16676 (output_weakrefs): Update.
16678 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16679 Martin Jambor <mjambor@suse.cz>
16681 PR middle-end/57276
16682 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16683 value that corresponds to the given aggval is found in values vector.
16685 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16687 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16688 sse, sse2, sse3, ssse3 and sse4a flags to options.
16690 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16692 * gengtype-state.c: (s_expr_writer): New class, to handle
16693 prettifying of output layout of s-expressions.
16694 (state_writer): New class, to write out gtype.state.
16695 (state_written_type_count): Move this variable into member data of
16697 (s_expr_writer::s_expr_writer): New code: constructor for new class
16698 (state_writer::state_writer(): ditto
16699 (s_expr_writer::write_new_line): New function
16700 (s_expr_writer::write_any_indent): ditto
16701 (s_expr_writer::begin_s_expr): ditto
16702 (s_expr_writer::end_s_expr): ditto
16703 (write_state_fileloc): convert to method of state_writer...
16704 (state_writer:: write_state_fileloc): ...and use methods of
16705 s_expr_writer to write indentation into the gtype.state output file
16706 to visually represent the hierarchical structure of the list
16708 (write_state_fields): ditto, renaming to...
16709 (state_writer::write_state_fields)
16710 (write_state_a_string): ditto, renaming to...
16711 (state_writer::write_state_a_string)
16712 (write_state_string_option): ditto, renaming to...
16713 (state_writer::write_state_string_option)
16714 (write_state_type_option): ditto, renaming to...
16715 (state_writer::write_state_type_option)
16716 (write_state_nested_option): ditto, renaming to...
16717 (state_writer::write_state_nested_option)
16718 (write_state_option): ditto, renaming to...
16719 (state_writer::write_state_option)
16720 (write_state_options): ditto, renaming to...
16721 (state_writer::write_state_options)
16722 (write_state_lang_bitmap): ditto, renaming to...
16723 (state_writer::write_state_lang_bitmap)
16724 (write_state_version): ditto, renaming to...
16725 (state_writer::write_state_version)
16726 (write_state_scalar_type): ditto, renaming to...
16727 (state_writer::write_state_scalar_type)
16728 (write_state_string_type): ditto, renaming to...
16729 (state_writer::write_state_string_type)
16730 (write_state_undefined_type): ditto, renaming to...
16731 (state_writer::write_state_undefined_type)
16732 (write_state_struct_union_type): ditto, renaming to...
16733 (state_writer::write_state_struct_union_type)
16734 (write_state_struct_type): ditto, renaming to...
16735 (state_writer::write_state_struct_type)
16736 (write_state_user_struct_type): ditto, renaming to...
16737 (state_writer::write_state_user_struct_type)
16738 (write_state_lang_struct_type): ditto, renaming to...
16739 (state_writer::write_state_lang_struct_type)
16740 (write_state_param_struct_type): ditto, renaming to...
16741 (state_writer::write_state_param_struct_type)
16742 (write_state_pointer_type): ditto, renaming to...
16743 (state_writer::write_state_pointer_type)
16744 (write_state_array_type): ditto, renaming to...
16745 (state_writer::write_state_array_type)
16746 (write_state_gc_used): ditto, renaming to...
16747 (state_writer::write_state_gc_used)
16748 (write_state_common_type_content): ditto, renaming to...
16749 (state_writer::write_state_common_type_content)
16750 (write_state_type): ditto, renaming to...
16751 (state_writer::write_state_type)
16752 (write_state_pair_list): ditto, renaming to...
16753 (state_writer::write_state_pair_list)
16754 (write_state_pair): ditto, renaming to...
16755 (state_writer::write_state_pair)
16756 (write_state_typedefs): ditto, renaming to...
16757 (state_writer::write_state_typedefs)
16758 (write_state_structures): ditto, renaming to...
16759 (state_writer::write_state_structures)
16760 (write_state_param_structs): ditto, renaming to...
16761 (state_writer::write_state_param_structs)
16762 (write_state_variables): ditto, renaming to...
16763 (state_writer::write_state_variables)
16764 (write_state_srcdir): ditto, renaming to...
16765 (state_writer::write_state_srcdir)
16766 (write_state_files_list): ditto, renaming to...
16767 (state_writer::write_state_files_list)
16768 (write_state_languages): ditto, renaming to...
16769 (state_writer::write_state_languages)
16770 (write_state): create a state_writer instance and use it when
16771 writing out the state file
16773 2013-05-17 Mike Stump <mikestump@comcast.net>
16775 PR rtl-optimization/57304
16776 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16777 accessing DF_REF_REAL_LOC.
16779 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16781 PR rtl-optimization/57281
16782 PR rtl-optimization/57300
16783 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16784 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16785 what the other splitter did if the registers are dead.
16787 2013-05-17 Richard Biener <rguenther@suse.de>
16789 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16792 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16794 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16797 2013-05-17 Marek Polacek <polacek@redhat.com>
16799 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16800 length when doing non-zero store of storing '\0' to '\0'.
16802 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16804 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16805 vect_external_def oprnd1 with loop_vinfo, try to emit
16806 optional cast, negation and and stmts on the loop preheader
16807 edge instead of into the pattern def seq.
16809 PR tree-optimization/57051
16810 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16811 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16813 2013-05-16 Nick Clifton <nickc@redhat.com>
16815 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16816 (rl78_is_naked_func): New function.
16817 (rl78_expand_prologue): Skip prologue generation for naked functions.
16818 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16819 * doc/extend.texi (naked): Add RL78 to the list of processors
16820 that supports this attribute.
16822 2013-05-16 Jeff Law <law@redhat.com>
16824 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16826 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16828 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16829 cache parameters using detect_caches_amd also for CYRIX,
16830 NSC and TM2 signatures.
16832 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16833 Dzianis Kahanovich <mahatma@eu.by>
16837 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16838 VIA/Centaur processors and determine their cache parameters
16839 using detect_caches_amd.
16841 2013-05-16 Teresa Johnson <tejohnson@google.com>
16843 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16844 (rtl_verify_edges): New function.
16845 (rtl_verify_bb_insns): Ditto.
16846 (rtl_verify_bb_pointers): Ditto.
16847 (rtl_verify_bb_insn_chain): Ditto.
16848 (rtl_verify_fallthru): Ditto.
16849 (rtl_verify_bb_layout): Ditto.
16850 (rtl_verify_flow_info_1): Outline checks into new functions.
16851 (rtl_verify_flow_info): Ditto.
16853 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16855 * cfghooks.c (copy_bbs): Add update_dominance argument.
16856 * cfghooks.h (copy_bbs): Update prototype.
16857 * tree-cfg.c (gimple_duplicate_sese_region):
16858 Add update_dominance argument.
16859 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16860 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16861 gimple_duplicate_sese_region call.
16862 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16863 Update copy_bbs call.
16864 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16865 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16867 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16869 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16870 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16871 vect_recog_rotate_pattern.
16872 (vect_recog_rotate_pattern): New function.
16874 2013-05-16 Jason Merrill <jason@redhat.com>
16876 * Makefile.in (LLINKER): New variable.
16877 (mostlyclean): Remove link mutex.
16878 * configure.ac: Handle --enable-link-mutex.
16879 * lock-and-run.sh: New script.
16881 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16884 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16887 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16889 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16891 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16893 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16894 * config/arm/arm.c (next_consecutive_mem): New function.
16895 (gen_movmem_ldrd_strd): Likewise.
16896 * config/arm/arm.md (movmemqi): Update condition and code.
16897 (unaligned_loaddi, unaligned_storedi): New patterns.
16899 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16901 * config.gcc: Obsolete *-*-solaris2.9*.
16902 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16904 2013-05-16 Richard Biener <rguenther@suse.de>
16906 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16907 earlier, after GRAPHITE transforms and IV canonicalization.
16909 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16911 * omp-low.c (extract_omp_for_data): For collapsed loops,
16912 if at least one of the loops is known at compile time to
16913 iterate zero times, set count to 0.
16914 (expand_omp_regimplify_p): New function.
16915 (expand_omp_for_generic): For collapsed loops, if at least
16916 one of the loops isn't known to iterate at least once,
16917 add runtime check with setting count to 0.
16918 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16919 For unsigned types if it isn't known at compile time that
16920 the loop will iterate at least once, add runtime check to bypass
16921 the whole loop if initial condition isn't true.
16923 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16925 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16927 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16929 PR middle-end/57286
16930 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16931 transformations to avoid an infinite loop.
16933 2013-05-16 Marek Polacek <polacek@redhat.com>
16935 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16937 2013-05-15 Leif Ekblad <leif@rdos.net>
16939 * config/i386/i386.c (ix86_decompose_address): Use
16940 DEFAULT_TLS_SEG_REG to access TLS segment register.
16941 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16942 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16943 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16945 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16948 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16949 sibling calls to functions that would normally be lazily bound,
16950 unless $gp is call-clobbered.
16952 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16954 * config/i386/i386.c (ix86_option_override_internal): Update
16955 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16956 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16957 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16958 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16960 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16962 2013-05-15 Andreas Schwab <schwab@suse.de>
16964 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16965 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16968 2013-05-15 Teresa Johnson <tejohnson@google.com>
16970 * loop-unroll.c (report_unroll_peel): Check decision before
16971 emitting unroll/peel message.
16973 2013-05-15 Teresa Johnson <tejohnson@google.com>
16975 * function.h (has_bb_partition): New rtl_data flag.
16976 (bb_reorder_complete): Ditto.
16977 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16978 instead of flag_reorder_blocks_and_partition.
16979 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16980 with some enhancements.
16981 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16982 * bb-reorder.c (connect_traces): Check for has_bb_partition
16983 instead of flag_reorder_blocks_and_partition.
16984 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16985 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16986 verify_hot_cold_block_grouping.
16987 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16989 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16992 * config/arm/predicates.md (call_insn_operand): New predicate.
16993 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16994 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16995 if insn is not a tail call.
16996 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16998 * config/arm/arm.h (enum reg_class): New caller save register class.
16999 (REG_CLASS_NAMES): Likewise.
17000 (REG_CLASS_CONTENTS): Likewise.
17001 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17004 2013-05-15 Richard Biener <rguenther@suse.de>
17006 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17007 of MSG_OPTIMIZED_LOCATIONS.
17008 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17009 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17011 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17012 of MSG_OPTIMIZED_LOCATIONS.
17013 (execute_vect_slp): Likewise.
17014 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17015 (vect_create_cond_for_alias_checks): Likewise.
17016 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17017 (vect_recog_widen_mult_pattern): Likewise.
17018 (vect_recog_widen_sum_pattern): Likewise.
17019 (vect_recog_over_widening_pattern): Likewise.
17020 (vect_recog_widen_shift_pattern): Likewise.
17021 (vect_recog_vector_vector_shift_pattern): Likewise.
17022 (vect_recog_divmod_pattern): Likewise.
17023 (vect_recog_mixed_size_cond_pattern): Likewise.
17024 (vect_recog_bool_pattern): Likewise.
17025 (vect_pattern_recog_1): Likewise.
17027 2013-05-15 Martin Jambor <mjambor@suse.cz>
17029 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17030 non-functions to builtin_unreachable.
17031 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17032 correct when new direct edges were discovered.
17034 2013-05-15 Martin Jambor <mjambor@suse.cz>
17036 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17037 header, print symbol order instead of node uid, print more information
17038 about indirect edge targets.
17039 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17041 (ipa_make_edge_direct_to_target): Likewise.
17042 (remove_described_reference): Likewise.
17043 (propagate_controlled_uses): Likewise.
17044 (ipa_print_node_params): Also print symbol order.
17045 (ipcp_transform_function): Print symbol order instead of node uids.
17046 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17047 (cgraph_get_create_real_symbol_node): Likewise.
17048 * ipa-cp.c (print_lattice): Likewise.
17049 (print_all_lattices): Likewise.
17050 (determine_versionability): Likewise.
17051 (initialize_node_lattices): Likewise.
17052 (estimate_local_effects): Likewise.
17053 (update_profiling_info): Likewise.
17054 (create_specialized_node): Likewise.
17055 (perhaps_add_new_callers): Likewise.
17056 (decide_about_value): Likewise.
17057 (decide_whether_version_node): Likewise.
17058 (identify_dead_nodes): Likewise.
17059 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17060 (dump_inline_summary): Likewise.
17061 (estimate_node_size_and_time): Likewise.
17062 (inline_analyze_function): Likewise.
17063 * ipa-inline.c (report_inline_failed_reason): Likewise.
17064 (want_early_inline_function_p): Likewise.
17065 (edge_badness): Likewise.
17066 (update_edge_key): Likewise.
17067 (inline_small_functions): Likewise. Add dumping of order to two other
17069 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17070 instead of node uids.
17071 (propagate_pure_const): Likewise.
17072 (propagate_pure_const): Likewise.
17073 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17074 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17075 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17077 * tree-pretty-print.c (dump_function_header): Likewise.
17078 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17079 Print symbol order instead of node uids.
17081 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17083 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17084 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17086 2013-05-15 Richard Biener <rguenther@suse.de>
17088 PR tree-optimization/57275
17089 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17090 return value for fail to do runtime alias checks for gather loads.
17092 2013-05-15 Jan Hubicka <jh@suse.cz>
17096 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17097 weakrefs are not external.
17098 (lto_symtab_merge_decls): Fix thinko when dealing with
17099 non-lto_symtab decls.
17100 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17101 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17102 * varpool.c (dump_varpool_node): Dump more flags.
17104 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17106 * config/i386/i386.c (processor_alias_table): Add instruction
17107 FSGSBASE for AMD bdver3 architecture.
17109 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17111 * tree.c (warn_deprecated_use): Print file:line using locus color.
17112 * diagnostic.c (diagnostic_report_current_module): Print file:line
17113 and file:line:column using locus color.
17115 2013-05-14 Mike Stump <mikestump@comcast.net>
17117 * gdbinit.in: Add __null.
17119 2013-05-14 Mike Stump <mikestump@comcast.net>
17121 * recog.h: Rename struct recog_data to Recog_data.
17122 * recog.c: Likewise.
17123 * reload.c (can_reload_into): Likewise.
17124 * config/picochip/picochip.c: Likewise.
17126 2013-05-14 Mike Stump <mikestump@comcast.net>
17128 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17130 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17132 * resource.h (struct resources): Remove unch_memory member.
17133 (CLEAR_RESOURCE): Don't clear unch_memory.
17134 * resource.c (mark_referenced_resources): Don't set it.
17135 (mark_set_resources): Likewise.
17136 (mark_target_live_regs): Don't clear it.
17137 (init_resource_info): Likewise.
17138 * reorg.c (resource_conflicts_p): Don't compare it.
17139 (redundant_insn): Don't set it.
17141 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17143 * emit-rtl.c (next_label): Remove unused function.
17144 (skip_consecutive_labels, link_cc0_insns): Move to ...
17145 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17146 only place where these functions are used, and make them static.
17148 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17150 * fold-const.c (fold_negate_expr): Handle vectors.
17151 (fold_truth_not_expr): Make it static.
17152 (fold_invert_truthvalue): New static function.
17153 (invert_truthvalue_loc): Handle vectors. Do not call
17154 fold_truth_not_expr directly.
17155 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17156 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17157 (fold_comparison): Handle vector constants.
17158 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17159 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17160 * tree.h (fold_truth_not_expr): Remove declaration.
17162 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17164 * config/aarch64/aarch64-simd.md
17165 (aarch64_vcond_internal<mode>): Rename to...
17166 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17167 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17168 float modes. Clarify all iterator modes.
17169 (vcond<mode><mode>): Use new name for vcond expanders.
17170 (vcond<v_cmp_result><mode>): Likewise.
17171 (vcondu<mode><mode>: Likewise.
17172 * config/aarch64/iterators.md (VDQF_COND): New.
17174 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17177 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17178 variable for the shift amount. Check that we shift by non-negative
17181 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17184 * config/arm/arm.h (EPILOGUE_USES): Only return true
17185 for LR_REGNUM after epilogue_completed.
17187 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17189 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17190 is error_mark_node.
17192 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17195 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17196 and Solaris 11+/x86 with gld.
17197 * configure: Regenerate.
17199 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17201 * expmed.c (expand_shift_1): Canonicalize rotates by
17202 constant bitsize / 2 to bitsize - 1.
17203 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17204 case ROTATERT>: Likewise.
17207 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17209 * config/i386/i386.md (rotateinv): New code attr.
17210 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17211 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17212 roll $31, %eax, etc.
17214 2013-05-14 Richard Biener <rguenther@suse.de>
17216 PR middle-end/57235
17217 * tree-eh.c (sink_clobbers): Give up for successors with
17218 multiple predecessors and no virtual uses.
17220 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17222 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17223 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17225 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17227 PR middle-end/57251
17228 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17229 the case when both op0 and op1 have VOIDmode.
17231 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17233 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17234 in multiply-accumulate mode.
17236 2013-05-13 Guozhi Wei <carrot@google.com>
17238 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17240 2013-05-13 Kai Tietz <ktietz@redhat.com>
17243 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17244 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17245 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17246 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17247 (ix86_expand_prologue): Likewise.
17248 (ix86_expand_split_stack_prologue): Likewise.
17249 (legitimate_pic_address_disp_p): Likewise.
17250 (legitimize_pic_address): Likewise.
17251 (legitimize_tls_address): Likewise.
17252 (legitimize_pe_coff_symbol): Likewise.
17253 (output_pic_addr_const): Likewise.
17254 (construct_plt_address): Likewise.
17255 (ix86_expand_call): Likewise.
17256 (x86_output_mi_thunk): Likewise.
17257 (x86_function_profiler): Likewise.
17259 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17261 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17262 similar switch cases.
17263 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17264 (aarch64_simd_mov_to_<mode>low): Delete.
17265 (aarch64_simd_mov_to_<mode>high): Delete.
17266 (move_lo_quad_<mode>): Add w<-r alternative.
17267 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17268 (aarch64_simd_mov_from_*): Update type attribute.
17269 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17272 2013-05-13 Jan Hubicka <jh@suse.cz>
17274 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17275 * config/i386/i386.c (ix86_compute_frame_layout,
17276 ix86_expand_epilogue, emit_i387_cw_initialization,
17277 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17278 ix86_local_alignment): Fix use of size/speed predicates.
17280 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17282 PR tree-optimization/45216
17283 PR tree-optimization/57157
17284 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17285 the (-Y) & (B - 1) variant if OP is |.
17286 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17287 return shifted. Use (-op1) & (prec - 1) as other_amount
17288 instead of prec - op1.
17290 2013-05-13 Martin Jambor <mjambor@suse.cz>
17292 PR middle-end/42371
17293 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17294 (ipa_constant_data): New type.
17295 (ipa_jump_func): Use ipa_constant_data to hold information about
17296 constant jump functions.
17297 (ipa_get_jf_constant): Adjust to jump function type changes.
17298 (ipa_get_jf_constant_rdesc): New function.
17299 (ipa_param_descriptor): New field controlled_uses.
17300 (ipa_get_controlled_uses): New function.
17301 (ipa_set_controlled_uses): Likewise.
17302 * ipa-ref.h (ipa_find_reference): Declare.
17303 * ipa-prop.c (ipa_cst_ref_desc): New type.
17304 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17306 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17307 New parameter cs. Adjust all callers.
17308 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17309 (remove_described_reference): New function.
17310 (jfunc_rdesc_usable): Likewise.
17311 (try_make_edge_direct_simple_call): Decrement controlled use count,
17312 attempt to remove reference if it hits zero.
17313 (combine_controlled_uses_counters): New function.
17314 (propagate_controlled_uses): Likewise.
17315 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17316 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17317 (ipa_print_node_params): Print described use counter.
17318 (ipa_write_jump_function): Adjust to jump function type changes.
17319 (ipa_read_jump_function): New parameter CS, pass it to
17320 ipa_set_jf_constant. Adjust caller.
17321 (ipa_write_node_info): Stream controlled use count
17322 (ipa_read_node_info): Likewise.
17323 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17325 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17326 count. Remove cloning-added reference if it reaches zero.
17327 * ipa-ref.c (ipa_find_reference): New function.
17329 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17331 * config/i386/i386.c (processor_target_table): Modified default
17332 alignment values for AMD BD and BT architectures.
17334 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17336 * tree-vect-generic.c (uniform_vector_p): Move ...
17337 * tree.c (uniform_vector_p): ... here.
17338 * tree.h (uniform_vector_p): Declare it.
17339 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17342 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17344 PR tree-optimization/57230
17345 * tree-ssa-strlen.c (handle_char_store): Record length for
17346 array store from STRING_CST.
17348 PR tree-optimization/57230
17349 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17352 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17354 * config/epiphany/epiphany.c (epiphany_init): Check size of
17355 NUM_MODES_FOR_MODE_SWITCHING.
17356 (epiphany_expand_prologue):
17357 Remove CONFIG_REGNUM initial value handling code.
17358 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17359 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17360 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17361 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17362 Don't return 1 for FP_MODE_NONE.
17363 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17364 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17365 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17366 * config/epiphany/epiphany.md (save_config): New pattern.
17368 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17370 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17372 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17374 * config/i386/i386.md (memory): Handle sseishft1.
17375 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17376 (*vec_extractv2di_1): Ditto.
17378 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17380 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17383 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17385 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17386 Add mthumb/march=armv7-a multilib.
17387 Add mthumb/march=armv7-r multilib.
17388 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17390 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17392 * config/v850/t-rtems: Add more multilibs.
17394 2013-05-10 Richard Biener <rguenther@suse.de>
17396 PR tree-optimization/57214
17397 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17398 not propagate from SSA names that occur in abnormal PHI nodes.
17400 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17402 * stor-layout.c (element_precision): New function.
17403 * machmode.h (element_precision): Declare it.
17404 * tree.c (build_minus_one_cst): New function.
17405 (element_precision): Likewise.
17406 * tree.h (build_minus_one_cst): Declare new function.
17407 (element_precision): Likewise.
17408 * fold-const.c (operand_equal_p): Use element_precision.
17409 (fold_binary_loc): Handle vector types.
17410 * convert.c (convert_to_integer): Use element_precision.
17411 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17414 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17416 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17417 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17418 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17419 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17420 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17421 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17422 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17423 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17424 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17425 (Uuw8): New constraints.
17426 (Usb4): Move into alphabetical order.
17427 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17428 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17429 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17430 previously unnamed patterns.
17431 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17432 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17433 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17434 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17435 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17436 of specific lengths.
17438 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17440 * config/i386/i386.md (rotateinv): New code attr.
17441 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17442 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17443 roll $31, %eax, etc.
17445 PR tree-optimization/45216
17446 PR tree-optimization/57157
17447 * tree-ssa-forwprop.c (simplify_rotate): New function.
17448 (ssa_forward_propagate_and_combine): Call it.
17450 2013-05-10 Richard Biener <rguenther@suse.de>
17452 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17453 disable peeling when we version for aliasing.
17454 (vector_alignment_reachable_p): Honor explicit user alignment.
17455 (vect_supportable_dr_alignment): Likewise.
17456 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17457 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17458 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17459 then peeling to arrange for the cost-model check to come first.
17461 2013-05-10 Alan Modra <amodra@gmail.com>
17463 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17464 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17465 * configure: Regenerate.
17467 2013-05-10 Alan Modra <amodra@gmail.com>
17470 * varasm.c (default_elf_select_section): Move !DECL_P check..
17471 (get_named_section): ..to here before calling get_section_name.
17473 (default_section_type_flags): Add DECL_P check.
17474 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17475 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17477 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17479 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17480 When using gen_stack_adjust_str with a register offset, add a
17481 REG_FRAME_RELATED_EXPR note.
17483 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17485 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17486 (*vec_extractv4si_zext_mem): Ditto.
17487 (*vec_extractv2di): Add 0->x and x->x alternatives.
17488 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17489 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17491 2013-05-09 Jason Merrill <jason@redhat.com>
17493 N3639 C++1y VLA support
17494 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17497 * tree.c (build_constructor_va): New.
17498 * tree.h: Declare it.
17500 2013-05-09 Martin Jambor <mjambor@suse.cz>
17503 * gimple-fold.c (canonicalize_constructor_val): Call
17504 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17506 2013-05-09 Jan Hubicka <jh@suse.cz>
17507 Richard Biener <rguenther@suse.de>
17510 * symtab.c (symtab_make_decl_local): Do not add private names.
17512 2013-05-09 Jan Hubicka <jh@suse.cz>
17515 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17516 (unlink_from_assembler_name_hash): Likewise.
17517 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17518 symtab_unregister_node, symtab_initialize_asm_name_hash,
17519 change_decl_assembler_name): Update.
17521 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17523 * config/aarch64/aarch64.md: New movtf split.
17524 (*movtf_aarch64): Update.
17525 (aarch64_movdi_tilow): Handle TF modes and rename to
17526 aarch64_movdi_<mode>low.
17527 (aarch64_movdi_tihigh): Handle TF modes and rename to
17528 aarch64_movdi_<mode>high
17529 (aarch64_movtihigh_di): Handle TF modes and rename to
17530 aarch64_mov<mode>high_di
17531 (aarch64_movtilow_di): Handle TF modes and rename to
17532 aarch64_mov<mode>low_di
17533 (aarch64_movtilow_tilow): Remove spurious whitespace.
17534 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17536 (aarch64_print_operand): Update.
17538 2013-05-09 Alan Modra <amodra@gmail.com>
17540 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17542 * configure: Regenerate.
17544 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17546 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17547 splitter preparation statements.
17548 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17549 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17552 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17554 * gimple-ssa-strength-reduction.c (count_candidates): Change
17555 return value to int.
17556 (analyze_candidates_and_replace): Change type of length to int.
17558 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17560 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17561 (*vec_extract<mode>): Use VI12_128 mode iterator.
17562 (*vec_extract<mode>_mem): Ditto.
17563 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17566 2013-05-08 Diego Novillo <dnovillo@google.com>
17571 2012-08-17 Diego Novillo <dnovillo@google.com>
17574 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17575 * config.in: Regenerate.
17576 * configure: Regenerate.
17577 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17579 2013-05-08 Jan Hubicka <jh@suse.cz>
17582 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17583 * cgraph.h (symtab_node_base): Add unique_name.
17584 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17585 input_overwrite_node, input_varpool_node): Stream unique_name.
17586 * cgraphclones.c (cgraph_create_virtual_clone,
17587 cgraph_function_versioning): Set unique_name.
17588 * ipa.c (function_and_variable_visibility): Set unique_name.
17590 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17592 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17593 (alloc_cand_and_find_basis): Restrict conditional candidate
17594 processing to CAND_MULTs.
17596 2013-05-08 Jan Hubicka <jh@suse.cz>
17599 lto-symtab.c (lto_symtab_symbol_p): New function.
17600 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17601 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17602 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17603 Skip static symbols.
17605 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17607 PR tree-optimization/57200
17608 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17609 Only call inform if the preceding warning_at returns true.
17611 2013-05-07 Han Shen <shenhan@google.com>
17613 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17614 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17615 * common.opt (fstack-protector-strong): New option.
17616 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17617 * doc/invoke.texi (Optimization Options): Document
17618 "-fstack-protector-strong".
17619 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17621 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17623 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17625 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17627 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17628 it is not lhs of the insn.
17629 (setup_reg_equiv): Remove insn having equiv memory which it is not
17631 * lra-constraints.c (process_address): Try to improve generation
17632 code for address base + disp.
17633 (lra_constraints): Make correct the code for checking insn setting
17634 up backward equivalence. Remove insn only if it is in the init
17636 * lra-eliminations.c (update_reg_eliminate): Change return value.
17637 (lra_eliminate): Use the result.
17639 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17641 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17642 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17643 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17644 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17645 register target operands.
17646 (*vec_extractv8hi_sse2): New pattern.
17647 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17648 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17649 (*vec_extract<mode>_mem): New insn and split pattern.
17651 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17653 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17654 (TARGET_ASAN_SHADOW_OFFSET): Define.
17655 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17656 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17658 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17660 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17661 (incr_vec_index): Return -1 if increment not found.
17662 (create_add_on_incoming_edge): Assert if increment not found.
17663 (record_increment): Limit number of increments recorded.
17664 (all_phi_incrs_profitable): Return false if an increment not found.
17665 (replace_profitable_candidates): Don't process increments that were
17667 (analyze_candidates_and_replace): Limit size of incr_vec.
17669 2013-05-07 Richard Biener <rguenther@suse.de>
17671 * calls.c (special_function_p): setjmp-like functions are leaf.
17672 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17673 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17675 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17677 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17678 (aarch64_simd_mov<mode>): New expander.
17679 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17680 (aarch64_simd_mov_to_<mode>high): Likewise.
17681 (aarch64_simd_mov_from_<mode>low): Likewise.
17682 (aarch64_simd_mov_from_<mode>high): Likewise.
17683 (aarch64_dup_lane<mode>): Update.
17684 (aarch64_dup_lanedi): New instruction pattern.
17685 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17686 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17688 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17690 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17691 (replace_mult_candidate): Remove unnecessary argument; remove
17692 unnecessary parameter from call to introduce_cast_before_cand.
17693 (replace_unconditional_candidate): Remove unnecessary parameter
17694 from call to replace_mult_candidate.
17695 (replace_conditional_candidate): Likewise.
17696 (insert_initializers): Use make_temp_ssa_name.
17697 (introduce_cast_before_cand): Remove unnecessary argument; use
17698 make_temp_ssa_name.
17699 (replace_one_candidate): Remove unnecessary argument; remove
17700 unnecessary parameter from calls to introduce_cast_before_cand.
17701 (replace_profitable_candidates): Remove unnecessary parameters
17702 from calls to replace_one_candidate.
17704 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17706 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17707 phi def as possibly hiding a basis for a CAND_ADD whose operands
17708 have been commuted in the analysis.
17709 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17711 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17713 * config/aarch64/aarch64.md
17714 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17715 shift value between 0-4.
17717 2013-05-07 Richard Biener <rguenther@suse.de>
17719 * double-int.h (rshift): New overload.
17720 * double-int.c (rshift): New function.
17721 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17722 (create_reference_ops_from_ref): Remove.
17723 (vn_reference_insert): Use shared ops for constructing the
17724 reference and copy it.
17726 2013-05-07 Richard Biener <rguenther@suse.de>
17728 PR middle-end/57190
17729 * tree-eh.c (sink_clobbers): Properly propagate
17730 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17732 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17734 PR tree-optimization/57149
17735 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17736 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17737 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17738 uninit_undefined_value_p instead of ssa_undefined_value_p.
17741 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17742 for modifier == EXPAND_INITIALIZER.
17744 2013-05-07 Anton Blanchard <anton@samba.org>
17746 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17747 for powerpc64 little endian.
17748 * configure: Regenerate.
17750 2013-05-06 Graham Stott <grahams@btinternet.com>
17752 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17753 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17754 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17755 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17757 2013-05-06 Graham Stott <grahams@btinternet.com>
17759 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17760 codes which allow non-lvalues.
17762 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17764 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17765 components are all 1s.
17766 (integer_minus_onep): New function.
17767 * tree.h (integer_minus_onep): Declare it.
17768 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17769 integer_minus_onep instead of integer_all_onesp.
17771 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17774 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17775 variations of these patterns.
17777 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17779 * config/i386/i386.md (isa): Add x64_sse4 member.
17780 (enabled): Handle x64_sse4.
17781 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17782 instruction for 64bit SSE4_1 targets. Update insn attributes.
17783 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17784 instruction for SSE4_1 targets. Update insn attributes.
17785 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17786 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17787 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17789 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17790 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17791 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17793 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17796 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17798 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17800 * final.c (do_assembler_dialects): Don't handle curly braces and
17801 vertical bar escaped by % as dialect delimiters.
17802 (output_asm_insn): Print curly braces and vertical bar if escaped
17803 by % and ASSEMBLER_DIALECT defined.
17804 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17805 * doc/tm.texi: Regenerated.
17807 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17809 * config/mips/mips.c: Include tree-pass.h.
17810 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17811 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17812 (pass_mips_machine_reorg2): New machine specific pass.
17813 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17814 (mips_option_override): Register the new pass.
17815 * rtl.h (cleanup_barriers): Remove prototype.
17816 (dbr_schedule): Likewise.
17817 * jump.c (cleanup_barriers): Make static.
17818 * reorg.c (dbr_schedule): Likewise.
17820 2013-05-06 Richard Biener <rguenther@suse.de>
17822 PR tree-optimization/57185
17823 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17824 SSA names properly.
17826 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17829 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17830 between "first_arg" and "insn", not "last" and "insn".
17832 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17834 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17835 (find_candidates_in_block): Re-enable slsr_process_phi.
17836 (create_phi_basis): Fix double counting of candidate adjustment.
17838 2013-05-06 Richard Biener <rguenther@suse.de>
17840 PR middle-end/57147
17841 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17842 the edge is also fallthru, preserve it and just clear the
17844 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17845 also complex, preserve that and just clear the fallthru flag.
17846 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17847 update virtual operands.
17849 2013-05-06 Alan Modra <amodra@gmail.com>
17851 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17852 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17853 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17854 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17855 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17856 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17858 2013-05-06 Alan Modra <amodra@gmail.com>
17860 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17861 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17862 (DEFAULT_ASM_ENDIAN): Define.
17863 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17864 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17865 Update -K PIC clause from sysv4.h.
17866 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17867 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17869 2013-05-06 Alan Modra <amodra@gmail.com>
17871 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17872 twice for little-endian.
17873 (ashrdi3_no_power, ashrdi3): Support little-endian.
17875 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17878 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17879 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17881 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17882 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17883 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17886 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17887 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17889 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17890 * config/pa/pa.opt: Make mbig-switch a no-op.
17891 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17892 (CASE_VECTOR_MODE): Always return SImode.
17893 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17894 for the !TARGET_BIG_SWITCH case.
17895 * config/pa/pa-linux.h: Likewise.
17896 * config/pa/pa-openbsd.h: Likewise.
17897 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17898 * config/pa/pa.md (short_jump): Remove define_insn.
17899 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17900 (casesi0): Remove define_insn.
17901 (type): Remove btable_branch.
17902 (pa_combine_type): Likewise.
17903 (in_nullified_branch_delay): Likewise.
17904 (in_call_delay): Likewise.
17905 (define_delay): Likewise.
17906 (define_insn_reservation "Z3"): Likewise.
17907 (define_insn_reservation "Z4"): Likewise.
17908 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17909 (pa_adjust_insn_length): Remove adjustment for btable branches.
17910 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17913 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17915 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17916 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17917 Add m->r,x alternatives.
17918 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17919 splitters using SWI48x mode iterator.
17920 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17921 TARGET_64BIT. Add m->x alternative.
17922 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17923 Add o->x alternative. Enable for TARGET_SSE.
17924 (sse_storeq): Remove expander.
17925 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17926 with memory input operand.
17927 (*vec_extractv2di_1 splitter): New.
17928 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17929 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17931 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17933 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17934 (INT_LOWPART): Delete.
17935 (extract_MB): Adjust.
17936 (extract_ME): Adjust.
17937 (print_operand): Adjust.
17939 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17941 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17942 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17943 (reg_or_logical_cint_operand, easy_fp_constant,
17944 logical_const_operand): Delete "CONST_DOUBLE" case.
17945 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17946 "HOST_BITS_PER_WIDE_INT == 64" test.
17947 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17948 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17949 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17950 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17952 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17953 CONST_DOUBLE DImode/VOIDmode case.
17954 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17955 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17956 CONST_DOUBLE VOIDmode case.
17957 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17958 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17959 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17960 Delete CONST_DOUBLE case.
17961 (splitters for mov FMOVE64 const_double): Delete
17962 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17963 "HOST_BITS_PER_WIDE_INT >= 64" test.
17964 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17966 (mov DI const_double): Delete.
17968 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17970 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17971 on op shows all bits zero in mode of a lowpart subreg, return zero.
17973 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17976 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17977 to save TFmode registers and DImode to save TImode registers for
17978 caller save operations.
17979 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17980 mark being partially clobbered since they only use the first
17983 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17984 and TDmode only use the upper 64-bits of each VSX register.
17986 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17988 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17989 (find_candidates_in_block): Disable slsr_process_phi.
17991 2013-05-03 Guozhi Wei <carrot@google.com>
17993 * coverage.c (coverage_obj_init): Move the construction of gcov
17995 (build_init_ctor): ... here.
17997 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17999 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18000 (slsr_cand_d): Redefine def_phi.
18001 (stride_status, phi_adjust_status, count_phis_status): New enums.
18002 (find_phi_def): New.
18003 (find_basis_for_base_expr): New.
18004 (find_basis_for_candidate): Handle hidden bases.
18005 (alloc_cand_and_find_basis): Handle phi candidates.
18006 (slsr_process_phi): New.
18007 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18008 (create_mul_imm_cand): Likewise.
18009 (create_add_ssa_cand): Exclude phi base candidates.
18010 (create_add_imm_cand): Likewise.
18011 (slsr_process_cast): Likewise.
18012 (slsr_process_copy): Likewise.
18013 (find_candidates_in_block): Handle phi candidates.
18014 (dump_candidate): Likewise.
18015 (unconditional_cands): Delete.
18016 (unconditional_cands_with_known_stride_p): Delete.
18017 (phi_dependent_cand_p): New.
18018 (cand_increment): Handle phi-dependent candidates.
18019 (replace_dependent): Delete.
18020 (replace_mult_candidate): New.
18021 (replace_unconditional_candidate): New.
18022 (incr_vec_index): Move to avoid forward reference.
18023 (create_add_on_incoming_edge): New.
18024 (create_phi_basis): New.
18025 (replace_dependents): Delete.
18026 (replace_conditional_candidate): New.
18027 (phi_add_costs): New.
18028 (replace_uncond_cands_and_profitable_phis): New.
18029 (record_increment): Handle phi adjustments.
18030 (record_phi_increments): New.
18031 (record_increments): Handle phi adjustments.
18032 (phi_incr_cost): New.
18033 (lowest_cost_path): Handle phis.
18034 (total_savings): Likewise.
18035 (analyze_increments): Likewise.
18036 (ncd_with_phi): New.
18037 (ncd_of_cand_and_phis): New.
18038 (nearest_common_dominator_for_cands): Handle phi increments.
18039 (all_phi_incrs_profitable): New.
18040 (replace_profitable_candidates): Handle phi-dependent candidates.
18041 (analyze_candidates_and_replace): Likewise.
18043 2013-05-03 Teresa Johnson <tejohnson@google.com>
18046 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18047 do not exceed REG_BR_PROB_BASE.
18049 2013-05-03 Jeff Law <law@redhat.com>
18051 PR tree-optimization/57144
18052 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18053 operand of the condition will bit into the new type when eliminating
18054 a cast feeding a condition.
18056 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18058 PR rtl-optimization/57130
18059 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18060 of COMPARE as in_code to the recursive call if needed.
18062 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18064 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18065 (enabled): Handle new members.
18066 * config/i386/sse.md (*vec_concatv2si): Merge from
18067 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18068 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18070 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18072 PR tree-optimization/57027
18073 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18074 for fnms opportunity, check we got the prerequisite kind
18075 of tree / gimple before using accessor functions.
18077 2013-05-03 Richard Biener <rguenther@suse.de>
18079 * double-int.h (lshift): New overload without precision
18080 and arith argument.
18081 (operator *=, operator +=, operator -=): Move ...
18082 * double-int.c (operator *=, operator +=, operator -=): ... here
18083 and implement more efficiently.
18084 (mul_double_with_sign): Remove.
18085 (lshift_double): Adjust to take unsinged shift argument, push
18086 dispatching code to callers.
18087 (mul_double_wide_with_sign): Add early out for callers that
18088 are not interested in high parts or overflow.
18089 (lshift): New function.
18090 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18091 dispatch code here.
18092 (lrotate, rrotate): Use logical shifts.
18093 * expr.c (get_inner_reference): Use lshift.
18094 * fixed-value.c (do_fixed_divide): Likewise.
18095 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18096 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18097 (indirect_refs_may_alias_p): Likewise.
18098 (stmt_kills_ref_p_1): Likewise.
18100 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18102 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18104 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18106 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18107 scalar form of FABD instruction.
18109 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18111 * lra-constraints.c (process_alt_operands): Add checking alt
18112 number to choose the best alternative.
18114 2013-05-02 Richard Biener <rguenther@suse.de>
18116 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18117 bitmap and its handling.
18118 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18120 2013-05-02 Richard Biener <rguenther@suse.de>
18122 PR middle-end/57140
18123 * tree-inline.c (copy_loops): Properly handle removed loops.
18124 (copy_cfg_body): Mark destination loops for fixup if source
18125 loops needed fixup.
18127 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18130 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18131 generating simple_return for naked functions.
18133 2013-05-02 Martin Jambor <mjambor@suse.cz>
18135 PR middle-end/56988
18136 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18137 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18139 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18140 ipa_agg_replacement_value structures.
18141 (known_aggs_to_agg_replacement_list): Likewise.
18142 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18143 (read_agg_replacement_chain): Likewise.
18144 (ipcp_transform_function): Also check that by_ref flags match.
18146 2013-05-02 Richard Biener <rguenther@suse.de>
18148 * graphds.h (struct graph): Add obstack member.
18149 * graphds.c (new_graph): Initialize obstack and allocate
18151 (add_edge): Allocate edge from the obstack.
18152 (free_graph): Free the obstack instead of all edges and vertices.
18154 2013-05-02 Teresa Johnson <tejohnson@google.com>
18156 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18158 * cfg.c (update_bb_profile_for_threading): Ditto.
18159 * tree-inline.c (copy_bb): Ditto.
18160 (copy_edges_for_bb): Ditto.
18161 (initialize_cfun): Ditto.
18162 (copy_cfg_body): Ditto.
18163 (expand_call_inline): Ditto.
18164 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18165 (estimate_node_size_and_time): Ditto.
18166 (inline_merge_summary): Ditto.
18167 * cgraphclones.c (cgraph_clone_edge): Ditto.
18168 (cgraph_clone_node): Ditto.
18169 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18170 (compute_trg_info): Ditto.
18172 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18174 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18175 S reg when fp attribute set.
18176 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18178 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18180 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18182 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18183 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18184 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18186 2013-05-02 Richard Biener <rguenther@suse.de>
18188 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18189 (struct instantiate_cache_entry): New type.
18190 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18191 (struct instantiate_cache_type): New type.
18192 (set_instantiated_value, get_instantiated_value): Remove.
18193 (get_instantiated_value_entry): New function.
18194 (instantiate_scev_name): Use the new cache and adjust.
18195 (instantiate_scev_poly): Adjust.
18196 (instantiate_scev_binary): Likewise.
18197 (instantiate_array_ref): Likewise.
18198 (instantiate_scev_convert): Likewise.
18199 (instantiate_scev_not): Likewise.
18200 (instantiate_scev_3): Likewise.
18201 (instantiate_scev_2): Likewise.
18202 (instantiate_scev_r): Likewise.
18203 (instantiate_scev): Likewise.
18204 (resolve_mixers): Likewise.
18206 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18209 * lra-constraints.c (best_small_class_operands_num): Remove.
18210 (process_alt_operands): Remove small_class_operands_num. Take
18211 small classes operands into losers and only if the operand is not
18212 matched. Modify debugging output.
18213 (curr_insn_transform): Remove best_small_class_operands_num.
18216 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18218 * config/aarch64/aarch64-builtins.c
18219 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18220 * config/aarch64/aarch64-simd-builtins.def
18221 (reduc_splus_): Add new modes.
18222 (reduc_uplus_): New.
18223 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18224 (reduc_uplus_v4sf): Likewise.
18225 (reduc_splus_v4sf): Likewise.
18226 (aarch64_addv<mode>): Likewise.
18227 (reduc_uplus_<mode>): Likewise.
18228 (reduc_splus_<mode>): Likewise.
18229 (aarch64_addvv2di): Likewise.
18230 (reduc_uplus_v2di): Likewise.
18231 (reduc_splus_v2di): Likewise.
18232 (aarch64_addvv2si): Likewise.
18233 (reduc_uplus_v2si): Likewise.
18234 (reduc_splus_v2si): Likewise.
18235 (reduc_<sur>plus_<mode>): New.
18236 (reduc_<sur>plus_v2di): Likewise.
18237 (reduc_<sur>plus_v2si): Likewise.
18238 (reduc_<sur>plus_v4sf): Likewise.
18239 (aarch64_addpv4sf): Likewise.
18240 * config/aarch64/arm_neon.h
18241 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18242 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18243 add UNSPEC_SADDV, UNSPEC_UADDV.
18245 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18247 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18249 * config/aarch64/arm_neon.h
18250 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18252 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18254 * config/aarch64/aarch64-builtins
18255 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18257 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18259 * config/aarch64/aarch64-simd-builtins.def
18260 (reduc_smax_): New.
18261 (reduc_smin_): Likewise.
18262 (reduc_umax_): Likewise.
18263 (reduc_umin_): Likewise.
18264 (reduc_smax_nan_): Likewise.
18265 (reduc_smin_nan_): Likewise.
18268 (smax): Update for V2SF, V4SF and V2DF modes.
18271 (smin_nan): Likewise.
18272 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18273 (<su><maxmin><mode>3): ...This, refactor.
18274 (s<maxmin><mode>3): New.
18275 (<maxmin_uns><mode>3): Likewise.
18276 (reduc_<maxmin_uns>_<mode>): Refactor.
18277 (reduc_<maxmin_uns>_v4sf): Likewise.
18278 (reduc_<maxmin_uns>_v2si): Likewise.
18279 (aarch64_<fmaxmin><mode>: Remove.
18280 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18282 (vmin<q>_f<32,64>): Likewise.
18283 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18285 (su): Add mappings for smax, smin, umax, umin.
18287 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18288 (FMAXMIN): Rename as...
18289 (FMAXMIN_UNS): ...This.
18291 (fmaxminv): Likewise.
18292 (fmaxmin): Likewise.
18294 (maxmin_uns_op): Likewise.
18296 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18298 * config/aarch64/arm_neon.h
18299 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18300 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18301 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18303 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18305 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18306 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18308 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18310 * config/aarch64/aarch64-simd.md
18311 (vcond<mode>_internal): Handle special cases for constant masks.
18312 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18313 (vcondu<mode><mode>): Likewise.
18314 (vcond<v_cmp_result><mode>): New.
18316 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18318 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18319 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18320 * config/aarch64/aarch64-simd-builtins.def
18321 (cmeq): Update to BUILTIN_VALLDI.
18326 * config/aarch64/arm_neon.h
18327 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18328 to builtins or C as appropriate.
18330 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18332 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18334 (cmhi): Rename to...
18336 * config/aarch64/aarch64-simd.md
18337 (simd_mode): Add SF.
18338 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18339 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18340 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18341 (cstore<mode>_neg): ...This.
18342 * config/aarch64/iterators.md
18344 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18345 (COMPARISONS): New.
18346 (UCOMPARISONS): Likewise.
18347 (optab): Add missing comparisons.
18353 (VCMP_S): Likewise.
18354 (VCMP_U): Likewise.
18355 (V_cmp_result): Add DF, SF modes.
18356 (v_cmp_result): Likewise.
18358 (vmtype): Likewise.
18359 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18361 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18363 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18364 define_insn to define_insn_and_split.
18365 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18366 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18367 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18368 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18369 (thumb2_negscc): Likewise.
18371 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18373 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18375 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18377 * config/arm/thumb2.md: Remove trailing whitespaces.
18379 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18381 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18382 Use gen_int_mode rather than GEN_INT.
18384 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18386 * value-prof.c (stream_in_histogram_value): Remove the strayed
18389 2013-04-30 Richard Biener <rguenther@suse.de>
18391 PR middle-end/57122
18392 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18394 2013-04-30 Richard Biener <rguenther@suse.de>
18396 PR middle-end/57107
18397 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18399 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18401 PR rtl-optimization/56957
18402 PR rtl-optimization/57105
18403 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18404 variable. Use just INSN_UID for determining whether an insn
18405 should be only disconnected from the insn stream.
18406 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18408 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18410 PR tree-optimization/57104
18411 * tsan.c (instrument_expr): Don't instrument accesses to
18412 DECL_HARD_REGISTER VAR_DECLs.
18414 2013-04-30 Richard Biener <rguenther@suse.de>
18416 * function.h (loops_for_fn): New inline function.
18417 (set_loops_for_fn): Likewise.
18418 * cfgloop.h (place_new_loop): Add struct function parameter.
18419 (get_loop): Likewise.
18420 (get_loops): Likewise.
18421 (number_of_loops): Likewise.
18422 (fel_next): Adjust.
18423 (fel_init): Likewise.
18424 * cfg.c (get_loop_copy): Adjust.
18425 * cfgloop.c (flow_loops_dump): Likewise.
18426 (record_loop_exits): Likewise.
18427 (verify_loop_structure): Likewise.
18428 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18429 (estimate_reg_pressure_cost): Likewise.
18430 (mark_loop_exit_edges): Likewise.
18431 * cfgloopmanip.c (place_new_loop): Likewise.
18432 (add_loop): Likewise.
18433 (duplicate_loop): Likewise.
18434 * graph.c (draw_cfg_nodes): Likewise.
18435 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18436 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18437 (extract_affine_chrec): Likewise.
18438 (build_scop_iteration_domain): Likewise.
18439 * graphite.c (graphite_initialize): Likewise.
18440 * ira-build.c (create_loop_tree_nodes): Likewise.
18441 (more_one_region_p): Likewise.
18442 (rebuild_regno_allocno_maps): Likewise.
18443 (mark_loops_for_removal): Likewise.
18444 (mark_all_loops_for_removal): Likewise.
18445 (remove_unnecessary_regions): Likewise.
18446 (ira_build): Likewise.
18447 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18448 * loop-init.c (fix_loop_structure): Likewise.
18449 (gate_rtl_move_loop_invariants): Likewise.
18450 (gate_rtl_unswitch): Likewise.
18451 (gate_rtl_unroll_and_peel_loops): Likewise.
18452 (rtl_doloop): Likewise.
18453 * lto-streamer-in.c (input_cfg): Likewise.
18454 * lto-streamer-out.c (output_cfg): Likewise.
18455 * modulo-sched.c (sms_schedule): Likewise.
18456 * predict.c (tree_estimate_probability): Likewise.
18457 (tree_estimate_probability_driver): Likewise.
18458 (estimate_loops): Likewise.
18459 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18460 (move_sese_region_to_fn): Likewise.
18461 (debug_loop_num): Likewise.
18462 * tree-chrec.c (chrec_evaluate): Likewise.
18463 (hide_evolution_in_other_loops_than_loop): Likewise.
18464 (chrec_component_in_loop_num): Likewise.
18465 (reset_evolution_in_loop): Likewise.
18466 (evolution_function_is_invariant_rec_p): Likewise.
18467 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18468 * tree-inline.c (copy_loops): Likewise.
18469 (copy_cfg_body): Likewise.
18470 (tree_function_versioning): Likewise.
18471 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18472 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18474 (add_to_evolution_1): Likewise.
18475 (scev_const_prop): Likewise.
18476 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18477 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18478 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18479 (tree_ssa_lim_initialize): Likewise.
18480 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18481 (verify_loop_closed_ssa): Likewise.
18482 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18483 (tree_ssa_loop_im): Likewise.
18484 (tree_ssa_loop_unswitch): Likewise.
18485 (tree_vectorize): Likewise.
18486 (check_data_deps): Likewise.
18487 (tree_ssa_loop_ivcanon): Likewise.
18488 (tree_ssa_loop_bounds): Likewise.
18489 (tree_complete_unroll): Likewise.
18490 (tree_complete_unroll_inner): Likewise.
18491 (tree_parallelize_loops): Likewise.
18492 (tree_ssa_loop_prefetch): Likewise.
18493 (tree_ssa_loop_ivopts): Likewise.
18494 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18495 * tree-vectorizer.c (vectorize_loops): Likewise.
18497 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18499 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18500 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18501 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18502 with EABI_LINK_SPEC.
18504 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18507 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18510 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18513 * lra-constraints.c (process_alt_operands): Discourage a bit more
18514 using memory for pseudos. Print cost dump for alternatives.
18515 Modify cost values for conflicts with early clobbers.
18516 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18518 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18521 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18523 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18525 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18526 from/to S register.
18527 (movdi_aarch64): Support LDR/STR from/to D register.
18529 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18531 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18532 or higher by default.
18534 2013-04-29 Richard Biener <rguenther@suse.de>
18536 PR middle-end/57075
18537 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18538 even if not adding abnormal edges for calls that can make
18541 2013-04-29 Richard Biener <rguenther@suse.de>
18543 PR middle-end/57103
18544 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18546 (move_stmt_r): Remove redundant checking.
18548 2013-04-29 Teresa Johnson <tejohnson@google.com>
18551 * basic-block.h (apply_scale): New function.
18552 (apply_probability): Use apply_scale.
18553 * gimple-streamer-in.c (input_bb): Ditto.
18554 * lto-streamer-in.c (input_cfg): Ditto.
18555 * lto-cgraph.c (merge_profile_summaries): Ditto.
18556 * tree-optimize.c (execute_fixup_cfg): Ditto.
18557 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18558 (copy_edges_for_bb): Ditto.
18559 (copy_cfg_body): Ditto.
18561 2013-04-29 Tom de Vries <tom@codesourcery.com>
18563 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18564 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18565 (tail_merge_optimize): Handle current_loops == NULL.
18567 2013-04-26 Jeff Law <law@redhat.com>
18569 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18570 (simplify_cond_using_ranges): Generalize code to simplify
18571 COND_EXPRs where one argument is a constant and the other
18572 is an SSA_NAME created by an integral type conversion.
18574 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18576 * config/arm/arm.md (store_minmaxsi): Use only when
18577 optimize_insn_for_size_p.
18579 2013-04-29 Christian Bruel <christian.bruel@st.com>
18582 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18584 2013-04-29 Richard Biener <rguenther@suse.de>
18586 PR middle-end/57089
18587 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18588 loop tree make sure to schedule a fixup for the child as well.
18589 (expand_omp_for_generic): Properly add loops.
18590 (expand_omp_for_static_nochunk): Likewise.
18591 (expand_omp_for_static_chunk): Likewise.
18592 (expand_omp_for): For the degenerate case fixup loops.
18593 (expand_omp_sections): Fix default bb placement in loops.
18594 (expand_omp_atomic_pipeline): Properly add loops.
18596 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18598 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18600 2013-04-29 Tom de Vries <tom@codesourcery.com>
18602 * tree-ssa-tail-merge.c: Update header comment.
18604 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18606 * config/aarch64/arm_neon.h
18607 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18608 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18609 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18610 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18611 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18612 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18613 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18614 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18616 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18618 * config/aarch64/aarch64-simd.md
18619 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18620 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18621 fix_trunc, fixuns_trunc.
18622 (ftrunc<VDQF:mode>2): New.
18623 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18624 (fix_trunc_optab): New.
18626 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18628 * config/aarch64/aarch64-builtins.c
18629 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18630 iceilf, lround, iroundf.
18632 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18635 * config/i386/i386.h (enum ix86_tune_indices)
18636 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18637 New, split from X86_TUNE_INTER_UNIT_MOVES.
18638 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18639 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18640 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18641 (TARGET_INTER_UNIT_MOVES): Remove.
18642 * config/i386/i386.c (initial_ix86_tune_features): Update.
18643 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18644 (ix86_expand_convert_uns_didf_sse): Use
18645 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18646 (ix86_expand_vector_init_one_nonzero): Ditto.
18647 (ix86_expand_vector_init_interleave): Ditto.
18648 (inline_secondary_memory_needed): Return true for moves from SSE class
18649 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18650 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18651 * config/i386/constraints.md (Yi, Ym): Depend on
18652 TARGET_INTER_UNIT_MOVES_TO_VEC.
18653 (Yj, Yn): New constraints.
18654 * config/i386/i386.md (*movdi_internal): Change constraints of
18655 operand 1 from Yi to Yj and from Ym to Yn.
18656 (*movsi_internal): Ditto.
18657 (*movdf_internal): Ditto.
18658 (*movsf_internal): Ditto.
18659 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18660 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18661 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18662 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18663 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18664 * config/i386/sse.md (movdi_to_sse): Ditto.
18665 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18666 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18667 TARGET_INTER_UNIT_MOVES.
18668 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18669 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18670 instead of TARGET_INTER_UNIT_MOVES.
18671 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18672 operand 1 from Yi to Yj and from Ym to Yn.
18674 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18676 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18677 (float_truncate_hi_): Likewise.
18678 (float_extend_lo_): Likewise.
18679 (float_truncate_lo_): Likewise.
18680 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18681 (aarch64_float_extend_lo_v2df): Likewise.
18682 (vec_unpacks_hi_v4sf): Likewise.
18683 (aarch64_float_truncate_lo_v2sf): Likewise.
18684 (aarch64_float_truncate_hi_v4sf): Likewise.
18685 (vec_pack_trunc_v2df): Likewise.
18686 (vec_pack_trunc_df): Likewise.
18688 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18690 * config/aarch64/aarch64-builtins.c
18691 (aarch64_fold_builtin): Fold float conversions.
18692 * config/aarch64/aarch64-simd-builtins.def
18693 (floatv2si, floatv4si, floatv2di): New.
18694 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18695 * config/aarch64/aarch64-simd.md
18696 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18697 * config/aarch64/iterators.md (FLOATUORS): New.
18698 (optab): Add float, floatuns.
18699 (su_optab): Likewise.
18701 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18703 * config/aarch64/aarch64-builtins.c
18704 (aarch64_builtin_vectorized_function): Use new names for
18706 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18707 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18708 (fcvtzu): Split as...
18709 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18710 (fcvtas): Split as...
18711 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18712 (fcvtau): Split as...
18713 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18714 (fcvtps): Split as...
18715 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18716 (fcvtpu): Split as...
18717 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18718 (fcvtms): Split as...
18719 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18720 (fcvtmu): Split as...
18721 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18722 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18723 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18724 (lfrintnusf, lfrintnudf): Likewise.
18725 * config/aarch64/aarch64-simd.md
18726 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18728 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18729 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18730 (fcvt_pattern): Likewise.
18732 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18734 * config/aarch64/aarch64-simd.md
18735 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18736 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18738 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18740 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18741 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18742 (vrnd<a,m,n,p>_f32): Implement using builtins.
18743 (vrnd<i,x><q>_f<32, 64>): New.
18745 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18747 * config/aarch64/aarch64-builtins.c
18748 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18749 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18750 (frintz): Rename to...
18752 (frintp): Rename to...
18754 (frintm): Rename to...
18756 (frinti): Rename to...
18757 (nearbyint): ...this.
18758 (frintx): Rename to...
18760 (frinta): Rename to...
18762 * config/aarch64/aarch64-simd.md
18763 (aarch64_frint<frint_suffix><mode>): Delete.
18764 (<frint_pattern><mode>2): Convert to insn.
18765 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18766 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18767 (frint_pattern): Likewise.
18768 (frint_suffix): Likewise.
18770 2013-04-29 Richard Biener <rguenther@suse.de>
18772 PR tree-optimization/57081
18773 * loop-init.c: Include tree-flow.h.
18774 (loop_optimizer_finalize): Free number of iteration estimates.
18775 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18777 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18779 PR tree-optimization/57083
18780 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18781 non-singleton shift count range, zero extend low_bound for uns case.
18783 * config/i386/predicates.md (general_vector_operand): New predicate.
18784 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18785 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18786 if they aren't nonimmediate operands. If their original values
18787 satisfy const_vector_equal_evenodd_p, don't shift them.
18788 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18789 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18790 if not nonimmediate_operand.
18791 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18792 instead of register_operand.
18793 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18795 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18797 * stor-layout.c (finalize_size_functions): Allocate a structure and
18798 reset cfun before dumping the functions.
18800 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18802 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18805 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18806 use xop_pmacsdqh if uns_p.
18807 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18808 the immediate rotate count.
18810 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18812 * rtl.h (struct rtx_def): Add comment for field jump.
18813 (LRA_SUBREG_P): New macro.
18814 * recog.c (register_operand): Check LRA_SUBREG_P.
18815 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18817 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18818 (lra_final_code_change): Skip subreg change for operators.
18819 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18820 if there are no operand changes.
18821 * lra-constraints.c (curr_insn_set): New.
18822 (match_reload): Set LRA_SUBREG_P.
18823 (emit_spill_move): Ditto.
18824 (check_and_process_move): Use curr_insn_set. Process only single
18825 set insns. Don't initialize sec_mem_p and change_p.
18826 (simplify_operand_subreg): Use LRA_SUBREG_P.
18827 (reg_in_class_p): New function.
18828 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18829 of #ifdef. Add code to remove cycling.
18830 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18831 non-null disp. Reload inner instead of disp when base and index
18832 are null. Try to put lo_sum into register.
18833 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18834 (check_and_process_move): Move code for move cost check to
18835 simple_move_p. Remove equiv_substitution.
18836 (simple_move_p): New function.
18837 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18838 curr_insn_set. Call check_and_process_move only for single set
18839 insns. Use the new function. Move call of check_and_process_move
18840 after operand equiv substitution and address process.
18842 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18845 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18846 with nonlocal goto receivers or returns twice calls, ignore
18847 unininitialized values from abnormal edges to nl goto receiver
18848 or returns twice call.
18850 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18852 PR tree-optimization/57051
18853 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18854 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18857 2013-04-26 Richard Biener <rguenther@suse.de>
18859 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18860 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18861 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18862 (fixup_loop_arrays_after_move): New function.
18863 (move_sese_region_to_fn): Properly outline the loop tree parts
18864 of the SESE region.
18866 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18868 * config/i386/i386.md (type, unit): Fix long lines.
18870 2013-04-26 Richard Biener <rguenther@suse.de>
18872 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18873 (lto-streamer-out.o): Likewise.
18874 * cfgloop.c (init_loops_structure): Export, add struct function
18875 argument and adjust.
18876 (flow_loops_find): Adjust.
18877 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18878 (init_loops_structure): Declare.
18879 * lto-streamer-in.c: Include cfgloop.h.
18880 (input_cfg): Input the loop tree.
18881 * lto-streamer-out.c: Include cfgloop.h.
18882 (output_cfg): Output the loop tree.
18883 (output_struct_function_base): Do not drop PROP_loops.
18885 2013-03-26 Richard Biener <rguenther@suse.de>
18887 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18888 (pass_build_cfg): Provide PROP_loops.
18889 (move_sese_region_to_fn): Remove loops that are outlined into fn
18891 * tree-inline.c: Include cfgloop.h.
18892 (initialize_cfun): Do not drop PROP_loops.
18893 (copy_loops): New function.
18894 (copy_cfg_body): Copy loop structure.
18895 (tree_function_versioning): Initialize destination loop tree.
18896 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18897 (pass_parallelize_loops): Do IL verification.
18898 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18899 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18900 the CFG make sure we fixup loops as well.
18901 * tree-ssa-tail-merge.c: Include cfgloop.h.
18902 (replace_block_by): When merging loop latches mark loops for fixup.
18903 * lto-streamer-out.c (output_struct_function_base): Drop
18904 PROP_loops for now.
18905 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18906 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18907 * ipa-split.c: Include cfgloop.h.
18908 (split_function): Add the new return block to the loop tree root.
18909 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18910 whether we have removed the forwarder block.
18911 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18912 * cfgloop.h (place_new_loop): Declare.
18913 * cfgloopmanip.c (place_new_loop): Export.
18914 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18915 (tree-switch-conversion.o): Likewise.
18916 (tree-complex.o): Likewise.
18917 (tree-inline.o): Likewise.
18918 (tree-ssa-tailmerge.o): Likewise.
18919 (ipa-split.o): Likewise.
18920 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18921 (tree-ssa-copy.o): Likewise.
18922 * tree-switch-conversion.c: Include cfgloop.h
18923 (process_switch): If we emit a bit-test cascade, schedule loops
18925 * tree-complex.c: Include cfgloop.h.
18926 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18927 * asan.c: Include cfgloop.h.
18928 (create_cond_insert_point): Properly add new basic-blocks to
18929 loops, schedule loop fixup.
18930 * cfgloop.c (verify_loop_structure): Check that looks are not
18932 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18934 (expand_omp_for_generic): Likewise.
18935 (expand_omp_sections): Likewise.
18936 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18937 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18938 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18939 is initialized, not when loops are present.
18940 * tree-parloops.c (parallelize_loops): Remove checking here.
18941 * passes.c (init_optimization_passes): Schedule a copy-propagation
18942 pass before complete unrolling of inner loops.
18944 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18946 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18947 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18948 (colorize_init): Add argument to _WIN32 version.
18949 * toplev.c: Include diagnostic-color.h.
18950 (process_options): Default to -fdiagnostics-color=auto if
18951 GCC_COLORS env var is in the environment.
18952 * common.opt (fdiagnostics-color=): Add Var and Init.
18953 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18954 env var is in the environment, the default is auto rather than never.
18956 * diagnostic.h (file_name_as_prefix): Add context argument.
18957 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18958 the string as locus.
18959 * langhooks.c (lhd_print_error_function): Adjust caller.
18961 2013-04-25 Lawrence Crowl <crowl@google.com>
18963 * var-tracking.c (shared_hash_def::htab):
18964 Change type to hash_table. Update dependent calls and types.
18966 2013-04-25 Lawrence Crowl <crowl@google.com>
18968 * Makefile.in: Update as needed below.
18970 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18971 Move declaration to after the type's method definitons.
18973 * attribs.c (htab_t scoped_attributes::attribute_hash):
18974 Change type to hash_table. Update dependent calls and types.
18976 * bitmap.c (htab_t bitmap_desc_hash):
18977 Change type to hash_table. Update dependent calls and types.
18979 * cselib.c (htab_t cselib_hash_table):
18980 Change type to hash_table. Update dependent calls and types.
18982 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18983 (hash_string_slot_node): Move implementation into lto-streamer.h
18984 struct string_slot_hasher.
18985 (eq_string_slot_node): Likewise.
18987 * data-streamer-out.c: Update output_block::string_hash_table
18988 dependent calls and types.
18990 * dwarf2cfi.c (htab_t trace_index):
18991 Change type to hash_table. Update dependent calls and types.
18993 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18994 Change type to hash_table. Update dependent calls and types.
18995 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18996 (htab_t optimize_external_refs::map): Likewise.
18997 (htab_t output_comp_unit::extern_map): Likewise.
18998 (htab_t output_comdat_type_unit::extern_map): Likewise.
18999 (htab_t output_macinfo::macinfo_htab): Likewise.
19000 (htab_t optimize_location_lists::htab): Likewise.
19001 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19003 * except.c (htab_t ehspec_hash_type):
19004 Change type to hash_table. Update dependent calls and types.
19005 (assign_filter_values::ttypes): Likewise.
19006 (assign_filter_values::ehspec): Likewise.
19007 (sjlj_assign_call_site_values::ar_hash): Likewise.
19008 (convert_to_eh_region_ranges::ar_hash): Likewise.
19010 * gcse.c (htab_t pre_ldst_table):
19011 Change type to hash_table. Update dependent calls and types.
19013 * ggc-common.c (htab_t saving_htab):
19014 Change type to hash_table. Update dependent calls and types.
19015 (htab_t loc_hash): Likewise.
19016 (htab_t ptr_hash): Likewise.
19017 (call_count): Rename ggc_call_count.
19018 (call_alloc): Rename ggc_call_alloc.
19019 (loc_descriptor): Rename make_loc_descriptor.
19020 (add_statistics): Rename ggc_add_statistics.
19022 * ggc-common.c (saving_htab):
19023 Change type to hash_table. Update dependent calls and types.
19025 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19026 (push_gimplify_context): Likewise.
19027 (pop_gimplify_context): Likewise.
19028 (struct gimple_temp_hash_elt): Added.
19029 (struct gimplify_hasher): Likewise.
19030 (struct gimplify_ctx.temp_htab):
19031 Change type to hash_table. Update dependent calls and types.
19033 * gimple-fold.c: Include gimplify-ctx.h.
19035 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19036 Change type to hash_table. Update dependent calls and types.
19037 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19038 avoid potential global name collision.
19040 * gimplify.c: Include gimplify-ctx.h.
19041 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19042 (htab_t gimplify_ctx::temp_htab):
19043 Update dependent calls and types for new type hash_table.
19044 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19045 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19047 * gimplify-ctx.h: New.
19048 (struct gimple_temp_hash_elt): Move from gimplify.c.
19049 (class gimplify_hasher): New.
19050 (struct gimplify_ctx): Move from gimple.h.
19051 (htab_t gimplify_ctx::temp_htab):
19052 Change type to hash_table. Update dependent calls and types.
19054 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19055 (htab_t ivs_params::newivs_index):
19056 Change type to hash_table. Update dependent calls and types.
19057 (htab_t ivs_params::params_index): Likewise.
19058 (htab_t print_generated_program::params_index): Likewise.
19059 (htab_t gloog::newivs_index): Likewise.
19060 (htab_t gloog::params_index): Likewise.
19062 * graphite.c: Include graphite-htab.h.
19063 4htab_t graphite_transform_loops::bb_pbb_mapping):
19064 Change type to hash_table. Update dependent calls and types.
19066 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19067 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19068 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19070 * graphite-dependences.c: Include graphite-htab.h.
19071 (loop_is_parallel_p): Change hash table type of parameter.
19073 * graphite-htab.h: New.
19074 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19075 (extern find_pbb_via_hash): Move from graphite-poly.h.
19076 (extern loop_is_parallel_p): Move from graphite-poly.h.
19077 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19079 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19080 (extern loop_is_parallel_p): Move to graphite-htab.h.
19081 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19083 * haifa-sched.c (htab_t delay_htab):
19084 Change type to hash_table. Update dependent calls and types.
19085 (htab_t delay_htab_i2): Likewise.
19087 * ira-color.c (htab_t allocno_hard_regs_htab):
19088 Change type to hash_table. Update dependent calls and types.
19090 * ira-costs.c (htab_t cost_classes_htab):
19091 Change type to hash_table. Update dependent calls and types.
19093 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19094 Change type to hash_table. Update dependent calls and types.
19096 * loop-iv.c (htab_t bivs):
19097 Change type to hash_table. Update dependent calls and types.
19099 * loop-unroll.c (htab_t opt_info::insns_to_split):
19100 Change type to hash_table. Update dependent calls and types.
19101 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19103 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19104 (struct string_slot_hasher): New.
19105 (htab_t output_block::string_hash_table):
19106 Change type to hash_table. Update dependent calls and types.
19108 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19109 (htab_t file_name_hash_table):
19110 Change type to hash_table. Update dependent calls and types.
19112 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19115 * lto-streamer.c (htab_t tree_htab):
19116 Change type to hash_table. Update dependent calls and types.
19118 * omp-low.c: Include gimplify-ctx.h.
19120 * passes.c (htab_t name_to_pass_map):
19121 Change type to hash_table. Update dependent calls and types.
19122 (pass_traverse): Rename to passes_pass_traverse.
19124 * plugin.c (htab_t event_tab):
19125 Change type to hash_table. Update dependent calls and types.
19127 * postreload-gcse.c (htab_t expr_table):
19128 Change type to hash_table. Update dependent calls and types.
19129 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19131 * sese.c (debug_rename_map_1): Make extern.
19132 (htab_t copy_bb_and_scalar_dependences::rename_map):
19133 Change type to hash_table. Update dependent calls and types.
19135 * sese.h (extern debug_rename_map): Move to .c file.
19137 * store-motion.c (htab_t store_motion_mems_table):
19138 Change type to hash_table. Update dependent calls and types.
19140 * trans-mem.c (htab_t tm_new_mem_hash):
19141 Change type to hash_table. Update dependent calls and types.
19143 * tree-browser.c (htab_t TB_up_ht):
19144 Change type to hash_table. Update dependent calls and types.
19146 * tree-cfg.c (htab_t discriminator_per_locus):
19147 Change type to hash_table. Update dependent calls and types.
19149 * tree-complex.c: Include tree-hasher.h
19150 (htab_t complex_variable_components):
19151 Change type to hash_table. Update dependent calls and types.
19153 * tree-eh.c (htab_t finally_tree):
19154 Change type to hash_table. Update dependent calls and types.
19156 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19157 struct int_tree_hasher.
19158 (extern int_tree_map_eq): Likewise.
19159 (uid_decl_map_hash): Removed.
19160 (extern decl_tree_map_eq): Likewise.
19162 * tree-hasher.h: New.
19163 (struct int_tree_hasher): New.
19164 (typedef int_tree_htab_type): New.
19166 * tree-inline.c: Include gimplify-ctx.h.
19168 * tree-mudflap.c: Include gimplify-ctx.h.
19170 * tree-parloops.c: Include tree-hasher.h.
19171 (htab_t eliminate_local_variables_stmt::decl_address):
19172 Change type to hash_table. Update dependent calls and types.
19173 (htab_t separate_decls_in_region::decl_copies): Likewise.
19175 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19176 Change type to hash_table. Update dependent calls and types.
19178 * tree-sra.c (candidates):
19179 Change type to hash_table. Update dependent calls and types.
19181 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19183 (int_tree_map_hash): Likewise.
19185 * tree-ssa-dom.c (htab_t avail_exprs):
19186 Change type to hash_table. Update dependent calls and types.
19188 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19189 Change type to hash_table. Update dependent calls and types.
19191 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19192 Change type to hash_table. Update dependent calls and types.
19194 * tree-ssa-phiopt.c (seen_ssa_names):
19195 Change type to hash_table. Update dependent calls and types.
19197 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19198 Change type to hash_table. Update dependent calls and types.
19200 * tree-ssa-uncprop.c (equiv):
19201 Change type to hash_table. Update dependent calls and types.
19203 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19205 PR rtl-optimization/57003
19206 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19207 call note_stores with kill_clobbered_value callback again after
19208 killing regs_invalidated_by_call.
19210 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19212 * config/aarch64/aarch64-simd.md
19213 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19214 (aarch64_simd_bsl<mode>): Likewise.
19215 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19217 2013-04-25 Marek Polacek <polacek@redhat.com>
19219 PR tree-optimization/57066
19220 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19222 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19224 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19226 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19228 * config/aarch64/aarch64-builtins.c
19229 (aarch64_fold_builtin): New.
19230 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19231 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19232 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19233 * config/aarch64/arm_neon.h
19234 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19236 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19237 Tejas Belagod <tejas.belagod@arm.com>
19239 * config/aarch64/aarch64-builtins.c
19240 (aarch64_gimple_fold_builtin): New.
19241 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19242 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19243 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19244 (addvv4sf): Update.
19245 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19247 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19249 * config/aarch64/aarch64.md
19250 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19252 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19254 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19255 (*ngcsi_uxtw): New pattern.
19257 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19258 Julian Brown <julian@codesourcery.com>
19260 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19261 (TB_DREG): Add T_V4HF.
19262 (v4hf_UP): New macro.
19263 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19264 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19265 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19267 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19268 (arm_vector_mode_supported_p): Handle V4HF.
19269 (arm_mangle_map): Handle V4HFmode.
19270 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19271 * config/arm/arm_neon_builtins.def: Add entries for
19272 vcvtv4hfv4sf, vcvtv4sfv4hf.
19273 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19274 (neon_vcvtv4hfv4sf): Likewise.
19275 * config/arm/neon-gen.ml: Handle half-precision floating point
19277 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19278 * config/arm/arm_neon.h: Regenerate.
19279 * config/arm/neon.ml (type elts): Add F16.
19280 (type vectype): Add T_float16x4, T_floatHF.
19281 (type vecmode): Add V4HF.
19282 (type features): Add Requires_FP_bit feature.
19283 (elt_width): Handle F16.
19284 (elt_class): Likewise.
19285 (elt_of_class_width): Likewise.
19286 (mode_of_elt): Refactor.
19287 (type_for_elt): Handle F16, fix error messages.
19288 (vectype_size): Handle T_float16x4.
19289 (vcvt_sh): New function.
19290 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19291 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19292 (string_of_mode): Handle V4HF.
19293 * doc/arm-neon-intrinsics.texi: Regenerate.
19295 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19297 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19298 format specifier in 'X' case.
19300 2013-04-25 Alan Modra <amodra@gmail.com>
19303 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19304 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19305 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19306 Repeat for many other rotate/shift and mask patterns using subregs.
19307 Name lshiftrt insns.
19308 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19309 on WORDS_BIG_ENDIAN.
19311 2013-04-25 Alan Modra <amodra@gmail.com>
19313 * config.gcc: Support little-endian powerpc-linux targets.
19314 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19315 (LINK_OS_LINUX_SPEC): Define.
19316 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19317 Preserve MASK_LITTLE_ENDIAN.
19318 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19319 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19320 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19321 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19322 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19323 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19324 Correct fp word order for little-endian. Don't shift toc entries
19325 smaller than a word for little-endian.
19326 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19327 (bswapdi2 splits): Correct low-part subreg for little-endian.
19328 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19329 low/high where such is correct only for be.
19330 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19331 little-endian for -mcall-aixdesc.
19333 2013-04-25 Alan Modra <amodra@gmail.com>
19335 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19336 replace_equiv_address_nv.
19338 2013-04-25 Alan Modra <amodra@gmail.com>
19340 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19342 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19345 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19346 * rtl.h (struct rtx_def): ...
19348 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19350 PR rtl-optimizations/57046
19351 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19352 for multi-reg splits.
19354 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19356 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19358 2013-04-24 Sterling Augustine <saugustine@google.com>
19360 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19361 (comp_dir_string, debug_str_dwo_section): New.
19362 (DEBUG_STR_DWO_SECTION): Rename to ...
19363 (DEBUG_DWO_STR_SECTION): ... this.
19364 (DEBUG_NORM_STR_SECTION): Delete.
19365 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19366 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19367 (find_AT_string): Move most logic to ...
19368 (find_AT_string_in_table): ... here. New.
19369 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19370 add_skeleton_AT_string. Delete logic.
19371 (output_skeleton_debug_sections): Remove call to
19372 add_top_level_skeleton_die_attrs.
19373 (add_comp_dir_attribute): Move logic to comp_dir_string.
19374 (dwarf2out_init): Initialize debug_str_dwo_section.
19375 (output_indirect_string): Call find_string_form.
19376 (output_indirect_strings): Rewrite.
19377 (prune_unused_types): Empty skeleton_debug_str_hash.
19378 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19379 (dwarf2out_finish): Call output_indirect_strings.
19381 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19383 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19385 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19387 * rtl.h (struct rtx_def): Add comment for field jump.
19388 (LRA_SUBREG_P): New macro.
19389 * recog.c (register_operand): Check LRA_SUBREG_P.
19390 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19392 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19393 (lra_final_code_change): Skip subreg change for operators.
19394 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19395 if there are no operand changes.
19396 * lra-constraints.c (curr_insn_set): New.
19397 (match_reload): Set LRA_SUBREG_P.
19398 (emit_spill_move): Ditto.
19399 (check_and_process_move): Use curr_insn_set. Process only single
19400 set insns. Don't initialize sec_mem_p and change_p.
19401 (simplify_operand_subreg): Use LRA_SUBREG_P.
19402 (reg_in_class_p): New function.
19403 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19404 of #ifdef. Add code to remove cycling.
19405 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19406 non-null disp. Reload inner instead of disp when base and index
19407 are null. Try to put lo_sum into register.
19408 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19409 (check_and_process_move): Move code for move cost check to
19410 simple_move_p. Remove equiv_substitution.
19411 (simple_move_p): New function.
19412 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19413 curr_insn_set. Call check_and_process_move only for single set
19414 insns. Use the new function. Move call of check_and_process_move
19415 after operand equiv substitution and address process.
19417 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19419 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19420 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19421 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19423 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19425 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19427 2013-04-24 Marek Polacek <polacek@redhat.com>
19429 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19430 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19431 (select_loops_exit_conditions): Likewise.
19432 (number_of_iterations_for_all_loops): Likewise.
19433 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19434 (scev_analysis): Likewise.
19436 2013-04-02 Catherine Moore <clm@codesourcery.com>
19437 Chao-ying Fu <fu@mips.com>
19439 * config/mips/micromips.md (jraddiusp): New pattern.
19440 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19441 instruction if possible.
19443 2013-04-24 Alan Modra <amodra@gmail.com>
19445 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19447 2013-04-24 Julian Brown <julian@codesourcery.com>
19448 Chung-Lin Tang <cltang@codesourcery.com>
19450 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19451 dependency behavior in enumeration type DIE generation. Add TODO note
19452 to comments about future DW_FORM_sdata/udata re-work of related code.
19454 2013-04-23 Lawrence Crowl <crowl@google.com>
19456 * Makefile.in: Update as needed below.
19458 * hash-table.h (class hash_table):
19459 Correct many methods with parameter types compare_type to the correct
19460 value_type. (Correct code was unlikely to notice the change.)
19461 (hash_table::elements_with_deleted) New.
19462 (class hashtable::iterator): New.
19463 (hashtable::begin()): New.
19464 (hashtable::end()): New.
19465 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19467 * statistics.c (statistics_hashes):
19468 Change type to hash_table. Update dependent calls and types.
19470 * tree-into-ssa.c (var_infos):
19471 Change type to hash_table. Update dependent calls and types.
19473 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19474 Change type to hash_table. Update dependent calls and types.
19476 * tree-ssa-loop-im.c (struct mem_ref.refs):
19477 Change type to hash_table. Update dependent calls and types.
19479 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19480 Change type to hash_table. Update dependent calls and types.
19482 * tree-ssa-sccvn.c (vn_tables_s::nary):
19483 Change type to hash_table. Update dependent calls and types.
19484 (vn_tables_s::phis): Likewise.
19485 (vn_tables_s::references): Likewise.
19487 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19488 (vn_reference_eq): Update parameter and return types.
19490 * tree-ssa-structalias.c (pointer_equiv_class_table):
19491 Change type to hash_table. Update dependent calls and types.
19492 (location_equiv_class_table): Likewise.
19494 * tree-vect-data-refs.c: Consequential changes for making
19495 peeling a hash_table.
19497 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19498 (destroy_loop_vec_info): Dependent hash_table update.
19500 * tree-vectorizer.h (peeling_htab):
19501 Change type to hash_table. Update dependent calls and types.
19503 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19505 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19506 to check the register content is equal or not.
19507 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19508 to assign register content record.
19509 * lra-eliminations.c (update_reg_eliminate): Use
19510 lra_update_reg_val_offset to update register content offset.
19511 * lra-int.h (struct lra_reg): Add offset member.
19512 (lra_reg_val_equal_p): New static inline function.
19513 (lra_update_reg_val_offset): New static inline function.
19514 (lra_assign_reg_val): New static inline function.
19515 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19516 to assign register content record.
19517 (initialize_lra_reg_info_element): Initial offset to zero.
19519 2013-04-23 Catherine Moore <clm@codesourcery.com>
19521 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19522 operands. Record compression.
19524 2013-04-23 Xinliang David Li <davidxl@google.com>
19526 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19528 2013-04-23 Richard Biener <rguenther@suse.de>
19530 PR middle-end/57036
19531 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19532 parameter, only add abnormal goto edges from the copied body
19533 if the call could perform abnormal gotos.
19534 (copy_cfg_body): Adjust.
19536 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19538 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19540 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19542 * coretypes.h (gimple_stmt_iterator): Add struct to make
19545 2013-04-23 Richard Biener <rguenther@suse.de>
19547 PR tree-optimization/57026
19548 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19549 from SSA names occuring in abnormal PHI nodes.
19551 2013-04-22 Andi Kleen <ak@linux.intel.com>
19553 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19555 2013-04-22 Andi Kleen <ak@linux.intel.com>
19557 * lto/lto.c (print_lto_report_1): Declare early.
19558 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19560 2013-04-22 Andi Kleen <ak@linux.intel.com>
19562 * common.opt (-flto-report-wpa): Add.
19563 * doc/invoke.texi (-flto-report-wpa): Add.
19564 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19567 2013-04-22 Xinliang David Li <davidxl@google.com>
19569 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19570 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19571 * Makefile.in: New dependency
19573 David Daney <ddaney.cavm@gmail.com>
19575 * configure.ac (gcc_cv_as_micromips_support): Use the
19576 --fatal-warnings option.
19577 * configure: Regenerate.
19579 2013-04-22 Marek Polacek <polacek@redhat.com>
19582 * tsan.c (instrument_expr): Don't instrument expression
19583 in case its size is zero.
19585 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19589 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19591 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19593 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19595 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19596 (gimple_stmt_iterator): New typedef.
19597 * gimple.h (gimple_stmt_iterator): Rename to...
19598 (gimple_stmt_iterator_d): ... This.
19599 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19600 trees be valid for GIMPLE and GENERIC.
19601 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19602 * gimple-fold.c (gimple_fold_call): Call target hook
19603 gimple_fold_builtin.
19604 * hooks.c (hook_bool_gsiptr_false): New.
19605 * hooks.h (hook_bool_gsiptr_false): New.
19606 * target.def (fold_stmt): New.
19607 * doc/tm.texi: Regenerate.
19609 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19612 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19613 a set sp if no stack realignment.
19615 2013-04-22 Nick Clifton <nickc@redhat.com>
19617 * config.gcc (tilegx-linux): Extend extra_objs rather than
19619 (tilepro-linux): Likewise.
19621 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19623 * config/aarch64/aarch64-builtins.c
19625 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19626 (VAR<1-12>): Add MAP parameter.
19627 (BUILTIN_*): Likewise.
19628 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19629 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19630 (aarch64_ushl_n<mode>): Likewise.
19631 (aarch64_sshr_n<mode>): Likewise.
19632 (aarch64_ushr_n<mode>): Likewise.
19633 (aarch64_<maxmin><mode>): Likewise.
19634 (aarch64_sqrt<mode>): Likewise.
19635 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19636 (vshr<q>_n_*): Likewise.
19638 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19640 * config/aarch64/aarch64-builtins.c
19641 (aarch64_simd_builtin_type_mode): Handle SF types.
19643 (BUILTIN_GPF): Define.
19644 (aarch64_init_simd_builtins): Handle SF types.
19645 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19646 (frecps): Likewise.
19647 (frecpx): Likewise.
19648 * config/aarch64/aarch64-simd.md
19649 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19650 (aarch64_frecpe<mode>): New.
19651 (aarch64_frecps<mode>): Likewise.
19652 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19653 (v8type): Add frecp<esx>.
19654 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19655 (aarch64_frecps<mode>): Likewise.
19656 * config/aarch64/iterators.md (FRECP): New.
19657 (frecp_suffix): Likewise.
19658 * config/aarch64/arm_neon.h
19659 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19661 2013-04-22 Christian Bruel <christian.bruel@st.com>
19664 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19665 (REG_CLASS_NAMES): Idem.
19666 (REG_CLASS_CONTENTS): Idem.
19667 (REGCLASS_HAS_FP_REG): Idem.
19668 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19669 (sh_conditional_register_usage): Idem.
19671 2013-04-21 Jeff Law <law@redhat.com>
19673 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19674 (ssa_forward_propagate_and_combine): Use it.
19676 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19678 * lra.c: Update the flow chart diagram.
19680 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19682 PR rtl-optimization/56847
19683 * lra-constraints.c (process_alt_operands): Discourage alternative
19684 with non-matche doffsettable memory constraint fro memory with
19687 2013-04-19 Richard Biener <rguenther@suse.de>
19689 PR tree-optimization/56982
19690 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19692 * gimplify.c (gimplify_call_expr): Notice special calls.
19693 (gimplify_modify_expr): Likewise.
19694 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19695 abnormal control flow receivers.
19696 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19697 in the same way as cfun->has_nonlocal_labels.
19698 (gimple_purge_dead_abnormal_call_edges): Likewise.
19699 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19700 receivers start a basic-block.
19702 2013-04-19 Richard Biener <rguenther@suse.de>
19704 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19706 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19707 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19708 (SLP_TREE_LOAD_PERMUTATION): Add.
19709 (vect_transform_slp_perm_load): Adjust prototype.
19710 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19711 (vect_free_slp_instance): Likewise.
19712 (vect_create_new_slp_node): Likewise.
19713 (vect_supported_slp_permutation_p): Remove.
19714 (vect_slp_rearrange_stmts): Adjust.
19715 (vect_supported_load_permutation_p): Likewise. Inline
19716 vect_supported_slp_permutation_p here.
19717 (vect_analyze_slp_instance): Compute load permutations per
19718 slp node instead of per instance.
19719 (vect_get_slp_defs): Adjust.
19720 (vect_transform_slp_perm_load): Likewise.
19721 (vect_schedule_slp_instance): Remove redundant code.
19722 (vect_schedule_slp): Remove hack for PR56270, add it ...
19723 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19724 CSE loads for SLP. Adjust.
19726 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19728 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19729 spelling in two comments.
19731 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19734 * config/arm/arm.c (load_multiple_sequence): Require SP
19735 as base register for loads if SP is in the register list.
19737 2013-04-19 Martin Jambor <mjambor@suse.cz>
19739 PR tree-optimization/56718
19740 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19741 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19742 and made public. Adjusted all callers.
19743 (ipa_intraprocedural_devirtualization): New function.
19744 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19745 (ipa_intraprocedural_devirtualization): Likewise.
19746 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19748 2013-04-19 Richard Biener <rguenther@suse.de>
19750 PR tree-optimization/57000
19751 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19753 2013-04-19 Terry Guo <terry.guo@arm.com>
19755 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19757 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19758 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19759 (cortex_m4_fmacs): Use new reservations.
19760 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19762 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19764 PR rtl-optimization/56999
19765 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19767 (lra_coalesce): Remove split_origin_bitmap and related code.
19768 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19769 ranges if necessary.
19771 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19773 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19775 (ix86_expand_call): Remove clobbered_registers array and use
19776 x86_64_ms_sysv_extra_clobbered_registers instead.
19777 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19779 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19781 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19782 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19783 (*call_value_rex64_ms_sysv): Ditto.
19785 2013-04-18 Cary Coutant <ccoutant@google.com>
19787 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19788 parent instead of die_mark.
19790 2013-04-18 Diego Novillo <dnovillo@google.com>
19792 * gimple.c (create_gimple_tmp): New.
19793 (get_expr_type): New.
19794 (build_assign): New.
19795 (build_type_cast): New.
19796 * gimple.h (enum ssa_mode): Define.
19797 (gimple_seq_set_location): New.
19798 * asan.c (build_check_stmt): Change some gimple_build_* calls
19799 to use build_assign and build_type_cast.
19801 2013-04-18 Richard Biener <rguenther@suse.de>
19803 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19804 handle negative step. Remove redundant checks.
19805 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19806 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19807 for negative step and grouped loads fail to vectorize.
19809 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19811 * emit-rtl.c (reset_insn_used_flags): New function.
19812 (reset_all_used_flags): Use it.
19813 (verify_insn_sharing): New function.
19814 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19816 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19818 PR tree-optimization/56984
19819 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19820 and (x >> M) >= N don't register any assertion if N << M is the
19823 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19825 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19826 do not delete_insn insns that have not yet been emitted, only
19827 unlink them with remove_insn.
19828 * df-scan.c (df_insn_delete): Revert r197492.
19830 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19832 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19833 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19835 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19837 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19838 define_insn_and_split.
19839 (and_scc,ior_scc,negscc): Likewise.
19840 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19842 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19844 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19845 can benefit from using a sequence of LDRD instructions in epilogue
19846 instead of a single LDM instruction.
19848 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19851 * doc/extend.texi: Fix typo.
19853 2013-04-17 Richard Biener <rguenther@suse.de>
19855 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19856 (vect_build_slp_tree): ... here.
19857 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19858 match. Remove special-casing of mismatched complex loads.
19859 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19860 re-try the match with swapped commutative operands.
19861 (vect_supported_load_permutation_p): Remove special-casing of
19862 mismatched complex loads.
19863 (vect_analyze_slp_instance): Adjust.
19865 2013-04-17 Richard Biener <rguenther@suse.de>
19867 PR rtl-optimization/56921
19868 * cfgloop.h (struct loop): Add simple_loop_desc member.
19869 (struct niter_desc): Mark with GTY(()).
19870 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19871 * loop-iv.c (get_simple_loop_desc): Likewise.
19872 (free_simple_loop_desc): Likewise.
19875 2013-04-16 Richard Biener <rguenther@suse.de>
19877 PR rtl-optimization/56921
19878 * loop-init.c (pass_rtl_move_loop_invariants): Add
19879 TODO_do_not_ggc_collect to todo_flags_finish.
19880 (pass_rtl_unswitch): Same.
19881 (pass_rtl_unroll_and_peel_loops): Same.
19882 (pass_rtl_doloop): Same.
19884 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19886 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19887 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19888 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19890 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19891 * tree-streamer.c (record_common_node): Adjust reference in comment.
19893 2013-04-17 Terry Guo <terry.guo@arm.com>
19895 * config/arm/cortex-m4.md: Add a new bypass.
19897 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19899 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19901 (*subs_<optab><mode>_multp2): New pattern.
19902 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19903 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19905 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19907 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19908 (*subs_mul_imm_<mode>): New pattern.
19910 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19913 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19914 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19915 (vsx_movti_32bit): Change j->wa to O->wa.
19917 2013-04-16 Richard Biener <rguenther@suse.de>
19919 PR rtl-optimization/56921
19920 * loop-init.c (pass_rtl_move_loop_invariants): Add
19921 TODO_do_not_ggc_collect to todo_flags_finish.
19922 (pass_rtl_unswitch): Same.
19923 (pass_rtl_unroll_and_peel_loops): Same.
19924 (pass_rtl_doloop): Same.
19926 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19928 * config/arm/arm.c (emit_multi_reg_push): New declaration
19929 for an existing function.
19930 (arm_emit_strd_push): New function.
19931 (arm_expand_prologue): Used here.
19932 (arm_emit_ldrd_pop): New function.
19933 (arm_expand_epilogue): Used here.
19934 (arm_get_frame_offsets): Update condition.
19935 (arm_emit_multi_reg_pop): Add a special case for load of a single
19936 register with writeback.
19938 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19940 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19943 2013-04-16 Richard Biener <rguenther@suse.de>
19945 PR tree-optimization/56756
19946 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19947 (first_mem_ref_loc): New.
19948 (execute_sm): Place the load temporarily before a previous
19949 access instead of in the latch edge to ensure its SSA dependencies
19950 are defined at points dominating the load.
19952 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19954 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19955 correct fix by moving header and footer insn to the footer of
19956 the merged basic block. Clear BB_END of the merged-away block.
19958 PR middle-end/43631
19959 * emit-rtl.c (make_note_raw): New function.
19960 (link_insn_into_chain): New static inline function.
19961 (add_insn): Use it.
19962 (add_insn_before, add_insn_after): Factor insn chain linking code...
19963 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19964 using link_insn_into_chain.
19965 (note_outside_basic_block_p): New helper function for emit_note_after
19966 and emit_note_before.
19967 (emit_note_after): Use nobb variant of add_insn_after if the note
19968 should not be contained in a basic block.
19969 (emit_note_before): Use nobb variant of add_insn_before if the note
19970 should not be contained in a basic block.
19971 (emit_note_copy): Use make_note_raw.
19972 (emit_note): Likewise.
19973 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19974 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19975 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19976 the moved barrier the tail of the basic block it follows.
19977 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19979 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19981 PR tree-optimization/56962
19982 * gimple-ssa-strength-reduction.c (record_increment): Only set
19983 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19984 either rhs1 or rhs2 is equal to c->base_expr.
19986 2013-04-15 Richard Biener <rguenther@suse.de>
19988 PR tree-optimization/56933
19989 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19991 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19992 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19993 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19994 dependence check ...
19995 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19997 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19998 GROUP_READ_WRITE_DEPENDENCE.
20000 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20002 * emit-rtl.c (reset_all_used_flags): New function.
20003 (verify_rtl_sharing): Call reset_all_used_flags before and after
20004 performing the checks.
20006 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20008 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20009 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20010 * config/arm/constraints.md (De): New constraint.
20011 * config/arm/neon.md (anddi3_neon): Delete.
20012 (neon_vand<mode>): Expand to standard anddi3 pattern.
20013 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20014 Move earlier in the file.
20015 (neon_inv_logic_op2): Likewise.
20016 (arm_anddi_operand_neon): New predicate.
20018 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20020 * configure.ac (gcc_cv_ld_as_needed): Set
20021 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20022 Use -z ignore, -z record on *-*-solaris2*.
20023 (HAVE_LD_AS_NEEDED): Update comment.
20024 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20025 * configure: Regenerate.
20026 * config.in: Regenerate.
20027 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20028 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20029 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20030 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20031 equivalents. Fix markup.
20032 * doc/tm.texi: Regenerate.
20034 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20036 * config/i386/i386.opt: New option mstack-protector-guard=.
20037 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20038 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20039 TARGET_SSP_TLS_GUARD.
20040 * config/i386/i386.c (ix86_option_override_internal): Set
20041 ix86_stack_protector_guard.
20042 * config/i386/i386.md (stack_protect_set): Enable for
20043 TARGET_SSP_TLS_GUARD only.
20044 (stack_protect_set_<mode>): Ditto.
20045 (stack_protect_test): Ditto.
20046 (stack_protect_test_<mode>): Ditto.
20047 * doc/invoke.texi (i386 Option): Document.
20049 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20052 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20053 (S_MODES): Set H_MODE bit.
20054 (SF_MODES): Set only S_MODE and SF_MODE bits.
20055 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20056 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20057 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20058 <MODE_FLOAT>: Likewise.
20060 2013-04-15 Joey Ye <joey.ye@arm.com>
20062 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20064 2013-04-15 Joey Ye <joey.ye@arm.com>
20066 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20068 (thumb_far_jump_used_p): Count instruction size and set
20071 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20073 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20074 * resource.c (mark_target_live_regs): Reformat conditional block.
20076 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20078 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20079 notes, they are emitted only just before final.
20080 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20082 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20084 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20085 * cfgrtl.c (delete_insn): Call it here instead.
20086 * lra-spills.c (lra_final_code_change): Use delete_insn.
20087 * haifa-sched.c (sched_remove_insn): Likewise.
20088 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20089 returning to the nop pool.
20090 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20091 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20093 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20095 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20096 * doc/tm.texi: Regenerated.
20098 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20100 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20103 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20105 * df-core.c (df_find_def): Compare register numbers.
20106 (df_find_use): Likewise.
20108 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20111 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20112 lra_in_progress for return.
20114 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20116 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20117 define_insn into define_insn_and_split and emit movsicc patterns.
20119 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20121 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20123 2013-04-12 Richard Biener <rguenther@suse.de>
20125 * tree-pass.h (TODO_do_not_ggc_collect): New.
20126 * passes.c (execute_one_ipa_transform_pass): Honor
20127 TODO_do_not_ggc_collect.
20128 (execute_one_pass): Likewise.
20131 2013-04-10 Richard Biener <rguenther@suse.de>
20133 * passes.c (init_optimization_passes): Remove reload pass.
20134 * ira.c (do_reload): Merge into ...
20136 (rest_of_handle_reload): Remove.
20137 (pass_reload): Likewise.
20138 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20139 of reload for vzeroupper pass placement.
20141 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20143 PR tree-optimization/56918
20144 PR tree-optimization/56920
20145 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20146 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20147 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20148 use wide_mul_with_sign method.
20150 2013-04-12 Richard Biener <rguenther@suse.de>
20152 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20153 not be considered a gimple constant.
20155 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20157 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20158 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20161 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20162 Jakub Jelinek <jakub@redhat.com>
20164 * opts.c: Include diagnostic-color.h.
20165 (common_handle_option): Handle OPT_fdiagnostics_color_.
20166 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20167 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20168 (diagnostic-color.o): New.
20169 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20170 (diagnostic_color_rule): New enum.
20171 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20172 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20173 the location string.
20174 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20175 either NULL, or color kind.
20176 * diagnostic-color.c: New file.
20177 * diagnostic-color.h: New file.
20178 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20180 * doc/invoke.texi (-fdiagnostics-color): Document.
20181 * pretty-print.h (pp_show_color): Define.
20182 (struct pretty_print_info): Add show_color field.
20183 * diagnostic.c: Include diagnostic-color.h.
20184 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20185 macros. Colorize error:, warning: etc. strings and also the location
20187 (diagnostic_show_locus): Colorize the caret line.
20188 * pretty-print.c: Include diagnostic-color.h.
20189 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20190 inside of %< %> quotes or quoted through q format modifier.
20192 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20194 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20196 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20198 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20199 code in CC_NZ mode.
20200 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20203 2013-04-11 Marek Polacek <polacek@redhat.com>
20205 PR tree-optimization/48184
20206 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20208 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20210 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20211 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20212 (skip_simple_arithmetic): Tidy up.
20213 * tree.h (skip_simple_constant_arithmetic): Declare.
20215 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20217 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20219 2013-04-11 Richard Biener <rguenther@suse.de>
20221 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20222 generate vector constants.
20224 2013-04-11 Richard Biener <rguenther@suse.de>
20226 PR tree-optimization/56878
20227 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20228 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20230 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20231 Prefer to align the DR with the most invariant base address.
20233 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20235 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20238 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20240 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20241 floating-point vector comparisons against 0.
20243 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20245 PR tree-optimization/56899
20246 * fold-const.c (extract_muldiv_1): Apply distributive law
20247 only if TYPE_OVERFLOW_WRAPS (ctype).
20249 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20252 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20253 loading has side effect.
20255 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20257 * config/sparc/sparc.c: Include tree-pass.h.
20258 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20259 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20260 head of file. Change return type. Split off gate function.
20261 (sparc_gate_work_around_errata): New function.
20262 (pass_work_around_errata): New pass definition.
20263 (insert_pass_work_around_errata) New pass insert definition to
20264 insert pass_work_around_errata just after delayed-branch scheduling.
20265 (sparc_option_override): Insert the pass.
20266 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20268 2013-04-10 David S. Miller <davem@davemloft.net>
20270 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20271 or -mcpu=hypersparc.
20273 * target.def (cstore_mode): New hook.
20274 * target.h: Include insn-codes.h
20275 * targhooks.c: Likewise.
20276 (default_cstore_mode): New function.
20277 * targhooks.h: Declare it.
20278 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20279 * doc/tm.texi: Rebuild.
20280 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20281 target hook, rather than inspecting the insn_data.
20282 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20283 (TARGET_CSTORE_MODE): Redefine.
20284 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20286 * config/sparc/predicates.md (cstore_result_operand): New special
20288 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20289 Use it for operand 0.
20290 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20291 (*snesi_special): Likewise.
20292 (*snesi_zero): Likewise.
20293 (*seqsi_zero): Likewise.
20294 (*sltu_insn): Likewise.
20295 (*sgeu_insn): Likewise.
20296 (*seqdi_special): Make operand 0 and comparison operation be of
20298 (*snedi_special): Likewise.
20299 (*snedi_special_vis3): Likewise.
20300 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20301 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20302 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20303 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20304 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20305 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20306 (*sltu_extend_sp64): Likewise.
20307 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20308 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20309 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20310 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20311 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20313 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20315 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20316 (aarch64_start_file): Use the new function.
20318 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20319 Jason Merrill <jason@redhat.com>
20321 * common.opt: Add -gdwarf.
20322 * opts.c (common_handle_option): Handle it.
20323 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20325 2013-04-10 Richard Biener <rguenther@suse.de>
20327 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20328 (execute_one_ipa_transform_pass): But unconditionally here.
20329 (execute_one_pass): And here.
20330 (init_optimization_passes): Remove reload pass.
20331 * tree-pass.h (TODO_ggc_collect): Remove.
20332 (pass_reload): Likewise.
20333 * ira.c (do_reload): Merge into ...
20335 (rest_of_handle_reload): Remove.
20336 (pass_reload): Likewise.
20337 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20338 of reload for vzeroupper pass placement.
20339 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20340 and todo_flags_finish of all passes.
20342 2013-04-10 Richard Biener <rguenther@suse.de>
20344 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20345 first_const_oprnd field, rename first_def_type to first_op_type.
20346 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20347 (vect_get_and_check_slp_defs): Always use the type of the
20348 operand. Allow mixed vect_external_def, vect_constant_def types.
20349 (vect_get_constant_vectors): Handle mixed vect_external_def,
20350 vect_constant_def types.
20352 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20354 PR tree-optimization/55524
20355 * tree-ssa-math-opts.c
20356 (convert_mult_to_fma): Don't use an fms construct
20357 when we don't have an fms operation, but fnma, and it looks
20358 likely that we'll be able to use the latter.
20360 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20362 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20364 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20365 inline fail caused by overwritable functions.
20367 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20369 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20370 unnecessary bits in the constant power of two case.
20372 2013-04-10 Richard Biener <rguenther@suse.de>
20374 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20375 broken code swapping operands.
20376 (vect_build_slp_tree): Do not compute load permutations here.
20377 (vect_analyze_slp_instance): Compute load permutations here,
20378 after building the SLP tree.
20380 2013-04-09 Christian Bruel <christian.bruel@st.com>
20382 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20383 of next/prev_real_insn.
20385 2013-04-09 Jan Hubicka <jh@suse.cz>
20387 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20388 Drop aliased parameter.
20389 (function_and_variable_visibility): Do not handle alias pairs.
20390 * cgraph.c (varpool_externally_visible_p): Update prototype.
20391 * varpool.c (varpool_add_new_variable): Update.
20393 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20395 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20397 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20399 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20401 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20403 2013-04-09 Marek Polacek <polacek@redhat.com>
20405 PR tree-optimization/48762
20406 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20408 2013-04-09 Richard Biener <rguenther@suse.de>
20410 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20412 (vect_build_slp_tree): Likewise.
20413 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20414 calculating the cost of a SLP instance.
20415 (vect_analyze_slp_instance): Use it from here, after building
20418 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20420 PR middle-end/56883
20421 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20422 expand_omp_for_static_chunk): Use simple_p = true in
20423 force_gimple_operand_gsi calls when assigning to addressable decls.
20425 2013-04-09 Jeff Law <law@redhat.com>
20427 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20428 when the boolean was created by converting a wider object which
20429 had a boolean range.
20431 2013-04-09 Richard Biener <rguenther@suse.de>
20433 * tree-vectorizer.h (slp_void_p): Remove.
20434 (slp_tree): Typedef before _slp_tree declaration.
20435 (struct _slp_tree): Use a vector of slp_tree as children.
20436 (vect_get_place_in_interleaving_chain): Remove.
20437 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20439 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20441 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20442 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20443 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20444 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20445 Use slp_node instead of slp_void_p and adjust.
20447 2013-04-09 Richard Biener <rguenther@suse.de>
20449 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20450 work that is not necessary.
20452 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20454 PR tree-optimization/56854
20455 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20456 forward into clobber stmts if it would change MEM_REF lhs into
20459 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20461 * tree.c (type_hash_lookup, type_hash_add): Make static.
20462 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20464 2013-04-09 Richard Biener <rguenther@suse.de>
20466 * tree.h (unsave_expr_now): Remove.
20467 * tree-inline.c (mark_local_for_remap_r): Remove.
20468 (unsave_expr_1): Likewise.
20469 (unsave_r): Likewise.
20470 (unsave_expr_now): Likewise.
20471 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20472 (propagate_tree_value): Likewise.
20474 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20476 * doc/rtl.texi (sequence): Rewrite documentation to match the
20477 current use of SEQUENCE rtl objects.
20478 * rtl.def (SEQUENCE): Likewise.
20480 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20481 Update documentation.
20482 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20483 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20485 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20487 2013-04-08 Teresa Johnson <tejohnson@google.com>
20489 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20490 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20492 (estimate_edge_size_and_time): Add comment to suggest using rounding
20494 (estimate_node_size_and_time): Ditto.
20495 (remap_edge_change_prob): Use helper rounding divide methods.
20496 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20497 (gimple_mod_pow2_value_transform): Ditto.
20498 (gimple_mod_subtract_transform): Ditto.
20499 (gimple_ic_transform): Ditto.
20500 (gimple_stringops_transform): Ditto.
20501 * stmt.c (conditional_probability): Ditto.
20502 (emit_case_dispatch_table): Ditto.
20503 * lto-cgraph.c (merge_profile_summaries): Ditto.
20504 * tree-optimize.c (execute_fixup_cfg): Ditto.
20505 * cfgcleanup.c (try_forward_edges): Ditto.
20506 * cfgloopmanip.c (scale_loop_profile): Ditto.
20508 (duplicate_loop_to_header_edge): Ditto.
20509 (lv_adjust_loop_entry_edge): Ditto.
20510 * tree-vect-loop.c (vect_transform_loop): Ditto.
20511 * profile.c (compute_branch_probabilities): Ditto.
20512 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20513 * lto-streamer-in.c (input_cfg): Ditto.
20514 * gimple-streamer-in.c (input_bb): Ditto.
20515 * ipa-cp.c (update_profiling_info): Ditto.
20516 (update_specialized_profile): Ditto.
20517 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20518 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20520 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20521 (compute_trg_info): Ditto.
20522 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20523 (purge_dead_edges): Ditto.
20524 * loop-unswitch.c (unswitch_loop): Ditto.
20525 * cgraphclones.c (cgraph_clone_edge): Ditto.
20526 (cgraph_clone_node): Ditto.
20527 * tree-inline.c (copy_bb): Ditto.
20528 (copy_edges_for_bb): Ditto.
20529 (initialize_cfun): Ditto.
20530 (copy_cfg_body): Ditto.
20531 (expand_call_inline): Ditto.
20533 2013-04-08 Kai Tietz <ktietz@redhat.com>
20535 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20536 TARGET_CYGWIN64 by TARGET_64BIT.
20538 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20540 * config/epiphany/epiphany.md (GPR_1): New constant.
20541 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20542 * config/epiphany/epiphany.c (gen_compare_reg):
20543 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20544 is already in place.
20545 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20546 Don't require being called during rtl expansion; If y operlaps r0,
20548 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20549 (epiphany_expand_epilogue): Likewise.
20551 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20552 Don't use CC_FPmode for ORDERED / UNORDERED.
20553 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20555 * config/epiphany/constraints.md (CnL): New constraint.
20556 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20557 * config/epiphany/predicates.md (add_operand): Allow 1024.
20559 * config/epiphany/epiphany.md (logical_op): New code iterator.
20560 (op_mnc): New code attribute.
20561 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20562 (mov_f+1, mov_f+2): New peephole2 patterns.
20564 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20565 (cstoresi4): Also allow re-use of zero result when doing a NE
20566 comparison to a non-zero operand.
20567 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20569 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20570 Use gen_addsi3_i / gen_subsi3_i.
20572 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20576 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20577 contain anything but clobbers, at most one __builtin_stack_restore,
20578 optionally debug stmts and final resx, and if it has at least one
20579 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20580 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20581 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20582 which isn't defaut definition, remove them.
20583 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20584 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20585 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20586 with MEM_REF LHS with SSA_NAME address.
20588 2013-04-08 Jeff Law <law@redhat.com>
20590 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20592 2013-04-08 Richard Biener <rguenther@suse.de>
20594 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20596 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20597 determined vector type.
20598 (vect_analyze_data_refs): Likewise.
20599 (vect_get_new_vect_var): Adjust.
20600 (vect_create_destination_var): Preserve SSA name versions.
20601 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20602 not dump anything here.
20604 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20606 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20607 Add member lr_slot_known.
20608 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20610 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20611 Remove code that sets lr_slot_offset according to what a previous
20612 version of epiphany_emit_save_restore used to do.
20613 (epiphany_emit_save_restore): When doing an lr save or restore,
20614 set/verify lr_slot_known and lr_slot_offset.
20616 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20619 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20622 2013-04-08 Richard Biener <rguenther@suse.de>
20624 * alias.c (find_base_term): Fix thinko in previous change.
20626 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20628 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20629 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20630 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20631 if possible to compute val.
20632 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20633 For QImode integers don't require anything about precision. Use
20634 const_with_all_bytes_same to find out if the constant doesn't have
20635 repeated bytes in it.
20637 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20639 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20642 2013-04-08 Marek Polacek <polacek@redhat.com>
20644 PR rtl-optimization/48182
20645 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20648 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20651 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20652 nuses, make sure we have a label.
20654 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20657 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20658 (rs6000_emit_swdiv_low_precision): Remove.
20659 (rs6000_emit_swdiv): Rewrite to handle between one and four
20660 iterations of Newton-Raphson generally; modify required number of
20661 iterations for some cases.
20662 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20664 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20666 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20667 set-but-unused variable.
20669 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20670 basic blocks of released function bodies garbage-collectable.
20672 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20673 (struct rtl_opt_pass): Add TODO_df_finish.
20675 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20677 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20679 * config/arm/constraints.md (q): New constraint.
20680 * config/arm/ldrdstrd.md: New file.
20681 * config/arm/arm.md (ldrdstrd.md) New include.
20682 (arm_movdi): Use "q" instead of "r" constraint
20683 for double-word memory access.
20684 (movdf_soft_insn): Likewise.
20685 * config/arm/vfp.md (movdi_vfp): Likewise.
20686 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20687 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20688 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20689 (mem_ok_for_ldrd_strd): Likewise.
20690 (output_move_double): Update assertion.
20692 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20694 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20696 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20698 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20699 define_insn_and_split.
20700 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20702 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20704 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20705 define_insn_and_split.
20706 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20707 (shiftsi3_compare): New pattern.
20708 (rrx): New pattern.
20709 * config/arm/unspecs.md (UNSPEC_RRX): New.
20711 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20713 * config/arm/arm.md (negdi_extendsidi): New pattern.
20714 (negdi_zero_extendsidi): Likewise.
20716 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20718 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20719 define_insn_and_split.
20720 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20721 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20723 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20725 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20726 define_insn_and_split.
20727 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20728 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20730 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20732 * config/arm/arm.md (subsi3_carryin): New pattern.
20733 (subsi3_carryin_const): Likewise.
20734 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20735 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20737 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20739 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20741 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20743 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20744 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20746 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20748 * config/arm/arm.c (arm_expand_builtin): Change fcode
20749 type to unsigned int.
20751 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20753 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20755 2013-04-04 Ian Lance Taylor <iant@google.com>
20757 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20760 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20762 PR middle-end/56729
20763 * df-scan.c (df_insn_delete): Disable failing assert.
20765 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20767 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20768 New function prototype.
20769 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20770 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20771 (arm_builtin_vectorized_function): New function.
20773 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20775 * config/arm/arm_neon_builtins.def: New file.
20776 * config/arm/arm.c (neon_builtin_data): Move contents to
20777 arm_neon_builtins.def.
20778 (enum arm_builtins): Include neon builtin definitions.
20779 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20780 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20782 2013-04-04 Marek Polacek <polacek@redhat.com>
20784 PR tree-optimization/48186
20785 * predict.c (maybe_hot_frequency_p): Return false if
20786 HOT_BB_FREQUENCY_FRACTION is 0.
20787 (cgraph_maybe_hot_edge_p): Likewise.
20789 2013-04-04 Richard Biener <rguenther@suse.de>
20791 PR tree-optimization/56826
20792 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20795 2013-04-04 Richard Biener <rguenther@suse.de>
20797 PR tree-optimization/56213
20798 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20799 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20801 2013-04-04 Richard Biener <rguenther@suse.de>
20803 PR tree-optimization/56837
20804 * tree-loop-distribution.c (classify_partition): For non-zero
20805 values require that the value has the same precision as its
20806 mode to be useful as memset value.
20808 2013-04-03 Nick Clifton <nickc@redhat.com>
20810 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20811 (fmssf4): Use fmsf.s on E3V5 architectures.
20812 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20813 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20815 2013-04-03 Jeff Law <law@redhat.com>
20817 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20818 (lra-eliminations.o): Likewise.
20820 2013-04-03 Teresa Johnson <tejohnson@google.com>
20822 * gcov-io.c (compute_working_sets): Moved most of body of old
20823 compute_working_sets here from profile.c.
20824 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20825 (gcov_working_set_t): Moved typedef here from basic-block.h
20826 (compute_working_set): Declare.
20827 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20828 (get_working_sets): Renamed from compute_working_set,
20829 replace most of body with call to new compute_working_sets.
20830 (get_exec_counts): Replace call to compute_working_sets
20831 to get_working_sets.
20832 * profile.h (get_working_sets): Renamed from compute_working_set.
20833 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20834 to get_working_sets.
20835 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20836 * gcov-dump.c (dump_working_sets): New function.
20838 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20840 * hwint.c (sext_hwi, zext_hwi): New functions.
20841 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20842 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20843 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20844 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20845 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20846 (sext_hwi, zext_hwi): New functions.
20848 2013-04-03 Jeff Law <law@redhat.com>
20850 PR tree-optimization/56799
20851 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20852 back test for widening conversion erroneously dropped in prior change.
20854 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20857 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20858 instead of next_real_insn.
20860 2013-04-03 Marek Polacek <polacek@redhat.com>
20863 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20865 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20868 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20870 (thumb1_output_casesi): Likewise.
20871 (thumb2_output_casesi): Likewise.
20873 2013-04-03 Richard Biener <rguenther@suse.de>
20875 PR tree-optimization/56817
20876 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20878 (tree_unroll_loops_completely_1): ... new function to manually
20879 walk the loop tree, properly defering outer loops of unrolled
20880 loops to later iterations.
20882 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20884 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20885 (vectorizable_load): Likewise.
20886 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20887 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20889 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20891 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20894 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20896 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20898 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20900 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20902 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20904 PR tree-optimization/56790
20905 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20908 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20910 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20912 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20913 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20916 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20919 * tree.h (force_folding_builtin_constant_p): New decl.
20920 * builtins.c (force_folding_builtin_constant_p): New variable.
20921 (fold_builtin_constant_p): Fold immediately also if
20922 force_folding_builtin_constant_p.
20924 2013-04-03 Richard Biener <rguenther@suse.de>
20926 PR tree-optimization/56812
20927 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20928 DRs of the same interleaving chain are independent.
20930 2013-04-02 Jason Merrill <jason@redhat.com>
20932 * gdbinit.in (pbb): Use debug fn.
20934 2013-04-02 Lawrence Crowl <crowl@google.com>
20936 * sese.h (struct ivtype_map_elt_s): Remove unused.
20937 (extern debug_ivtype_map): Remove unused.
20938 (extern eq_ivtype_map_elts): Remove unused.
20939 * sese.c (debug_ivtype_map): Removed unused.
20940 (debug_ivtype_map_1): Removed unused.
20941 (debug_ivtype_elt): Remove unused.
20942 (eq_ivtype_map_elts): Remove unused.
20945 2013-04-02 Kai Tietz <ktietz@redhat.com>
20948 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20949 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20950 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20952 (legitimize_pe_coff_symbol): Likewise.
20953 (is_imported_p): New helper-function.
20954 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20955 for Windows x64 targets.
20956 (ix86_expand_prologue): Optimize for pe-coff targets.
20957 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20958 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20959 medium/large code-model.
20960 (legitimize_pic_address): Likewise.
20961 (legitimize_tls_address): Likewise.
20962 (ix86_expand_call): Likewise.
20963 (x86_output_mi_thunk): Likewise.
20964 (get_dllimport_decl): Add new beimport argument.
20965 (construct_plt_address): Don't assert for x64 pe-coff targets.
20966 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20968 (SYMBOL_FLAG_STUBVAR): New macro.
20969 (SYMBOL_REF_STUBVAR_P): Likewise.
20970 * config/i386/winnt.c (stub_list): New structure.
20971 (stub_head): New local variable.
20972 (i386_pe_record_stub): New function.
20973 (i386_pe_file_end): Emit refptr-stubs.
20975 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20977 PR rtl-optimization/56745
20978 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20979 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20982 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20983 and both of them are MEM_REFs, just compare first argument for
20984 equality and attempt to deal even with differing offsets.
20987 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20988 of gimple_clobber_p to be MEM_REF.
20989 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20990 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20991 after gimplification.
20992 * asan.c (get_mem_ref_of_assignment): Don't instrument
20993 gimple_clobber_p stmts.
20994 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20995 gimple_clobber_p stmt if they have MEM_REF lhs and
20996 are dead because of another gimple_clobber_p stmt.
20997 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20998 gimple_clobber_p stmts like debug stmts.
20999 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21000 that refer to unused VAR_DECLs or uninitialized values.
21001 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21002 gimple_clobber_p stmts if they refer to removed parameters.
21003 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21006 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21008 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21009 using SWI48 mode attribute.
21011 2013-04-02 Wei Mi <wmi@google.com>
21013 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21014 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21015 *<rotate_insn><mode>3_mask in i386.md.
21017 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21019 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21021 2013-04-02 Richard Biener <rguenther@suse.de>
21023 PR tree-optimization/56778
21024 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21025 Runtime alias tests are not supported for gather loads.
21026 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21027 stmts referenced from SSA operands before updating SSA form.
21029 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21030 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21032 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21033 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21034 * config/arm/cortex-a53.md: New file.
21035 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21036 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21037 * config/arm/arm.c (arm_issue_rate): Likewise.
21038 * config/arm/arm-tune.md: Regenerate
21039 * config/arm/arm-tables.opt: Regenerate.
21040 * config/arm/arm-cores.def: Add cortex-a53.
21042 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21044 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21047 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21049 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21050 scalar load/store operations using B/H registers.
21051 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21053 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21055 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21057 * config/aarch64/aarch64.c
21058 (aarch64_simd_scalar_immediate_valid_for_move): New.
21059 * config/aarch64/aarch64-protos.h
21060 (aarch64_simd_scalar_immediate_valid_for_move): New.
21061 * config/aarch64/constraints.md (Dh, Dq): New.
21062 * config/aarch64/iterators.md (hq): New.
21064 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21066 * reorg.c (get_branch_condition): Deal with conditional returns.
21067 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21069 2013-04-01 Wei Mi <wmi@google.com>
21071 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21072 Truncate operand 2 using %b asm operand modifier.
21073 (*<shift_insn><mode>3_mask): Ditto.
21074 (*<rotate_insn><mode>3_mask): Ditto.
21076 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21078 PR middle-end/56798
21079 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21081 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21083 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21085 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21087 2013-03-30 Lawrence Crowl <crowl@google.com>
21089 * dse.c (clear_alias_sets): Remove never set.
21090 (disqualified_clear_alias_sets): Remove never set.
21091 (clear_alias_mode_pool): Remove never set.
21092 (dse_step0): Remove condition that is never true.
21093 (canon_address): Remove condition that is never true.
21094 (dse_step7): Remove condition that is never true.
21095 (rest_of_handle_dse): Remove condition that is never true.
21096 (rest_of_handle_dse::did_global): Remove never read from above.
21097 (dse_step2_spill): Remove never called from above.
21098 (dse_step5_spill): Remove never called from above.
21100 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21102 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21103 JUMP_TABLE_DATA changes.
21104 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21105 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21106 (Insns) <jump_table_data>: New entry.
21107 * doc/tm.texi: Regenerate.
21109 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21111 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21112 for table jump at the end of a basic block using tablejump_p.
21113 * targhooks.c (default_invalid_within_doloop): Likewise.
21114 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21115 target hook implementation that is identical to the default hook.
21116 (rs6000_invalid_within_doloop): Remove.
21118 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21119 unused variable from tablejump_p call.
21121 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21122 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21123 (INSN_DELETED_P): Likewise.
21124 (emit_jump_table_data): New prototype.
21125 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21126 after 4th as unused.
21127 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21128 * sched-vis.c (print_insn): Likewise.
21129 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21130 insn for compatibility with back ends that use next_active_insn to
21131 identify jump table data.
21132 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21133 (remove_insn): Likewise.
21134 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21136 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21137 (emit_jump_table_data): New function.
21139 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21140 basic block, a JUMP_TABLE_DATA never is.
21141 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21142 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21143 off from code handling real insns.
21144 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21145 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21146 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21147 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21148 is not a NONDEBUG_INSN_P.
21149 * ira-costs.c (scan_one_insn): Likewise.
21150 * jump.c (mark_all_labels): Likewise.
21151 (mark_jump_label_1): Likewise.
21152 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21153 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21155 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21156 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21157 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21158 JUMP_TABLE_DATA_P insns.
21159 (calculate_elim_costs_all_insns): Likewise.
21160 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21161 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21162 (delete_output_reload): Code style fixups.
21163 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21164 insn flags on this non-insn.
21165 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21166 as scheduling barriers, for pre-change compatibility.
21167 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21168 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21170 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21171 redundant JUMP_TABLE_DATA_P test.
21172 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21173 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21174 (frv_for_each_packet): Likewise.
21175 * config/i386/i386.c (min_insn_size): Likewise.
21176 (ix86_avoid_jump_mispredicts): Likewise.
21177 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21178 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21179 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21180 (mips16_insn_length): Robustify.
21181 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21182 (mips16_split_long_branches): Likewise.
21183 * config/pa/pa.c (pa_combine_instructions): Likewise.
21184 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21185 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21186 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21187 as contributing to pool range lengths.
21188 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21189 Remove redundant JUMP_TABLE_DATA_P test.
21190 (sh_loop_align): Likewise.
21191 (split_branches): Likewise.
21192 (sh_insn_length_adjustment): Likewise.
21193 * config/spu/spu.c (get_branch_target): Likewise.
21195 2013-03-29 Jan Hubicka <jh@suse.cz>
21197 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21198 gcov streaming; stream hot bb threshold to ltrans.
21199 * predict.c (get_hot_bb_threshold): Break out from ....
21200 (maybe_hot_count_p): ... here.
21201 (set_hot_bb_threshold): New function.
21202 * lto-section-in.c (lto_section_name): Add profile.
21203 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21204 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21205 and data-streamer.h
21206 (histogram_entry): New structure.
21207 (histogram, histogram_pool): New global vars.
21208 (histogram_hash): New structure.
21209 (histogram_hash::hash): New method.
21210 (histogram_hash::equal): Likewise.
21211 (account_time_size): New function.
21212 (cmp_counts): New function.
21213 (dump_histogram): New function.
21214 (ipa_profile_generate_summary): New function.
21215 (ipa_profile_write_summary): New function.
21216 (ipa_profile_read_summary): New function.
21217 (ipa_profile): Decide on threshold.
21218 (pass_ipa_profile): Add ipa_profile_write_summary and
21219 ipa_profile_read_summary.
21220 * Makefile.in (ipa.o): Update dependencies.
21221 * lto-streamer.h (LTO_section_ipa_profile): New section.
21223 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21225 * tree.h (VAR_P): New.
21227 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21230 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21232 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21234 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21235 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21236 (control_flow_insn_p): Likewise.
21237 * cfgrtl.c (duplicate_insn_chain): Likewise.
21238 * final.c (get_attr_length_1): Likewise.
21239 (shorten_branches): Likewise.
21240 (final_scan_insn): Likewise.
21241 * function.c (instantiate_virtual_regs): Likewise.
21242 * gcse.c (insert_insn_end_basic_block): Likewise.
21243 * ira-costs.c (scan_one_insn): Likewise.
21244 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21245 * lra.c (check_rtl): Likewise.
21246 * reload1.c (elimination_costs_in_insn): Likewise.
21247 * reorg.c (follow_jumps): Likewise.
21249 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21250 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21251 (thumb_far_jump_used_p): Likewise.
21252 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21253 (workaround_speculation): Likewise.
21254 (add_sched_insns_for_speculation): Likewise.
21255 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21256 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21257 (frv_for_each_packet): Likewise.
21258 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21259 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21260 (final_emit_insn_group_barriers): Likewise.
21261 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21262 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21263 (mips16_insn_length): Likewise.
21264 * config/pa/pa.c (pa_reorg): Likewise.
21265 (pa_combine_instructions): Likewise.
21266 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21267 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21268 (sh_reorg): Likewise.
21269 (split_branches): Likewise.
21270 * config/spu/spu.c (get_branch_target): Likewise.
21272 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21275 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21277 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21278 Fix declaration name.
21280 2013-03-28 Lawrence Crowl <crowl@google.com>
21282 * graphds.h (struct graph.indicies): Remove unused.
21283 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21284 (SCOP_ORIGINAL_PDDRS): Remove unused.
21285 * sese.h (extern insert_loop_close_phis): Removed unused.
21286 (extern insert_guard_phis): Removed unused.
21287 (extern ivtype_map_elt_info): Removed unused.
21288 (new_ivtype_map_elt): Removed unused.
21289 * sese.c (ivtype_map_elt_info): Removed unused.
21291 2013-03-28 Lawrence Crowl <crowl@google.com>
21293 * Makefile.in: Add several missing include dependences.
21295 (test-dump.o): New. This object is not added to any executable,
21296 but is present for ad-hoc testing.
21298 (debug (const bitmap_head_def &)): New.
21299 (debug (const bitmap_head_def *)): New.
21301 (extern debug (const bitmap_head_def &)): New.
21302 (extern debug (const bitmap_head_def *)): New.
21304 (debug (edge_def &)): New.
21305 (debug (edge_def *)): New.
21307 (debug (basic_block_def &)): New.
21308 (debug (basic_block_def *)): New.
21310 (dump_node (const_tree, int, FILE *)): Correct source file.
21312 (debug (die_struct &)): New.
21313 (debug (die_struct *)): New.
21315 (extern debug (die_struct &)): New.
21316 (extern debug (die_struct *)): New.
21317 * gimple-pretty-print.c
21318 (debug (gimple_statement_d &)): New.
21319 (debug (gimple_statement_d *)): New.
21320 * gimple-pretty-print.h
21321 (extern debug (gimple_statement_d &)): New.
21322 (extern debug (gimple_statement_d *)): New.
21324 (debug (ira_allocno_copy &)): New.
21325 (debug (ira_allocno_copy *)): New.
21326 (debug (ira_allocno &)): New.
21327 (debug (ira_allocno *)): New.
21329 (extern debug (ira_allocno_copy &)): New.
21330 (extern debug (ira_allocno_copy *)): New.
21331 (extern debug (ira_allocno &)): New.
21332 (extern debug (ira_allocno *)): New.
21334 (debug (live_range &)): New.
21335 (debug (live_range *)): New.
21337 (debug (lra_live_range &)): New.
21338 (debug (lra_live_range *)): New.
21340 (debug (lra_live_range &)): New.
21341 (debug (lra_live_range *)): New.
21343 (debug (omega_pb_d &)): New.
21344 (debug (omega_pb_d *)): New.
21346 (extern debug (omega_pb_d &)): New.
21347 (extern debug (omega_pb_d *)): New.
21349 (debug (const rtx_def &)): New.
21350 (debug (const rtx_def *)): New.
21352 (debug_tree (tree): Move within file.
21353 (debug_raw (const tree_node &)): New.
21354 (debug_raw (const tree_node *)): New.
21355 (dump_tree_via_hooks (const tree_node *, int)): New.
21356 (debug (const tree_node &)): New.
21357 (debug (const tree_node *)): New.
21358 (debug_verbose (const tree_node &)): New.
21359 (debug_verbose (const tree_node *)): New.
21360 (debug_head (const tree_node &)): New.
21361 (debug_head (const tree_node *)): New.
21362 (debug_body (const tree_node &)): New.
21363 (debug_body (const tree_node *)): New.
21364 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21365 (debug (vec<tree, va_gc> &)): New.
21366 (debug (vec<tree, va_gc> *)): New.
21368 (extern debug (const rtx_def &)): New.
21369 (extern debug (const rtx_def *)): New.
21371 (debug_raw (simple_bitmap_def &)): New.
21372 (debug_raw (simple_bitmap_def *)): New.
21373 (debug (simple_bitmap_def &)): New.
21374 (debug (simple_bitmap_def *)): New.
21376 (extern debug (simple_bitmap_def &)): New.
21377 (extern debug (simple_bitmap_def *)): New.
21378 (extern debug_raw (simple_bitmap_def &)): New.
21379 (extern debug_raw (simple_bitmap_def *)): New.
21381 (debug (vinsn_def &)): New.
21382 (debug (vinsn_def *)): New.
21383 (debug_verbose (vinsn_def &)): New.
21384 (debug_verbose (vinsn_def *)): New.
21385 (debug (expr_def &)): New.
21386 (debug (expr_def *)): New.
21387 (debug_verbose (expr_def &)): New.
21388 (debug_verbose (expr_def *)): New.
21389 (debug (vec<rtx> &)): New.
21390 (debug (vec<rtx> *)): New.
21392 (extern debug (vinsn_def &)): New.
21393 (extern debug (vinsn_def *)): New.
21394 (extern debug_verbose (vinsn_def &)): New.
21395 (extern debug_verbose (vinsn_def *)): New.
21396 (extern debug (expr_def &)): New.
21397 (extern debug (expr_def *)): New.
21398 (extern debug_verbose (expr_def &)): New.
21399 (extern debug_verbose (expr_def *)): New.
21400 (extern debug (vec<rtx> &)): New.
21401 (extern debug (vec<rtx> *)): New.
21403 (_list_iter_cond_expr): Make inline instead of static.
21405 (debug (sreal &)): New.
21406 (debug (sreal *)): New.
21408 (extern debug (sreal &)): New.
21409 (extern debug (sreal *)): New.
21411 (extern debug_raw (const tree_node &)): New.
21412 (extern debug_raw (const tree_node *)): New.
21413 (extern debug (const tree_node &)): New.
21414 (extern debug (const tree_node *)): New.
21415 (extern debug_verbose (const tree_node &)): New.
21416 (extern debug_verbose (const tree_node *)): New.
21417 (extern debug_head (const tree_node &)): New.
21418 (extern debug_head (const tree_node *)): New.
21419 (extern debug_body (const tree_node &)): New.
21420 (extern debug_body (const tree_node *)): New.
21421 (extern debug (vec<tree, va_gc> &)): New.
21422 (extern debug (vec<tree, va_gc> *)): New.
21424 (debug (struct loop &)): New.
21425 (debug (struct loop *)): New.
21426 (debug_verbose (struct loop &)): New.
21427 (debug_verbose (struct loop *)): New.
21428 * tree-dump.c: Add header dependence.
21430 (extern debug (struct loop &)): New.
21431 (extern debug (struct loop *)): New.
21432 (extern debug_verbose (struct loop &)): New.
21433 (extern debug_verbose (struct loop *)): New.
21435 (debug (data_reference &)): New.
21436 (debug (data_reference *)): New.
21437 (debug (vec<data_reference_p> &)): New.
21438 (debug (vec<data_reference_p> *)): New.
21439 (debug (vec<ddr_p> &)): New.
21440 (debug (vec<ddr_p> *)): New.
21442 (extern debug (data_reference &)): New.
21443 (extern debug (data_reference *)): New.
21444 (extern debug (vec<data_reference_p> &)): New.
21445 (extern debug (vec<data_reference_p> *)): New.
21446 (extern debug (vec<ddr_p> &)): New.
21447 (extern debug (vec<ddr_p> *)): New.
21449 (debug (pt_solution &)): New.
21450 (debug (pt_solution *)): New.
21452 (extern debug (pt_solution &)): New.
21453 (extern debug (pt_solution *)): New.
21455 (debug (_var_map &)): New.
21456 (debug (_var_map *)): New.
21457 (debug (tree_live_info_d &)): New.
21458 (debug (tree_live_info_d *)): New.
21460 (extern debug (_var_map &)): New.
21461 (extern debug (_var_map *)): New.
21462 (extern debug (tree_live_info_d &)): New.
21463 (extern debug (tree_live_info_d *)): New.
21465 2013-03-28 Jan Hubicka <jh@suse.cz>
21467 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21469 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21471 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21472 record only when desired or required.
21474 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21476 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21477 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21479 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21481 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21482 (*andsi3_compare0_uxtw): New pattern.
21483 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21484 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21486 2013-03-28 Jan Hubicka <jh@suse.cz>
21488 * data-streamer-in.c (streamer_read_gcov_count): New function.
21489 * gimple-streamer-out.c: Include value-prof.h.
21490 (output_gimple_stmt): Output histogram.
21491 (output_bb): Use streamer_write_gcov_count.
21492 * value-prof.c: Include data-streamer.h
21493 (dump_histogram_value): Add HIST_TYPE_MAX.
21494 (stream_out_histogram_value): New function.
21495 (stream_in_histogram_value): New function.
21496 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21497 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21498 * data-streamer-out.c (streamer_write_gcov_count): New function.
21499 (streamer_write_gcov_count_stream): New function.
21500 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21501 (lto_output_node): Likewise.
21502 (input_node, input_edge): Likewise.
21503 * lto-streamer-out.c (output_cfg): Update streaming.
21504 * lto-streamer-in.c (input_cfg): Likewise.
21505 * data-streamer.h (streamer_write_gcov_count,
21506 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21507 * gimple-streamer-in.c: Include value-prof.h
21508 (input_gimple_stmt): Input histograms.
21509 (input_bb): Update profile streaming.
21511 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21513 * genmodes.c (emit_max_int): New function.
21514 (emit_insn_modes_h): Added call to emit_max_function.
21515 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21517 * machmode.def: Fixed comment.
21519 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21521 * combine.c (try_combine): Removed useless assert.
21522 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21524 2013-03-28 Marek Polacek <polacek@redhat.com>
21525 Richard Biener <rguenther@suse.de>
21527 PR tree-optimization/56695
21528 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21529 build signed result of a vector comparison.
21530 * tree-cfg.c (verify_gimple_comparison): Check that a result
21531 of a vector comparison has signed type.
21533 2013-03-28 Richard Biener <rguenther@suse.de>
21535 PR tree-optimization/37021
21536 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21537 do not restrict gaps between groups.
21538 * tree-vect-stmts.c (vectorizable_load): Properly account for
21539 a gap between groups.
21541 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21543 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21544 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21547 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21549 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21550 * doc/extend.texi (Named Address Spaces): Ditto.
21551 (Variable Attributes): Ditto.
21553 2013-03-27 Kai Tietz <ktietz@redhat.com>
21555 * config.build: Add support for cygwin x64 target.
21556 * config.gcc: Likewise.
21557 * config.host: Likewise.
21558 * configure.ac: Likewise
21559 * configure: Regenerated.
21561 2013-03-27 Kai Tietz <ktietz@redhat.com>
21563 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21564 * config/i386/t-cygwin-w64: New file.
21565 * config/i386/cygwin-w64.h: New file.
21566 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21567 and add support for x64-cygwin target.
21568 (CPP_SPEC): Likewise.
21569 (CXX_WRAP_SPEC_LIST): Undefine before define.
21570 (LIBGCJ_SONAME): Use 15 as version.
21572 2013-03-27 Richard Biener <rguenther@suse.de>
21574 PR tree-optimization/56716
21575 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21576 dumping for ref nodes.
21578 2013-03-27 Martin Jambor <mjambor@suse.cz>
21580 PR tree-optimization/55334
21581 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21582 restricted pointers to arrays.
21584 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21586 * Makefile.in (.SUFFIXES): Add .cc.
21587 (.c.o): Apply same recipe for implicit rule .cc.o.
21589 2013-03-27 Richard Biener <rguenther@suse.de>
21591 PR tree-optimization/37021
21592 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21593 REALPART/IMAGPART_EXPRs around the supported refs.
21594 * tree-ssa-structalias.c (find_func_aliases): Assume that
21595 floating-point values are not used to transfer pointers.
21597 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21599 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21600 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21601 * doc/tm.texi: Regenerate.
21602 * targhooks.h (default_has_ifunc_p): New.
21603 * targhooks.c (default_has_ifunc_p): Ditto.
21604 * config/linux-protos.h: New file.
21605 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21606 hook for linux which disables support of indirect functions in android.
21607 * config/linux-android.c: New file.
21608 * config/t-linux-android.c: Ditto.
21609 * config.gcc: Added new object file linux-android.o.
21610 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21611 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21612 * varasm.c (do_assemble_alias): Likewise.
21613 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21614 doesn't support indirect functions.
21615 * configure: Regenerate.
21617 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21620 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21621 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21624 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21626 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21628 2013-03-27 Terry Guo <terry.guo@arm.com>
21630 * config/arm/arm-cores.def: Added core cortex-r7.
21631 * config/arm/arm-tune.md: Regenerated.
21632 * config/arm/arm-tables.opt: Regenerated.
21633 * doc/invoke.texi: Added entry for core cortex-r7.
21635 2013-03-27 Walter Lee <walt@tilera.com>
21637 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21638 double-decrement of next_scratch_regno.
21640 2013-03-27 Walter Lee <walt@tilera.com>
21642 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21644 (insn_v1mulus): Ditto.
21645 (insn_v2muls): Ditto.
21647 2013-03-27 Walter Lee <walt@tilera.com>
21649 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21650 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21652 2013-03-27 Walter Lee <walt@tilera.com>
21654 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21655 (*sibcall_value): Ditto.
21657 2013-03-27 Walter Lee <walt@tilera.com>
21659 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21660 (insn_mnz_v8qi): ... this ...
21661 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21663 (insn_v<n>mnz): Replaced by ...
21664 (insn_v1mnz): ... this ...
21665 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21667 (insn_mz_<mode>): Replaced by ...
21668 (insn_mz_v8qi): ... this ...
21669 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21671 (insn_v<n>mz): Replaced by ...
21672 (insn_v1mz): ... this ...
21673 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21676 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21678 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21680 2013-03-26 Roland McGrath <mcgrathr@google.com>
21682 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21683 than fprintf with a non-constant, non-format string.
21685 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21687 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21688 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21689 operand 0 predicate.
21690 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21691 attribute. Use general_x64nomem_operand as operand 1 predicate.
21692 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21693 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21694 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21695 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21696 (mov<mode>_insv_1): Remove expander. Merge insn with
21697 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21698 Use general_x64nomem_operand as operand 1 predicate.
21699 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21700 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21701 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21702 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21703 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21704 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21705 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21706 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21707 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21708 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21709 (general_x64nomem_operand): Ditto.
21711 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21713 * config/rtems.opt: Add -pthread option.
21715 2013-03-26 Richard Biener <rguenther@suse.de>
21717 * alias.c (find_base_term): Avoid redundant and not used recursion.
21718 (base_alias_check): Get the initial base term from the caller.
21719 (true_dependence_1): Compute and pass base terms to base_alias_check.
21720 (write_dependence_p): Likewise.
21721 (may_alias_p): Likewise.
21723 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21725 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21726 PC-relative load in SI modes and above only.
21728 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21730 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21731 * config/arm/iwmmxt.md (WCGR0): Update.
21732 (WCGR1, WCGR2, WCGR3): Likewise.
21734 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21736 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21737 Use x64 and nox64 isa attributes.
21739 2013-03-26 Richard Biener <rguenther@suse.de>
21741 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21742 alignment computations and rely on get_object_alignment_1
21743 for the !TYPE_P case.
21744 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21746 2013-03-26 Walter Lee <walt@tilera.com>
21748 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21749 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21751 2013-03-25 Jeff Law <law@redhat.com>
21753 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21754 check for INTEGRAL_TYPE_P that was missing due to checking in
21755 wrong version of prior patch.
21757 2013-03-25 Walter Lee <walt@tilera.com>
21759 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21760 TILEGX_INSN_SHUFFLEBYTES1.
21761 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21763 (tilegx_builtins): Ditto.
21764 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21766 2013-03-25 Walter Lee <walt@tilera.com>
21768 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21769 (floatunssisf2): New pattern.
21770 (floatsidf2): New pattern.
21771 (floatunssidf2): New pattern.
21773 2013-03-25 Walter Lee <walt@tilera.com>
21775 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21776 tests for constraint J, K, N, P.
21778 2013-03-25 Walter Lee <walt@tilera.com>
21780 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21781 Use indirect/pcrel encoding.
21782 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21785 2013-03-25 Steve Ellcey <sellcey@mips.com>
21787 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21788 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21789 * config/mips/mips.c (mips_option_override): Set IMADD default.
21790 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21791 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21792 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21793 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21794 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21796 2013-03-25 Jeff Law <law@redhat.com>
21798 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21799 slightly to avoid creating and folding useless trees. Simplify
21800 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21802 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21804 * config/i386/i386.md (*zero_extendsidi2): Merge with
21805 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21806 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21807 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21808 targets. Clarify comment.
21810 2013-03-25 Martin Jambor <mjambor@suse.cz>
21812 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21813 pass-through jump functions differently.
21814 (ipa_read_jump_function): Likewise. Also use setter functions to set
21817 2013-03-25 Martin Jambor <mjambor@suse.cz>
21819 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21820 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21822 (ipa_get_indirect_edge_target): New function.
21823 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21824 ipa_get_indirect_edge_target. Update all callers.
21825 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21826 ipa_get_indirect_edge_target_1 instead of calling
21827 ipa_get_indirect_edge_target.
21828 (create_specialized_node): Pass aggvlas to
21829 ipcp_discover_new_direct_edges.
21831 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21833 * config/arm/arm.md (f_sels, f_seld): New types.
21834 (*cmov<mode>): New pattern.
21835 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21838 2013-03-25 Kai Tietz <ktietz@redhat.com>
21840 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21841 POSIX-printf for mingw-hosted builds.
21843 2013-03-25 Richard Biener <rguenther@suse.de>
21845 PR middle-end/56694
21846 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21847 must-not-throw stmt location.
21849 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21851 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21852 Emit load-acquire versions when acq is true.
21853 (arm_emit_store_exclusive): Add rel parameter.
21854 Emit store-release versions when rel is true.
21855 (arm_split_compare_and_swap): Use acquire-release instructions
21857 of barriers when appropriate.
21858 (arm_split_atomic_op): Likewise.
21859 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21860 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21861 (VUNSPEC_SLX): Likewise.
21862 (VUNSPEC_LDA): Likewise.
21863 (VUNSPEC_STL): Likewise.
21864 * config/arm/sync.md (atomic_load<mode>): New pattern.
21865 (atomic_store<mode>): Likewise.
21866 (arm_load_acquire_exclusive<mode>): Likewise.
21867 (arm_load_acquire_exclusivesi): Likewise.
21868 (arm_load_acquire_exclusivedi): Likewise.
21869 (arm_store_release_exclusive<mode>): Likewise.
21871 2013-03-25 Catherine Moore <clm@codesourcery.com>
21873 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21874 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21875 * config/mip/predicates.md (lwsp_swsp_operand,
21876 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21877 sb16_operand, db4_operand, db7_operand, ib3_operand,
21878 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21879 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21880 andi16_operand): New predicates.
21881 * config/mips/mips.md (compression): New attribute.
21882 (enabled): New attribute.
21883 (length): Consider compression in computing length.
21884 (shift_compression): New code attribute.
21885 (*add<mode>3): New operands. Record compression.
21886 (sub<mode>3): Likewise.
21887 (one_cmpl<mode>2): Likewise.
21888 (*and<mode>3): Likewise.
21889 (*ior<mode>3): Likewise.
21890 (unnamed pattern for xor): Likewise.
21891 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21892 (*<optab><mode>3): Likewise.
21893 (*mov<mode>_internal: Likewise.
21894 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21895 (mips_unsigned_immediate_p): New.
21896 (umips_lwsp_swsp_address_p): New.
21897 (m16_based_address_p): New.
21898 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21899 (mips_unsigned_immediate_p): New prototype.
21900 (lwsp_swsp_address_p): New prototype.
21901 (m16_based_address_p): New prototype.
21902 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21903 (mips_signed_immediate_p): New function.
21904 (m16_based_address_p): New function.
21905 (lwsp_swsp_address_p): New function.
21906 (mips_print_operand_punctuation): Recognize short delay slot insns
21907 for microMIPS.add<mode>3"
21909 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21912 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21913 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21915 2013-03-25 Richard Biener <rguenther@suse.de>
21917 PR tree-optimization/56689
21918 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21921 2013-03-25 Richard Biener <rguenther@suse.de>
21923 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21925 (memory_references): Likewise.
21926 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21927 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21928 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21929 (gather_mem_refs_in_loops): Fold into ...
21930 (analyze_memory_references): ... this. Move initialization
21931 to tree_ssa_lim_initialize.
21932 (fill_always_executed_in): Rename to ...
21933 (fill_always_executed_in_1): ... this.
21934 (fill_always_executed_in): Move contains_call computation to
21935 this new function from ...
21936 (tree_ssa_lim_initialize): ... here.
21937 (tree_ssa_lim): Call fill_always_executed_in.
21939 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21941 * postreload.c (reload_combine): Fix code detecting returns.
21943 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21945 * function.c (emit_use_return_register_into_block): On cc0 targets,
21946 do not emit the sequence between cc0 setter and user.
21948 2013-03-25 Kai Tietz <ktietz@redhat.com>
21950 * config/i386/predicates.md (local_symbolic_operand): Interpret
21951 dll-imported symbols as none-local.
21953 2013-03-25 Richard Biener <rguenther@suse.de>
21955 * tree-ssa-loop-im.c (struct depend): Remove.
21956 (struct lim_aux_data): Make depends a vec of gimples.
21957 (free_lim_aux_data): Adjust.
21958 (add_dependency): Likewise.
21959 (set_level): Likewise.
21961 2013-03-25 Richard Biener <rguenther@suse.de>
21963 PR middle-end/56434
21964 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21965 the pointer returned by calls with ECF_MALLOC set.
21967 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21969 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21971 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21973 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21974 using MMXMODE mode iterator.
21975 (*move<mode>_internal): Merge with *movv2sf_internal and
21976 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21978 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21980 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21981 (record_last_mem_set_info): Likewise.
21983 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21984 of XNEWVEC followed by memset.
21985 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21987 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21989 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21990 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21991 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21992 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21993 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21994 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21995 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21996 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21997 BARRIER_P instead of GET_CODE.
21999 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22001 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22002 inaccuracy in the probing code.
22004 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22005 (ctrapdi4): Likewise.
22007 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22009 * calls.c (expand_call): Add missing guard to code handling return
22010 of non-BLKmode structures in MSB.
22011 * function.c (expand_function_end): Likewise.
22013 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22015 * combine.c (try_combine): Adjust comment. Do not add the set of
22016 insn #0 if the destination indirectly is set or dies in insn #2.
22017 Tidy up code to distribute a new note.
22019 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22021 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22022 also for alternatives 16 and 17.
22024 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22026 * config/i386/sse.md (*mov<mode>_internal): Merge with
22027 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22028 Emit insn template depending on type attribute. Use
22029 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22030 movd instead of movq mnemonic for interunit moves. Rewrite mode
22031 attribute calculation. Remove unit attribute calculation.
22032 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22033 Set prefix_data16 attribute for DImode ssemov types.
22034 Use Ym instead of y for SSE-MMX conversion alternatives.
22035 Reorder operand constraints.
22037 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22039 * df.h (df_insn_delete): Adjust prototype.
22040 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22041 and let it decide whether mark the basic block dirty.
22042 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22043 * df-scan.c (df_insn_info_delete): New helper function, split
22044 off from df_insn_delete.
22045 (df_scan_free_bb_info): Use it.
22046 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22048 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22049 that the insn is actually an insn and it has a non-NULL basic block.
22050 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22052 2013-03-22 Richard Biener <rguenther@suse.de>
22054 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22056 (mem_ref_alloc): Do not allocate them.
22057 (refs_independent_p): Do not query or maintain a cache.
22059 2013-03-22 Richard Biener <rguenther@suse.de>
22061 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22062 (gather_mem_refs_in_loops): Do not compute it.
22063 (analyze_memory_references): Do not allocate it.
22064 (tree_ssa_lim_finalize): Do not free it.
22065 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22067 2013-03-22 Richard Biener <rguenther@suse.de>
22069 * is-a.h (as_a): Use gcc_checking_assert.
22071 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22073 * config/aarch64/aarch64.c (aarch64_print_operand): New
22074 format specifier for printing a constant in hex.
22075 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22076 format specifier for printing second operand.
22078 2013-03-22 Richard Biener <rguenther@suse.de>
22080 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22082 (gather_mem_refs_in_loops): Perform store accumulation here.
22083 (create_vop_ref_mapping_loop): Remove.
22084 (create_vop_ref_mapping): Likewise.
22085 (analyze_memory_references): Initialize refs_stored_in_loop.
22086 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22087 (record_indep_loop): Remove.
22088 (record_dep_loop): New function.
22089 (ref_indep_loop_p_1): Adjust to only walk over references
22090 in the loop, not its subloops.
22091 (ref_indep_loop_p): Rename to ...
22092 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22093 maintaining a more fine-grained cache.
22094 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22095 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22097 2013-03-22 Richard Biener <rguenther@suse.de>
22099 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22100 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22101 aggregate mem_ref_loc.
22102 (free_mem_ref_locs): Inline into ...
22103 (memref_free): ... this and adjust.
22104 (mem_ref_alloc): Adjust.
22105 (mem_ref_locs_alloc): Remove.
22106 (record_mem_ref_loc): Adjust.
22107 (get_all_locs_in_loop): Rewrite into ...
22108 (for_all_locs_in_loop): ... this iterator.
22109 (rewrite_mem_ref_loc): New functor.
22110 (rewrite_mem_refs): Use for_all_locs_in_loop.
22111 (sm_set_flag_if_changed): New functor.
22112 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22113 (ref_always_accessed): New functor.
22114 (ref_always_accessed_p): Use for_all_locs_in_loop.
22116 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22118 * tree-pass.h (PROP_gimple_lvec): New.
22119 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22120 (init_optimization_passes): Move pass_lower_vector.
22121 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22123 (pass_lower_vector): Provide PROP_gimple_lvec.
22124 (pass_lower_vector_ssa): Likewise.
22125 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22127 2013-03-21 Mark Wielaard <mjw@redhat.com>
22129 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22131 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22133 * config/i386/i386.md (*movdi_internal): Disparage slightly
22134 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22135 conversion alternatives.
22137 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22139 PR middle-end/48087
22140 * diagnostic.def (DK_WERROR): New kind.
22141 * diagnostic.h (werrorcount): Define.
22142 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22143 promoted to DK_ERROR, increment DK_WERROR counter instead of
22145 * toplev.c (toplev_main): Call print_ignored_options even if
22146 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22147 even if just werrorcount is non-zero.
22150 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22152 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22153 (string_cst_pool_decl): New function.
22154 (optimize_one_addr_into_implicit_ptr): New function.
22155 (resolve_addr_in_expr): Optimize DWARF location expression
22156 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22157 which doesn't live in memory, but has DW_AT_location or
22158 DW_AT_const_value, or refers to a string literal, into
22159 DW_OP_GNU_implicit_pointer.
22160 (optimize_location_into_implicit_ptr): New function.
22161 (resolve_addr): If removing DW_AT_location of a variable because
22162 it was DW_OP_addr of address of the variable, but the variable doesn't
22163 live in memory, try to emit const value attribute for the initializer.
22165 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22167 * tree.h (VECTOR_TYPE_P): New macro.
22168 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22169 TYPE_MODE): Use it.
22170 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22171 VEC_COND_EXPR cannot be lvalues.
22172 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22174 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22176 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22177 Restrict the transformation to equal modes.
22179 2013-03-21 Richard Biener <rguenther@suse.de>
22181 PR tree-optimization/39326
22182 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22183 (MEM_ANALYZABLE): Adjust.
22184 (record_mem_ref_loc): Move bitmap ops ...
22185 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22186 unanalyzable refs, do not record locations for it.
22187 (analyze_memory_references): Allocate ref zero as shared
22189 (refs_independent_p): Do not test for unanalyzed mems here.
22190 (ref_indep_loop_p_1): Special-case disambiguation against
22191 the unanalyzed ref.
22192 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22194 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22196 * config/arm/arm-protos.h (tune_params): Add
22197 prefer_neon_for_64bits field.
22198 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22199 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22200 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22201 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22202 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22203 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22204 (arm_option_override): Handle -mneon-for-64bits new option.
22205 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22206 (prefer_neon_for_64bits): Declare new variable.
22207 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22208 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22209 (arch_enabled): Handle new arch types. Remove support for onlya8
22211 (one_cmpldi2): Use new arch names.
22212 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22213 * config/arm/arm.opt (mneon-for-64bits): Add option.
22214 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22215 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22216 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22218 * doc/invoke.texi (-mneon-for-64bits): Document.
22220 2013-03-21 Richard Biener <rguenther@suse.de>
22222 PR tree-optimization/39326
22223 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22224 (sort_bbs_in_loop_postorder_cmp): New function.
22225 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22227 2013-03-21 Richard Biener <rguenther@suse.de>
22229 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22230 (vect_insert_into_interleaving_chain): Likewise.
22231 (vect_drs_dependent_in_basic_block): Inline ...
22232 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22234 (vect_analyze_data_ref_dependence): ... here. Simplify.
22235 (vect_check_interleaving): Simplify.
22236 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22237 (vect_slp_analyze_data_ref_dependences): ... this new function.
22238 (dr_group_sort_cmp): New function.
22239 (vect_analyze_data_ref_accesses): Compute data-reference groups
22240 here instead of in vect_analyze_data_ref_dependence. Use
22241 a more efficient algorithm.
22242 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22243 vect_slp_analyze_data_ref_dependences. Call
22244 vect_analyze_data_ref_accesses earlier.
22245 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22246 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22247 (vect_slp_analyze_data_ref_dependences): New prototype.
22249 2013-03-21 Richard Biener <rguenther@suse.de>
22251 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22252 ref is stored in the loop.
22253 (find_refs_for_sm): Walk only over all stores.
22254 (store_motion_loop): Allocate from lim_bitmap_obstack.
22255 (store_motion): Likewise.
22257 2013-03-21 Richard Biener <rguenther@suse.de>
22259 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22260 Update virtual SSA form.
22262 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22264 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22265 * configure: Regenerate.
22266 * config.in: Regenerate.
22267 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22268 if !HAVE_LD_EH_FRAME_CIEV3.
22270 2013-03-21 Richard Biener <rguenther@suse.de>
22272 * tree-cfg.c (verify_expr_no_block): New function.
22273 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22274 nor DECL_VALUE_EXPR have locations with associated blocks.
22275 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22276 (clear_unused_block_pointer): Remove code dealing with
22277 blocks in DECL_DEBUG_EXPR locations.
22279 2013-03-21 Richard Biener <rguenther@suse.de>
22281 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22282 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22283 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22284 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22285 instead of DECL_DEBUG_EXPR_IS_FROM.
22286 * gimplify.c (gimplify_modify_expr): Likewise.
22287 * tree-cfg.c (verify_expr_location_1): Likewise.
22288 * tree-complex.c (create_one_component_var): Likewise.
22289 * tree-sra.c (create_access_replacement): Likewise.
22290 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22291 (clear_unused_block_pointer): Likewise.
22292 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22293 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22294 * var-tracking.c (var_debug_decl): Likewise.
22295 (track_expr_p): Likewise.
22296 * tree-inline.c (add_local_variables): Likewise. Set
22297 DECL_HAS_DEBUG_EXPR_P after copying it.
22298 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22299 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22301 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22304 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22305 * configure: Regenerate.
22306 * config.in: Regenerate.
22307 * config/i386/i386.md (*movdf_internal): Use
22308 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22309 movd instead of movq mnemonic for interunit moves.
22310 (*movdi_internal): Ditto.
22312 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22314 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22315 (abd<mode>_3): New pattern.
22316 (aba<mode>_3): New pattern.
22317 (fabd<mode>_3): New pattern.
22319 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22321 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22322 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22323 occurrence of REGISTER_PREFIX as its empty string.
22325 2013-03-20 Jeff Law <law@redhat.com>
22327 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22328 addititional equivalences for equality comparisons between an SSA_NAME
22329 and a constant where the SSA_NAME was set from a widening conversion.
22331 2013-03-20 Walter Lee <walt@tilera.com>
22333 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22335 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22337 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22338 depending on type attribute.
22339 (*movti_internal): Ditto.
22340 (*movtf_internal): Ditto.
22341 (*movxf_internal): Ditto.
22342 (*movdf_internal): Ditto.
22343 (*movsf_internal): Ditto.
22345 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22347 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22348 maybe_vex for sselog1 and ssemov types.
22349 (*movdi_internal): Reorder operand constraints.
22350 (*movsi_internal): Ditto. Set prefix attribute to
22351 maybe_vex for sselog1 and ssemov types.
22352 (*movtf_internal): Set prefix attribute to maybe_vex
22353 for sselog1 and ssemov types.
22354 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22355 DImode ssemov types. Reorder operand constraints.
22356 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22357 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22358 attribute for SImode ssemov types. Reorder operand constraints.
22360 2013-03-20 Martin Jambor <mjambor@suse.cz>
22362 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22363 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22365 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22367 * config/rs6000/predicates.md (indexed_address, update_address_mem
22368 update_indexed_address_mem): New predicates.
22369 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22370 attribute for load/store instructions.
22371 * config/rs6000/dfp.md (movsd_store): Likewise.
22372 (movsd_load): Likewise.
22373 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22374 (unnamed HI->DI extend define_insn): Likewise.
22375 (unnamed SI->DI extend define_insn): Likewise.
22376 (unnamed QI->SI extend define_insn): Likewise.
22377 (unnamed QI->HI extend define_insn): Likewise.
22378 (unnamed HI->SI extend define_insn): Likewise.
22379 (unnamed HI->SI extend define_insn): Likewise.
22380 (extendsfdf2_fpr): Likewise.
22381 (movsi_internal1): Likewise.
22382 (movsi_internal1_single): Likewise.
22383 (movhi_internal): Likewise.
22384 (movqi_internal): Likewise.
22385 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22386 attribute for load/store instructions.
22387 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22389 (mov<mode>_softfloat): Likewise.
22390 (mov<mode>_hardfloat32): Likewise.
22391 (mov<mode>_hardfloat64): Likewise.
22392 (mov<mode>_softfloat64): Likewise.
22393 (movdi_internal32): Likewise.
22394 (movdi_internal64): Likewise.
22395 (probe_stack_<mode>): Likewise.
22397 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22399 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22400 floating point, and decimal floating point to reload iterator.
22402 * config/rs6000/constraints.md (wl constraint): New constraints to
22403 return FLOAT_REGS if certain options are used to reduce the number
22404 of separate patterns that exist in the file.
22405 (wx constraint): Likewise.
22406 (wz constraint): Likewise.
22408 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22409 -mdebug=reg, print wg, wl, wx, and wz constraints.
22410 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22411 Initialize the reload functions for 64-bit binary/decimal floating
22413 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22414 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22415 create the buffer on the stack to overcome not having a 32-bit
22417 (rs6000_emit_move): Likewise.
22418 (rs6000_secondary_memory_needed_rtx): Likewise.
22419 (rs6000_alloc_sdmode_stack_slot): Likewise.
22420 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22421 via xxlxor, just like DFmode 0.0.
22423 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22424 define as 1 if we are running on a power7 or newer.
22425 (enum r6000_reg_class_enum): Add new constraints.
22427 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22428 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22429 with other moves by using conditional constraits (wg). Use LFIWZX
22430 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22431 (movsd splitter): Likewise.
22432 (movsd_hardfloat): Likewise.
22433 (movsd_softfloat): Likewise.
22435 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22436 binary and decimal floating point moves.
22437 (fmove_ok): New attributes to combine binary and decimal floating
22438 point moves, and to combine power6x (mfpgpr) moves along normal
22440 (real_value_to_target): Likewise.
22441 (f32_lr): Likewise.
22442 (f32_lm): Likewise.
22443 (f32_li): Likewise.
22444 (f32_sr): Likewise.
22445 (f32_sm): Likewise.
22446 (f32_si): Likewise.
22447 (movsf): Combine binary and decimal floating point moves. Combine
22448 power6x (mfpgpr) moves with other moves by using conditional
22449 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22450 (mov<mode> for SFmode/SDmode); Likewise.
22451 (SFmode/SDmode splitters): Likewise.
22452 (movsf_hardfloat): Likewise.
22453 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22454 (movsf_softfloat): Likewise.
22455 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22457 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22458 wx and wz constraints.
22460 * config/rs6000/constraints.md (wg constraint): New constraint to
22461 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22463 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22466 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22467 -mdebug=reg, print wg, wl, wx, and wz constraints.
22468 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22469 Initialize the reload functions for 64-bit binary/decimal floating
22471 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22472 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22473 create the buffer on the stack to overcome not having a 32-bit
22475 (rs6000_emit_move): Likewise.
22476 (rs6000_secondary_memory_needed_rtx): Likewise.
22477 (rs6000_alloc_sdmode_stack_slot): Likewise.
22478 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22479 via xxlxor, just like DFmode 0.0.
22481 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22482 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22483 with other moves by using conditional constraits (wg). Use LFIWZX
22484 and STFIWX for loading SDmode on power7.
22485 (movdd splitters): Likewise.
22486 (movdd_hardfloat32): Likewise.
22487 (movdd_softfloat32): Likewise.
22488 (movdd_hardfloat64_mfpgpr): Likewise.
22489 (movdd_hardfloat64): Likewise.
22490 (movdd_softfloat64): Likewise.
22492 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22493 64-bit binary and decimal floating point moves.
22494 (FMOVE64X): Likewise.
22495 (movdf): Combine 64-bit binary and decimal floating point moves.
22496 Combine power6x (mfpgpr) moves with other moves by using
22497 conditional constraits (wg).
22498 (mov<mode> for DFmode/DDmode): Likewise.
22499 (DFmode/DDmode splitters): Likewise.
22500 (movdf_hardfloat32): Likewise.
22501 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22502 (movdf_softfloat32): Likewise.
22503 (movdf_hardfloat64_mfpgpr): Likewise.
22504 (movdf_hardfloat64): Likewise.
22505 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22506 (movdf_softfloat64): Likewise.
22507 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22508 (reload_<mode>_load): Move to later in the file so they aren't in
22509 the middle of the floating point move insns.
22510 (reload_<mode>_store): Likewise.
22512 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22515 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22516 constraint if -mdebug=reg.
22517 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22518 Enable using dd reload support if needed.
22520 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22521 binary and decimal floating point moves in rs6000.md.
22522 (movtd_internal): Likewise.
22524 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22525 decimal floating point moves.
22527 (movtf_internal): Likewise.
22528 (mov<mode>_internal, TDmode/TFmode): Likewise.
22529 (movtf_softfloat): Likewise.
22530 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22532 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22533 movdi_internal64, using wg constraint for move direct operations.
22534 (movdi_internal64): Likewise.
22536 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22537 MODES_TIEABLE_P for selected modes. Print the numerical value of
22538 the various virtual registers. Use GPR/FPR first/last values,
22539 instead of hard coding the register numbers. Print which modes
22540 have reload functions registered.
22541 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22542 settings before/after setting cpu, target and subtarget settings.
22543 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22544 and for secondary reload failures in rs6000_secondary_reload_inner.
22545 (rs6000_secondary_reload_fail): Likewise.
22546 (rs6000_secondary_reload_inner): Likewise.
22548 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22549 macros for first/last GPR and FPR registers.
22550 (LAST_GPR_REGNO): Likewise.
22551 (FIRST_FPR_REGNO): Likewise.
22552 (LAST_FPR_REGNO): Likewise.
22554 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22555 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22556 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22557 (vcond<mode><mode>): Likewise.
22558 (vcondu<mode><mode>): Likewise.
22559 (vector_gtu<mode>): Likewise.
22560 (vector_gte<mode>): Likewise.
22561 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22562 to prevent the compiler from converting DImode operations to TImode.
22563 (ior<mode>3): Likewise.
22564 (and<mode>3): Likewise.
22565 (one_cmpl<mode>2): Likewise.
22566 (nor<mode>3): Likewise.
22567 (andc<mode>3): Likewise.
22569 * config/rs6000/constraints.md (wt constraint): New constraint
22570 that returns VSX_REGS if TImode is allowed in VSX registers.
22572 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22573 constant under VSX.
22575 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22576 similar to TImode, but it is restricted to being in the GPRs.
22578 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22579 TImode to occupy a single VSX register.
22581 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22582 -mvsx-timode for power7/power8.
22583 (power7 cpu): Likewise.
22584 (power8 cpu): Likewise.
22586 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22587 sure that TFmode/TDmode take up two registers if they are ever
22588 allowed in the upper VSX registers.
22589 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22591 (rs6000_init_hard_regno_mode_ok): Likewise.
22592 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22593 constraint. Print if LRA is turned on.
22594 (rs6000_option_override_internal): Give an error if -mvsx-timode
22595 and VSX is not enabled.
22596 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22597 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22598 to reg+offset addressing. Use PTImode when checking offset
22599 addresses for validity.
22600 (reg_offset_addressing_ok_p): Likewise.
22601 (rs6000_legitimate_offset_address_p): Likewise.
22602 (rs6000_legitimize_address): Likewise.
22603 (rs6000_legitimize_reload_address): Likewise.
22604 (rs6000_legitimate_address_p): Likewise.
22605 (rs6000_eliminate_indexed_memrefs): Likewise.
22606 (rs6000_emit_move): Likewise.
22607 (rs6000_secondary_reload): Likewise.
22608 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22609 reloads to fpr registers to continue to use reg+offset addressing,
22610 but 64-bit reloads to altivec registers need reg+reg addressing.
22611 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22612 it. Treat LO_SUM like a PLUS operation.
22613 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22614 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22615 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22616 registers to share a register with a smaller sized type, since VSX
22617 puts scalars in the upper 64-bits.
22618 (print_operand): Add support for PTImode.
22619 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22620 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22621 registers, but don't have arithmetic support.
22622 (rs6000_memory_move_cost): Add test for VSX.
22623 (rs6000_opt_masks): Add -mvsx-timode.
22625 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22628 (VSr): Use wt constraint for TImode.
22629 (VSv): Drop TImode support.
22630 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22631 (vsx_movti_64bit): Likewise.
22632 (vsx_movti_32bit): Likewise.
22633 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22634 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22635 one '?' on the appropriate output constraint. Do not allow TImode
22636 logical operations on 32-bit systems.
22637 (vsx_ior<mode>3): Likewise.
22638 (vsx_xor<mode>3): Likewise.
22639 (vsx_one_cmpl<mode>2): Likewise.
22640 (vsx_nor<mode>3): Likewise.
22641 (vsx_andc<mode>3): Likewise.
22642 (vsx_concat_<mode>): Likewise.
22643 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22645 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22646 OPTION_MASK_VSX_TIMODE.
22647 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22648 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22650 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22651 (TI2 iterator): New iterator for TImode, PTImode.
22652 (wd mode attribute): Add values for vector types.
22653 (movti_string): Replace TI move operations with operations for TImode
22654 and PTImode. Add support for TImode being allowed in VSX registers.
22655 (mov<mode>_string, TImode/PTImode): Likewise.
22656 (movti_ppc64): Likewise.
22657 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22658 (TI mode splitters): Likewise.
22660 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22663 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22665 PR tree-optimization/56355
22666 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22667 Also handle integers with undefined overflow.
22669 2013-03-20 Catherine Moore <clm@codesourcery.com>
22670 Maciej W. Rozycki <macro@codesourcery.com>
22671 Tom de Vries <tom@codesourcery.com>
22672 Nathan Sidwell <nathan@codesourcery.com>
22673 Iain Sandoe <iain@codesourcery.com>
22674 Nathan Froyd <froydnj@codesourcery.com>
22675 Chao-ying Fu <fu@mips.com>
22677 * doc/extend.texi: (micromips, nomicromips, nocompression):
22678 Document new function attributes.
22679 * doc/invoke.texi (minterlink-compressed, mmicromips,
22680 m14k, m14ke, m14kec): Document new options.
22681 (minterlink-mips16): Update documentation.
22682 * doc/md.texi (ZC, ZD): Document new constraints.
22683 * configure.ac (gcc_cv_as_micromips): Check if linker
22684 supports the .set micromips directive.
22685 * configure: Regenerate.
22686 * config.in: Regenerate.
22687 * config/mips/mips-tables.opt: Regenerate.
22688 * config/mips/micromips.md: New file.
22689 * constraints.md (ZC, ZD): New constraints.
22690 * config/mips/predicates.md (movep_src_register): New predicate.
22691 (movep_src_operand): New predicate.
22692 (non_volatile_mem_operand): New predicate.
22693 * config/mips/mips.md (multimem): New type.
22694 (length): Differentiate between 17-bit and 18-bit branch offsets.
22695 (MOVEP1, MOVEP2): New mode iterator.
22696 (mov_<load>l): Use ZC constraint.
22697 (mov_<load>r): Likewise.
22698 (mov_<store>l): Likewise.
22699 (mov_<store>r): Likewise.
22700 (*branch_equality<mode>_inverted): Add microMIPS support.
22701 (*branch_equality<mode>): Likewise.
22702 (*jump_absolute): Likewise.
22703 (indirect_jump_<mode>): Likewise.
22704 (tablejump_<mode>): Likewise.
22705 (<optab>_internal): Likewise.
22706 (sibcall_internal): Likewise.
22707 (sibcall_value_internal): Likewise.
22708 (prefetch): Use constraint ZD.
22709 * config/mips/mips.opt (minterlink-compressed): New option.
22710 (minterlink-mips16): Now an alias for minterlink-compressed.
22711 (mmicromips): New option.
22712 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22713 (compare_and_swap_12): Likewise.
22714 (sync_add<mode>): Likewise.
22715 (sync_<optab>_12): Likewise.
22716 (sync_old_<optab>_12): Likewise.
22717 (sync_new_<optab>_12): Likewise.
22718 (sync_nand_12): Likewise.
22719 (sync_old_nand_12): Likewise.
22720 (sync_new_nand_12): Likewise.
22721 (sync_sub<mode>): Likewise.
22722 (sync_old_add<mode>): Likewise.
22723 (sync_old_sub<mode>): Likewise.
22724 (sync_new_add<mode>): Likewise.
22725 (sync_new_sub<mode>): Likewise.
22726 (sync_<optab><mode>): Likewise.
22727 (sync_old_<optab><mode>): Likewise.
22728 (sync_new_<optab><mode>): Likewise.
22729 (sync_nand<mode>): Likewise.
22730 (sync_old_nand<mode>): Likewise.
22731 (sync_new_nand<mode>): Likewise.
22732 (sync_lock_test_and_set<mode>): Likewise.
22733 (test_and_set_12): Likewise.
22734 (atomic_compare_and_swap<mode>): Likewise.
22735 (atomic_exchange<mode>_llsc): Likewise.
22736 (atomic_fetch_add<mode>_llsc): Likewise.
22737 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22738 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22739 (umips_save_restore_pattern_p): Likewise.
22740 (umips_load_store_pair_p): Likewise.
22741 (umips_output_load_store_pair): Likewise.
22742 (umips_movep_target_p): Likewise.
22743 (umips_12bit_offset_address_p): Likewise.
22744 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22745 (mips_base_mips16): Rename this...
22746 (mips_base_compression_flags): ...to this. Update all uses.
22747 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22748 (mips_mips16_decl_p): Delete.
22749 (mips_nomips16_decl_p): Delete.
22750 (mips_get_compress_on_flags): New function.
22751 (mips_get_compress_off_flags): New function.
22752 (mips_get_compress_mode): New function.
22753 (mips_get_compress_on_name): New function.
22754 (mips_get_compress_off_name): New function.
22755 (mips_insert_attributes): Support multiple compression types.
22756 (mips_merge_decl_attributes): Likewise.
22757 (umips_12bit_offset_address_p): New function.
22758 (mips_start_function_definition): Emit .set micromips directive.
22759 (mips_call_may_need_jalx_p): New function.
22760 (mips_function_ok_for_sibcall): Add microMIPS support.
22761 (mips_print_operand_punctuation): Support short delay slots and
22763 (umips_swm_mask, umips_swm_encoding): New.
22764 (umips_build_save_restore): New function.
22765 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22766 (was_mips16_p): Remove.
22767 (old_compression_mode): New.
22768 (mips_set_compression_mode): New function.
22769 (mips_set_current_function): Add microMIPS support.
22770 (mips_option_override): Likewise.
22771 (umips_save_restore_pattern_p): New function.
22772 (umips_output_save_restore): New function.
22773 (umips_load_store_pair_p_1): New function.
22774 (umips_load_store_pair_p): New function.
22775 (umips_output_load_store_pair_1): New function.
22776 (umips_output_load_store_pair): New function.
22777 (umips_movep_target_p) New function.
22778 (mips_prepare_pch_save): Add microMIPS support.
22779 * config/mips/mips.h (TARGET_COMPRESSION): New.
22780 (TARGET_CPU_CPP_BUILTINS): Update macro
22781 to use new compression flags and to support microMIPS.
22782 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22783 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22784 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22785 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22786 (ASM_SPEC): Support mmicromips and mno-micromips.
22787 (M16STORE_REG_P): New macro.
22788 (MIPS_CALL): Support TARGET_MICROMIPS.
22789 (MICROMIPS_J): New macro.
22790 (mips_base_mips16): Rename this...
22791 (mips_base_compression_flags): ...to this.
22792 (UMIPS_12BIT_OFFSET_P): New macro.
22793 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22794 (MULTILIB_DIRNAMES): Likewise.
22795 2013-03-20 Richard Biener <rguenther@suse.de>
22797 PR tree-optimization/56661
22798 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22799 the result does not have to be distinct.
22801 2013-03-20 Richard Biener <rguenther@suse.de>
22803 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22806 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22807 Steven Bosscher <steven@gcc.gnu.org>
22809 PR rtl-optimization/56605
22810 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22812 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22815 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22816 that require movd instead of movq.
22818 2013-03-20 Richard Biener <rguenther@suse.de>
22820 * tree-ssa-structalias.c (struct variable_info): Add pointer
22821 to the first field of an aggregate with sub-vars. Make
22822 this and the pointer to the next subfield its ID.
22823 (vi_next): New function.
22824 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22825 storedanything_id, integer_id): Increment by one.
22826 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22827 get_call_clobber_vi): Adjust.
22828 (solution_set_expand): Simplify and speedup.
22829 (solution_set_add): Inline into ...
22830 (set_union_with_increment): ... this. Adjust accordingly.
22831 (do_sd_constraint): Likewise.
22832 (do_ds_constraint): Likewise.
22833 (do_complex_constraint): Simplify.
22834 (build_pred_graph): Adjust.
22835 (solve_graph): Likewise. Simplify and speedup.
22836 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22837 get_constraint_for_component_ref, get_constraint_for_1,
22838 first_vi_for_offset, first_or_preceding_vi_for_offset,
22839 create_function_info_for, create_variable_info_for_1,
22840 create_variable_info_for, intra_create_variable_infos): Adjust.
22841 (init_base_vars): Push NULL for ID zero.
22842 (compute_points_to_sets): Adjust.
22844 2013-03-20 Richard Biener <rguenther@suse.de>
22846 * cfgloop.c (verify_loop_structure): Streamline and avoid
22847 ICEing on corrupt loop tree.
22848 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22851 2013-03-20 Richard Biener <rguenther@suse.de>
22853 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22854 check whether an SSA update is needed.
22856 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22858 * config/mips/constraints.md (T): Rename to...
22862 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22863 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22865 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22867 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22868 (*subsi3_carryin_uxtw): Likewise.
22870 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22872 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22873 (*rorsi3_insn_uxtw): Likewise.
22875 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22877 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22878 (*extrsi5_insn_uxtw): Likewise.
22880 2013-03-19 Richard Biener <rguenther@suse.de>
22882 PR tree-optimization/56273
22883 * passes.c (init_optimization_passes): Move second VRP after DOM.
22885 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22887 * config/i386/i386.md (*movti_internal): Merge from
22888 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22889 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22890 nox64 isa attributes.
22892 2013-03-18 Richard Biener <rguenther@suse.de>
22894 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22896 (merge_node_constraints): Likewise.
22897 (build_succ_graph): Likewise.
22898 (valid_graph_edge): Inline into single caller.
22899 (unify_nodes): Likewise. Use bitmap_set_bit return value
22901 (scc_visit): Fix formatting and variable use.
22902 (do_sd_constraint): Use gcc_checking_assert.
22903 (do_ds_constraint): Likewise.
22904 (do_complex_constraint): Likewise.
22905 (condense_visit): Likewise. Cleanup.
22906 (dump_pred_graph): New function.
22907 (perform_var_substitution): Dump the pred-graph before
22908 variable substitution.
22909 (find_equivalent_node): Use gcc_checking_assert.
22910 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22912 2013-03-18 Richard Biener <rguenther@suse.de>
22914 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22915 Remove cond_expr_stmt_list argument and do not gimplify the
22917 (vect_loop_versioning): Adjust.
22918 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22919 Cleanup to use less temporaries.
22920 (vect_create_data_ref_ptr): Cleanup.
22922 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22924 PR tree-optimization/56635
22925 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22926 require types_compatible_p types.
22928 2013-03-18 Nick Clifton <nickc@redhat.com>
22930 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22931 spurious backslash.
22933 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22934 Add missing line to comment describing function.
22936 2013-03-18 Richard Biener <rguenther@suse.de>
22938 PR tree-optimization/56210
22939 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22940 Handle string / character search functions.
22941 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22943 2013-03-18 Richard Biener <rguenther@suse.de>
22945 PR middle-end/56483
22946 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22947 and implement properly.
22948 * gimple.h (gimple_cond_single_var_p): Remove.
22950 2013-03-18 Richard Biener <rguenther@suse.de>
22952 * tree-data-ref.h (find_data_references_in_loop): Declare.
22953 * tree-data-ref.c (get_references_in_stmt): Use a stack
22954 vector pre-allocated in the callers.
22955 (find_data_references_in_stmt): Adjust.
22956 (graphite_find_data_references_in_stmt): Likewise.
22957 (create_rdg_vertices): Likewise.
22958 (find_data_references_in_loop): Export.
22959 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22960 Compute dependences here...
22961 (vect_analyze_data_refs): ...not here. When we encounter
22962 a non-vectorizable data reference in basic-block vectorization
22963 truncate the data reference vector. Do not bother to
22964 fixup data-dependence information for gather loads.
22965 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22966 of data references, as reported.
22968 2013-03-18 Richard Biener <rguenther@suse.de>
22970 PR tree-optimization/3713
22971 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22972 has_constants and expr.
22973 (stmt_has_constants): Properly valueize SSA names when deciding
22974 whether the stmt has constants.
22976 2013-03-18 Richard Biener <rguenther@suse.de>
22978 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22979 whole function when there is nothing to do.
22980 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22981 * tree-vectorizer.c (vectorize_loops): Update virtual and
22982 loop-closed SSA once.
22983 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22985 2013-03-18 Richard Biener <rguenther@suse.de>
22987 PR middle-end/56113
22988 * domwalk.c (bb_postorder): New global static.
22989 (cmp_bb_postorder): New function.
22990 (walk_dominator_tree): Replace scheme imposing an order for
22991 visiting dominator sons by one sorting them at the time they
22992 are pushed on the stack.
22994 2013-03-18 Richard Biener <rguenther@suse.de>
22996 PR tree-optimization/39326
22997 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22998 (struct mem_ref): Replace mem member with ao_ref typed member.
22999 (MEM_ANALYZABLE): Adjust.
23000 (memref_eq): Likewise.
23001 (mem_ref_alloc): Likewise.
23002 (gather_mem_refs_stmt): Likewise.
23003 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23004 (execute_sm_if_changed_flag_set): Adjust.
23005 (execute_sm): Likewise.
23006 (ref_always_accessed_p): Likewise.
23007 (refs_independent_p): Likewise.
23008 (can_sm_ref_p): Likewise.
23010 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23013 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23014 return 1 even for !unsignedp.
23016 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23018 * config/i386/i386.md (isa): Add x64 and nox64.
23019 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23020 (*pushtf): Enable *roF alternative for x64 isa only.
23021 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23022 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23023 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23024 (*movtf_internal): Merge from *movtf_internal_rex64 and
23025 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23026 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23027 nox64 isa attributes.
23028 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23029 nox64 isa attributes.
23030 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23032 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23034 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23036 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23039 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23040 class names. Remove trailing comma after "ALL_REGS".
23042 2013-03-16 Jan Hubicka <jh@suse.cz>
23044 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23045 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23046 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23047 of cgraph_get_create_node.
23048 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23050 2013-03-16 Jason Merrill <jason@redhat.com>
23053 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23054 with DW_AT_default_value.
23056 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23058 * BASE-VER: Set to 4.9.0.
23060 2013-03-14 Andi Kleen <ak@linux.intel.com>
23063 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23064 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23065 Document _x* TSX intrinsics.
23067 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23068 David Holsgrove <david.holsgrove@xilinx.com>
23070 * configure.ac: Add MicroBlaze TLS support detection.
23071 * configure: Regenerate.
23072 * config/microblaze/microblaze-protos.h
23073 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23074 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23075 * config/microblaze/microblaze.c (microblaze_address_type): Add
23076 ADDRESS_TLS and tls_reloc address types.
23077 (microblaze_address_info): Add tls_reloc.
23078 (TARGET_HAVE_TLS): Define.
23079 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23080 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23081 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23082 load_tls_operand, microblaze_call_tls_get_addr,
23083 microblaze_legitimize_tls_address): New functions.
23084 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23085 (get_base_reg): Use microblaze_tls_symbol_p.
23086 (microblaze_classify_address): Handle TLS.
23087 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23088 label_mentioned_p and microblaze_tls_referenced_p.
23089 (microblaze_legitimize_address): Handle TLS.
23090 (microblaze_address_insns): Handle ADDRESS_TLS.
23091 (pic_address_needs_scratch): Handle TLS.
23092 (print_operand_address): Handle TLS.
23093 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23094 (microblaze_expand_move): Handle TLS.
23095 (microblaze_legitimate_constant_p): Check
23096 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23097 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23098 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23099 (PIC_OFFSET_TABLE_REGNUM): Set.
23100 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23101 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23102 (addsi3, movsi_internal2, movdf_internal): Update constraints
23103 * config/microblaze/predicates.md (arith_plus_operand): Define
23104 (move_operand): Redefine as move_src_operand,
23105 check microblaze_tls_referenced_p.
23107 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23109 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23110 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23112 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23114 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23117 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23119 PR tree-optimization/53265
23120 * common.opt (Waggressive-loop-optimizations): New option.
23121 * tree-ssa-loop-niter.c: Include tree-pass.h.
23122 (do_warn_aggressive_loop_optimizations): New function.
23123 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23124 if number_of_latch_executions returned constant.
23125 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23126 early. If number_of_latch_executions returned constant, set
23127 nb_iterations_upper_bound back to it.
23128 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23130 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23131 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23133 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23134 (MULTILIB_OSDIRNAMES): Set.
23135 * genmultilib: If defaultosdirname doesn't start with :: , set
23136 defaultosdirname2 instead, clear it and emit two . multilib_raw
23137 entries instead of just one.
23139 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23141 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23142 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23143 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23144 (SUBTARGET_OVERRIDE_OPTIONS): New.
23146 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23149 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23150 (musermode): Convert to Var(TARGET_USERMODE).
23151 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23152 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23153 * config/sh/sh.c (sh_option_override): Use
23154 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23155 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23157 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23159 (udivsi3_i4_single, divsi3_i4_single): Use
23160 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23162 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23164 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23167 2013-03-13 Richard Biener <rguenther@suse.de>
23169 PR tree-optimization/56608
23170 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23171 calls when vectorizing basic-blocks.
23173 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23176 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23179 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23181 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23183 2013-03-11 Jan Hubicka <jh@suse.cz>
23186 * lto-streamer-out.c (output_symbol_p): Skip references from
23187 constructors of external variables.
23189 2013-03-11 Jan Hubicka <jh@suse.cz>
23191 PR middle-end/56571
23192 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23194 * emit-rtl.c (verify_rtx_sharing): Likewise.
23195 (copy_insn_1): Likewise.
23196 * rtl.c (copy_rtx): Likewise.
23198 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23201 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23202 output_operand_lossage message.
23204 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23207 * arm.c (shift_op): Validate RTL pattern on the fly.
23208 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23211 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23214 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23215 new call patterns that clobber %fr12.
23216 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23217 split and postreload patterns.
23218 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23219 registers %fr12 and %fr12R as call used.
23221 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23223 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23224 (canon_address, record_store, replace_read, check_mem_read_rtx,
23225 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23226 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23227 rest_of_handle_dse): Likewise.
23229 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23231 PR middle-end/56524
23232 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23234 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23235 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23236 (save_optabs_if_changed): Replace with...
23237 (init_tree_optimization_optabs): ...this.
23238 * optabs.c (save_optabs_if_changed): Rename to...
23239 (init_tree_optimization_optabs): ...this. Take the optimization node
23240 as argument. Do nothing if the base optabs are already correct.
23241 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23242 to recompute optabs.
23243 * function.h (function): Remove optabs field.
23244 * function.c (invoke_set_current_function_hook): Call
23245 init_tree_optimization_optabs. Use the result to initialize
23248 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23250 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23251 if GTMA_HAS_NO_INSTRUMENTATION.
23252 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23253 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23254 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23255 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23256 GTMA_HAS_NO_INSTRUMENTATION.
23258 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23260 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23263 2013-03-08 Marek Polacek <polacek@redhat.com>
23264 Jakub Jelinek <jakub@redhat.com>
23266 PR tree-optimization/56478
23267 * predict.c (is_comparison_with_loop_invariant_p): Change the
23268 type of loop_step to tree.
23269 (predict_loops): Adjust.
23270 (predict_iv_comparison): Perform the computations on double_ints.
23272 2013-03-08 Richard Biener <rguenther@suse.de>
23274 PR tree-optimization/56570
23275 * tree-cfg.c (verify_expr_location_1): Verify locations for
23277 * tree-sra.c (create_access_replacement): Strip locations
23278 from DECL_DEBUG_EXPRs.
23280 2013-03-08 Richard Biener <rguenther@suse.de>
23282 * tree-inline.c (expand_call_inline): Do not associate
23283 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23284 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23286 2013-03-08 Richard Biener <rguenther@suse.de>
23288 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23289 or block changes with -Og. Fix for location / block encoding
23290 changes and PHI arguments with locations.
23292 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23294 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23296 (struct output_info): Likewise.
23297 (register_overhead): Remove bad gcc_assert.
23298 (bitmap_find_bit): If there is only a single bitmap element, do not
23299 count a miss as a search.
23300 (print_statistics): Update for counter type changes.
23301 (dump_bitmap_statistics): Likewise. Print headers such that they
23302 are properly lined up with the printed counters.
23304 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23306 PR tree-optimization/56559
23307 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23308 check that it has only a single use.
23310 2013-03-07 Richard Biener <rguenther@suse.de>
23312 * doc/invoke.texi (fwhole-program): Discourage use in combination
23315 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23317 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23319 PR tree-optimization/56539
23320 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23321 instead of GSI_CONTINUE_LINKING as last argument to
23322 force_gimple_operand_gsi. Adjust function comment.
23324 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23327 PR middle-end/56548
23328 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23329 promoted mode, convert the result back to the original mode.
23331 2013-03-06 Richard Biener <rguenther@suse.de>
23333 PR middle-end/56294
23334 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23335 (insert_updated_phi_nodes_compare_uids): New function.
23336 (update_ssa): Sort symbols_to_rename after UID before
23337 traversing it to insert PHI nodes.
23339 2013-03-06 Richard Biener <rguenther@suse.de>
23341 PR middle-end/50494
23342 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23343 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23346 2013-02-13 Richard Biener <rguenther@suse.de>
23349 * varasm.c (output_constant_def_1): Get the decl representing
23350 the constant as argument.
23351 (output_constant_def): Wrap output_constant_def_1.
23352 (make_decl_rtl): Use output_constant_def_1 with the decl
23353 representing the constant.
23354 (build_constant_desc): Optionally re-use a decl already
23355 representing the constant.
23356 (tree_output_constant_def): Adjust.
23358 2013-03-06 Joey Ye <joey.ye@arm.com>
23361 * gcc.c (convert_white_space): New function.
23362 (main): Handles white space in function name.
23364 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23367 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23368 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23369 to SH_DIV_CALL_TABLE for TARGET_SH2.
23370 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23372 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23373 call-table options.
23375 2013-03-05 Sterling Augustine <saugustine@google.com>
23376 Cary Coutant <ccoutant@google.com>
23379 * dwarf2out.c (resolve_addr): Don't call
23380 remove_loc_list_addr_table_entries a second time for the same
23383 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23386 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23387 (avoid_complex_debug_insns): New function.
23388 (expand_debug_locations): Call it.
23390 PR rtl-optimization/56484
23391 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23392 lifetimes of hard registers on small register class machines.
23394 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23396 * config/microblaze/microblaze-protos.h: Rename
23397 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23398 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23400 (microblaze_fast_interrupt_function_p): New function.
23401 (microblaze_is_interrupt_handler): Rename to
23402 microblaze_is_interrupt_variant and add fast_interrupt check.
23403 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23404 (save_restore_insns): Likewise.
23405 (compute_frame_size): Likewise.
23406 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23407 (microblaze_globalize_label): Likewise.
23408 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23409 * config/microblaze/microblaze.md: Use wrapper
23410 microblaze_is_interrupt_variant.
23412 2013-03-05 Kai Tietz <ktietz@redhat.com>
23414 * sdbout.c (sdbout_one_type): Switch to current function's section
23415 supporting cold/hot.
23417 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23419 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23422 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23424 PR middle-end/56461
23425 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23426 if VALGRIND_GET_VBITS is defined, temporarily make object
23427 memory all defined, and restore previous valgrind addressability
23428 and definability afterwards. Free this_object at the end.
23430 PR middle-end/56461
23431 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23432 right before calling lra_create_live_ranges, also call it
23433 when clearing live_p. Only call lra_clear_live_ranges
23434 at the end if live_p.
23436 PR middle-end/56461
23437 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23439 2013-03-05 Richard Biener <rguenther@suse.de>
23441 PR tree-optimization/56521
23442 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23445 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23448 * except.h (remove_unreachable_eh_regions): New prototype.
23449 * except.c (remove_eh_handler_splicer): New function, split out
23450 of remove_eh_handler.
23451 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23452 warning about running it on many EH regions one at a time.
23453 (remove_unreachable_eh_regions_worker): New function, walk the
23454 EH tree in depth-first order and remove non-marked regions.
23455 (remove_unreachable_eh_regions): New function.
23456 * tree-eh.c (mark_reachable_handlers): New function, split out
23457 from remove_unreachable_handlers.
23458 (remove_unreachable_handlers): Use mark_reachable_handlers and
23459 remove_unreachable_eh_regions.
23460 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23461 and remove_unreachable_eh_regions.
23463 2013-03-05 Richard Biener <rguenther@suse.de>
23465 PR middle-end/56525
23466 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23467 not freeing them until the end.
23469 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23471 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23473 2013-03-05 Richard Biener <rguenther@suse.de>
23475 PR tree-optimization/56270
23476 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23477 of loads after scheduling an SLP instance.
23479 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23481 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23483 (check_gcc_parallelize): Run guality.exp as a separate job from
23484 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23485 struct-layout-1.exp with stackalign.exp.
23487 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23489 PR middle-end/56461
23490 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23491 load_index sbitmap even if some bit in it isn't set.
23493 PR middle-end/56461
23494 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23495 (discover_iteration_bound_by_body_walk): Change queues to
23496 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23497 spelling in comment. Call safe_push on queues[bound_index] directly.
23498 Release queues[queue_index] in every iteration unconditionally.
23499 Release bounds vector.
23501 PR middle-end/56461
23502 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23503 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23504 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23507 2013-03-05 Richard Biener <rguenther@suse.de>
23510 * tree-inline.c (remap_blocks_to_null): New function.
23511 (expand_call_inline): When expanding a call stmt without
23512 an associated block inline remap all callee blocks to NULL.
23514 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23516 PR rtl-optimization/56494
23517 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23518 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23519 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23521 PR middle-end/56461
23522 * sel-sched-ir.c (free_sched_pools): Release
23523 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23524 if succs_info_pool.max_top isn't -1.
23527 * opts.c (opts_obstack, opts_concat): Moved to...
23528 * opts-common.c (opts_obstack, opts_concat): ... here.
23530 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23532 PR middle-end/56461
23533 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23535 2013-03-04 Martin Jambor <mjambor@suse.cz>
23537 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23538 all appropriate places.
23540 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23542 PR tree-optimization/56424
23543 * ipa-split.c (split_function): Do not set the RSO flag if result is
23544 not by reference and its type is a register type.
23546 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23548 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23549 (microblaze_legitimate_pic_operand): Likewise
23550 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23551 new function microblaze_legitimate_pic_operand
23552 * config/microblaze/microblaze-protos.h
23553 (microblaze_legitimate_pic_operand): Declare.
23555 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23557 * config/microblaze/predicates.md (call_insn_simple_operand):
23558 New predicate for supported rtx code types.
23559 * config/microblaze/microblaze.md (call_internal1): Use
23560 call_insn_simple_operand predicate.
23562 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23564 PR middle-end/56461
23565 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23566 partitions.ordered_remove.
23568 PR middle-end/56461
23569 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23570 vec_oprnds0.create (1) for modifier == NONE.
23572 PR middle-end/56461
23573 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23574 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23575 vec_oprnds1 right before pushing anything to it for
23578 PR middle-end/56461
23579 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23580 set nbbs to 0 instead of having separate code path.
23581 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23582 instead of false as last argument if returning NULL.
23584 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23586 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23587 the attribute is now called "target" instead of "option".
23588 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23589 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23590 attribute/pragma name for TARGET_OPTION_VALID_P and
23591 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23592 * doc/tm.texi: Regenerated.
23594 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23596 * config/microblaze/microblaze.c:
23597 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23598 * config/microblaze/microblaze.h: Add -mxl-reorder to
23600 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23601 instructions emitted if TARGET_REORDER.
23602 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23603 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23606 2013-03-01 Xinliang David Li <davidxl@google.com>
23608 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23611 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23613 PR middle-end/56461
23614 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23615 vector even when returning true. Fix up function comment formatting.
23617 PR middle-end/56461
23618 * ira-build.c (ira_loop_nodes_count): New variable.
23619 (create_loop_tree_nodes): Initialize it.
23620 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23622 PR middle-end/56461
23623 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23624 method on dr_chain and result_chain.
23625 * tree-vect-stmts.c (vectorizable_store): Only call
23626 result_chain.create if j == 0.
23628 PR middle-end/56461
23629 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23630 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23631 before overwriting it.
23633 2013-03-01 Tobias Burnus <burnus@net-b.de>
23635 * doc/extended.texi (C Extensions): Change order in @menu
23637 (Other MIPS Built-in Functions): Move last MIPS entry before
23638 "picoChip Built-in Functions".
23639 (SH Built-in Functions): Move after RX Built-in Functions.
23640 * doc/gcc.texi (Introduction): Change order in @menu
23642 * doc/md.texi (Constraints): Ditto.
23643 * gty.texi (Type Information): Ditto.
23644 (User-provided marking routines for template types): Make
23646 * doc/invoke.texi (AArch64 Options): Move before
23647 "Adapteva Epiphany Options".
23649 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23650 Jakub Jelinek <jakub@redhat.com>
23653 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23654 no_address_safety_analysis attribute.
23655 * doc/extend.texi (no_address_safety_attribute): Rename to
23656 no_sanitize_address attribute, mention no_address_safety_analysis
23657 attribute as deprecated alias.
23659 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23661 PR middle-end/56461
23662 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23663 type to vec<vec<tree> > *.
23664 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23665 to be vec<tree> instead of vec<tree> *, set vec_defs
23666 to vNULL and call vec_defs.create (number_of_vects), adjust other
23668 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23669 vectorizable_condition): Adjust vect_get_slp_defs callers.
23671 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23673 * config/aarch64/aarch64.c
23674 (aarch64_float_const_representable): Remove unused variable.
23676 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23678 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23680 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23682 * config/aarch64/aarch64-builtins.c
23683 (aarch64_init_simd_builtins): Make static.
23685 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23687 * config/aarch64/aarch64.c
23688 (aarch64_simd_make_constant): Make static.
23690 2013-02-28 Martin Jambor <mjambor@suse.cz>
23692 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23693 with no initialization to the RHS of debug statements.
23695 2013-02-28 Martin Jambor <mjambor@suse.cz>
23697 PR tree-optimization/56294
23698 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23700 (get_access_replacement): Do not call create_access_replacement.
23701 Assert a replacement exists.
23702 (get_repl_default_def_ssa_name): Create the replacement declaration
23705 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23707 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23708 final_end_function.
23710 2013-02-28 Marek Polacek <polacek@redhat.com>
23712 PR rtl-optimization/56466
23713 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23714 if we're changing a loop.
23715 (peel_loops_completely): Likewise.
23717 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23720 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23722 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23725 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23726 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23727 INTX_FTYPE_FX, FX_FTYPE_INTX.
23728 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23730 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23732 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23733 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23734 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23735 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23736 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23737 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23738 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23739 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23740 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23741 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23742 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23743 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23744 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23745 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23746 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23747 (avrxmega6): Increase max flash segments from 5 to 6.
23748 * config/avr/t-multilib: Regenerate.
23749 * config/avr/avr-tables.opt: Regenerate.
23750 * doc/avr-mmcu.texi: Regenerate.
23752 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23754 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23755 (avr_device_to_arch): Rename to avr_device_to_ld.
23756 (avr_device_to_as): New prototype.
23757 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23758 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23759 * config/avr/driver-avr.c (avr_device_to_as): New.
23760 (avr_device_to_arch): Rename to avr_device_to_ld.
23762 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23764 PR middle-end/56461
23765 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23766 method on dr_chain and result_chain.
23768 PR middle-end/56461
23769 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23770 pointer_set_destroy on not_executed_last_iteration.
23772 PR middle-end/56461
23773 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23775 PR middle-end/56461
23776 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23777 FOR_EACH_DEFINED_FUNCTION when freeing state.
23779 PR middle-end/56461
23780 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23782 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23785 PR middle-end/56461
23786 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23787 known_aggs[i].items and release known_aggs vector.
23789 PR middle-end/56461
23790 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23792 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23794 * config/microblaze/microblaze.c (microblaze_emit_compare):
23795 Use xor for EQ/NE comparisions.
23796 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23797 (cbranchsf4): Adjust operator to comparison_operator.
23799 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23801 PR middle-end/56461
23802 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23804 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23805 vec_safe_push, always update *slot.
23806 (redirect_edge_var_map_clear): Use vec_free.
23807 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23808 (free_var_map_entry): Use vec_free.
23809 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23810 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23812 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23814 PR middle-end/45472
23815 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23816 when the may_trap_p bit of the exprs being merged differs.
23817 Reorder tests for speculativeness in the logical and operator.
23819 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23821 * incpath.c (add_standard_paths): Use reconcat instead of concat
23822 where appropriate and avoid leaking memory.
23824 * opts.h: Include obstack.h.
23825 (opts_concat): New prototype.
23826 (opts_obstack): New declaration.
23827 * opts.c (opts_concat): New function.
23828 (opts_obstack): New variable.
23829 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23830 (finish_options): Use opts_concat instead of concat
23831 and XOBNEWVEC instead of XNEWVEC.
23832 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23833 generate_option): Likewise.
23834 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23835 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23838 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23839 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23841 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23843 PR middle-end/56461
23844 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23846 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23848 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23849 (arm_block_move_unaligned_straight): Likewise.
23850 (arm_adjust_block_mem): Likewise.
23852 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23855 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23856 temp, cond and label.
23857 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23860 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23861 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23864 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23865 of prologue/epilogue functions.
23868 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23869 Remove unused variable cfa_offset.
23870 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23873 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23874 type promotion to unsigned.
23877 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23878 for HOST_WIDE_INT of 32 bit / same size as int.
23879 (arm_block_move_unaligned_straight): Likewise.
23880 (arm_adjust_block_mem): Likewise.
23883 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23886 2013-02-26 Marek Polacek <polacek@redhat.com>
23888 PR tree-optimization/56426
23889 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23891 2013-02-26 Richard Biener <rguenther@suse.de>
23894 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23895 unused variable loops.
23897 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23899 PR tree-optimization/56448
23900 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23901 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23902 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23903 later operands of the references, or even first operand for
23904 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23906 PR tree-optimization/56443
23907 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23908 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23909 to type_for_mode langhook.
23911 2013-02-25 Matt Turner <mattst88@gmail.com>
23913 * doc/invoke.texi: Document r4700.
23915 2013-02-25 Richard Biener <rguenther@suse.de>
23917 PR tree-optimization/56175
23918 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23920 (simplify_bitwise_binary): ... here. Also guard the conversion
23921 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23923 2013-02-25 Catherine Moore <clm@codesourcery.com>
23926 2013-02-24 Catherine Moore <clm@codesourcery.com>
23927 Maciej W. Rozycki <macro@codesourcery.com>
23928 Tom de Vries <tom@codesourcery.com>
23929 Nathan Sidwell <nathan@codesourcery.com>
23930 Iain Sandoe <iain@codesourcery.com>
23931 Nathan Froyd <froydnj@codesourcery.com>
23932 Chao-ying Fu <fu@mips.com>
23934 * doc/extend.texi: (micromips, nomicromips, nocompression):
23935 Document new function attributes.
23936 * doc/invoke.texi (minterlink-compressed, mmicromips,
23937 m14k, m14ke, m14kec): Document new options.
23938 (minterlink-mips16): Update documentation.
23939 * doc/md.texi (ZC, ZD): Document new constraints.
23940 * configure.ac (gcc_cv_as_micromips): Check if linker
23941 supports the .set micromips directive.
23942 * configure: Regenerate.
23943 * config.in: Regenerate.
23944 * config/mips/mips-tables.opt: Regenerate.
23945 * config/mips/micromips.md: New file.
23946 * constraints.md (ZC, AD): New constraints.
23947 * config/mips/predicates.md (movep_src_register): New predicate.
23948 (movep_src_operand): New predicate.
23949 (non_volatile_mem_operand): New predicate.
23950 * config/mips/mips.md (multimem): New type.
23951 (length): Differentiate between 17-bit and 18-bit branch offsets.
23952 (MOVEP1, MOVEP2): New mode iterator.
23953 (mov_<load>l): Use ZC constraint.
23954 (mov_<load>r): Likewise.
23955 (mov_<store>l): Likewise.
23956 (mov_<store>r): Likewise.
23957 (*branch_equality<mode>_inverted): Add microMIPS support.
23958 (*branch_equality<mode>): Likewise.
23959 (*jump_absolute): Likewise.
23960 (indirect_jump_<mode>): Likewise.
23961 (tablejump_<mode>): Likewise.
23962 (<optab>_internal): Likewise.
23963 (sibcall_internal): Likewise.
23964 (sibcall_value_internal): Likewise.
23965 (prefetch): Use constraint ZD.
23966 * config/mips/mips.opt (minterlink-compressed): New option.
23967 (minterlink-mips16): Now an alias for minterlink-compressed.
23968 (mmicromips): New option.
23969 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23970 (compare_and_swap_12): Likewise.
23971 (sync_add<mode>): Likewise.
23972 (sync_<optab>_12): Likewise.
23973 (sync_old_<optab>_12): Likewise.
23974 (sync_new_<optab>_12): Likewise.
23975 (sync_nand_12): Likewise.
23976 (sync_old_nand_12): Likewise.
23977 (sync_new_nand_12): Likewise.
23978 (sync_sub<mode>): Likewise.
23979 (sync_old_add<mode>): Likewise.
23980 (sync_old_sub<mode>): Likewise.
23981 (sync_new_add<mode>): Likewise.
23982 (sync_new_sub<mode>): Likewise.
23983 (sync_<optab><mode>): Likewise.
23984 (sync_old_<optab><mode>): Likewise.
23985 (sync_new_<optab><mode>): Likewise.
23986 (sync_nand<mode>): Likewise.
23987 (sync_old_nand<mode>): Likewise.
23988 (sync_new_nand<mode>): Likewise.
23989 (sync_lock_test_and_set<mode>): Likewise.
23990 (test_and_set_12): Likewise.
23991 (atomic_compare_and_swap<mode>): Likewise.
23992 (atomic_exchange<mode>_llsc): Likewise.
23993 (atomic_fetch_add<mode>_llsc): Likewise.
23994 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23995 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23996 (umips_save_restore_pattern_p): Likewise.
23997 (umips_load_store_pair_p): Likewise.
23998 (umips_output_load_store_pair): Likewise.
23999 (umips_movep_target_p): Likewise.
24000 (umips_12bit_offset_address_p): Likewise.
24001 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24002 (mips_base_mips16): Rename this...
24003 (mips_base_compression_flags): ...to this. Update all uses.
24004 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24005 (mips_mips16_decl_p): Delete.
24006 (mips_nomips16_decl_p): Delete.
24007 (mips_get_compress_on_flags): New function.
24008 (mips_get_compress_off_flags): New function.
24009 (mips_get_compress_mode): New function.
24010 (mips_get_compress_on_name): New function.
24011 (mips_get_compress_off_name): New function.
24012 (mips_insert_attributes): Support multiple compression types.
24013 (mips_merge_decl_attributes): Likewise.
24014 (umips_12bit_offset_address_p): New function.
24015 (mips_start_function_definition): Emit .set micromips directive.
24016 (mips_call_may_need_jalx_p): New function.
24017 (mips_function_ok_for_sibcall): Add microMIPS support.
24018 (mips_print_operand_punctuation): Support short delay slots and
24020 (umips_swm_mask, umips_swm_encoding): New.
24021 (umips_build_save_restore): New function.
24022 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24023 (was_mips16_p): Remove.
24024 (old_compression_mode): New.
24025 (mips_set_compression_mode): New function.
24026 (mips_set_current_function): Add microMIPS support.
24027 (mips_option_override): Likewise.
24028 (umips_save_restore_pattern_p): New function.
24029 (umips_output_save_restore): New function.
24030 (umips_load_store_pair_p_1): New function.
24031 (umips_load_store_pair_p): New function.
24032 (umips_output_load_store_pair_1): New function.
24033 (umips_output_load_store_pair): New function.
24034 (umips_movep_target_p) New function.
24035 (mips_prepare_pch_save): Add microMIPS support.
24036 * config/mips/mips.h (TARGET_COMPRESSION): New.
24037 (TARGET_CPU_CPP_BUILTINS): Update macro
24038 to use new compression flags and to support microMIPS.
24039 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24040 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24041 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24042 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24043 (ASM_SPEC): Support mmicromips and mno-micromips.
24044 (M16STORE_REG_P): New macro.
24045 (MIPS_CALL): Support TARGET_MICROMIPS.
24046 (MICROMIPS_J): New macro.
24047 (mips_base_mips16): Rename this...
24048 (mips_base_compression_flags): ...to this.
24049 (UMIPS_12BIT_OFFSET_P): New macro.
24050 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24051 (MULTILIB_DIRNAMES): Likewise.
24053 2013-02-25 Tom de Vries <tom@codesourcery.com>
24055 PR rtl-optimization/56131
24056 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24057 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24058 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24060 2013-02-25 Tobias Burnus <burnus@net-b.de>
24062 * doc/invoke.texi (-fsanitize=): Move from optimization
24063 to debugging options.
24065 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24067 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24069 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24070 Alexander Monakov <amonakov@ispras.ru>
24072 PR middle-end/56077
24073 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24074 flush pending lists also on non-jumps. Adjust comment.
24076 2013-02-24 Catherine Moore <clm@codesourcery.com>
24077 Maciej W. Rozycki <macro@codesourcery.com>
24078 Tom de Vries <tom@codesourcery.com>
24079 Nathan Sidwell <nathan@codesourcery.com>
24080 Iain Sandoe <iain@codesourcery.com>
24081 Nathan Froyd <froydnj@codesourcery.com>
24082 Chao-ying Fu <fu@mips.com>
24084 * doc/extend.texi: (micromips, nomicromips, nocompression):
24085 Document new function attributes.
24086 * doc/invoke.texi (minterlink-compressed, mmicromips,
24087 m14k, m14ke, m14kec): Document new options.
24088 (minterlink-mips16): Update documentation.
24089 * doc/md.texi (ZC, ZD): Document new constraints.
24090 * configure.ac (gcc_cv_as_micromips): Check if linker
24091 supports the .set micromips directive.
24092 * configure: Regenerate.
24093 * config.in: Regenerate.
24094 * config/mips/mips-tables.opt: Regenerate.
24095 * config/mips/micromips.md: New file.
24096 * constraints.md (ZC, AD): New constraints.
24097 * config/mips/predicates.md (movep_src_register): New predicate.
24098 (movep_src_operand): New predicate.
24099 (non_volatile_mem_operand): New predicate.
24100 * config/mips/mips.md (multimem): New type.
24101 (length): Differentiate between 17-bit and 18-bit branch offsets.
24102 (MOVEP1, MOVEP2): New mode iterator.
24103 (mov_<load>l): Use ZC constraint.
24104 (mov_<load>r): Likewise.
24105 (mov_<store>l): Likewise.
24106 (mov_<store>r): Likewise.
24107 (*branch_equality<mode>_inverted): Add microMIPS support.
24108 (*branch_equality<mode>): Likewise.
24109 (*jump_absolute): Likewise.
24110 (indirect_jump_<mode>): Likewise.
24111 (tablejump_<mode>): Likewise.
24112 (<optab>_internal): Likewise.
24113 (sibcall_internal): Likewise.
24114 (sibcall_value_internal): Likewise.
24115 (prefetch): Use constraint ZD.
24116 * config/mips/mips.opt (minterlink-compressed): New option.
24117 (minterlink-mips16): Now an alias for minterlink-compressed.
24118 (mmicromips): New option.
24119 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24120 (compare_and_swap_12): Likewise.
24121 (sync_add<mode>): Likewise.
24122 (sync_<optab>_12): Likewise.
24123 (sync_old_<optab>_12): Likewise.
24124 (sync_new_<optab>_12): Likewise.
24125 (sync_nand_12): Likewise.
24126 (sync_old_nand_12): Likewise.
24127 (sync_new_nand_12): Likewise.
24128 (sync_sub<mode>): Likewise.
24129 (sync_old_add<mode>): Likewise.
24130 (sync_old_sub<mode>): Likewise.
24131 (sync_new_add<mode>): Likewise.
24132 (sync_new_sub<mode>): Likewise.
24133 (sync_<optab><mode>): Likewise.
24134 (sync_old_<optab><mode>): Likewise.
24135 (sync_new_<optab><mode>): Likewise.
24136 (sync_nand<mode>): Likewise.
24137 (sync_old_nand<mode>): Likewise.
24138 (sync_new_nand<mode>): Likewise.
24139 (sync_lock_test_and_set<mode>): Likewise.
24140 (test_and_set_12): Likewise.
24141 (atomic_compare_and_swap<mode>): Likewise.
24142 (atomic_exchange<mode>_llsc): Likewise.
24143 (atomic_fetch_add<mode>_llsc): Likewise.
24144 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24145 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24146 (umips_save_restore_pattern_p): Likewise.
24147 (umips_load_store_pair_p): Likewise.
24148 (umips_output_load_store_pair): Likewise.
24149 (umips_movep_target_p): Likewise.
24150 (umips_12bit_offset_address_p): Likewise.
24151 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24152 (mips_base_mips16): Rename this...
24153 (mips_base_compression_flags): ...to this. Update all uses.
24154 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24155 (mips_mips16_decl_p): Delete.
24156 (mips_nomips16_decl_p): Delete.
24157 (mips_get_compress_on_flags): New function.
24158 (mips_get_compress_off_flags): New function.
24159 (mips_get_compress_mode): New function.
24160 (mips_get_compress_on_name): New function.
24161 (mips_get_compress_off_name): New function.
24162 (mips_insert_attributes): Support multiple compression types.
24163 (mips_merge_decl_attributes): Likewise.
24164 (umips_12bit_offset_address_p): New function.
24165 (mips_start_function_definition): Emit .set micromips directive.
24166 (mips_call_may_need_jalx_p): New function.
24167 (mips_function_ok_for_sibcall): Add microMIPS support.
24168 (mips_print_operand_punctuation): Support short delay slots and
24170 (umips_swm_mask, umips_swm_encoding): New.
24171 (umips_build_save_restore): New function.
24172 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24173 (was_mips16_p): Remove.
24174 (old_compression_mode): New.
24175 (mips_set_compression_mode): New function.
24176 (mips_set_current_function): Add microMIPS support.
24177 (mips_option_override): Likewise.
24178 (umips_save_restore_pattern_p): New function.
24179 (umips_output_save_restore): New function.
24180 (umips_load_store_pair_p_1): New function.
24181 (umips_load_store_pair_p): New function.
24182 (umips_output_load_store_pair_1): New function.
24183 (umips_output_load_store_pair): New function.
24184 (umips_movep_target_p) New function.
24185 (mips_prepare_pch_save): Add microMIPS support.
24186 * config/mips/mips.h (TARGET_COMPRESSION): New.
24187 (TARGET_CPU_CPP_BUILTINS): Update macro
24188 to use new compression flags and to support microMIPS.
24189 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24190 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24191 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24192 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24193 (ASM_SPEC): Support mmicromips and mno-micromips.
24194 (M16STORE_REG_P): New macro.
24195 (MIPS_CALL): Support TARGET_MICROMIPS.
24196 (MICROMIPS_J): New macro.
24197 (mips_base_mips16): Rename this...
24198 (mips_base_compression_flags): ...to this.
24199 (UMIPS_12BIT_OFFSET_P): New macro.
24200 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24201 (MULTILIB_DIRNAMES): Likewise.
24203 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24206 * target-globals.c (save_target_globals): For init_reg_sets and
24207 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24209 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24211 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24212 * config/aarch64/t-aarch64
24213 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24215 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24217 PR inline-asm/56148
24218 * lra-constraints.c (process_alt_operands): Reload operand
24219 conflicting with earlier clobber only if no more other conflicting
24222 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24225 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24226 if not linking a shared library.
24228 2013-02-22 Seth LaForge <sethml@google.com>
24230 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24232 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24234 * config/arm/arm.md (split for extendsidi): Update condition.
24235 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24236 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24237 (qhs_zextenddi_cstr): Likewise.
24239 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24241 PR middle-end/56420
24242 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24243 avoid signed wrapping.
24244 (expand_mult): Handle properly multiplication by
24245 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24246 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24247 in the compiler if coeff is HOST_WIDE_INT_MIN.
24248 (expand_divmod): Don't make ext_op1 static, change it's type to
24249 uhwi. Avoid undefined behavior in -INTVAL (op1).
24251 PR rtl-optimization/50339
24252 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24254 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24255 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24256 into splitting_ashiftrt field.
24257 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24259 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24262 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24264 PR middle-end/56108
24265 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24266 are sure to go irrevocable.
24268 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24270 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24271 scalars are valid operands.
24273 2013-02-21 Martin Jambor <mjambor@suse.cz>
24275 PR tree-optimization/56310
24276 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24277 only matching indices and non-negative final offsets.
24278 (intersect_aggregates_with_edge): Pass src_idx to
24279 agg_replacements_to_vector. Pass src_idx insstead of index to
24280 intersect_with_agg_replacements.
24282 2013-02-21 Martin Jambor <mjambor@suse.cz>
24284 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24285 instead of hard-wired defaults.
24287 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24289 * doc/invoke.texi (MIPS Options): Update documentation of the
24290 floating-point multiply-accumulate instruction restrictions.
24292 2013-02-21 Kostya Serebryany <kcc@google.com>
24294 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24295 asan_shadow_offset on x86_64 linux.
24297 2013-02-21 Richard Biener <rguenther@suse.de>
24299 PR tree-optimization/56415
24301 2013-02-11 Richard Biener <rguenther@suse.de>
24303 PR tree-optimization/56273
24304 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24307 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24310 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24313 PR inline-asm/56405
24314 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24315 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24317 2013-02-20 Jan Hubicka <jh@suse.cz>
24319 PR tree-optimization/56265
24320 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24321 when target is referenced for first time.
24323 2013-02-20 Richard Biener <rguenther@suse.de>
24325 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24326 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24327 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24328 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24329 not return anything.
24330 (rename_ssa_copies): Do not remove unused locals.
24331 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24332 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24333 * passes.c (execute_function_todo): Do not schedule unused locals
24334 removal if cleanup_tree_cfg did something.
24335 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24336 about the number of removed locals.
24338 2013-02-20 Richard Biener <rguenther@suse.de>
24340 PR tree-optimization/56398
24341 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24343 2013-02-20 Martin Jambor <mjambor@suse.cz>
24345 PR tree-optimization/55334
24346 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24347 restricted pointers to arrays.
24349 2013-02-20 Richard Biener <rguenther@suse.de>
24350 Jakub Jelinek <jakub@redhat.com>
24352 PR tree-optimization/56396
24353 * tree-ssa-ccp.c (n_const_val): New static variable.
24354 (get_value): Return NULL for SSA names we don't have a lattice
24356 (ccp_initialize): Initialize n_const_val.
24357 * tree-ssa-copy.c (n_copy_of): New static variable.
24358 (init_copy_prop): Initialize n_copy_of.
24359 (get_value): Return NULL_TREE for SSA names we don't have a
24362 2013-02-20 Martin Jambor <mjambor@suse.cz>
24364 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24366 2013-02-20 Richard Biener <rguenther@suse.de>
24368 * genpreds.c (write_lookup_constraint): Do not compare first
24369 letter of the constraint again.
24371 2013-02-20 Richard Biener <rguenther@suse.de>
24373 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24375 (get_use_iv_cost): Terminate hashtable walk when coming across
24378 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24380 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24381 reassociation for avx2 targets.
24383 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24385 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24386 Add version check for v8.10.a to enable microblaze_has_clz
24387 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24388 version and TARGET_PATTERN_COMPARE check
24389 * config/microblaze/microblaze.md: New clzsi2 instruction
24391 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24393 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24394 function before branching.
24396 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24398 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24400 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24402 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24404 PR middle-end/55889
24405 * sel-sched.c: Include ira.h.
24406 (implicit_clobber_conflict_p): New function.
24407 (moveup_expr): Use it.
24408 * Makefile.in (sel-sched.o): Depend on ira.h.
24410 2013-02-19 Richard Biener <rguenther@suse.de>
24412 PR tree-optimization/56384
24413 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24414 (vn_hash_type): Split out from ...
24415 (vn_hash_constant_with_type): ... here.
24416 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24417 (vn_phi_eq): Compare types from vn_phi_s structure.
24418 (vn_phi_lookup): Populate vn_phi_s type.
24419 (vn_phi_insert): Likewise.
24421 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24423 PR tree-optimization/56350
24424 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24425 if haven't found reduction or nested cycle operand, rather than
24426 asserting we must find it.
24428 PR tree-optimization/56381
24429 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24432 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24433 Jakub Jelinek <jakub@redhat.com>
24436 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24437 (swap_optab_enable): Same.
24438 (init_all_optabs): Use argument instead of global.
24439 * tree.h (struct tree_optimization_option): New field target_optabs.
24440 * expr.h (init_all_optabs): Add argument to prototype.
24441 (TREE_OPTIMIZATION_OPTABS): New.
24442 (save_optabs_if_changed): Protoize.
24443 * optabs.h: Declare this_fn_optabs.
24444 * optabs.c (save_optabs_if_changed): New.
24445 Declare this_fn_optabs.
24446 (init_optabs): Add argument to init_all_optabs() call.
24447 * function.c (invoke_set_current_function_hook): Handle per
24449 * function.h (struct function): New field optabs.
24450 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24451 optimization_current_node has changed.
24452 * target-globals.h (save_target_globals_default_opts): Protoize.
24453 * target-globals.c (save_target_globals_default_opts): New.
24455 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24458 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24459 registers %fr12 and %fr12R as call used.
24462 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24463 and HImode, require all displacements to be an integer multiple of
24465 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24466 only allow QImode and HImode when reload is in progress and strict is
24467 true. Likewise for symbolic addresses. Use base14_operand to check
24468 displacements in REG+BASE addresses.
24470 2013-02-18 Richard Biener <rguenther@suse.de>
24472 PR tree-optimization/56366
24473 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24474 handle sign-conversion of outer-loop initial induction value.
24476 2013-02-18 Richard Biener <rguenther@suse.de>
24478 PR middle-end/56349
24479 * cfghooks.c (merge_blocks): If we merge a latch into another
24480 block adjust references to it.
24481 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24482 (verify_loop_structure): Verify that a recorded latch is in fact
24485 2013-02-18 Richard Biener <rguenther@suse.de>
24487 PR tree-optimization/56321
24488 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24489 order SSA name release and virtual operand unlinking.
24491 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24493 * config/microblaze/microblaze.md (save_stack_block): Define.
24494 (restore_stack_block): Likewise.
24496 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24498 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24499 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24500 * config/microblaze/microblaze.c (microblaze_option_override):
24501 Bail out early for PIC modes when target does not support PIC.
24503 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24505 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24506 Replace with a microblaze version.
24507 (microblaze_trampoline_init): Adapt for microblaze.
24508 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24511 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24512 Dodji Seketeli <dodji@redhat.com>
24515 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24516 (instrument_mem_region_access): Do not forget to always put
24517 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24518 0) statement, even for cases where either 'base' or 'base + len'
24519 are not instrumented -- because they have been previously
24520 instrumented. Simplify the logic by putting all the statements
24521 instrument 'base + len' inside a sequence, and then insert that
24522 sequence right before the current insertion point. Then, to
24523 instrument 'base + len', just get an iterator on that statement.
24524 And do not forget to update the pointer to iterator the function
24525 received as argument.
24527 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24529 PR rtl-optimization/56348
24530 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24532 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24534 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24535 (clean_graph_dump_file): Pass base to start_graph_dump.
24537 2013-02-14 Richard Henderson <rth@redhat.com>
24540 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24542 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24544 * collect2-aix.h: Define F_LOADONLY.
24546 2013-02-14 Richard Biener <rguenther@suse.de>
24549 * varasm.c (output_constant_def_1): Get the decl representing
24550 the constant as argument.
24551 (output_constant_def): Wrap output_constant_def_1.
24552 (make_decl_rtl): Use output_constant_def_1 with the decl
24553 representing the constant.
24554 (build_constant_desc): Optionally re-use a decl already
24555 representing the constant.
24556 (tree_output_constant_def): Adjust.
24558 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24561 * asan.c (instrument_builtin_call): Really put the length of the
24562 second source argument into src1_len.
24564 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24566 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24567 argument. If it is false, don't create edge from then_bb to
24569 (insert_if_then_before_iter): Pass true to it.
24570 (build_check_stmt): Pass false to it.
24571 (transform_statements): Flush hash table only on extended basic
24572 block boundaries, rather than at the beginning of every bb.
24573 Don't flush hash table on nonfreeing_call_p calls.
24574 * tree-flow.h (nonfreeing_call_p): New prototype.
24575 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24577 2013-02-13 David S. Miller <davem@davemloft.net>
24579 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24581 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24584 * ira.c (max_regno_before_ira): Move from ...
24586 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24589 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24591 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24593 2013-02-13 Richard Biener <rguenther@suse.de>
24596 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24597 globals in MEM_REFs.
24599 2013-02-13 Richard Biener <rguenther@suse.de>
24601 * loop-init.c (loop_optimizer_init): Clear loop state when
24602 re-initializing preserved loops.
24603 * loop-unswitch.c (unswitch_single_loop): Return whether
24604 we unswitched the loop. Do not verify loop state here.
24605 (unswitch_loops): When we unswitched a loop discover new loops.
24607 2013-02-13 Kostya Serebryany <kcc@google.com>
24609 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24611 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24613 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24615 Avoid instrumenting duplicated memory access in the same basic block
24616 * Makefile.in (asan.o): Add new dependency on hash-table.h
24617 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24618 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24619 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24620 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24621 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24622 (get_mem_ref_of_assignment): New functions.
24623 (get_mem_refs_of_builtin_call): Extract from
24624 instrument_builtin_call and tweak a little bit to make it fit with
24626 (instrument_builtin_call): Use the new
24627 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24628 of is_gimple_builtin_call.
24629 (instrument_derefs, instrument_mem_region_access): Insert the
24630 instrumented memory reference into the hash table.
24631 (maybe_instrument_assignment): Renamed instrument_assignment into
24632 this, and change it to advance the iterator when instrumentation
24633 actually happened and return true in that case. This makes it
24634 homogeneous with maybe_instrument_assignment, and thus give a
24635 chance to callers to be more 'regular'.
24636 (transform_statements): Clear the memory reference hash table
24637 whenever we enter a new BB, when we cross a function call, or when
24638 we are done transforming statements. Use
24639 maybe_instrument_assignment instead of instrumentation. No more
24640 need to special case maybe_instrument_assignment and advance the
24641 iterator after calling it; it's now handled just like
24642 maybe_instrument_call. Update comment.
24644 2013-02-13 Richard Biener <rguenther@suse.de>
24646 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24647 Fix loop discovery code.
24649 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24651 PR inline-asm/56148
24652 * lra-constraints.c (process_alt_operands): Match early clobber
24653 operand with itself. Check conflicts with earlyclobber only if
24654 the operand is not reloaded. Prefer to reload conflicting operand
24655 if earlyclobber and matching operands are the same.
24657 2013-02-12 Richard Biener <rguenther@suse.de>
24660 * lto-streamer-out.c (write_symbol): Do not output symbols
24661 for hard register variables.
24663 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24666 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24667 (umulsidi3_insn, mulsidi3_insn): New insns.
24669 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24671 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24672 (struct tune_params): Add vec_costs field.
24673 * config/arm/arm.c (arm_builtin_vectorization_cost)
24674 (arm_add_stmt_cost): New functions.
24675 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24676 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24677 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24678 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24679 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24680 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24681 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24683 2013-02-12 Richard Biener <rguenther@suse.de>
24686 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24687 decls again if possible.
24689 2013-02-12 Richard Biener <rguenther@suse.de>
24691 PR middle-end/56288
24692 * tree-ssa.c (verify_ssa_name): Fix check, move
24693 SSA_NAME_IN_FREE_LIST check up.
24695 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24696 Steven Bosscher <steven@gcc.gnu.org>
24698 PR rtl-optimization/56151
24699 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24700 equal to op0 or op1, and last_insn pattern is CODE operation
24701 with MEM dest and one of the operands matches that MEM.
24703 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24705 * doc/extend.texi: Document Function Multiversioning and "default"
24706 parameter string to target attribute.
24707 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24708 target attribute parameter is "default".
24709 (ix86_compare_version_priority): Remove checks for target attribute.
24710 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24711 Remove check for target attribute equal to NULL. Add assert.
24712 (ix86_generate_version_dispatcher_body): Change error to sorry.
24714 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24715 Jack Howarth <howarth@bromo.med.uc.edu>
24716 Patrick Marlier <patrick.marlier@gmail.com>
24719 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24720 define ENDFILE_SPEC as TM_DESTRUCTOR.
24721 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24723 2013-02-11 Alexander Potapenko <glider@google.com>
24724 Jack Howarth <howarth@bromo.med.uc.edu>
24725 Jakub Jelinek <jakub@redhat.com>
24728 * config/darwin.c (cdtor_record): Rename ctor_record.
24729 (sort_cdtor_records): Rename sort_ctor_records.
24730 (finalize_dtors): New routine to sort destructors by
24731 priority before use in assemble_integer.
24732 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24734 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24736 PR rtl-optimization/56275
24737 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24738 offset is non-negative and less than cmode size before
24739 calling simplify_subreg.
24741 2013-02-11 Richard Biener <rguenther@suse.de>
24743 PR tree-optimization/56264
24744 * cfgloop.h (fix_loop_structure): Adjust prototype.
24745 * loop-init.c (fix_loop_structure): Return the number of
24746 newly discovered loops.
24747 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24748 are discovered, do a full loop-closed SSA rewrite.
24750 2013-02-11 Richard Biener <rguenther@suse.de>
24752 PR tree-optimization/56273
24753 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24755 (check_array_ref): Fix missing newline in dumps.
24756 (search_for_addr_array): Likewise.
24758 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24760 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24762 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24765 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24767 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24769 PR rtl-optimization/56246
24770 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24772 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24773 constraints are satisfied.
24775 2013-02-08 Jeff Law <law@redhat.com>
24778 * emit-rtl.c (reg_is_parm_p): New function.
24779 * regs.h (reg_is_parm_p): New prototype.
24780 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24781 callee-clobbered registers.
24783 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24786 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24787 If there is no implicit builtin declaration, just return NULL.
24789 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24791 * config/i386/sse.md (FMAMODEM): New mode iterator.
24792 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24793 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24795 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24797 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24798 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24799 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24801 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24803 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24804 (microblaze*-*-elf): Likewise.
24805 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24807 * config/microblaze/microblaze-c.c: Add builtin defines for
24808 _LITTLE_ENDIAN and _BIG_ENDIAN.
24809 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24810 add to TARGET_DEFAULT flags.
24811 Expand ASM_SPEC and LINK_SPEC.
24812 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24813 * config/microblaze/microblaze.md: Update extendsidi2 and
24814 movdi_internal instructions to use low-order / high-order reg
24816 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24817 options and inversemask / mask of LITTLE_ENDIAN.
24818 * config/microblaze/t-microblaze: Expand multilib options to
24819 include mlittle-endian (le) and update exceptions patterns.
24821 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24823 PR rtl-optimization/56195
24824 * lra-constraints.c (get_reload_reg): Don't reuse regs
24825 if they have smaller mode than requested, if they have
24826 wider mode than requested, try to return a SUBREG.
24828 PR tree-optimization/56250
24829 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24830 if type is unsigned and code isn't MULT_EXPR.
24832 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24834 PR tree-optimization/56064
24835 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24836 bits according to mode.
24837 * fixed-value.h (fixed_from_double_int)
24838 (const_fixed_from_double_int): Adjust comments.
24840 2013-02-08 Richard Biener <rguenther@suse.de>
24843 * lto-streamer.h (struct data_in): Remove current_file, current_line
24844 and current_col members.
24845 * lto-streamer-out.c (lto_output_location): Stream changed bits
24846 en-block for efficiency.
24847 * lto-streamer-in.c (clear_line_info): Remove.
24848 (lto_input_location): Cache current file, line and column
24849 globally via local statics. Read changed bits en-block.
24850 (input_function): Do not call clear_line_info.
24851 (lto_read_body): Likewise.
24852 (lto_input_toplevel_asms): Likewise.
24854 2013-02-08 Michael Matz <matz@suse.de>
24856 PR tree-optimization/52448
24857 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24858 (nt_call_phase): New static.
24859 (add_or_mark_expr): Only mark accesses with newer phase than any
24861 (nonfreeing_call_p): New.
24862 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24863 (nt_fini_block): Keep blocks marked as visited.
24864 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24866 2013-02-08 Richard Biener <rguenther@suse.de>
24868 * ira.c (ira): Free broken dominator information.
24870 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24872 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24874 2013-02-08 Marek Polacek <polacek@redhat.com>
24876 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24878 2013-02-08 Richard Biener <rguenther@suse.de>
24880 PR middle-end/56181
24881 * cfgloop.h (flow_loops_find): Adjust.
24882 (bb_loop_header_p): Declare.
24883 * cfgloop.c (bb_loop_header_p): New function split out from ...
24884 (flow_loops_find): ... here. Adjust function signature,
24885 support incremental loop structure update.
24886 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24887 * cfgloopmanip.c (fix_loop_structure): Move ...
24888 * loop-init.c (fix_loop_structure): ... here.
24889 (apply_loop_flags): Split out from ...
24890 (loop_optimizer_init): ... here.
24891 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24892 in incremental mode, only remove dead loops here.
24894 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24897 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24898 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24899 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24900 (*round<mode>3.libgcc): New insns for fixed-modes.
24901 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24902 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24903 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24904 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24905 implementations. Define to __builtin_avr_absFX,
24906 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24907 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24908 __builtin_avr_countlsFX, respectively.
24909 * config/avr/avr-c.c (target.h): Include it.
24910 (enum avr_builtin_id): New enum.
24911 (avr_resolve_overloaded_builtin): New static function.
24912 (avr_register_target_pragmas): Use it to set
24913 targetm.resolve_overloaded_builtin.
24914 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24915 tree nodes used by DEF_BUILTIN.
24916 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24917 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24918 <AVR_BUILTIN_xxBITS>: Same.
24920 2013-02-08 Richard Biener <rguenther@suse.de>
24922 * cfgloop.c (verify_loop_structure): Properly handle
24923 a loop exiting to another loop header.
24924 * ira-int.h (ira_loops): Remove.
24925 * ira.c (ira_loops): Remove.
24926 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24927 (do_reload): Use loop_optimizer_finalize.
24928 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24929 number_of_loops to access the loop tree.
24930 (more_one_region_p): Likewise.
24931 (finish_loop_tree_nodes): Likewise.
24932 (rebuild_regno_allocno_maps): Likewise.
24933 (mark_loops_for_removal): Likewise.
24934 (mark_all_loops_for_removal): Likewise.
24935 (remove_unnecessary_regions): Likewise.
24936 (ira_build): Likewise.
24937 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24939 2013-02-08 Richard Biener <rguenther@suse.de>
24941 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24942 * ipa-pure-const.c (analyze_function): Avoid calling
24943 mark_irreducible_loops twice.
24944 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24946 2013-02-07 David S. Miller <davem@davemloft.net>
24948 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24950 * var-tracking.c (vt_add_function_parameter): Test the presence of
24951 HAVE_window_save properly and do not remap argument registers when
24952 we have a leaf function.
24954 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24957 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24959 * config/i386/i386.c (ix86_print_operand): Ditto.
24961 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24963 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24965 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24967 PR rtl-optimization/56225
24968 * lra-constraints.c (process_alt_operands): Check that reload hard
24969 reg can hold value for strict_low_part.
24971 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24974 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24975 dwarf2out_end_function.
24976 (in_first_function_p, maybe_at_text_label_p,
24977 first_loclabel_num_not_at_text_label): New variables.
24978 (dwarf2out_var_location): In the first function find out
24979 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24980 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24983 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24985 PR rtl-optimization/56178
24986 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24987 SUBREG of a register. Tidy up related block of code.
24988 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24989 note if the source is a register or a SUBREG of a register.
24991 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24994 * config/rs6000/rs6000.md (ptrm): New mode attr.
24995 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24996 call_value_indirect_aix<pttrsize>,
24997 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25000 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25002 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25003 if -bnortl. Convert to strcmp and strncmp.
25005 2013-02-07 Alan Modra <amodra@gmail.com>
25008 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25009 addresses won't wrap when offsetting.
25010 (rs6000_secondary_reload): Provide secondary reloads needed for
25011 wrapping LO_SUM addresses.
25013 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25015 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25016 MACH, just __MACH__.
25018 2013-02-06 Richard Biener <rguenther@suse.de>
25020 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25021 instead of calling fix_loop_structure.
25023 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25025 PR middle-end/56217
25026 * omp-low.c (use_pointer_for_field): Return false if
25027 lower_send_shared_vars doesn't generate any copy-out code.
25029 2013-02-06 Tom de Vries <tom@codesourcery.com>
25031 PR rtl-optimization/56131
25032 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25033 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25034 of the label is NULL. Add comment.
25036 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25038 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25041 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25042 (STATIC_LIBTSAN_LIBS): Likewise.
25043 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25044 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25045 is defined, don't add anything else beyond that.
25046 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25047 (LINK_COMMAND_SPEC): Use them.
25049 PR tree-optimization/56205
25050 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25051 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25052 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25054 2013-02-05 Richard Biener <rguenther@suse.de>
25056 PR tree-optimization/53342
25057 PR tree-optimization/53185
25058 * tree-vectorizer.h (vect_check_strided_load): Remove.
25059 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25060 not disallow peeling for vectorized strided loads.
25061 (vect_check_strided_load): Make static and simplify.
25062 (vect_analyze_data_refs): Adjust.
25063 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25064 correctly when vectorizing strided loads.
25066 2013-02-05 Richard Biener <rguenther@suse.de>
25068 * doc/install.texi: Refer to ISL, not PPL.
25070 2013-02-05 Jan Hubicka <jh@suse.cz>
25072 PR tree-optimization/55789
25073 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25075 2013-02-05 Jan Hubicka <jh@suse.cz>
25077 PR tree-optimization/55789
25078 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25079 the dead call anyway.
25081 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25084 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25086 2013-02-04 Alexander Potapenko <glider@google.com>
25087 Jack Howarth <howarth@bromo.med.uc.edu>
25088 Jakub Jelinek <jakub@redhat.com>
25091 * config/darwin.c (sort_ctor_records): Stabilized qsort
25092 on constructor priority by using original position.
25093 (finalize_ctors): New routine to sort constructors by
25094 priority before use in assemble_integer.
25095 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25097 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25100 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25101 about visibility on artificial decls.
25102 * config/sol2.c (solaris_assemble_visibility): Likewise.
25104 2013-02-04 Kai Tietz <ktietz@redhat.com>
25107 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25108 argument and improve checking of return-argument types for 16-byte
25110 (ix86_function_value_1): Add additional valtype argument on call
25111 of function_value_64.
25112 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25113 handling infunction_value_64 function.
25115 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25117 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25119 2013-02-04 Richard Biener <rguenther@suse.de>
25121 PR tree-optimization/56188
25122 * tree-ssa-structalias.c (label_visit): Consider case with
25123 initially non-empty points-to set.
25124 (perform_var_substitution): Dump node mapping and clean up.
25126 2013-02-04 Richard Guenther <rguenther@suse.de>
25129 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25130 node prevail as last resort.
25131 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25132 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25134 2013-02-04 Richard Biener <rguenther@suse.de>
25136 PR tree-optimization/56113
25137 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25139 (equiv_class_lookup_or_add): ... this.
25140 (label_visit): Adjust and fix error in previous patch.
25141 (perform_var_substitution): Adjust.
25143 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25145 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25146 * config/sh/sh4-300.md: Likewise.
25147 * config/sh/sh4a.md: Likewise.
25148 * config/sh/constraints.md: Likewise.
25149 * config/sh/sh.md: Likewise.
25150 * config/sh/netbsd-elf.h: Likewise.
25151 * config/sh/predicates.md: Likewise.
25152 * config/sh/sh-protos.h: Likewise.
25153 * config/sh/ushmedia.h: Likewise.
25154 * config/sh/linux.h: Likewise.
25155 * config/sh/sh.c: Likewise.
25156 * config/sh/superh.h: Likewise.
25157 * config/sh/elf.h: Likewise.
25158 * config/sh/sh4.md: Likewise.
25159 * config/sh/sh.h: Likewise.
25161 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25163 * config/pa/constraints.md: Adjust unused letters. Change "T"
25164 constraint to match_test floating_point_store_memory_operand().
25165 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25166 (base14_operand): New.
25167 (floating_point_store_memory_operand): New.
25168 (integer_store_memory_operand): Revise to use base14_operand and
25169 reg_plus_base_memory_operand.
25170 (move_dest_operand): Allow symbolic_memory_operands.
25171 (symbolic_memory_operand): Check for LO_SOM.
25172 (symbolic_operand): Change default case to break.
25173 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25174 CONST_DOUBLE values to be reloaded by putting them into memory when
25175 the destination is a floating point register.
25176 (movdf): Remove code to handle CONST_DOUBLE.
25178 (reload_indf_r1): New.
25179 (reload_insf_r1): New.
25180 Consistently use "Q" and "T" constraints with integer and floating
25181 point move instructions, respectively.
25182 (movdi): Remove FAIL.
25183 Change predicate for source operand unamed DImode move from
25184 general_operand to move_src_operand.
25185 (umulsidi3): Change predicate for destination operand to
25187 Likewise for similar unamed patterns.
25188 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25189 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25190 (hppa_legitimize_address): Simplify mask calculation.
25191 (pa_emit_move_sequence): Revised handling of secondary reloads from
25192 REG+D addresses for floating point loads and stores. Directly handle
25193 loading CONST0_RTX (mode) to a floating point register.
25194 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25195 to floating point registers. Don't restrict secondary reloads to
25196 floating point registers to integer modes. Revise some comments and
25198 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25199 (pa_legitimate_address_p): New.
25200 (pa_legitimize_reload_address): New.
25201 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25202 (STRICT_REG_OK_FOR_BASE_P): New.
25203 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25204 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25206 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25207 Andrew Dixie <andrewd@gentrack.com>
25209 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25212 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25214 * expmed.c (extract_bit_field_1): Pass the full width of the
25215 structure to get_best_reg_extraction_insn.
25217 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25220 * configure.ac (use_cxa_atexit): Add AIX.
25221 * configure: Regenerate.
25223 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25225 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25228 * final.c (need_profile_function): New variable.
25229 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25230 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25231 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25232 notes, targetm.asm_out.function_prologue doesn't emit anything,
25233 HAVE_prologue and profiler should be emitted before prologue,
25234 set need_profile_function instead of emitting it.
25235 (final_scan_insn): If need_profile_function, emit
25236 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25237 NOTE_INSN_FUNCTION_BEG note.
25239 2013-02-01 Richard Henderson <rth@redhat.com>
25241 * config/rs6000/rs6000.md (smulditi3): New.
25244 * config/alpha/alpha.md (umulditi3): New.
25246 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25248 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25249 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25251 2013-02-01 Richard Biener <rguenther@suse.de>
25253 PR tree-optimization/56113
25254 * tree-ssa-structalias.c (label_visit): Reduce work for
25255 single-predecessor nodes.
25257 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25259 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25260 range isn't testing for zero.
25262 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25264 PR middle-end/56113
25265 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25267 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25268 Nick Clifton <nickc@redhat.com>
25270 * config/v850/constraints.md (Q): Define as a memory constraint.
25271 * config/v850/predicates.md (label_ref_operand): New predicate.
25272 (e3v5_shift_operand): New predicate.
25273 (ior_operator): New predicate.
25274 * config/v850/t-v850: Add e3v5 multilib.
25275 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25276 (v850_gen_movdi): Prototype.
25277 * config/v850/v850.c: Add support for e3v5 architecture.
25278 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25280 (construct_save_jarl): Add e3v5 long JARL support.
25281 (v850_adjust_insn_length): New function. Adjust length of call
25282 insns when using e3v5 instructions.
25283 (v850_gen_movdi): New function: Generate instructions to move a
25285 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25286 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25287 (TARGET_USE_FPU): Enable for e3v5.
25288 (CONST_OK_FOR_W): New macro.
25289 (ADJUST_INSN_LENGTH): Define.
25290 * config/v850/v850.md (UNSPEC_LOOP): Define.
25291 (attr cpu): Add v850e3v5.
25292 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25293 (movdi): New pattern.
25294 (movdi_internal): New pattern.
25295 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25296 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25297 (cstoresf4): Likewise.
25298 (cstoredf4): Likewise.
25299 (insv): New pattern.
25300 (rotlso3_a): New pattern.
25301 (rotlsi3_b): New pattern
25302 (rotlsi3_v850e3v5): New pattern.
25303 (doloop_begin): New pattern.
25304 (fix_loop_counter): New pattern.
25305 (doloop_end): New pattern.
25306 (branch_normal): Add e3v5 long branch support.
25307 (branch_invert): Likewise.
25308 (branch_z_normal): Likewise.
25309 (branch_z_invert): Likewise.
25310 (branch_nz_normal): Likewise.
25311 (branch_nz_invert): Likewise.
25312 (call_internal_short): Add e3v5 register-indirect JARL support.
25313 (call_internal_long): Likewise.
25314 (call_value_internal_short): Likewise.
25315 (call_value_internal_long): Likewise.
25316 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25317 (mloop): New option.
25318 * config.gcc: Add support for configuring v840e3v5 target.
25319 * doc/invoke.texi: Document new v850 specific command line options.
25321 2013-01-31 Paul Koning <ni1d@arrl.net>
25325 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25326 children if parent is a class.
25327 (prune_unused_types_prune): Don't add DW_AT_declaration.
25329 2013-01-31 Richard Biener <rguenther@suse.de>
25331 PR tree-optimization/56157
25332 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25333 match up operand with SLP child.
25335 2013-01-31 Jason Merrill <jason@redhat.com>
25338 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25339 parameters the first time.
25340 (gen_scheduled_generic_parms_dies): Check completeness here.
25342 2013-01-31 Richard Biener <rguenther@suse.de>
25344 PR middle-end/53073
25345 * common.opt (faggressive-loop-optimizations): New flag,
25346 enabled by default.
25347 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25348 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25349 infer_loop_bounds_from_undefined by it.
25351 2013-01-31 Richard Biener <rguenther@suse.de>
25353 PR tree-optimization/56150
25354 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25355 visit virtual operands.
25356 (find_uses_to_rename_bb): Likewise.
25358 2013-01-31 Richard Biener <rguenther@suse.de>
25360 PR tree-optimization/56150
25361 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25362 mixed store non-store stmts.
25364 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25367 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25368 LIBASAN_EARLY_SPEC is defined.
25369 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25370 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25372 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25375 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25376 invalid args instead of ICEing on it.
25377 (ix86_valid_target_attribute_tree): Return error_mark_node if
25378 ix86_valid_target_attribute_inner_p failed.
25379 (ix86_valid_target_attribute_p): Return false only if
25380 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25381 target("default") attribute.
25382 (sorted_attr_string): Change argument from const char * to tree,
25383 merge in all target attribute arguments rather than just one.
25384 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25385 instead of free. Avoid using strcat.
25386 (ix86_mangle_function_version_assembler_name): Mangle
25387 target("default") as if no target attribute is present. Adjust
25388 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25389 instead of xmalloc and XDELETEVEC instead of free.
25390 (ix86_function_versions): Don't return true if one of the decls
25391 doesn't have target attribute. If they don't and one of the decls
25392 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25393 sorted_attr_string caller. Use XDELETEVEC instead of free.
25394 (ix86_supports_function_versions): Remove.
25395 (make_name): Fix up formatting.
25396 (make_dispatcher_decl): Remove resolver_name and its initialization.
25397 Avoid leaking memory.
25398 (is_function_default_version): Return true if there is
25399 target("default") attribute rather than no target attribute at all.
25400 (make_resolver_func): Avoid leaking memory.
25401 (ix86_generate_version_dispatcher_body): Likewise.
25402 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25403 * target.def (supports_function_versions): Remove.
25404 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25405 * doc/tm.texi: Regenerated.
25407 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25409 PR rtl-optimization/56144
25410 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25411 for values with side effects.
25413 2013-01-30 Richard Biener <rguenther@suse.de>
25415 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25416 (sparseset_pop): Likewise.
25417 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25418 to be able to use quick_push in the worker loop.
25420 2013-01-30 Marek Polacek <polacek@redhat.com>
25422 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25424 2013-01-30 Richard Biener <rguenther@suse.de>
25427 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25429 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25431 PR tree-optimization/56064
25432 * fixed-value.c (fixed_from_double_int): New function.
25433 * fixed-value.h (fixed_from_double_int): New prototype.
25434 (const_fixed_from_double_int): New static inline function.
25435 * fold-const.c (native_interpret_fixed): New static function.
25436 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25437 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25438 (native_encode_fixed): New static function.
25439 (native_encode_expr) <FIXED_CST>: Use it.
25440 (native_interpret_int): Move double_int worker code to...
25441 * double-int.c (double_int::from_buffer): ...this new static method.
25442 * double-int.h (double_int::from_buffer): Prototype it.
25444 2013-01-30 Richard Biener <rguenther@suse.de>
25446 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25447 New pointer-map and obstack.
25448 (init_alias_vars): Allocate pointer-map and obstack.
25449 (delete_points_to_sets): Free them.
25450 (find_what_var_points_to): Cache result.
25451 (find_what_p_points_to): Adjust for changed interface of
25452 find_what_var_points_to.
25453 (compute_points_to_sets): Likewise.
25454 (ipa_pta_execute): Likewise.
25456 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25458 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25459 * configure: Regenerate.
25460 * config.in: Regenerate.
25461 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25462 #nobits/#progbits if supported.
25464 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25467 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25468 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25469 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25471 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25473 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25474 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25476 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25478 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25479 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25481 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25483 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25485 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25486 * config/arm/cortex-a7.md: New bypasses using
25487 arm_mac_accumulator_is_result.
25489 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25491 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25492 (cortex_a7_neon_mla): Likewise.
25493 (cortex_a7_fpfmad): New reservation.
25494 (cortex_a7_fpmacs): Use ffmas and update required units.
25495 (cortex_a7_fpmuld): Update required units and latency.
25496 (cortex_a7_fpmacd): Likewise.
25497 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25498 (cortex_a7_neon). Likewise.
25499 (bypass) Update participating units.
25501 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25503 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25504 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25506 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25507 (vfp_fmul): Use ffmad.
25508 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25509 (cortex_r4_fmacd): Use ffmad.
25510 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25511 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25512 (cortex_a9_fmacd): Use ffmad.
25513 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25514 (cortex_a8_vfp_macd): Use ffmad.
25515 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25516 (cortex_a5_fpmacd): Use ffmad.
25517 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25518 (cortex_a15_vfp_macd): Use ffmad.
25519 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25521 2013-01-29 Jason Merrill <jason@redhat.com>
25524 * varasm.c (default_assemble_visibility): Don't warn about
25525 visibility on artificial decls.
25527 2013-01-29 Richard Biener <rguenther@suse.de>
25529 PR tree-optimization/56113
25530 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25532 (label_visit): Free duplicate bitmaps and record the leader instead.
25533 (perform_var_substitution): Adjust.
25535 2013-01-29 Richard Biener <rguenther@suse.de>
25537 PR tree-optimization/55270
25538 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25539 the CFG, schedule loops for fixup.
25541 2013-01-29 Nick Clifton <nickc@redhat.com>
25543 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25546 2013-01-28 Leif Ekblad <leif@rdos.net>
25548 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25549 * config/i386/i386.h (TARGET_RDOS): New macro.
25550 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25551 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25552 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25553 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25554 DEFAULT_LARGE_SECTION_THRESHOLD.
25555 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25556 * config/i386/rdos.h: New file.
25557 * config/i386/rdos64.h: New file.
25559 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25562 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25565 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25567 PR rtl-optimization/56117
25568 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25569 call cselib_lookup_from_insn on the MEM before calling
25570 add_insn_mem_dependence.
25572 2013-01-28 Richard Biener <rguenther@suse.de>
25574 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25575 to a stmt that didn't have one.
25576 (copy_phis_for_bb): Likewise for PHI arguments.
25577 (copy_debug_stmt): Likewise for debug stmts.
25579 2013-01-28 Richard Biener <rguenther@suse.de>
25581 PR tree-optimization/56034
25582 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25583 (partition_builtin_p): Adjust.
25584 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25585 it is the last partition.
25586 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25587 up the vertex for the definition.
25588 (classify_partition): Classify whether a partition is a
25589 PKIND_REDUCTION, thus has uses outside of the loop.
25590 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25591 Merge all PKIND_REDUCTION partitions into the last partition.
25592 (tree_loop_distribution): Seed partitions from reductions as well.
25594 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25596 PR tree-optimization/56125
25597 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25598 pow(x,c) into sqrt(x) * powi(x, n/2) or
25599 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25600 optimizing for size.
25601 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25602 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25605 PR tree-optimization/56094
25606 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25607 to UNKNOWN_LOCATION while gimplifying expr.
25609 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25612 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25613 operand 0 in movabs insn template for -masm=intel asm alternative.
25614 (*movabs<mode>_2): Ditto for operand 1.
25616 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25619 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25622 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25624 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25627 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25629 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25630 Undef to avoid warning.
25632 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25634 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25635 * configure: Regenerate.
25637 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25639 PR tree-optimization/56098
25640 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25641 for stmts with volatile ops.
25642 (cond_store_replacement): Don't optimize if assign has volatile ops.
25643 (cond_if_else_store_replacement_1): Don't optimize if either
25644 then_assign or else_assign have volatile ops.
25645 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25648 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25650 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25652 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25654 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25655 missing ':' in asm example.
25657 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25659 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25660 entries into lane and laneq entries.
25661 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25662 Remove AdvSIMD scalar modes.
25663 (aarch64_sq<r>dmulh_laneq<mode>): New.
25664 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25666 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25667 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25668 * config/aarch64/iterators.md (VCOND): New.
25671 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25674 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25675 Add NULL LIBNAME argument to existing definitions.
25676 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25677 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25678 * config/avr/avr.c (DEF_BUILTIN): Same.
25679 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25680 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25681 implementation is available (DECL_ASSEMBLER_NAME is set).
25682 (avr_fold_absfx): New static function.
25683 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25684 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25685 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25686 AVR_BUILTIN_ABSLLK.
25687 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25688 (abshk, absk, abslk, absllk): Provide as static inline functions.
25690 2013-01-25 Marek Polacek <polacek@redhat.com>
25692 PR tree-optimization/56035
25693 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25695 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25697 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25698 (*movtf_internal_rex64): Add (!o,C) alternative
25699 (*movxf_internal_rex64): Ditto.
25700 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25702 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25704 * doc/invoke.texi: fix typo.
25705 * doc/objc.texi: fix typo.
25707 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25709 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25710 for the first two alternatives.
25712 2013-01-24 Diego Novillo <dnovillo@google.com>
25714 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25715 (ggc-zone.o): Remove.
25716 * configure.ac: Remove option --with-gc.
25717 * configure: Re-generate.
25718 * doc/install.texi: Remove documentation for --with-gc.
25719 * gengtype.c (write_enum_defn): Remove. Update all users.
25720 (write_Types_process_field): Remove generation of gt_e_* argument.
25721 (output_type_enum): Remove. Update all users.
25722 (write_enum_defn): Remove. Update all users.
25723 (enum alloc_zone): Remove. Update all users.
25724 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25725 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25726 Update all callers.
25727 (struct ptr_data): Remove field TYPE. Update all users.
25728 (gt_pch_note_object): Remove argument TYPE. Update all users.
25729 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25731 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25732 (struct alloc_zone): Remove.
25733 (ggc_internal_alloc_zone_stat): Remove.
25734 (ggc_internal_cleared_alloc_zone_stat): Remove.
25735 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25736 (ggc_pch_count_object): Remove last argument. Update all users.
25737 (ggc_pch_alloc_object): Remove last argument. Update all users.
25738 (struct alloc_zone): Remove.
25739 * ggc-zone.c: Remove.
25740 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25741 (struct alloc_zone): Remove.
25742 (ggc_alloc_typed_stat): Remove.
25743 (ggc_alloc_typed): Remove.
25744 (ggc_splay_alloc): Remove first argument.
25745 (rtl_zone): Remove. Update all users.
25746 (tree_zone): Remove. Update all users.
25747 (tree_id_zone): Remove. Update all users.
25748 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25749 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25750 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25751 * tree-ssanames.c: Remove references to zone allocator in comments.
25753 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25755 * config/avr/avr.c (avr_out_fract): Make register numbers that
25756 might be outside of source operand signed.
25758 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25760 * config/i386/constraints.md (Yf): New constraint.
25761 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25762 of f constraint to conditionaly disable x87 register preferences.
25763 (*movdf_internal): Ditto.
25764 (*movsf_internal): Ditto.
25766 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25768 PR inline-asm/55934
25769 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25770 that have operands with impossible constraints.
25771 Add a FIXME for a speed-up opportunity.
25772 * lra-constraints.c (process_alt_operands): Verify that a class
25773 selected from constraints on asms is valid for the operand mode.
25774 (curr_insn_transform): Remove incorrect comment.
25776 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25778 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25779 TOC operand is a valid symbol ref in the constant pool.
25781 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25783 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25785 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25788 * config/avr/stdfix.h: New file.
25789 * t-avr (stdfix-gcc.h): New rule to build it.
25790 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25792 2013-01-23 Kostya Serebryany <kcc@google.com>
25794 * config/darwin.h: remove dependency on
25795 CoreFoundation (asan on Mac OS).
25797 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25800 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25801 instead of cmpdi_operand for first comparison operand.
25802 Don't assert that comparison operands aren't both constants.
25804 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25806 * doc/install.texi (Downloading the Source): Update references to
25807 downloading separate components.
25809 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25811 * doc/extend.texi (__int128): Improve grammar.
25813 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25816 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25817 alternative to (o,r).
25818 (*movdi_internal_rex64): Remove (!o,n) alternative.
25819 (DImode immediate->memory splitter): Remove.
25820 (DImode immediate->memory peephole2): Remove.
25821 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25822 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25823 alternative to (!o,*r).
25824 (*movtf_internal_sse): New pattern.
25825 (*movxf_internal_rex64): New pattern.
25826 (*movxf_internal): Disable for TARGET_64BIT.
25827 (*movdf_internal_rex64): Remove (!o,F) alternative.
25829 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25831 PR middle-end/56074
25832 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25833 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25834 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25835 stmt locations where LOCATION_LOCUS of the stmt location is
25836 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25839 * config/i386/i386.md (UNSPEC_STOS): New.
25840 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25841 *strsetqi_1): Add UNSPEC_STOS.
25843 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25846 * doc/invoke.texi: Remove left over -Wsynth example.
25848 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25850 PR tree-optimization/56051
25851 * fold-const.c (fold_binary_loc): Don't fold
25852 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25853 a narrowing conversion, or widening conversion from signed
25856 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25858 PR rtl-optimization/56023
25859 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25860 dependent on debug instruction.
25862 2013-01-21 Martin Jambor <mjambor@suse.cz>
25864 PR middle-end/56022
25865 * function.c (allocate_struct_function): Call
25866 invoke_set_current_function_hook earlier.
25868 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25870 * reload1.c (init_reload): Only initialize reload_obstack
25871 during the first call.
25873 2013-01-21 Marek Polacek <polacek@redhat.com>
25875 * cfgloop.c (verify_loop_structure): Fix up grammar.
25877 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25879 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25880 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25882 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25885 * config/arm/marvell-pj4.md: Update copyright year.
25886 Fix up use of alu to alu_reg and simple_alu_imm.
25888 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25890 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25892 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25895 * lra-constraints.c (curr_insn_transform): Don't reuse original
25896 insn for secondary memory move when memory mode should be different.
25898 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25900 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25901 atomic_storedi_1): New patterns.
25903 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25905 btver2 pipeline descriptions.
25906 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25908 * config/i386/i386.md (btver2_decode): New type attributes.
25909 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25911 * config/i386/btver2.md: New file describing btver2 pipelines.
25913 2013-01-19 Andrew Pinski <apinski@cavium.com>
25915 PR tree-optimization/52631
25916 * tree-ssa-sccvn (visit_use): Before looking up the original
25917 statement, try looking up the simplified expression.
25919 2013-01-19 Anthony Green <green@moxielogic.com>
25921 * config/moxie/moxie.c (moxie_expand_prologue): Set
25922 current_function_static_stack_size.
25924 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25926 PR tree-optimization/56029
25927 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25928 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25930 2013-01-18 Sharad Singhai <singhai@google.com>
25932 PR tree-optimization/55995
25933 * dumpfile.c (dump_loc): Print location only if available.
25934 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25936 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25939 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25940 secondary memory move.
25941 (inherit_reload_reg): Use rclass instead of cl for
25942 check_secondary_memory_needed_p.
25944 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25946 PR middle-end/56015
25947 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25948 the case where writing real complex part of target modifies op1.
25950 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25952 * config/aarch64/aarch64-simd.md
25953 (aarch64_vcond_internal<mode>): Handle unordered cases.
25954 * config/aarch64/iterators.md (v_cmp_result): New.
25956 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25957 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25959 * config/arm/marvell-pj4.md: New file.
25960 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25961 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25962 (generic_vfp): Likewise.
25963 * config/arm/arm-cores.def: Add marvell-pj4.
25964 * config/arm/arm-tune.md: Regenerate.
25965 * config/arm/arm-tables.opt: Regenerate.
25966 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25967 * doc/invoke.texi: Document marvell-pj4.
25969 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25971 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25973 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25978 * var-tracking.c (negative_power_of_two_p): New.
25979 (global_get_addr_cache, local_get_addr_cache): New.
25980 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25981 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25983 (vt_stack_offset_p): Remove.
25984 (vt_canon_true_dep): Always canonicalize loc's address.
25985 (clobber_overlapping_mems): Make sure we have a MEM.
25986 (local_get_addr_clear_given_value): New.
25987 (val_reset): Clear local cached entries.
25988 (compute_bb_dataflow): Create and release the local cache.
25989 Disable duplicate MEMs clobbering.
25990 (emit_notes_in_bb): Clobber MEMs likewise.
25991 (vt_emit_notes): Create and release the local cache.
25992 (vt_initialize, vt_finalize): Create and release the global
25993 cache, respectively.
25994 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25996 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25998 PR libmudflap/53359
25999 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26000 not found in the symtab.
26002 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26005 PR rtl-optimization/55547
26006 PR rtl-optimization/53827
26009 * alias.c (offset_overlap_p): New, factored out of...
26010 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26011 the conservative special case for symbolic constants. Don't
26012 adjust zero sizes on alignment.
26014 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26016 PR rtl-optimization/52573
26017 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26018 REG_UNUSED for the same register.
26020 2013-01-17 Richard Biener <rguenther@suse.de>
26021 Marek Polacek <polacek@redhat.com>
26023 PR rtl-optimization/55833
26024 * loop-unswitch.c (unswitch_loops): Move loop verification...
26025 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26026 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26027 Set it to true when we're removing a loop from hierarchy tree in
26028 an irreducible region.
26029 (fix_bb_placements): Adjust caller.
26030 (fix_loop_placements): Likewise.
26032 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26034 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26035 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26036 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26037 Remove ID. Adjust comments.
26038 * config/avr/avr-c.c (avr_builtin_name): Remove.
26039 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26040 * config/avr/avr.c (avr_tolower): New static function.
26041 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26042 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26043 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26046 2013-01-17 Jan Hubicka <jh@suse.cz>
26048 PR tree-optimization/55273
26049 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26051 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26054 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26055 store through atomic_store<mode>_1.
26056 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26058 2013-01-17 Martin Jambor <mjambor@suse.cz>
26060 PR tree-optimizations/55264
26061 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26062 for virtual methods.
26063 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26064 virtual methods before inlining is over.
26065 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26067 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26070 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26072 PR rtl-optimization/56005
26073 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26074 pending reads for prefetch.
26076 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26078 * config/aarch64/aarch64.md
26079 (*cstoresi_neg_uxtw): New pattern.
26080 (*cmovsi_insn_uxtw): New pattern.
26081 (*<optab>si3_uxtw): New pattern.
26082 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26083 (*<optab>si3_insn_uxtw): New pattern.
26084 (*bswapsi2_uxtw): New pattern.
26086 2013-01-16 Richard Biener <rguenther@suse.de>
26088 * tree-inline.c (tree_function_versioning): Remove set but
26089 never used variable.
26091 2013-01-16 Richard Biener <rguenther@suse.de>
26093 PR tree-optimization/55964
26094 * tree-flow.h (rename_variables_in_loop): Remove.
26095 (rename_variables_in_bb): Likewise.
26096 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26097 (copy_loop_before): Adjust and delete update-ssa status.
26098 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26099 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26100 (rename_variables_in_loop): Remove.
26101 (slpeel_update_phis_for_duplicate_loop): Likewise.
26102 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26103 use available cfg machinery instead of duplicating it.
26104 Update PHI nodes and perform poor-mans SSA update here.
26105 (slpeel_tree_peel_loop_to_edge): Adjust.
26107 2013-01-16 Richard Biener <rguenther@suse.de>
26109 PR tree-optimization/54767
26110 PR tree-optimization/53465
26111 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26112 (vrp_visit_phi_node): For PHI arguments coming via backedges
26113 drop all symbolical range information.
26114 (execute_vrp): Compute backedges.
26116 2013-01-16 Richard Biener <rguenther@suse.de>
26118 * doc/install.texi: Update CLooG and ISL requirements to
26121 2013-01-16 Christian Bruel <christian.bruel@st.com>
26124 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26125 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26126 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26128 2013-01-16 DJ Delorie <dj@redhat.com>
26130 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26131 (UNSPECV_SP_SWITCH_E): New.
26132 (sp_switch_1): Change to an unspec.
26133 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26136 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26138 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26139 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26140 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26141 (expand_mem_thread_fence): Ditto.
26142 (expand_mem_signal_fence): Ditto.
26143 (expand_atomic_load): Ditto.
26144 (expand_atomic_store): Ditto.
26146 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26148 PR rtl-optimization/55547
26149 PR rtl-optimization/53827
26152 * alias.c (memrefs_conflict_p): Set sizes to negative after
26155 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26158 * function.c (thread_prologue_and_epilogue_insns): Always
26159 add crtl->drap_reg to set_up_by_prologue.set, even if
26160 stack_realign_drap is false.
26162 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26164 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26165 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26166 *call): Fix indention.
26168 2013-01-15 Tom de Vries <tom@codesourcery.com>
26171 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26174 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26176 PR rtl-optimization/55153
26177 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26179 2013-01-15 Martin Jambor <mjambor@suse.cz>
26181 PR tree-optimization/55920
26182 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26183 accesses as grp_to_be_debug_replaced.
26185 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26187 PR tree-optimization/55920
26188 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26189 there is non-useless type conversion needed from debug rhs to lhs,
26190 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26192 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26193 Mikael Pettersson <mikpe@it.uu.se>
26196 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26198 (ASM_OUTPUT_CASE_LABEL): Remove.
26199 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26200 * final.c (shorten_branches): Update alignment of labels before
26201 jump tables if CASE_VECTOR_SHORTEN_MODE.
26203 2013-01-15 Richard Biener <rguenther@suse.de>
26206 * system.h: Do not include gmp.h for building host tools.
26208 2013-01-15 Richard Biener <rguenther@suse.de>
26210 PR middle-end/55882
26211 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26212 account for bitpos when computing alignment.
26214 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26216 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26217 (ix86_target_macros_internal): Likewise.
26219 * config/i386/i386.c (m_CORE2I7): Removed.
26220 (m_CORE_HASWELL): New macro.
26221 (m_CORE_ALL): Likewise.
26222 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26223 (initial_ix86_arch_features): Likewise.
26224 (processor_target_table): Initializations for Core avx2.
26225 (cpu_names): New names "core-avx2".
26226 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26227 PROCESSOR_CORE_HASWELL.
26228 (ix86_issue_rate): New case.
26229 (ia32_multipass_dfa_lookahead): Likewise.
26230 (ix86_sched_init_global): Likewise.
26232 * config/i386/i386.h (TARGET_HASWELL): New macro.
26233 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26234 (processor_type): New PROCESSOR_HASWELL.
26236 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26238 PR tree-optimization/55955
26239 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26240 *SHIFT_EXPR and *ROTATE_EXPR codes.
26242 PR tree-optimization/48766
26243 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26244 -ftrapv disable -fwrapv.
26246 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26249 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26250 etc. to 1 and not to __flash.
26251 Use LL suffix for __INT24_MAX__ with -mint8.
26252 Use ULL suffix for __UINT24_MAX__ with -mint8.
26254 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26256 * config/avr/avr-arch.h
26257 (struct base_arch_s): Use typedef avr_arch_t instead.
26258 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26259 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26260 * config/avr/avr.c: Same.
26261 * config/avr/avr-devices.c: Same.
26262 * config/avr/driver-avr.c: Same.
26263 * config/avr/gen-avr-mmcu-texi.c: Same.
26264 * config/avr/avr-mcus.def: Adjust comment.
26266 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26268 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26269 * config/aarch64/iterators.md (VALLDI): New.
26271 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26272 Andi Kleen <ak@linux.intel.com>
26275 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26276 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26279 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26281 * config/avr/avr-stdint.h: Remove trailing blanks.
26282 * config/avr/avr-log.h: Same.
26283 * config/avr/avr-arch.h: Same.
26284 * config/avr/avr-devices.c: Same.
26285 * config/avr/avr-dimode.md: Same.
26286 * config/avr/predicates.md: Same.
26287 * config/avr/avr-c.c: Same. And fix typo.
26289 * config/avr/avr-protos.h: Same. And:
26290 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26291 (init_cumulative_args): Rename to avr_init_cumulative_args.
26292 (expand_prologue): Rename to avr_expand_prologue.
26293 (expand_epilogue): Rename to avr_expand_epilogue.
26294 (adjust_insn_length): Rename to avr_adjust_insn_length.
26295 (notice_update_cc): Rename to avr_notice_update_cc.
26296 (final_prescan_insn): Rename to avr_final_prescan_insn.
26297 * config/avr/avr.c: Same.
26298 * config/avr/avr.h: Same.
26299 * config/avr/avr.md: Remove trailing blanks.
26300 (prologue): Use avr_expand_prologue.
26301 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26303 2013-01-14 Richard Biener <rguenther@suse.de>
26305 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26306 verify_location, collect_subblocks): New functions.
26307 (verify_gimple_in_cfg): Verify that locations only reference
26308 BLOCKs in the functions BLOCK tree.
26310 2013-01-14 Richard Biener <rguenther@suse.de>
26312 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26314 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26316 (insert_out_of_ssa_copy_on_edge): Likewise.
26317 (rewrite_close_phi_out_of_ssa): Likewise.
26318 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26320 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26321 propagated constants.
26322 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26325 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26327 * config/avr/avr-modes.def: Add GPL copyright notice.
26329 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26331 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26332 MEMMODEL_MASK to determine memory model.
26333 (atomic_store<mode>): Ditto from operands[2].
26334 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26336 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26339 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26340 (fold_gimple_assign): Don't call unshare_expr here.
26341 (fold_ctor_reference): Call unshare_expr.
26343 2013-01-13 Terry Guo <terry.guo@arm.com>
26345 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26346 * doc/fragments.texi: Document MULTILIB_REUSE.
26347 * gcc.c (multilib_reuse): New internal spec.
26348 (set_multilib_dir): Also search multilib from multilib_reuse.
26349 * genmultilib (tmpmultilib3): Refactor code.
26350 (tmpmultilib4): Ditto.
26351 (multilib_reuse): New multilib argument.
26353 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26355 * Makefile.in: Update copyright.
26357 2013-01-12 Tom de Vries <tom@codesourcery.com>
26359 PR middle-end/55890
26360 * calls.c (expand_call): Check if arg_nr is valid.
26362 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26364 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26365 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26366 documentation. Add missing '__' in front of
26367 __builtin_ia32_packssdw256.
26369 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26372 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26373 NO_REGS for larl operands.
26374 (s390_reload_larl_operand): Use s390_load_address instead of
26377 2013-01-11 Richard Biener <rguenther@suse.de>
26379 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26380 (verify_node_sharing): ... here.
26381 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26383 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26385 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26386 Substitute TREECHECKING.
26387 * configure: Regenerate.
26388 * Makefile.in (TREECHECKING): New.
26390 2013-01-11 Richard Guenther <rguenther@suse.de>
26392 PR tree-optimization/44061
26393 * tree-vrp.c (extract_range_basic): Compute zero as
26394 value-range for __builtin_constant_p of function parameters.
26396 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26398 Update copyright years.
26400 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26402 PR rtl-optimization/55672
26403 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26404 const to be eliminable.
26406 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26408 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26409 * configure: Regenerate.
26411 2013-01-10 Richard Biener <rguenther@suse.de>
26413 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26415 2013-01-10 Richard Biener <rguenther@suse.de>
26418 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26419 locations for virtual PHI arguments.
26420 (rewrite_update_phi_arguments): Likewise.
26422 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26424 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26427 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26429 PR tree-optimization/55921
26430 * tree-complex.c (expand_complex_asm): New function.
26431 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26433 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26436 * config/s390/s390.c (s390_symref_operand_p)
26437 (s390_loadrelative_operand_p): Merge the two functions.
26438 (s390_check_qrst_address, print_operand_address): Add parameters
26439 to s390_loadrelative_operand_p invokation.
26440 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26441 (s390_reload_larl_operand, s390_secondary_reload): Use
26442 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26443 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26445 2013-01-09 Mike Stump <mikestump@comcast.net>
26447 * dse.c (record_store): Remove unnecessary assert.
26449 2013-01-09 Jan Hubicka <jh@suse.cz>
26451 PR tree-optimization/55569
26452 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26453 * cfgloop.h (scale_loop_profile): Likewise.
26455 2013-01-09 Jan Hubicka <jh@suse.cz>
26458 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26460 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26462 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26464 PR middle-end/55114
26465 * expr.h (maybe_emit_group_store): Declare.
26466 * expr.c (maybe_emit_group_store): New function.
26467 * builtins.c (expand_builtin_int_roundingfn): Call it.
26468 (expand_builtin_int_roundingfn_2): Likewise.
26470 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26472 PR rtl-optimization/55829
26473 * lra-constraints.c (match_reload): Add code for absent output.
26474 (curr_insn_transform): Add code for reloads of matched inputs
26477 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26479 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26480 attribute of movddup insn to DF.
26481 (*vec_interleave_lowv2df): Ditto.
26482 (vec_dupv2df): Ditto.
26484 2013-01-09 Jan Hubicka <jh@suse.cz>
26486 PR tree-optimiation/55875
26487 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26488 EVERY_ITERATION parameter.
26489 (number_of_iterations_exit): Check if exit is executed every iteration.
26490 (idx_infer_loop_bounds): Similarly here.
26491 (n_of_executions_at_most): Simplify
26492 to only test for cases where statement is dominated by the
26493 particular bound; handle correctly the "postdominance" test.
26494 (scev_probably_wraps_p): Use max loop iterations info
26495 as a global bound first.
26497 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26498 Nick Clifton <nickc@redhat.com>
26500 * config/v850/v850.md (cbranchsf4): New pattern.
26501 (cstoresf4): New pattern.
26502 (cbranchdf4): New pattern.
26503 (cstoredf4): New pattern.
26504 (movsicc): Disallow floating point comparisons.
26505 (cmpsf_le_insn): Fix order of operators.
26506 (cmpsf_lt_insn): Likewise.
26507 (cmpsf_eq_insn): Likewise.
26508 (cmpdf_le_insn): Likewise.
26509 (cmpdf_lt_insn): Likewise.
26510 (cmpdf_eq_insn): Likewise.
26511 (cmpsf_ge_insn): Use LE comparison.
26512 (cmpdf_ge_insn): Likewise.
26513 (cmpsf_gt_insn): Use LT comparison.
26514 (cmpdf_gt_insn): Likewise.
26515 (cmpsf_ne_insn): Delete pattern.
26516 (cmpdf_ne_insn): Delete pattern.
26517 * config/v850/v850.c (v850_gen_float_compare): Use
26518 gen_cmpdf_eq_insn for NE comparison.
26519 (v850_float_z_comparison_operator)
26520 (v850_float_nz_comparison_operator): Move from here ...
26521 * config/v850/predicates.md: ... to here. Move GT and GE
26522 comparisons into v850_float_z_comparison_operator.
26523 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26525 (v850_float_nz_comparison_operator): Likewise.
26527 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26529 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26530 with calls to gen_insvsi/gen_insvdi.
26532 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26534 * config/i386/i386.c (initial_ix86_tune_features): Set up
26535 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26537 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26538 Jakub Jelinek <jakub@redhat.com>
26540 PR tree-optimization/48189
26541 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26542 If nitercst is 0, don't predict the exit edge.
26544 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26546 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26547 in asm_fprintf with reg_names.
26548 (aarch64_print_operand_address): Likewise.
26549 (aarch64_return_addr): Likewise.
26550 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26552 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26554 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26555 (INT_U6_BITS): Likewise.
26556 * config/pa/predicates.md (uint6_operand): New predicate.
26557 (shift5_operand, shift6_operand): Likewise.
26558 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26560 (lshrdi3): Use shift6_operand.
26561 (shrpsi4, shrpdi4): New insn patterns.
26562 (extzv): Delete expander.
26563 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26564 predicates in unamed zero extract patterns. Tighten common constraint.
26565 (extv): Delete expander.
26566 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26567 predicates in unamed sign extract patterns. Tighten common constraint.
26568 (insv): Delete expander.
26569 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26570 predicates in unamed insert patterns. Tighten common constraint.
26571 Change uint32_operand predicate to uint6_operand predicate in unamed
26572 DImode pattern to insert constant values of type 1...1xxxx.
26574 2013-01-04 Jan Hubicka <jh@suse.cz>
26576 PR tree-optimization/55823
26577 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26580 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26581 Uros Bizjak <ubizjak@gmail.com>
26583 PR rtl-optimization/55845
26584 * df-problems.c (can_move_insns_across): Stop scanning at
26585 volatile_insn_p source instruction or give up if
26586 across_from .. across_to range contains any volatile_insn_p
26589 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26591 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26592 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26594 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26595 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26597 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26600 * asan.c (asan_clear_shadow): New function.
26601 (asan_emit_stack_protection): Use it.
26603 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26605 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26606 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26607 with tab instead of space.
26609 2013-01-08 Nick Clifton <nickc@redhat.com>
26611 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26612 register bank 0 at the start of an interrupt handler.
26613 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26616 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26618 * config/aarch64/aarch64-simd.md
26619 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26620 (aarch64_simd_bsl): Likewise.
26621 (aarch64_vcond_internal<mode>): Likewise.
26622 (vcond<mode><mode>): Likewise.
26623 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26624 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26626 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26628 * config/aarch64/aarch64-builtins.c
26629 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26631 2013-01-08 Martin Jambor <mjambor@suse.cz>
26634 * tree-sra.c (analyze_access_subtree): Return true also after
26635 potentially creating a debug-only replacement.
26637 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26639 PR middle-end/55890
26640 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26642 PR tree-optimization/54120
26643 * tree-vrp.c (range_fits_type_p): Don't allow
26644 src_precision < precision from signed vr to unsigned_p
26645 if vr->min or vr->max is negative.
26646 (simplify_float_conversion_using_ranges): Test can_float_p
26647 against CODE_FOR_nothing.
26649 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26650 Richard Biener <rguenther@suse.de>
26652 PR middle-end/55851
26653 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26654 types instead of just INTEGER_TYPE types.
26656 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26658 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26661 2013-01-07 Steve Ellcey <sellcey@mips.com>
26664 * config/mips/mips.opt: Change mad to mmad to match documentation.
26666 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26669 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26670 .progmemx.data now.
26672 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26675 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26676 (avr_addrspace_t): Add .section_name field.
26677 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26679 (avr_addrspace): Same. Initialize .section_name. Remove last
26680 NULL entry. Put __memx into .progmemx.data.
26681 (progmem_section_prefix): Remove.
26682 (avr_asm_init_sections): No need to initialize progmem_section.
26683 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26684 section name prefix.
26685 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26686 retrieve the progmem section.
26687 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26688 boundary to run over avr_addrspace[].
26689 (avr_register_target_pragmas): Ditto.
26691 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26693 * varasm.c (output_constant_def_contents): For asan_protect_global
26694 protected strings, adjust DECL_ALIGN if needed, before testing for
26696 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26697 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26699 (output_object_block): For asan protected decls, emit asan padding
26700 after their contents.
26701 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26702 (asan_finish_file): Test it here instead.
26704 2013-01-07 Nick Clifton <nickc@redhat.com>
26705 Matthias Klose <doko@debian.org>
26706 Doug Kwan <dougkwan@google.com>
26707 H.J. Lu <hongjiu.lu@intel.com>
26710 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26712 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26714 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26716 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26718 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26720 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26723 * doc/install.texi (Cross-Compiler-Specific Options): Document
26726 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26728 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26729 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26730 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26731 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26732 vqmovun_high_s64): Fix source operand number and update copyright.
26734 2013-01-07 Richard Biener <rguenther@suse.de>
26736 PR middle-end/55890
26737 * gimple.h (gimple_call_builtin_p): New overload.
26738 * gimple.c (validate_call): New function.
26739 (gimple_call_builtin_p): Likewise.
26740 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26741 Use gimple_call_builtin_p.
26742 (find_func_clobbers): Likewise.
26743 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26744 (strlen_optimize_stmt): Likewise.
26746 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26748 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26749 (vld1q_dup_*): Likewise.
26750 (vld1_*): Likewise.
26751 (vld1q_*): Likewise.
26752 (vld1_lane_*): Likewise.
26753 (vld1q_lane_*): Likewise.
26755 2013-01-07 Richard Biener <rguenther@suse.de>
26757 * lto-streamer.h (LTO_minor_version): Bump to 2.
26759 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26761 * config/aarch64/aarch64-protos.h
26762 (aarch64_const_double_zero_rtx_p): Rename to...
26763 (aarch64_float_const_zero_rtx_p): ...this.
26764 (aarch64_float_const_representable_p): New.
26765 (aarch64_output_simd_mov_immediate): Likewise.
26766 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26767 move immediate case.
26768 * config/aarch64/aarch64.c
26769 (aarch64_const_double_zero_rtx_p): Rename to...
26770 (aarch64_float_const_zero_rtx_p): ...this.
26771 (aarch64_print_operand): Allow printing of new constants.
26772 (aarch64_valid_floating_const): New.
26773 (aarch64_legitimate_constant_p): Check for valid floating-point
26775 (aarch64_simd_valid_immediate): Likewise.
26776 (aarch64_vect_float_const_representable_p): New.
26777 (aarch64_float_const_representable_p): Likewise.
26778 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26779 (aarch64_output_simd_mov_immediate): New.
26780 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26781 (*movdf_aarch64): Likewise.
26782 * config/aarch64/constraints.md (Ufc): New.
26783 (Y): call aarch64_float_const_zero_rtx.
26784 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26786 2013-01-07 Richard Biener <rguenther@suse.de>
26788 PR tree-optimization/55888
26789 PR tree-optimization/55862
26790 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26791 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26792 not if it is contained therein.
26794 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26796 * config/avr/t-avr: Typo.
26798 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26801 * config/avr/t-avr: Don't automatically rebuild
26802 $(srcdir)/config/avr/t-multilib
26803 $(srcdir)/config/avr/avr-tables.opt
26804 $(srcdir)/doc/avr-mmcu.texi
26805 (avr-mcus): New phony target to build them on request.
26806 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26807 * avr/avr-mcus.def: Adjust comments.
26809 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26811 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26813 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26815 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26817 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26819 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26821 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26823 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26824 to generate profiling.
26825 * config/rs6000/aix64.h (LIB_SPEC): Same.
26827 2013-01-04 Andrew Pinski <apinski@cavium.com>
26829 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26831 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26833 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26835 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26837 (ix86_expand_move): Ditto.
26838 (ix86_zero_extend_to_Pmode): Ditto.
26839 (ix86_expand_call): Ditto.
26840 (ix86_expand_special_args_builtin): Ditto.
26841 (ix86_expand_builtin): Ditto.
26843 2013-01-04 Richard Biener <rguenther@suse.de>
26845 PR tree-optimization/55862
26846 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26847 translating them through PHI nodes.
26849 2013-01-04 Martin Jambor <mjambor@suse.cz>
26851 PR tree-optimization/55755
26852 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26853 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26855 2013-01-04 Marek Polacek <polacek@redhat.com>
26857 PR middle-end/55859
26858 * opts.c (default_options_optimization): Clarify error message.
26860 2013-01-04 Richard Biener <rguenther@suse.de>
26862 PR middle-end/55863
26863 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26866 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26869 * config/pa/pa.md (movsi): Revert previous change.
26870 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26873 2013-01-03 Richard Henderson <rth@redhat.com>
26875 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26876 after eliminating TLS symbols.
26878 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26881 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26882 * graphite-poly.c (debug_gmp_value): Likewise.
26884 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26887 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26888 selected code model, define __code_mode_small__, __code_model_medium__,
26889 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26890 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26891 xchg temporary register with %k. Declare temporary register as
26893 [__x86_64__]: For medium and large code models, preserve %rbx register.
26895 2013-01-03 Richard Biener <rguenther@suse.de>
26897 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26898 (dump_subscript): Adjust.
26899 (finalize_ddr_dependent): Do not dump redundant info.
26900 (analyze_siv_subscript): Adjust.
26901 (subscript_dependence_tester): Likewise.
26902 (compute_affine_dependence): Likewise.
26904 2013-01-03 Richard Biener <rguenther@suse.de>
26907 2013-01-03 Richard Biener <rguenther@suse.de>
26909 PR tree-optimization/55857
26910 * tree-vect-stmts.c (vectorizable_load): Do not setup
26911 re-alignment for invariant loads.
26913 2013-01-02 Richard Biener <rguenther@suse.de>
26915 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26916 invariant load do not generate a vector load from the scalar location.
26918 2013-01-03 Richard Biener <rguenther@suse.de>
26920 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26921 for not vectorizing.
26922 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26923 not build INDIRECT_REFs, call get_name once only.
26924 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26925 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26927 2013-01-03 Richard Biener <rguenther@suse.de>
26929 PR tree-optimization/55857
26930 * tree-vect-stmts.c (vectorizable_load): Do not setup
26931 re-alignment for invariant loads.
26933 2013-01-03 Richard Biener <rguenther@suse.de>
26936 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26937 prefer a built-in decl.
26939 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26941 * gcc.c (process_command): Update copyright notice dates.
26942 * gcov.c (print_version): Likewise.
26943 * gcov-dump.c (print_version): Likewise.
26945 PR rtl-optimization/55838
26946 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26947 iv0.step, iv1.step and step.
26949 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26950 Marc Glisse <marc.glisse@inria.fr>
26952 PR tree-optimization/55832
26953 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26954 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26955 integer_{one,zero}_node.
26957 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26960 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26961 * var-tracking.c (reverse_op): Don't add reverse ops to
26962 VALUEs that have already
26963 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26965 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26967 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26969 2013-01-02 Teresa Johnson <tejohnson@google.com>
26971 * dumpfile.c (dump_loc): Print filename with location.
26972 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26973 new location_t parameter to emit complete unroll message with
26974 new dump framework.
26975 (canonicalize_loop_induction_variables): Compute loops location
26976 and pass to try_unroll_loop_completely.
26977 * loop-unroll.c (report_unroll_peel): New function.
26978 (peel_loops_completely): Use new dump format with location
26979 for main dumpfile message, and invoke report_unroll_peel on success.
26980 (decide_unrolling_and_peeling): Ditto.
26981 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26982 by report_unroll_peel.
26983 (decide_peel_completely): Ditto.
26984 (decide_unroll_constant_iterations): Ditto.
26985 (decide_unroll_runtime_iterations): Ditto.
26986 (decide_peel_simple): Ditto.
26987 (decide_unroll_stupid): Ditto.
26988 * cfgloop.c (get_loop_location): New function.
26989 * cfgloop.h (get_loop_location): Declare.
26991 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26993 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26996 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26998 PR middle-end/55198
26999 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27000 BLKmode objects when EXPAND_MEMORY is specified.
27002 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27004 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27006 (fold_builtin_cpu): Do not share cpu model decls across statements.
27008 2013-01-02 Jason Merrill <jason@redhat.com>
27011 * tree.c (build_array_type_1): Revert earlier change.
27013 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27015 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27017 * config/aarch64/aarch64-tune.md: Re-generate.
27019 2013-01-02 Richard Biener <rguenther@suse.de>
27021 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27022 invariant load do not generate a vector load from the scalar location.
27024 2013-01-02 Richard Biener <rguenther@suse.de>
27027 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27028 * configure: Regenerate.
27030 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27032 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27033 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27034 (expand_builtin_int_roundingfn_2): Keep the original target around
27035 for the fallback case.
27037 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27039 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27040 to be clear for sign changes.
27042 2013-01-01 Jan Hubicka <jh@suse.cz>
27044 * ipa-inline-analysis.c: Fix formatting.
27046 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27048 PR tree-optimization/55831
27049 * tree-vect-loop.c (get_initial_def_for_induction): Use
27050 gsi_after_labels instead of gsi_start_bb.
27052 Copyright (C) 2013 Free Software Foundation, Inc.
27054 Copying and distribution of this file, with or without modification,
27055 are permitted in any medium without royalty provided the copyright
27056 notice and this notice are preserved.