1 2013-10-18 Teresa Johnson <tejohnson@google.com>
3 * predict.c (probably_never_executed): Compare frequency-based
4 count to number of training runs.
5 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
7 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9 * config/arm/arm.c (cortexa9_extra_costs): New table.
10 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
12 2013-10-18 Jeff Law <law@redhat.com>
14 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
16 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
17 (register_jump_thread): Add "registered_jump_thread" debug counter support.
18 * dbgcnt.def (registered_jump_thread): New debug counter.
20 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
22 * config/rs6000/rs6000.c: Include cgraph.h.
24 2013-10-18 Teresa Johnson <tejohnson@google.com>
26 * tree-ssa-tail-merge.c (replace_block_by): Update edge
27 weights during merging.
29 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
31 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
32 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
33 * tree-ssa-operands.h: Remove prototype.
34 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
35 * gimple.c (virtual_operand_p): Relocate from gimple.c.
36 * gimple.h: Add prototype.
37 * gimple-ssa.h: Include tree-ssa-operands.h.
38 * tree-dump.c: Add tree-cfg.h to include list.
39 * tree-ssa-alias.c: Add ipa-reference.h to include list.
40 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
41 * config/i386/i386.c: Don't include tree-flow.h.
42 * config/rs6000/rs6000.c: Likewise.
44 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
46 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
48 2013-10-18 Richard Biener <rguenther@suse.de>
50 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
51 or TYPE_UNSIGNED of integral types.
52 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
53 NULL_TREE for zero-precision integral types.
55 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
57 * config/aarch64/arm_neon.h
58 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
60 2013-10-17 Sriraman Tallam <tmsriram@google.com>
62 * opth-gen.awk: Define target_flags_explicit.
64 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
66 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
67 fields to the reg_addr array that describes the valid addressing
68 mode for any register, general purpose registers, floating point
69 registers, and Altivec registers.
70 (FIRST_RELOAD_REG_CLASS): Likewise.
71 (LAST_RELOAD_REG_CLASS): Likewise.
72 (struct reload_reg_map_type): Likewise.
73 (reload_reg_map_type): Likewise.
74 (RELOAD_REG_VALID): Likewise.
75 (RELOAD_REG_MULTIPLE): Likewise.
76 (RELOAD_REG_INDEXED): Likewise.
77 (RELOAD_REG_OFFSET): Likewise.
78 (RELOAD_REG_PRE_INCDEC): Likewise.
79 (RELOAD_REG_PRE_MODIFY): Likewise.
81 (mode_supports_pre_incdec_p): New helper functions to say whether
82 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
83 (mode_supports_pre_modify_p): Likewise.
84 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
85 print the valid address mode bits for each mode.
86 (rs6000_debug_print_mode): Likewise.
87 (rs6000_debug_reg_global): Likewise.
88 (rs6000_setup_reg_addr_masks): New function to set up the address
89 mask bits for each type.
90 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
91 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
92 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
93 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
94 PRE_MODIFY are supported.
95 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
96 registers, instead of {src,dest}_av_p.
97 (rs6000_print_options_internal): Tweak the debug output slightly.
99 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
101 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
104 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
106 * tree-flow.h (struct omp_region): Move to omp-low.c.
107 Remove omp_ prototypes and variables.
108 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
109 (copy_var_decl): Relocate prototype from tree-flow.h.
110 * gimple.c (copy_var_decl): Relocate from omp-low.c.
111 * tree.h: Move prototype to omp-low.h.
112 * omp-low.h: New File. Relocate prototypes here.
113 * omp-low.c (struct omp_region): Make local here.
114 (root_omp_region): Make static.
115 (copy_var_decl) Move to gimple.c.
116 (new_omp_region): Make static.
117 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
118 * tree-cfg.c: Include omp-low.h.
119 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
120 * gimplify.c: Include omp-low.h.
121 * tree-parloops.c: Likewise.
123 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
125 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
126 operands are in memory, prefer to force non-matched operand 1 to
129 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
132 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
133 restrict TImode addresses to single indirect registers if both
134 -mquad-memory and -mvsx-timode are used.
135 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
136 we should emit load/store quad. Remove using %y for quad memory
139 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
140 constraints to allow load/store quad on machines where TImode is
141 not allowed in VSX registers. Use 'n' instead of 'F' constraint
142 for TImode to load integer constants.
144 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
146 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
148 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
150 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
151 handling of STACK_REG.
153 2013-10-17 Richard Biener <rguenther@suse.de>
155 PR tree-optimization/58143
156 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
158 (rewrite_to_defined_overflow): Likewise.
159 (move_computations_dom_walker::before_dom): Rewrite stmts
160 with undefined signed overflow that are not always executed
161 into unsigned arithmetic.
163 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
166 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
167 explicit isa flag to be an options variable, instead of using
168 global_options_set. Remove define from rs6000.h.
169 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
171 * config/rs6000/rs6000.c (rs6000_option_override_internal):
172 Initialize rs6000_isa_flags_explicit.
173 (rs6000_function_specific_save): Add gcc_options* parameter, so
174 that the powerpc builds after the 2013-10-15 changes.
175 (rs6000_function_specific_restore): Likewise.
177 2013-10-16 DJ Delorie <dj@redhat.com>
179 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
180 op is a REG before checking REGNO.
181 (rl78_alloc_physical_registers): Verify pattern is a SET before
184 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
186 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
188 (vec_unpacks_lo_v4sf): Likewise.
189 (vec_unpacks_float_hi_v4si): Likewise.
190 (vec_unpacks_float_lo_v4si): Likewise.
191 (vec_unpacku_float_hi_v4si): Likewise.
192 (vec_unpacku_float_lo_v4si): Likewise.
194 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
196 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
197 (vsx_concat_v2sf): Likewise.
199 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
201 * config/aarch64/aarch64.md
202 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
204 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
206 PR tree-optimization/58697
207 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
208 estimated_loop_iterations_int.
209 (max_stmt_executions_int): Call get_max_loop_iterations_int.
210 (get_max_loop_iterations_int): New. HWINT version of
211 get_max_loop_iterations.
212 * cfgloop.h: Add prototypes.
213 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
214 * loop-unroll.c (decide_peel_once_rolling): Call
215 get_estimated_loop_iterations_int.
216 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
217 * tree-ssa-loop-niter.h: Tweak prototypes.
219 2013-10-16 David Malcolm <dmalcolm@redhat.com>
221 * gengtype-parse.c (struct_field_seq): Ignore access-control
222 keywords ("public:" etc).
224 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
226 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
227 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
229 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
231 * config/arm/arm.opt (mlra): New option.
232 * config/arm/arm.c (arm_lra_p): New function.
233 (TARGET_LRA_P): Define.
235 2013-10-16 Paulo Matos <pmatos@broadcom.com>
237 * tree-core.h (tree_code_name): Remove.
238 * tree.h (get_tree_code_name): New prototype.
239 * tree.c (tree_code_name): Make static.
240 (get_tree_code_name): New function.
241 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
242 tree_class_check_failed, tree_range_check_failed,
243 tree_not_class_check_failed, omp_clause_check_failed,
244 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
245 wrapper get_tree_code_name instead of calling tree_code_name directly.
246 * tree-vrp.c (dump_asserts_for): Likewise.
247 * tree-dump.c (dequeue_and_dump): Likewise.
248 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
249 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
250 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
251 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
252 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
253 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
254 dump_gimple_omp_for): Likewise.
255 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
256 * tree-ssa-pre.c (print_pre_expr): Likewise.
257 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
258 * print-tree.c (print_node_brief, print_node): Likewise.
259 * gimple.c (gimple_check_failed): Likewise.
260 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
261 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
262 * config/mep/mep.c (mep_validate_vliw): Likewise.
263 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
264 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
266 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
268 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
271 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
273 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
274 (MULTILIB_MATCHES): Add multilib for -march=v8.
276 2013-10-15 Sriraman Tallam <tmsriram@google.com>
279 * optc-save-gen.awk: Add extra parameter to the save and restore
281 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
282 * tree.c (build_optimization_node): New parameter. Add extra parameter
283 to call to cl_optimization_save.
284 (build_target_option_node): New parameter. Add extra parameter
285 to call to cl_target_option_save.
286 * tree.h (build_optimization_node): New parameter.
287 (build_target_option_node): New parameter.
288 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
289 build_optimization_node and build_target_option_node.
290 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
291 (handle_pragma_push_options): Ditto.
292 * toplev.c (process_options): Ditto.
293 * opts.c (init_options_struct): Check for opts_set non-null.
294 * target.def (target_option.save): New parameter.
295 (target_option.restore): New parameter.
297 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
298 (ix86_pragma_target_parse): Ditto.
299 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
301 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
302 to build_optimization_node and build_target_option_node.
303 (rs6000_valid_attribute_p): Ditto.
304 (rs6000_pragma_target_parse): Ditto.
305 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
307 * config/i386/i386.h:
308 TARGET_64BIT_P: New Macro
309 TARGET_MMX_P: New Macro.
310 TARGET_3DNOW_P: New Macro.
311 TARGET_3DNOW_A_P: New Macro.
312 TARGET_SSE_P: New Macro.
313 TARGET_SSE2_P: New Macro.
314 TARGET_SSE3_P: New Macro.
315 TARGET_SSSE3_P: New Macro.
316 TARGET_SSE4_1_P: New Macro.
317 TARGET_SSE4_2_P: New Macro.
318 TARGET_AVX_P: New Macro.
319 TARGET_AVX2_P: New Macro.
320 TARGET_AVX512F_P: New Macro.
321 TARGET_AVX512PF_P: New Macro.
322 TARGET_AVX512ER_P: New Macro.
323 TARGET_AVX512CD_P: New Macro.
324 TARGET_FMA_P: New Macro.
325 TARGET_SSE4A_P: New Macro.
326 TARGET_FMA4_P: New Macro.
327 TARGET_XOP_P: New Macro.
328 TARGET_LWP_P: New Macro.
329 TARGET_ABM_P: New Macro.
330 TARGET_BMI_P: New Macro.
331 TARGET_BMI2_P: New Macro.
332 TARGET_LZCNT_P: New Macro.
333 TARGET_TBM_P: New Macro.
334 TARGET_POPCNT_P: New Macro.
335 TARGET_SAHF_P: New Macro.
336 TARGET_MOVBE_P: New Macro.
337 TARGET_CRC32_P: New Macro.
338 TARGET_AES_P: New Macro.
339 TARGET_PCLMUL_P: New Macro.
340 TARGET_CMPXCHG16B_P: New Macro.
341 TARGET_FSGSBASE_P: New Macro.
342 TARGET_RDRND_P: New Macro.
343 TARGET_F16C_P: New Macro.
344 TARGET_RTM_P: New Macro.
345 TARGET_HLE_P: New Macro.
346 TARGET_RDSEED_P: New Macro.
347 TARGET_PRFCHW_P: New Macro.
348 TARGET_ADX_P: New Macro.
349 TARGET_FXSR_P: New Macro.
350 TARGET_XSAVE_P: New Macro.
351 TARGET_XSAVEOPT_P: New Macro.
352 TARGET_LP64_P: New Macro.
353 TARGET_X32_P: New Macro.
354 TARGET_FPMATH_DEFAULT_P: New Macro.
355 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
356 * config/i386/i386.c (ix86_option_override_internal): New parameters.
358 Change ix86_tune_string to access opts->x_ix86_tune_string.
359 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
360 Change ix86_arch_string to access opts->x_ix86_arch_string.
361 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
362 Change ix86_pmode to access opts->x_ix86_pmode.
363 Change ix86_abi to access opts->x_ix86_abi.
364 Change ix86_cmodel to access opts->x_ix86_cmodel.
365 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
366 Change ix86_isa_flags_explicit to access
367 opts->x_ix86_isa_flags_explicit.
368 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
369 Change ix86_regparm to access opts->x_ix86_regparm.
370 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
371 Change ix86_preferred_stack_boundary_arg to access
372 opts->x_ix86_preferred_stack_boundary_arg.
373 Change ix86_force_align_arg_pointer to access
374 opts->x_ix86_force_align_arg_pointer.
375 Change ix86_incoming_stack_boundar_arg to access
376 opts->x_ix86_incoming_stack_boundar_arg.
377 Change ix86_fpmath to access opts->x_ix86_fpmath.
378 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
379 Change ix86_recip_name to access opts->x_ix86_recip_name.
380 Change ix86_stack_protector_guard to access
381 opts->x_ix86_stack_protector_guard.
382 Change ix86_tune_memcpy_strategy to access
383 opts->x_ix86_tune_memcpy_strategy.
384 Change ix86_tune_memset_strategy to access
385 opts->x_ix86_tune_memset_strategy.
386 Change global_options to access opts.
387 Change global_options_set to access opts_set.
388 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
389 Change TARGET_MMX to TARGET_MMX_P (opts->...).
390 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
391 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
392 Change TARGET_SSE to TARGET_SSE_P (opts->...).
393 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
394 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
395 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
396 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
397 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
398 Change TARGET_AVX to TARGET_AVX_P (opts->...).
399 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
400 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
401 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
402 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
403 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
404 Change TARGET_FMA to TARGET_FMA_P (opts->...).
405 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
406 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
407 Change TARGET_XOP to TARGET_XOP_P (opts->...).
408 Change TARGET_LWP to TARGET_LWP_P (opts->...).
409 Change TARGET_ABM to TARGET_ABM_P (opts->...).
410 Change TARGET_BMI to TARGET_BMI_P (opts->...).
411 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
412 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
413 Change TARGET_TBM to TARGET_TBM_P (opts->...).
414 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
415 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
416 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
417 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
418 Change TARGET_AES to TARGET_AES_P (opts->...).
419 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
420 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
421 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
422 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
423 Change TARGET_F16C to TARGET_F16C_P (opts->...).
424 Change TARGET_RTM to TARGET_RTM_P (opts->...).
425 Change TARGET_HLE to TARGET_HLE_P (opts->...).
426 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
427 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
428 Change TARGET_ADX to TARGET_ADX_P (opts->...).
429 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
430 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
431 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
432 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
433 Change TARGET_X32 to TARGET_X32_P (opts->...).
434 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
435 Change TARGET_FLOAT_RETURNS_IN_80387 to
436 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
437 (ix86_function_specific_save): New parameter. Use opts-> fields
438 to replace global fields.
439 (ix86_function_specific_restore): Ditto.
440 (ix86_valid_target_attribute_inner_p): New parameters.
442 Fix call to ix86_handle_option and set_option.
443 (ix86_valid_target_attribute_tree): New parameters.
444 Change global_options to access opts.
445 Change global_options_set to access opts_set.
446 Fix call to ix86_valid_target_attribute_inner_p.
447 Change ix86_tune_string to access opts->x_ix86_tune_string.
448 Change ix86_arch_string to access opts->x_ix86_arch_string.
449 Change ix86_fpmath to access opts->x_ix86_fpmath
450 Fix call to ix86_option_override_internal.
451 Fix call to ix86_add_new_builtins.
452 Fix calls to build_optimization_node and build_target_option_node.
453 (ix86_valid_target_attribute_p): Remove access to global_options.
454 Use new gcc_options structure func_options.
455 Fix call to ix86_valid_target_attribute_tree.
456 Fix call to build_optimization_node.
457 (get_builtin_code_for_version): Fix call to
458 ix86_valid_target_attribute_tree.
460 2013-10-15 David Malcolm <dmalcolm@redhat.com>
462 * Makefile.in (PICFLAG): New.
463 (enable_host_shared): New.
464 (INTERNAL_CFLAGS): Use PICFLAG.
465 (LIBIBERTY): Use pic build of libiberty.a if configured with
466 --enable-host-shared.
467 * configure.ac: Add --enable-host-shared, setting up new
469 * configure: Regenerate.
470 * doc/install.texi (--enable-shared): Add note contrasting it with ...
471 (--enable-host-shared): New option.
473 2013-10-15 Richard Biener <rguenther@suse.de>
475 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
476 for built-in functions.
478 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
480 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
481 (optimize_range_tests_1): New function,
482 extracted from optimize_range_tests.
483 (optimize_range_tests_xor): Similarly.
484 (optimize_range_tests_diff): New function.
485 (optimize_range_tests): Use optimize_range_tests_1.
487 2013-10-15 Cong Hou <congh@google.com>
489 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
490 requirement of the reduction pattern so that one operand of the
491 reduction operation can come from outside of the loop.
493 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
495 * config/arm/neon-schedgen.ml: Remove.
496 * config/arm/cortex-a9-neon.md: Remove comment regarding
499 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
501 * config/arm/types: Remove old neon types.
503 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
505 * config/arm/cortex-a7.md
506 (cortex_a7_neon_type): New.
507 (cortex_a7_neon_mul): Update for new types.
508 (cortex_a7_neon_mla): Likewise.
509 (cortex_a7_neon): Likewise.
511 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
513 * config/arm/cortex-a15-neon.md
514 (cortex_a15_neon_type): New,
516 (cortex_a15_neon_int_1): Remove.
517 (cortex_a15_neon_int_2): Likewise.
518 (cortex_a15_neon_int_3): Likewise.
519 (cortex_a15_neon_int_4): Likewise.
520 (cortex_a15_neon_int_5): Likewise.
521 (cortex_a15_neon_vqneg_vqabs): Likewise.
522 (cortex_a15_neon_vmov): Likewise.
523 (cortex_a15_neon_vaba): Likewise.
524 (cortex_a15_neon_vaba_qqq): Likewise.
525 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
526 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
527 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
529 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
530 (cortex_a15_neon_mla_qqq_8_16): Likewise.
531 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
532 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
533 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
534 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
535 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
536 (cortex_a15_neon_shift_1): Likewise.
537 (cortex_a15_neon_shift_2): Likewise.
538 (cortex_a15_neon_shift_3): Likewise.
539 (cortex_a15_neon_vshl_ddd): Likewise.
540 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
541 (cortex_a15_neon_vsra_vrsra): Likewise.
542 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
543 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
544 (cortex_a15_neon_bp_3cycle): Likewise.
545 (cortex_a15_neon_ldm_2): Likewise.
546 (cortex_a15_neon_stm_2): Likewise.
547 (cortex_a15_neon_mcr): Likewise.
548 (cortex_a15_neon_mrc): Likewise.
549 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
550 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
551 (cortex_a15_neon_fp_vmul_ddd): Likewise.
552 (cortex_a15_neon_fp_vmul_qqd): Likewise.
553 (cortex_a15_neon_fp_vmla_ddd): Likewise.
554 (cortex_a15_neon_fp_vmla_qqq): Likewise.
555 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
556 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
557 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
558 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
559 (cortex_a15_neon_bp_simple): Likewise.
560 (cortex_a15_neon_bp_2cycle): Likewise.
561 (cortex_a15_neon_bp_3cycle): Likewise.
562 (cortex_a15_neon_vld1_1_2_regs): Likewise.
563 (cortex_a15_neon_vld1_3_4_regs): Likewise.
564 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
565 (cortex_a15_neon_vld2_4_regs): Likewise.
566 (cortex_a15_neon_vld3_vld4): Likewise.
567 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
568 (cortex_a15_neon_vst1_3_4_regs): Likewise.
569 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
570 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
571 (cortex_a15_neon_vst3_vst4): Rename to...
572 (cortex_a15_neon_vst4): This, update for new attributes.
573 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
574 (cortex_a15_neon_vld3_vld4_lane): Likewise.
575 (cortex_a15_neon_vst1_vst2_lane): Likewise.
576 (cortex_a15_neon_vst3_vst4_lane): Likewise.
577 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
578 (cortex_a15_neon_ldm_2): Likewise.
579 (cortex_a15_neon_stm_2): Likewise.
580 (cortex_a15_neon_mcr): Likewise.
581 (cortex_a15_neon_mcr_2_mcrr): Likewise.
582 (cortex_a15_neon_mrc): Likewise.
583 (cortex_a15_neon_mrrc): Likewise.
585 (cortex_a15_neon_abd): New.
586 (cortex_a15_neon_abd_q): Likewise.
587 (cortex_a15_neon_aba): Likewise.
588 (cortex_a15_neon_aba_q): Likewise.
589 (cortex_a15_neon_acc): Likewise.
590 (cortex_a15_neon_acc_q): Likewise.
591 (cortex_a15_neon_arith_basic): Likewise.
592 (cortex_a15_neon_arith_complex): Likewise.
593 (cortex_a15_neon_multiply): Likewise.
594 (cortex_a15_neon_multiply_q): Likewise.
595 (cortex_a15_neon_mla): Likewise.
596 (cortex_a15_neon_mla_q): Likewise.
597 (cortex_a15_neon_sat_mla_long): Likewise.
598 (cortex_a15_neon_shift_acc): Likewise.
599 (cortex_a15_neon_shift_imm_basic): Likewise.
600 (cortex_a15_neon_shift_imm_complex): Likewise.
601 (cortex_a15_neon_shift_reg_basic): Likewise.
602 (cortex_a15_neon_shift_reg_basic_q): Likewise.
603 (cortex_a15_neon_shift_reg_complex): Likewise.
604 (cortex_a15_neon_shift_reg_complex_q): Likewise.
605 (cortex_a15_neon_fp_negabs): Likewise
606 (cortex_a15_neon_fp_arith): Likewise
607 (cortex_a15_neon_fp_arith_q): Likewise
608 (cortex_a15_neon_fp_cvt_int): Likewise
609 (cortex_a15_neon_fp_cvt_int_q): Likewise
610 (cortex_a15_neon_fp_cvt_16): Likewise
611 (cortex_a15_neon_fp_mul): Likewise
612 (cortex_a15_neon_fp_mul_q): Likewise
613 (cortex_a15_neon_fp_mla): Likewise
614 (cortex_a15_neon_fp_mla_q): Likewise
615 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
616 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
617 (cortex_a15_neon_bitops): Likewise.
618 (cortex_a15_neon_bitops_q): Likewise.
619 (cortex_a15_neon_from_gp): Likewise.
620 (cortex_a15_neon_from_gp_q): Likewise.
621 (cortex_a15_neon_tbl3_tbl4): Likewise.
622 (cortex_a15_neon_zip_q): Likewise.
623 (cortex_a15_neon_to_gp): Likewise.
624 (cortex_a15_neon_load_a): Likewise.
625 (cortex_a15_neon_load_b): Likewise.
626 (cortex_a15_neon_load_c): Likewise.
627 (cortex_a15_neon_load_d): Likewise.
628 (cortex_a15_neon_load_e): Likewise.
629 (cortex_a15_neon_load_f): Likewise.
630 (cortex_a15_neon_store_a): Likewise.
631 (cortex_a15_neon_store_b): Likewise.
632 (cortex_a15_neon_store_c): Likewise.
633 (cortex_a15_neon_store_d): Likewise.
634 (cortex_a15_neon_store_e): Likewise.
635 (cortex_a15_neon_store_f): Likewise.
636 (cortex_a15_neon_store_g): Likewise.
637 (cortex_a15_neon_store_h): Likewise.
638 (cortex_a15_vfp_to_from_gp): Likewise.
640 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
642 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
644 (cortex_a9_neon_vshl_ddd): Remove.
645 (cortex_a9_neon_vst3_vst4): Likewise.
646 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
648 (cortex_a9_neon_bit_ops_q): New.
650 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
651 (cortex_a9_neon_int_2): Likewise.
652 (cortex_a9_neon_int_3): Likewise.
653 (cortex_a9_neon_int_4): Likewise.
654 (cortex_a9_neon_int_5): Likewise.
655 (cortex_a9_neon_vqneg_vqabs): Likewise.
656 (cortex_a9_neon_vmov): Likewise.
657 (cortex_a9_neon_vaba): Likewise.
658 (cortex_a9_neon_vaba_qqq): Likewise.
659 (cortex_a9_neon_shift_1): Likewise.
660 (cortex_a9_neon_shift_2): Likewise.
661 (cortex_a9_neon_shift_3): Likewise.
662 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
663 (cortex_a9_neon_vsra_vrsra): Likewise.
664 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
665 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
666 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
668 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
669 (cortex_a9_neon_mla_qqq_8_16): Likewise.
670 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
672 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
673 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
674 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
675 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
676 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
677 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
678 (cortex_a9_neon_fp_vsum): Likewise.
679 (cortex_a9_neon_fp_vmul_ddd): Likewise.
680 (cortex_a9_neon_fp_vmul_qqd): Likewise.
681 (cortex_a9_neon_fp_vmla_ddd): Likewise.
682 (cortex_a9_neon_fp_vmla_qqq): Likewise.
683 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
684 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
685 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
686 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
687 (cortex_a9_neon_bp_simple): Likewise.
688 (cortex_a9_neon_bp_2cycle): Likewise.
689 (cortex_a9_neon_bp_3cycle): Likewise.
690 (cortex_a9_neon_ldr): Likewise.
691 (cortex_a9_neon_str): Likewise.
692 (cortex_a9_neon_vld1_1_2_regs): Likewise.
693 (cortex_a9_neon_vld1_3_4_regs): Likewise.
694 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
695 (cortex_a9_neon_vld2_4_regs): Likewise.
696 (cortex_a9_neon_vld3_vld4): Likewise.
697 (cortex_a9_neon_vld1_vld2_lane): Likewise.
698 (cortex_a9_neon_vld3_vld4_lane): Likewise.
699 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
700 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
701 (cortex_a9_neon_vst1_3_4_regs): Likewise.
702 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
703 (cortex_a9_neon_vst1_vst2_lane): Likewise.
704 (cortex_a9_neon_vst3_vst4_lane): Likewise.
705 (cortex_a9_neon_mcr): Likewise.
706 (cortex_a9_neon_mcr_2_mcrr): Likewise.
707 (cortex_a9_neon_mrc): Likewise.
708 (cortex_a9_neon_mrrc): Likewise.
710 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
712 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
714 (cortex_a8_neon_vshl_ddd): Remove.
715 (cortex_a8_neon_vst3_vst4): Likewise.
716 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
718 (cortex_a8_neon_bit_ops_q): New.
720 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
721 (cortex_a8_neon_int_2): Likewise..
722 (cortex_a8_neon_int_3): Likewise.
723 (cortex_a8_neon_int_5): Likewise.
724 (cortex_a8_neon_vqneg_vqabs): Likewise.
725 (cortex_a8_neon_int_4): Likewise.
726 (cortex_a8_neon_vaba): Likewise.
727 (cortex_a8_neon_vaba_qqq): Likewise.
728 (cortex_a8_neon_shift_1): Likewise.
729 (cortex_a8_neon_shift_2): Likewise.
730 (cortex_a8_neon_shift_3): Likewise.
731 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
732 (cortex_a8_neon_vsra_vrsra): Likewise.
733 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
734 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
735 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
737 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
738 (cortex_a8_neon_mla_qqq_8_16): Likewise.
739 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
741 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
742 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
743 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
744 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
745 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
746 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
747 (cortex_a8_neon_fp_vsum): Likewise.
748 (cortex_a8_neon_fp_vmul_ddd): Likewise.
749 (cortex_a8_neon_fp_vmul_qqd): Likewise.
750 (cortex_a8_neon_fp_vmla_ddd): Likewise.
751 (cortex_a8_neon_fp_vmla_qqq): Likewise.
752 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
753 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
754 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
755 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
756 (cortex_a8_neon_bp_simple): Likewise.
757 (cortex_a8_neon_bp_2cycle): Likewise.
758 (cortex_a8_neon_bp_3cycle): Likewise.
759 (cortex_a8_neon_ldr): Likewise.
760 (cortex_a8_neon_str): Likewise.
761 (cortex_a8_neon_vld1_1_2_regs): Likewise.
762 (cortex_a8_neon_vld1_3_4_regs): Likewise.
763 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
764 (cortex_a8_neon_vld2_4_regs): Likewise.
765 (cortex_a8_neon_vld3_vld4): Likewise.
766 (cortex_a8_neon_vld1_vld2_lane): Likewise.
767 (cortex_a8_neon_vld3_vld4_lane): Likewise.
768 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
769 (cortex_a8_neon_vst1_3_4_regs): Likewise.
770 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
771 (cortex_a8_neon_vst1_vst2_lane): Likewise.
772 (cortex_a8_neon_vst3_vst4_lane): Likewise.
773 (cortex_a8_neon_mcr): Likewise.
774 (cortex_a8_neon_mcr_2_mcrr): Likewise.
775 (cortex_a8_neon_mrc): Likewise.
776 (cortex_a8_neon_mrrc): Likewise.
778 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
780 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
782 * config/aarch64/aarch64-simd.md (neon_type): Remove.
783 (aarch64_simd_dup<mode>): Add "type" attribute.
784 (aarch64_dup_lane<mode>): Likewise.
785 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
786 (*aarch64_simd_mov<mode>): Likewise.
787 (aarch64_simd_mov_from_<mode>low): Likewise.
788 (aarch64_simd_mov_from_<mode>high): Likewise.
789 (orn<mode>3): Likewise.
790 (bic<mode>3): Likewise.
791 (add<mode>3): Likewise.
792 (sub<mode>3): Likewise.
793 (mul<mode>3): Likewise.
794 (*aarch64_mul3_elt<mode>): Likewise.
795 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
796 (*aarch64_mul3_elt_to_128df): Likewise.
797 (*aarch64_mul3_elt_to_64v2df): Likewise.
798 (neg<mode>2): Likewise.
799 (abs<mode>2): Likewise.
800 (abd<mode>_3): Likewise.
801 (aba<mode>_3): Likewise.
802 (fabd<mode>_3): Likewise.
803 (*fabd_scalar<mode>3): Likewise.
804 (and<mode>3): Likewise.
805 (ior<mode>3): Likewise.
806 (xor<mode>3): Likewise.
807 (one_cmpl<mode>2): Likewise.
808 (aarch64_simd_vec_set<mode>): Likewise.
809 (aarch64_simd_lshr<mode>): Likewise.
810 (aarch64_simd_ashr<mode>): Likewise.
811 (aarch64_simd_imm_shl<mode>): Likewise.
812 (aarch64_simd_reg_sshl<mode): Likewise.
813 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
814 (aarch64_simd_reg_shl<mode>_signed): Likewise.
815 (aarch64_simd_vec_setv2di): Likewise.
816 (aarch64_simd_vec_set<mode>): Likewise.
817 (aarch64_mla<mode>): Likewise.
818 (*aarch64_mla_elt<mode>): Likewise.
819 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
820 (aarch64_mls<mode>): Likewise.
821 (*aarch64_mls_elt<mode>): Likewise.
822 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
823 (<su><maxmin><mode>3): Likewise.
824 (move_lo_quad_<mode>): Likewise.
825 (aarch64_simd_move_hi_quad_<mode>): Likewise.
826 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
827 (vec_pack_trunc_<mode>): Likewise.
828 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
829 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
830 (*aarch64_<su>mlal_lo<mode>): Likewise.
831 (*aarch64_<su>mlal_hi<mode>): Likewise.
832 (*aarch64_<su>mlsl_lo<mode>): Likewise.
833 (*aarch64_<su>mlsl_hi<mode>): Likewise.
834 (*aarch64_<su>mlal<mode>): Likewise.
835 (*aarch64_<su>mlsl<mode>): Likewise.
836 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
837 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
838 (add<mode>3): Likewise.
839 (sub<mode>3): Likewise.
840 (mul<mode>3): Likewise.
841 (div<mode>3): Likewise.
842 (neg<mode>2): Likewise.
843 (abs<mode>2): Likewise.
844 (fma<mode>4): Likewise.
845 (*aarch64_fma4_elt<mode>): Likewise.
846 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
847 (*aarch64_fma4_elt_to_128df): Likewise.
848 (*aarch64_fma4_elt_to_64v2df): Likewise.
849 (fnma<mode>4): Likewise.
850 (*aarch64_fnma4_elt<mode>): Likewise.
851 (*aarch64_fnma4_elt_<vswap_width_name><mode>
852 (*aarch64_fnma4_elt_to_128df): Likewise.
853 (*aarch64_fnma4_elt_to_64v2df): Likewise.
854 (<frint_pattern><mode>2): Likewise.
855 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
856 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
857 (vec_unpacks_lo_v4sf): Likewise.
858 (aarch64_float_extend_lo_v2df): Likewise.
859 (vec_unpacks_hi_v4sf): Likewise.
860 (aarch64_float_truncate_lo_v2sf): Likewise.
861 (aarch64_float_truncate_hi_v4sf): Likewise.
862 (aarch64_vmls<mode>): Likewise.
863 (<su><maxmin><mode>3): Likewise.
864 (<maxmin_uns><mode>3): Likewise.
865 (reduc_<sur>plus_<mode>): Likewise.
866 (reduc_<sur>plus_v2di): Likewise.
867 (reduc_<sur>plus_v2si): Likewise.
868 (reduc_<sur>plus_<mode>): Likewise.
869 (aarch64_addpv4sf): Likewise.
870 (clz<mode>2): Likewise.
871 (reduc_<maxmin_uns>_<mode>): Likewise.
872 (reduc_<maxmin_uns>_v2di): Likewise.
873 (reduc_<maxmin_uns>_v2si): Likewise.
874 (reduc_<maxmin_uns>_<mode>): Likewise.
875 (reduc_<maxmin_uns>_v4sf): Likewise.
876 (aarch64_simd_bsl<mode>_internal): Likewise.
877 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
878 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
879 (aarch64_get_lane<mode>): Likewise.
880 (*aarch64_combinez<mode>): Likewise.
881 (aarch64_combine<mode>): Likewise.
882 (aarch64_simd_combine<mode>): Likewise.
883 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
884 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
885 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
886 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
887 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
888 (aarch64_<sur>h<addsub><mode>): Likewise.
889 (aarch64_<sur><addsub>hn<mode>): Likewise.
890 (aarch64_<sur><addsub>hn2<mode>): Likewise.
891 (aarch64_pmul<mode>): Likewise.
892 (aarch64_<su_optab><optab><mode>): Likewise.
893 (aarch64_<sur>qadd<mode>): Likewise.
894 (aarch64_sqmovun<mode>): Likewise.
895 (aarch64_<sur>qmovn<mode>): Likewise.
896 (aarch64_s<optab><mode>): Likewise.
897 (aarch64_sq<r>dmulh<mode>): Likewise.
898 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
899 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
900 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
901 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
902 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
903 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
904 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
905 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
906 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
907 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
908 (aarch64_sqdmull<mode>): Likewise.
909 (aarch64_sqdmull_lane<mode>_internal): Likewise.
910 (aarch64_sqdmull_n<mode>): Likewise.
911 (aarch64_sqdmull2<mode>_internal): Likewise.
912 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
913 (aarch64_sqdmull2_n<mode>_internal): Likewise.
914 (aarch64_<sur>shl<mode>): Likewise.
915 (aarch64_<sur>q<r>shl<mode>
916 (aarch64_<sur>shll_n<mode>): Likewise.
917 (aarch64_<sur>shll2_n<mode>): Likewise.
918 (aarch64_<sur>shr_n<mode>): Likewise.
919 (aarch64_<sur>sra_n<mode>): Likewise.
920 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
921 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
922 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
923 (aarch64_cm<optab><mode>): Likewise.
924 (aarch64_cm<optab>di): Likewise.
925 (aarch64_cm<optab><mode>): Likewise.
926 (aarch64_cm<optab>di): Likewise.
927 (aarch64_cmtst<mode>): Likewise.
928 (aarch64_cmtstdi): Likewise.
929 (aarch64_cm<optab><mode>): Likewise.
930 (*aarch64_fac<optab><mode>): Likewise.
931 (aarch64_addp<mode>): Likewise.
932 (aarch64_addpdi): Likewise.
933 (sqrt<mode>2): Likewise.
934 (vec_load_lanesoi<mode>): Likewise.
935 (vec_store_lanesoi<mode>): Likewise.
936 (vec_load_lanesci<mode>): Likewise.
937 (vec_store_lanesci<mode>): Likewise.
938 (vec_load_lanesxi<mode>): Likewise.
939 (vec_store_lanesxi<mode>): Likewise.
940 (*aarch64_mov<mode>): Likewise.
941 (aarch64_ld2<mode>_dreg): Likewise.
942 (aarch64_ld2<mode>_dreg): Likewise.
943 (aarch64_ld3<mode>_dreg): Likewise.
944 (aarch64_ld3<mode>_dreg): Likewise.
945 (aarch64_ld4<mode>_dreg): Likewise.
946 (aarch64_ld4<mode>_dreg): Likewise.
947 (aarch64_tbl1<mode>): Likewise.
948 (aarch64_tbl2v16qi): Likewise.
949 (aarch64_combinev16qi): Likewise.
950 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
951 (aarch64_st2<mode>_dreg): Likewise.
952 (aarch64_st2<mode>_dreg): Likewise.
953 (aarch64_st3<mode>_dreg): Likewise.
954 (aarch64_st3<mode>_dreg): Likewise.
955 (aarch64_st4<mode>_dreg): Likewise.
956 (aarch64_st4<mode>_dreg): Likewise.
957 (*aarch64_simd_ld1r<mode>): Likewise.
958 (aarch64_frecpe<mode>): Likewise.
959 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
960 (aarch64_frecps<mode>): Likewise.
962 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
964 * config/arm/iterators.md (V_elem_ch): New.
966 (VQH_type): Likewise.
967 * config/arm/arm.md (is_neon_type): New.
968 (conds): Use is_neon_type.
969 (anddi3_insn): Update type attribute.
970 (xordi3_insn): Likewise.
971 (one_cmpldi2): Likewise.
972 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
973 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
974 (*movmisalign<mode>_neon_store): Likewise.
975 (*movmisalign<mode>_neon_load): Likewise.
976 (vec_set<mode>_internal): Likewise.
977 (vec_set<mode>_internal): Likewise.
978 (vec_setv2di_internal): Likewise.
979 (vec_extract<mode>): Likewise.
980 (vec_extract<mode>): Likewise.
981 (vec_extractv2di): Likewise.
982 (*add<mode>3_neon): Likewise.
983 (adddi3_neon): Likewise.
984 (*sub<mode>3_neon): Likewise.
985 (subdi3_neon): Likewise.
986 (fma<VCVTF:mode>4): Likewise.
987 (fma<VCVTF:mode>4_intrinsic): Likewise.
988 (*fmsub<VCVTF:mode>4): Likewise.
989 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
990 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
991 (ior<mode>3): Likewise.
992 (and<mode>3): Likewise.
993 (orn<mode>3_neon): Likewise.
994 (orndi3_neon): Likewise.
995 (bic<mode>3_neon): Likewise.
996 (bicdi3_neon): Likewise.
997 (xor<mode>3): Likewise.
998 (one_cmpl<mode>2): Likewise.
999 (abs<mode>2): Likewise.
1000 (neg<mode>2): Likewise.
1001 (negdi2_neon): Likewise.
1002 (*umin<mode>3_neon): Likewise.
1003 (*umax<mode>3_neon): Likewise.
1004 (*smin<mode>3_neon): Likewise.
1005 (*smax<mode>3_neon): Likewise.
1006 (vashl<mode>3): Likewise.
1007 (vashr<mode>3_imm): Likewise.
1008 (vlshr<mode>3_imm): Likewise.
1009 (ashl<mode>3_signed): Likewise.
1010 (ashl<mode>3_unsigned): Likewise.
1011 (neon_load_count): Likewise.
1012 (ashldi3_neon_noclobber): Likewise.
1013 (ashldi3_neon): Likewise.
1014 (signed_shift_di3_neon): Likewise.
1015 (unsigned_shift_di3_neon): Likewise.
1016 (ashrdi3_neon_imm_noclobber): Likewise.
1017 (lshrdi3_neon_imm_noclobber): Likewise.
1018 (<shift>di3_neon): Likewise.
1019 (widen_ssum<mode>3): Likewise.
1020 (widen_usum<mode>3): Likewise.
1021 (quad_halves_<code>v4si): Likewise.
1022 (quad_halves_<code>v4sf): Likewise.
1023 (quad_halves_<code>v8hi): Likewise.
1024 (quad_halves_<code>v16qi): Likewise.
1025 (reduc_splus_v2di): Likewise.
1026 (neon_vpadd_internal<mode>): Likewise.
1027 (neon_vpsmin<mode>): Likewise.
1028 (neon_vpsmax<mode>): Likewise.
1029 (neon_vpumin<mode>): Likewise.
1030 (neon_vpumax<mode>): Likewise.
1031 (*ss_add<mode>_neon): Likewise.
1032 (*us_add<mode>_neon): Likewise.
1033 (*ss_sub<mode>_neon): Likewise.
1034 (*us_sub<mode>_neon): Likewise.
1035 (neon_vadd<mode>_unspec): Likewise.
1036 (neon_vaddl<mode>): Likewise.
1037 (neon_vaddw<mode>): Likewise.
1038 (neon_vhadd<mode>): Likewise.
1039 (neon_vqadd<mode>): Likewise.
1040 (neon_vaddhn<mode>): Likewise.
1041 (neon_vmul<mode>): Likewise.
1042 (neon_vfms<VCVTF:mode>): Likewise.
1043 (neon_vmlal<mode>): Likewise.
1044 (neon_vmls<mode>): Likewise.
1045 (neon_vmlsl<mode>): Likewise.
1046 (neon_vqdmulh<mode>): Likewise.
1047 (neon_vqdmlal<mode>): Likewise.
1048 (neon_vqdmlsl<mode>): Likewise.
1049 (neon_vmull<mode>): Likewise.
1050 (neon_vqdmull<mode>): Likewise.
1051 (neon_vsub<mode>_unspec): Likewise.
1052 (neon_vsubl<mode>): Likewise.
1053 (neon_vsubw<mode>): Likewise.
1054 (neon_vqsub<mode>): Likewise.
1055 (neon_vhsub<mode>): Likewise.
1056 (neon_vsubhn<mode>): Likewise.
1057 (neon_vceq<mode>): Likewise.
1058 (neon_vcge<mode>): Likewise.
1059 (neon_vcgeu<mode>): Likewise.
1060 (neon_vcgt<mode>): Likewise.
1061 (neon_vcgtu<mode>): Likewise.
1062 (neon_vcle<mode>): Likewise.
1063 (neon_vclt<mode>): Likewise.
1064 (neon_vcage<mode>): Likewise.
1065 (neon_vcagt<mode>): Likewise.
1066 (neon_vtst<mode>): Likewise.
1067 (neon_vabd<mode>): Likewise.
1068 (neon_vabdl<mode>): Likewise.
1069 (neon_vaba<mode>): Likewise.
1070 (neon_vabal<mode>): Likewise.
1071 (neon_vmax<mode>): Likewise.
1072 (neon_vmin<mode>): Likewise.
1073 (neon_vpaddl<mode>): Likewise.
1074 (neon_vpadal<mode>): Likewise.
1075 (neon_vpmax<mode>): Likewise.
1076 (neon_vpmin<mode>): Likewise.
1077 (neon_vrecps<mode>): Likewise.
1078 (neon_vrsqrts<mode>): Likewise.
1079 (neon_vqabs<mode>): Likewise.
1080 (neon_vqneg<mode>): Likewise.
1081 (neon_vcls<mode>): Likewise.
1082 (clz<mode>2): Likewise.
1083 (popcount<mode>2): Likewise.
1084 (neon_vrecpe<mode>): Likewise.
1085 (neon_vrsqrte<mode>): Likewise.
1086 (neon_vget_lane<mode>_sext_internal): Likewise.
1087 (neon_vget_lane<mode>_zext_internal): Likewise.
1088 (neon_vdup_n<mode>): Likewise.
1089 (neon_vdup_n<mode>): Likewise.
1090 (neon_vdup_nv2di): Likewise.
1091 (neon_vdup_lane<mode>_interal): Likewise.
1092 (*neon_vswp<mode>): Likewise.
1093 (neon_vcombine<mode>): Likewise.
1094 (float<mode><V_cvtto>2): Likewise.
1095 (floatuns<mode><V_cvtto>2): Likewise.
1096 (fix_trunc<mode><V_cvtto>2): Likewise.
1097 (fixuns_trunc<mode><V_cvtto>2
1098 (neon_vcvt<mode>): Likewise.
1099 (neon_vcvt<mode>): Likewise.
1100 (neon_vcvtv4sfv4hf): Likewise.
1101 (neon_vcvtv4hfv4sf): Likewise.
1102 (neon_vcvt_n<mode>): Likewise.
1103 (neon_vcvt_n<mode>): Likewise.
1104 (neon_vmovn<mode>): Likewise.
1105 (neon_vqmovn<mode>): Likewise.
1106 (neon_vqmovun<mode>): Likewise.
1107 (neon_vmovl<mode>): Likewise.
1108 (neon_vmul_lane<mode>): Likewise.
1109 (neon_vmul_lane<mode>): Likewise.
1110 (neon_vmull_lane<mode>): Likewise.
1111 (neon_vqdmull_lane<mode>): Likewise.
1112 (neon_vqdmulh_lane<mode>): Likewise.
1113 (neon_vqdmulh_lane<mode>): Likewise.
1114 (neon_vmla_lane<mode>): Likewise.
1115 (neon_vmla_lane<mode>): Likewise.
1116 (neon_vmlal_lane<mode>): Likewise.
1117 (neon_vqdmlal_lane<mode>): Likewise.
1118 (neon_vmls_lane<mode>): Likewise.
1119 (neon_vmls_lane<mode>): Likewise.
1120 (neon_vmlsl_lane<mode>): Likewise.
1121 (neon_vqdmlsl_lane<mode>): Likewise.
1122 (neon_vext<mode>): Likewise.
1123 (neon_vrev64<mode>): Likewise.
1124 (neon_vrev32<mode>): Likewise.
1125 (neon_vrev16<mode>): Likewise.
1126 (neon_vbsl<mode>_internal): Likewise.
1127 (neon_vshl<mode>): Likewise.
1128 (neon_vqshl<mode>): Likewise.
1129 (neon_vshr_n<mode>): Likewise.
1130 (neon_vshrn_n<mode>): Likewise.
1131 (neon_vqshrn_n<mode>): Likewise.
1132 (neon_vqshrun_n<mode>): Likewise.
1133 (neon_vshl_n<mode>): Likewise.
1134 (neon_vqshl_n<mode>): Likewise.
1135 (neon_vqshlu_n<mode>): Likewise.
1136 (neon_vshll_n<mode>): Likewise.
1137 (neon_vsra_n<mode>): Likewise.
1138 (neon_vsri_n<mode>): Likewise.
1139 (neon_vsli_n<mode>): Likewise.
1140 (neon_vtbl1v8qi): Likewise.
1141 (neon_vtbl2v8qi): Likewise.
1142 (neon_vtbl3v8qi): Likewise.
1143 (neon_vtbl4v8qi): Likewise.
1144 (neon_vtbl1v16qi): Likewise.
1145 (neon_vtbl2v16qi): Likewise.
1146 (neon_vcombinev16qi): Likewise.
1147 (neon_vtbx1v8qi): Likewise.
1148 (neon_vtbx2v8qi): Likewise.
1149 (neon_vtbx3v8qi): Likewise.
1150 (neon_vtbx4v8qi): Likewise.
1151 (*neon_vtrn<mode>_insn): Likewise.
1152 (*neon_vzip<mode>_insn): Likewise.
1153 (*neon_vuzp<mode>_insn): Likewise.
1154 (neon_vld1<mode>): Likewise.
1155 (neon_vld1_lane<mode>): Likewise.
1156 (neon_vld1_lane<mode>): Likewise.
1157 (neon_vld1_dup<mode>): Likewise.
1158 (neon_vld1_dup<mode>): Likewise.
1159 (neon_vld1_dupv2di): Likewise.
1160 (neon_vst1<mode>): Likewise.
1161 (neon_vst1_lane<mode>): Likewise.
1162 (neon_vst1_lane<mode>): Likewise.
1163 (neon_vld2<mode>): Likewise.
1164 (neon_vld2<mode>): Likewise.
1165 (neon_vld2_lane<mode>): Likewise.
1166 (neon_vld2_lane<mode>): Likewise.
1167 (neon_vld2_dup<mode>): Likewise.
1168 (neon_vst2<mode>): Likewise.
1169 (neon_vst2<mode>): Likewise.
1170 (neon_vst2_lane<mode>): Likewise.
1171 (neon_vst2_lane<mode>): Likewise.
1172 (neon_vld3<mode>): Likewise.
1173 (neon_vld3qa<mode>): Likewise.
1174 (neon_vld3qb<mode>): Likewise.
1175 (neon_vld3_lane<mode>): Likewise.
1176 (neon_vld3_lane<mode>): Likewise.
1177 (neon_vld3_dup<mode>): Likewise.
1178 (neon_vst3<mode>): Likewise.
1179 (neon_vst3qa<mode>): Likewise.
1180 (neon_vst3qb<mode>): Likewise.
1181 (neon_vst3_lane<mode>): Likewise.
1182 (neon_vst3_lane<mode>): Likewise.
1183 (neon_vld4<mode>): Likewise.
1184 (neon_vld4qa<mode>): Likewise.
1185 (neon_vld4qb<mode>): Likewise.
1186 (neon_vld4_lane<mode>): Likewise.
1187 (neon_vld4_lane<mode>): Likewise.
1188 (neon_vld4_dup<mode>): Likewise.
1189 (neon_vst4<mode>): Likewise.
1190 (neon_vst4qa<mode>): Likewise.
1191 (neon_vst4qb<mode>): Likewise.
1192 (neon_vst4_lane<mode>): Likewise.
1193 (neon_vst4_lane<mode>): Likewise.
1194 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1195 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1196 (neon_vec_<US>mult_lo_<mode>): Likewise.
1197 (neon_vec_<US>mult_hi_<mode>): Likewise.
1198 (neon_vec_<US>shiftl_<mode>): Likewise.
1199 (neon_unpack<US>_<mode>): Likewise.
1200 (neon_vec_<US>mult_<mode>): Likewise.
1201 (vec_pack_trunc_<mode>): Likewise.
1202 (neon_vec_pack_trunc_<mode>): Likewise.
1203 (neon_vabd<mode>_2): Likewise.
1204 (neon_vabd<mode>_3): Likewise.
1206 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1208 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1209 (load_pair): Update type attribute.
1210 (store_pair): Update type attribute.
1211 * config/aarch64/iterators.md (q): New.
1213 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1215 * config/arm/types.md: Add new types for Neon insns.
1217 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1218 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1219 Sergey Lega <sergey.s.lega@intel.com>
1220 Anna Tikhonova <anna.tikhonova@intel.com>
1221 Ilya Tocar <ilya.tocar@intel.com>
1222 Andrey Turetskiy <andrey.turetskiy@intel.com>
1223 Ilya Verbin <ilya.verbin@intel.com>
1224 Kirill Yukhin <kirill.yukhin@intel.com>
1225 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1227 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1228 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1229 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1231 (srcp14<mode>): Ditto.
1232 (rsqrt14<mode>): Ditto.
1233 (rsqrt14<mode>): Ditto.
1234 (avx512f_vmscalef<mode>): Ditto.
1235 (avx512f_scalef<mode>): Ditto.
1236 (avx512f_getexp<mode>): Ditto.
1237 (avx512f_sgetexp<mode>): Ditto.
1238 (avx512f_fixupimm<mode>): Ditto.
1239 (avx512f_sfixupimm<mode>): Ditto.
1240 (avx512f_rndscale<mode>): Ditto.
1241 (*avx512er_exp2<mode>): Ditto.
1242 (*avx512er_rcp28<mode>): Ditto.
1243 (avx512er_rsqrt28<mode>): Ditto.
1244 (avx512f_getmant<mode>): Ditto.
1245 (avx512f_getmant<mode>): Ditto.
1246 (avx512f_rndscale<mode>): Fix formatting.
1248 2013-10-15 Martin Jambor <mjambor@suse.cz>
1250 * ipa-utils.h (ipa_edge_within_scc): Declare.
1251 * ipa-cp.c (edge_within_scc): Moved...
1252 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1254 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1255 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1256 Sergey Lega <sergey.s.lega@intel.com>
1257 Anna Tikhonova <anna.tikhonova@intel.com>
1258 Ilya Tocar <ilya.tocar@intel.com>
1259 Andrey Turetskiy <andrey.turetskiy@intel.com>
1260 Ilya Verbin <ilya.verbin@intel.com>
1261 Kirill Yukhin <kirill.yukhin@intel.com>
1262 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1264 * config/i386/predicates.md (const_8_to_15_operand): New.
1265 (const_16_to_31_operand): Ditto.
1266 * config/i386/sse.md (V8FI): New.
1268 (reduc_splus_v8df): Ditto.
1269 (reduc_splus_v16sf): Ditto.
1270 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1271 (vec_extract_hi_<mode>): Ditto.
1272 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1273 (vec_set_lo_<mode>): Ditto.
1274 (vec_set_hi_<mode>): Ditto.
1275 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1276 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1277 (avx512f_pshufd_1): Ditto.
1278 (avx512f_broadcast<mode>): Ditto.
1279 (avx512f_broadcast<mode>): Ditto.
1280 (define_split): Split vec_extract_lo into move.
1281 (ssequartermode): Ditto.
1282 (ssedoublemode): Extened with wider modes.
1283 (vec_extract_lo_<mode>): Ditto.
1285 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1286 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1287 Sergey Lega <sergey.s.lega@intel.com>
1288 Anna Tikhonova <anna.tikhonova@intel.com>
1289 Ilya Tocar <ilya.tocar@intel.com>
1290 Andrey Turetskiy <andrey.turetskiy@intel.com>
1291 Ilya Verbin <ilya.verbin@intel.com>
1292 Kirill Yukhin <kirill.yukhin@intel.com>
1293 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1295 * config/i386/predicates.md (register_or_constm1_operand): New.
1296 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1297 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1298 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1299 UNSPEC_SCATTER_PREFETCH
1301 (avx512f_ucmp<mode>3): Ditto.
1302 (avx512f_vternlog<mode>): Ditto.
1303 (avx512f_align<mode>): Ditto.
1304 (<shift_insn><mode>3): Ditto.
1305 (avx512f_<rotate>v<mode>): Ditto.
1306 (avx512f_<rotate><mode>): Ditto.
1307 (avx512f_eq<mode>3): Ditto.
1308 (avx512f_eq<mode>3_1): Ditto.
1309 (avx512f_gt<mode>3): Ditto.
1310 (avx512f_testm<mode>3): Ditto.
1311 (avx512f_testnm<mode>3): Ditto.
1312 (avx512pf_gatherpf<mode>): Ditto.
1313 (*avx512pf_gatherpf<mode>_mask): Ditto.
1314 (*avx512pf_gatherpf<mode>): Ditto.
1315 (avx512pf_scatterpf<mode>): Ditto.
1316 (*avx512pf_scatterpf<mode>_mask): Ditto.
1317 (*avx512pf_scatterpf<mode>): Ditto.
1318 (avx512f_vec_dup_gpr<mode>): Ditto.
1319 (clz<mode>2): Ditto.
1320 (conflict<mode>): Ditto.
1321 (REDUC_SMINMAX_MODE): Extened with wider modes.
1322 (reduc_<code>_<mode>): Ditto.
1323 (vlshr<mode>3): Ditto.
1324 (vashl<mode>3): Ditto.
1326 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1327 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1328 Sergey Lega <sergey.s.lega@intel.com>
1329 Anna Tikhonova <anna.tikhonova@intel.com>
1330 Ilya Tocar <ilya.tocar@intel.com>
1331 Andrey Turetskiy <andrey.turetskiy@intel.com>
1332 Ilya Verbin <ilya.verbin@intel.com>
1333 Kirill Yukhin <kirill.yukhin@intel.com>
1334 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1336 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1339 (avx512fmaskmode): Ditto.
1340 (bcstscalarsuff): Ditto.
1341 (avx512f_blendm<mode>): Ditto.
1342 (cmp_imm_predicate): Ditto.
1343 (avx512f_cmp<mode>3): Ditto.
1344 (avx512f_vec_dup<mode>): Ditto.
1345 (avx512f_vec_dup_mem<mode>): Ditto.
1346 (avx512f_vpermi2var<mode>3): Ditto.
1347 (avx512f_vpermt2var<mode>3): Ditto.
1348 (vec_init<mode>): Ditto.
1349 (avx512f_gathersi<mode>): Ditto.
1350 (*avx512f_gathersi<mode>): Ditto.
1351 (*avx512f_gathersi<mode>_2): Ditto.
1352 (avx512f_gatherdi<mode>): Ditto.
1353 (*avx512f_gatherdi<mode>): Ditto.
1354 (*avx512f_gatherdi<mode>_2): Ditto.
1355 (avx512f_scattersi<mode>): Ditto.
1356 (*avx512f_scattersi<mode>): Ditto.
1357 (avx512f_scatterdi<mode>): Ditto.
1358 (*avx512f_scatterdi<mode>): Ditto.
1359 (sseintprefix): Extened with wider modes.
1360 (VEC_GATHER_IDXSI): Ditto.
1361 (VEC_GATHER_IDXDI): Ditto.
1362 (VEC_GATHER_SRCDI): Ditto.
1364 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1365 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1367 * config/arm/t-aprofile: New file.
1368 * config.gcc: Handle --with-multilib-list option.
1370 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1372 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1375 2013-10-15 Richard Biener <rguenther@suse.de>
1377 * tree-loop-distribution.c (build_empty_rdg): Inline into
1379 (rdg_flag_vertex): Inline into single user.
1380 (rdg_flag_vertex_and_dependent): Likewise.
1381 (build_rdg_partition_for_vertex): Remove processed bitmap.
1382 (rdg_build_partitions): Simplify.
1384 2013-10-15 Richard Biener <rguenther@suse.de>
1386 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1387 Restructure forwarding through conversions and copies to
1388 avoid performing copy-propagation the wrong way. Adjust
1389 recursion invocations.
1390 (forward_propagate_addr_expr): Add argument stating if we
1391 are recursing from a single-use.
1392 (ssa_forward_propagate_and_combine): Adjust.
1394 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1396 * dumpfile.h (gcc::dump_manager): New class, to hold state
1397 relating to dumpfile management.
1398 (get_dump_file_name): Remove in favor of method of dump_manager.
1399 (dump_initialized_p): Likewise.
1400 (dump_start): Likewise.
1401 (dump_finish): Likewise.
1402 (dump_switch_p): Likewise.
1403 (dump_register): Likewise.
1404 (get_dump_file_info): Likewise.
1405 * context.c (gcc::context::context): Construct the dump_manager
1407 * context.h (gcc::context::get_dumps): New.
1408 (gcc::context::m_dumps): New.
1409 * coverage.c (coverage_init): Port to dump_manager API.
1410 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1411 (extra_dump_files_in_use): Likewise.
1412 (extra_dump_files_alloced): Likewise.
1413 (gcc::dump_manager::dump_manager): New.
1414 (dump_register): Convert to...
1415 (gcc::dump_manager::dump_register): ...method, replacing
1416 function-static next_dump with m_next_dump field.
1417 (get_dump_file_info): Convert to...
1418 (gcc::dump_manager::get_dump_file_info): ...method.
1419 (get_dump_file_name): Convert to...
1420 (gcc::dump_manager::get_dump_file_name): ...method.
1421 (dump_start): Convert to...
1422 (gcc::dump_manager::dump_start): ...method.
1423 (dump_finish): Convert to...
1424 (gcc::dump_manager::dump_finish): ...method.
1425 (dump_begin): Replace body with...
1426 (gcc::dump_manager::dump_begin): ...new method.
1427 (dump_phase_enabled_p): Convert to...
1428 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1429 (dump_phase_enabled_p): Convert to...
1430 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1431 (dump_initialized_p): Convert to...
1432 (gcc::dump_manager::dump_initialized_p): ...method.
1433 (dump_flag_name): Replace body with...
1434 (gcc::dump_manager::dump_flag_name): ...new method.
1435 (dump_enable_all): Convert to...
1436 (gcc::dump_manager::dump_enable_all): ...new method.
1437 (opt_info_enable_passes): Convert to...
1438 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1439 (dump_switch_p_1): Convert to...
1440 (gcc::dump_manager::dump_switch_p_1): ...new method.
1441 (dump_switch_p): Convert to...
1442 (gcc::dump_manager::dump_switch_p): ...new method.
1443 (opt_info_switch_p): Port to dump_manager API.
1444 (enable_rtl_dump_file): Likewise.
1445 * opts-global.c (handle_common_deferred_options): Port to new
1447 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1448 (pass_manager::register_one_dump_file): Likewise.
1449 (pass_manager::register_pass): Likewise.
1450 (pass_init_dump_file): Likewise.
1451 (pass_fini_dump_file): Likewise.
1452 * statistics.c (statistics_early_init): Likewise.
1454 2013-10-14 Richard Biener <rguenther@suse.de>
1456 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1457 iterative_hash_canonical_type, gimple_canonical_type_hash,
1458 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1459 gimple_register_canonical_type, print_gimple_types_stats,
1460 free_gimple_type_tables): Move to lto/lto.c
1461 (gt-gimple.h): Do not include.
1462 * gimple.h (gimple_register_canonical_type,
1463 print_gimple_types_stats, free_gimple_type_tables): Remove.
1464 * Makefile.in (GTFILES): Remove gimple.c.
1466 2013-10-14 Travis Snoozy <quandary@remstate.com>
1469 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1470 scanning for msp430x targets.
1472 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1475 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1476 (registers_ok_for_ldd_peep): Move around.
1477 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1478 * config/sparc/sparc.md (widening peepholes): Use it.
1480 2013-10-14 Richard Biener <rguenther@suse.de>
1484 * gimple.c (iterative_hash_canonical_type): Make sure to
1485 record the hash into the correct hashtable slot.
1487 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1489 PR rtl-optimization/58662
1490 * combine.c (try_combine): Take into account death nodes on I2 when
1491 splitting a PARALLEL of two independent SETs. Fix dump message.
1493 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1496 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1497 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1498 * config/sh/t-sh (sh_treg_combine.o): New entry.
1499 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1500 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1501 (register_sh_passes): New function. Register sh_treg_combine pass.
1502 (sh_option_override): Invoke it.
1503 (sh_canonicalize_comparison): Handle op0_preserve_value.
1504 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1505 opportunities. Canonicalize branch condition.
1506 (nott): Allow only if pseudos can be created for non-SH2A.
1508 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1511 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1512 (ix86_expand_movmem): Replace copy_addr_to_reg with
1513 ix86_copy_addr_to_reg.
1514 (ix86_expand_setmem): Likewise.
1516 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1518 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1519 provided by ix86_fp_compare_mode instead of CCFPUmode.
1521 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1523 * config/aarch64/arm_neon.h
1524 (vtbx<1,3>_<psu>8): Fix register constriants.
1526 2013-10-11 Jeff Law <law@redhat.com>
1528 PR tree-optimization/58640
1529 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1530 threading paths that cross over two loop entry points.
1532 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1534 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1535 handle vector float as well.
1536 (*vsx_le_perm_load_v4si): Likewise.
1537 (*vsx_le_perm_store_v2di): Likewise.
1538 (*vsx_le_perm_store_v4si): Likewise.
1540 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1542 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1543 directly to circumvent subtract from splat{31} workaround.
1544 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1546 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1547 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1548 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1549 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1550 separate big and little endian logic.
1551 (*altivec_vperm_<mode>_internal): New define_insn.
1552 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1553 separate big and little endian logic.
1554 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1555 (vec_permv16qi): Add little endian logic.
1557 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1559 * doc/extend.texi (returns_nonnull): Remove arguments.
1561 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1562 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1563 Sergey Lega <sergey.s.lega@intel.com>
1564 Anna Tikhonova <anna.tikhonova@intel.com>
1565 Ilya Tocar <ilya.tocar@intel.com>
1566 Andrey Turetskiy <andrey.turetskiy@intel.com>
1567 Ilya Verbin <ilya.verbin@intel.com>
1568 Kirill Yukhin <kirill.yukhin@intel.com>
1569 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1571 * config/i386/sse.md (VI48F_256_512): New.
1572 (avx2_permvar<mode>): Change to ...
1573 (<avx2_avx512f>_permvar<mode>): This.
1575 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1576 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1577 Sergey Lega <sergey.s.lega@intel.com>
1578 Anna Tikhonova <anna.tikhonova@intel.com>
1579 Ilya Tocar <ilya.tocar@intel.com>
1580 Andrey Turetskiy <andrey.turetskiy@intel.com>
1581 Ilya Verbin <ilya.verbin@intel.com>
1582 Kirill Yukhin <kirill.yukhin@intel.com>
1583 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1585 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1586 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1587 * config/i386/sse.md (VI4_AVX): New.
1588 (sf2simodelower): Ditto.
1589 (sse2_cvtps2dq): Change to ...
1590 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1592 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1593 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1594 Sergey Lega <sergey.s.lega@intel.com>
1595 Anna Tikhonova <anna.tikhonova@intel.com>
1596 Ilya Tocar <ilya.tocar@intel.com>
1597 Andrey Turetskiy <andrey.turetskiy@intel.com>
1598 Ilya Verbin <ilya.verbin@intel.com>
1599 Kirill Yukhin <kirill.yukhin@intel.com>
1600 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1602 * config/i386/sse.md (V_512): New.
1604 (vcond<V_512:mode><VF_512:mode>): Ditto.
1605 (vcond<V_512:mode><VI_512:mode>): Ditto.
1606 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1608 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1609 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1610 Sergey Lega <sergey.s.lega@intel.com>
1611 Anna Tikhonova <anna.tikhonova@intel.com>
1612 Ilya Tocar <ilya.tocar@intel.com>
1613 Andrey Turetskiy <andrey.turetskiy@intel.com>
1614 Ilya Verbin <ilya.verbin@intel.com>
1615 Kirill Yukhin <kirill.yukhin@intel.com>
1616 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1618 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1619 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1621 (fma<mode>4): Removed condition.
1622 (fms<mode>4): Ditto.
1623 (fnma<mode>4): Ditto.
1624 (fnms<mode>4): Ditto.
1625 (fma4i_fmadd_<mode>): Ditto.
1626 (*fma_fmadd_<mode>): Ditto.
1627 (*fma_fmsub_<mode>): Ditto.
1628 (*fma_fnmadd_<mode>): Ditto.
1629 (*fma_fnmsub_<mode>): Ditto.
1630 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1631 (*fma_fmaddsub_<mode>): Ditto.
1632 (*fma_fmsubadd_<mode>): Ditto.
1633 (*fmai_fmadd_<mode>): Ditto.
1634 (*fmai_fmsub_<mode>): Ditto.
1635 (*fmai_fnmadd_<mode>): Ditto.
1636 (*fmai_fnmsub_<mode>): Ditto.
1638 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1639 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1640 Sergey Lega <sergey.s.lega@intel.com>
1641 Anna Tikhonova <anna.tikhonova@intel.com>
1642 Ilya Tocar <ilya.tocar@intel.com>
1643 Andrey Turetskiy <andrey.turetskiy@intel.com>
1644 Ilya Verbin <ilya.verbin@intel.com>
1645 Kirill Yukhin <kirill.yukhin@intel.com>
1646 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1648 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1649 (VI124_256): Changed to ...
1650 (VI124_256_48_512): This.
1651 (ssepackmode): Extended with wider modes.
1652 (<code><mode>3): Changed iterator.
1653 (*avx2_<code><mode>3): Ditto.
1654 (vec_pack_trunc_<mode>): Ditto.
1656 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1657 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1658 Sergey Lega <sergey.s.lega@intel.com>
1659 Anna Tikhonova <anna.tikhonova@intel.com>
1660 Ilya Tocar <ilya.tocar@intel.com>
1661 Andrey Turetskiy <andrey.turetskiy@intel.com>
1662 Ilya Verbin <ilya.verbin@intel.com>
1663 Kirill Yukhin <kirill.yukhin@intel.com>
1664 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1666 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1667 (VI8F_256_512): Ditto.
1668 (abs<mode>2): Changed iterator.
1669 (avx2_perm<mode>): Changed to ...
1670 (<avx2_avx512f>_perm<mode>): This.
1671 (avx2_perm<mode>_1): Changed to ...
1672 (<avx2_avx512f>_perm<mode>_1): This.
1674 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1675 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1676 Sergey Lega <sergey.s.lega@intel.com>
1677 Anna Tikhonova <anna.tikhonova@intel.com>
1678 Ilya Tocar <ilya.tocar@intel.com>
1679 Andrey Turetskiy <andrey.turetskiy@intel.com>
1680 Ilya Verbin <ilya.verbin@intel.com>
1681 Kirill Yukhin <kirill.yukhin@intel.com>
1682 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1684 * config/i386/sse.md (VI48_AVX512F): New.
1685 (VI48_AVX2): Changed to ...
1686 (VI48_AVX2_48_AVX512F): This.
1687 (avx2_ashrv<mode>): Changed to ...
1688 (<avx2_avx512f>_ashrv<mode>): This.
1689 (avx2_<shift_insn>v<mode>): Changed to ...
1690 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1692 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1693 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1694 Sergey Lega <sergey.s.lega@intel.com>
1695 Anna Tikhonova <anna.tikhonova@intel.com>
1696 Ilya Tocar <ilya.tocar@intel.com>
1697 Andrey Turetskiy <andrey.turetskiy@intel.com>
1698 Ilya Verbin <ilya.verbin@intel.com>
1699 Kirill Yukhin <kirill.yukhin@intel.com>
1700 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1702 * config/i386/sse.md (VI4_AVX512F): New.
1703 (VI8_AVX2_AVX512F): Ditto.
1704 (mul<mode>3): Extended with wider modes.
1705 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1706 (mul<mode>3): Ditto.
1707 (vec_widen_<s>mult_odd_<mode>): Ditto.
1709 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1710 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1711 Sergey Lega <sergey.s.lega@intel.com>
1712 Anna Tikhonova <anna.tikhonova@intel.com>
1713 Ilya Tocar <ilya.tocar@intel.com>
1714 Andrey Turetskiy <andrey.turetskiy@intel.com>
1715 Ilya Verbin <ilya.verbin@intel.com>
1716 Kirill Yukhin <kirill.yukhin@intel.com>
1717 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1719 * config/i386/sse.md (VI2_AVX512F): New.
1720 (VI124_AVX512F): Ditto.
1721 (sseunpackmode): Extended with wider modes.
1722 (sseunpackfltmode): Ditto.
1723 (vec_unpacks_float_hi_<mode>): Ditto.
1724 (vec_unpacks_float_lo_<mode>): Ditto.
1725 (vec_unpacku_float_hi_<mode>): Ditto.
1726 (vec_unpacku_float_lo_<mode>): Ditto.
1727 (vec_unpacks_lo_<mode>): Ditto.
1728 (vec_unpacks_hi_<mode>): Ditto.
1729 (vec_unpacku_lo_<mode>): Ditto.
1730 (vec_unpacku_hi_<mode>): Ditto.
1732 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1733 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1734 Sergey Lega <sergey.s.lega@intel.com>
1735 Anna Tikhonova <anna.tikhonova@intel.com>
1736 Ilya Tocar <ilya.tocar@intel.com>
1737 Andrey Turetskiy <andrey.turetskiy@intel.com>
1738 Ilya Verbin <ilya.verbin@intel.com>
1739 Kirill Yukhin <kirill.yukhin@intel.com>
1740 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1742 * config/i386/i386.md (multdiv): New.
1743 (multdiv_mnemonic): Ditto.
1744 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1745 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1746 (<sse>_vmdiv<mode>3): Removed.
1748 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1749 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1750 Sergey Lega <sergey.s.lega@intel.com>
1751 Anna Tikhonova <anna.tikhonova@intel.com>
1752 Ilya Tocar <ilya.tocar@intel.com>
1753 Andrey Turetskiy <andrey.turetskiy@intel.com>
1754 Ilya Verbin <ilya.verbin@intel.com>
1755 Kirill Yukhin <kirill.yukhin@intel.com>
1756 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1758 * config/i386/sse.md (V): Extended with wider modes.
1760 (ssehalfvecmode): Ditto.
1762 (ssepackfltmode): Ditto.
1763 (avx_vec_concat<mode>): Ditto.
1764 (V_256_512): New iterator.
1765 (VF2_512_256): Ditto.
1766 (si2dfmode): New attribute.
1767 (si2dfmodelower): Ditto.
1769 (concat_tg_mode): Ditto.
1770 (floatv4siv4df2): Changed to ...
1771 (float<si2dfmodelower><mode>2): This.
1772 (avx_cvtps2pd256): Changed to ...
1773 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1774 (vec_pack_trunc_v4df): Changed to ...
1775 (vec_pack_trunc_<mode>): This.
1776 (avx_vpermil<mode>): Changed to ...
1777 (<sse2_avx_avx512f>_vpermil<mode>): This.
1778 (<fixsuffix>fix_truncv8dfv8si2): New.
1779 (vec_pack_sfix_trunc_v8df): Ditto.
1780 (avx512f_rndscale<mode>): Ditto.
1781 (avx512f_roundpd512): Ditto.
1782 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1784 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1785 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1786 Sergey Lega <sergey.s.lega@intel.com>
1787 Anna Tikhonova <anna.tikhonova@intel.com>
1788 Ilya Tocar <ilya.tocar@intel.com>
1789 Andrey Turetskiy <andrey.turetskiy@intel.com>
1790 Ilya Verbin <ilya.verbin@intel.com>
1791 Kirill Yukhin <kirill.yukhin@intel.com>
1792 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1794 * config/i386/i386.md (any_fix): New iterator.
1795 (fixsuffix): New attribute.
1796 * config/i386/sse.md (VF1): Extened with wider modes.
1800 (sseintvecmodelower): Ditto.
1801 (ssescalarmode): Ditto.
1802 (ssescalarnum): Ditto.
1804 (ssexmmmode): Ditto.
1805 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1806 (<sse>_rcp<mode>2): Change iterator.
1807 (rsqrt<mode>2): Ditto.
1808 (<sse>_rsqrt<mode>2): Ditto.
1809 (avx2_vec_dup<mode>): Ditto.
1810 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1811 (round<mode>2_sfix): Ditto.
1812 (avx2_pbroadcast<mode>): Ditto.
1813 (*andnot<mode>3): Handle XI mode.
1814 (*<code><mode>3): Ditto.
1815 (AVXTOSSEMODE): Removed.
1816 (avx_vpermil<mode>): Changed to ...
1817 (<sse2_avx_avx512f>_vpermil<mode>): This.
1819 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1820 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1821 Sergey Lega <sergey.s.lega@intel.com>
1822 Anna Tikhonova <anna.tikhonova@intel.com>
1823 Ilya Tocar <ilya.tocar@intel.com>
1824 Andrey Turetskiy <andrey.turetskiy@intel.com>
1825 Ilya Verbin <ilya.verbin@intel.com>
1826 Kirill Yukhin <kirill.yukhin@intel.com>
1827 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1829 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1830 (<sse>_comi): Ditto.
1831 (<sse>_ucomi): Ditto.
1832 (sse_cvtss2siq_2): Ditto.
1833 (sse2_cvtsd2si): Ditto.
1834 (sse2_cvtsd2siq): Ditto.
1835 (sse2_cvttsd2si): Ditto.
1836 (sse2_cvttsd2siq): Ditto.
1837 (<shift_insn><mode>3): Ditto.
1838 (sse2_cvtsi2sdq): Update constraint and prefix.
1839 (sse_cvtsi2ss): Update prefix.
1840 (sse_cvtsi2ssq): Ditto.
1842 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1844 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1845 ignore internal calls.
1847 2013-10-11 Richard Biener <rguenther@suse.de>
1849 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1850 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1851 TREE_OVERFLOW is set.
1853 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1855 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1857 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1858 * gimple.h: Likewise.
1860 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1862 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1865 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1867 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1868 * configure: Regenerate.
1870 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1872 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1873 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1874 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1875 similarly pp_right_paren.
1876 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1877 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1878 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1879 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1881 (omp_declare_simd_clauses_equal,
1882 omp_remove_redundant_declare_simd_attrs): New functions.
1883 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1884 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1885 * tree.h (OMP_LOOP_CHECK): Define.
1886 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1887 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1888 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1889 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1890 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1891 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1892 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1893 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1894 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1895 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1896 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1897 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1898 (omp_remove_redundant_declare_simd_attrs): New prototype.
1899 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1900 GIMPLE_OMP_TEAMS): New codes.
1901 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1902 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1904 (target_nesting_level): New variable.
1905 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1906 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1907 for collapse > 1 static schedule unless ordered.
1908 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1909 (determine_parallel_type): Adjust get_ws_args_for caller.
1910 (install_var_field): Handle mask & 4 for double indirection.
1911 (scan_sharing_clauses): Ignore shared clause on teams construct.
1912 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1913 (create_omp_child_function): If inside target or declare target
1914 constructs, set "omp declare target" attribute on the child function.
1915 (find_combined_for): New function.
1916 (scan_omp_parallel): Handle combined loops.
1917 (scan_omp_target, scan_omp_teams): New functions.
1918 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1919 restrictions and set ctx->cancellable for cancellable constructs.
1920 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1921 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1922 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1923 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1924 (omp_clause_aligned_alignment): New function.
1925 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
1926 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1927 on teams constructs. Handle user defined reductions and new
1929 (lower_reduction_clauses): Don't set placeholder to address of ref
1930 if it has already the right type.
1931 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1932 (expand_parallel_call): Use the new non-_start suffixed builtins,
1933 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1934 and GOMP_parallel_end after the call.
1935 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1936 (expand_omp_for_init_counts): Handle combined loops.
1937 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1939 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1940 end of cancellable loops.
1941 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1942 Likewise. Handle collapse > 1 loops.
1943 (expand_omp_simd): Handle combined loops.
1944 (expand_omp_for): Add inner_stmt argument, adjust callers of
1945 expand_omp_for* functions, use expand_omp_for_static*chunk even
1946 for collapse > 1 unless ordered.
1947 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1948 of cancellable sections.
1949 (expand_omp_single): Remove need_barrier variable, just rely on
1950 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1951 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1952 (expand_omp_atomic_load, expand_omp_atomic_store,
1953 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1954 (expand_omp_target): New function.
1955 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1956 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1957 (build_omp_regions_1): Immediately close region for
1958 GF_OMP_TARGET_KIND_UPDATE.
1959 (maybe_add_implicit_barrier_cancel): New function.
1960 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1962 (lower_omp_single): Likewise. Add clobber after the barrier.
1963 (lower_omp_taskgroup): New function.
1964 (lower_omp_for): Handle combined loops. Adjust
1965 lower_rec_input_clauses caller. Handle cancellation.
1966 (lower_depend_clauses): New function.
1967 (lower_omp_taskreg): Lower depend clauses. Adjust
1968 lower_rec_input_clauses caller. Add clobber after the call. Handle
1970 (lower_omp_target, lower_omp_teams): New functions.
1971 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1972 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1973 and GOMP_cancellation_point calls.
1974 (lower_omp): Fold stmts inside of target region.
1975 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1976 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1977 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1978 (BT_FN_VOID_OMPFN_PTR_UINT,
1979 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1980 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1981 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1982 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1983 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1984 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1985 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1986 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1987 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1988 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1989 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1990 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1991 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1992 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1993 BUILT_IN_GOMP_PARALLEL_END.
1994 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1995 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1996 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1997 GF_OMP_FOR_KIND_DISTRIBUTE.
1998 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1999 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2000 (dump_gimple_omp_return): Print lhs if it has any.
2001 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2002 gimple_omp_atomic_seq_cst_p.
2003 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2004 and GIMPLE_OMP_TEAMS.
2005 * langhooks.c (lhd_omp_mappable_type): New function.
2006 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2007 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2009 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2010 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2011 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2012 (struct gimplify_omp_ctx): Add combined_loop field.
2013 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2014 on stmts inside of target region.
2015 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2016 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2017 ORT_TARGET and ORT_TARGET_DATA.
2018 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2020 (omp_notice_threadprivate_variable): Complain about threadprivate
2021 variables in target region.
2022 (omp_notice_variable): Complain about vars with non-mappable type
2023 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2024 (omp_check_private): Ignore ORT_TARGET* regions.
2025 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2026 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2027 (find_combined_omp_for): New function.
2028 (gimplify_omp_for): Handle gimplification of combined loops.
2029 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2031 (gimplify_omp_target_update): New function.
2032 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2033 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2034 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2035 (gimplify_body): If fndecl has "omp declare target" attribute, add
2036 implicit ORT_TARGET context around it.
2037 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2038 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2039 * tree-nested.c (convert_nonlocal_reference_stmt,
2040 convert_local_reference_stmt, convert_gimple_call): Handle
2041 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2042 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2043 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2044 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2045 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2046 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2047 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2048 BUILT_IN_GOMP_LOOP_END_CANCEL,
2049 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2050 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2051 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2052 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2053 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2054 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2055 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2056 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2057 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2058 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2059 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2060 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2061 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2062 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2063 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2064 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2065 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2066 gimple_build_omp_teams): New functions.
2067 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2068 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2069 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2070 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2071 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2072 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2073 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2074 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2075 GF_OMP_ATOMIC_SEQ_CST): New.
2076 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2077 gimple_build_omp_teams): New prototypes.
2078 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2079 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2080 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2081 GIMPLE_OMP_SINGLE as end of range.
2082 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2083 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2084 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2085 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2086 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2087 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2088 gimple_omp_target_kind, gimple_omp_target_set_kind,
2089 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2090 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2091 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2092 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2093 gimple_omp_teams_set_clauses): New inlines.
2094 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2095 and GIMPLE_OMP_TASKGROUP.
2096 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2097 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2098 enum omp_clause_proc_bind_kind): New.
2099 (union omp_clause_subcode): Add depend_kind, map_kind and
2100 proc_bind_kind fields.
2101 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2102 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2103 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2104 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2105 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2107 2013-10-10 Teresa Johnson <tejohnson@google.com>
2109 * predict.c (tree_estimate_probability): Add new parameter
2110 for estimate_bb_frequencies.
2111 (estimate_bb_frequencies): Add new parameter to force estimation.
2112 (rebuild_frequencies): When max frequency in function is small,
2113 recompute counts from frequencies.
2114 * predict.h (estimate_bb_frequencies): New parameter.
2116 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2118 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2119 optimizations are disabled.
2121 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2123 * coverage.c (coverage_finish): Fix leak of da_file_name.
2125 2013-10-10 Jan Hubicka <jh@suse.cz>
2127 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2128 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2131 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2134 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2135 if any labels are in FALLTHRU_BB, use a special label emitted
2136 immediately after the asm goto insn rather than label_rtx
2138 (expand_asm_stmt): Adjust caller.
2139 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2140 edge if the last insn in predecessor is a jump with single successor,
2141 but it isn't simplejump_p.
2143 2013-10-10 Richard Biener <rguenther@suse.de>
2145 PR tree-optimization/58656
2146 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2148 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2150 * gimplify.c: Include expr.h and tm_p.h for targets with special
2151 va-arg padding requirements.
2153 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2155 * tree-flow.h: Move some prototypes to gimple.h.
2156 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2157 * gimple.h: Relocate some prototypes from tree-flow.h
2158 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2160 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2161 (build_va_arg_indirect_ref): Relocate and make static.
2162 (std_gimplify_va_arg_expr): Relocate here.
2163 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2164 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2166 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2168 * doc/md.texi: Document the mnemonic attribute.
2170 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2173 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2174 (set_attr_alternative x ...) when searching for user defined
2177 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2179 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2181 2013-10-09 Easwaran Raman <eraman@google.com>
2183 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2184 * cfgexpand.c (defer_stack_allocation): ...use here
2185 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2187 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2189 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2190 (operand_equal_for_value_replacement): New function, extracted from
2191 value_replacement and enhanced to catch more cases.
2192 (value_replacement): Use operand_equal_for_value_replacement.
2194 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2196 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2197 get_max_loop_iterations.
2199 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2201 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2202 Place comment above function.
2204 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2206 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2207 are not related to tree-cfg.c.
2208 * tree-ssa-address.h: New file. Relocate prototypes.
2209 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2210 (addr_for_mem_ref): New. Combine call to get_address_description and
2211 return addr_for_mem_ref.
2212 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2213 * tree-ssa-live.h: Adjust prototypes.
2214 * passes.c: Include tree-ssa-live.h.
2215 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2216 * graphite.c (graphite_transform_loops): Make static.
2217 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2218 make_pass_graphite, pass_data_graphite_transforms,
2219 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2220 * ipa-pure-const.c (warn_function_noreturn): Make static.
2221 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2222 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2223 Relocate from tree-cfg.c
2224 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2226 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2227 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2228 Move to ipa-pure-const.c.
2229 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2230 Relocate from tree-optimize.c.
2231 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2232 make_pass_fixup_cfg): Move to tree-cfg.c.
2233 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2234 Relocate some prototypes.
2235 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2236 * tree-dump.c (dump_function_to_file): Remove prototype.
2237 Add tree-flow.h to the include file.
2238 * tree-dump.h: Remove prototype.
2239 * tree-parloops.h: New File. Add prototypes.
2240 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2241 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2242 from tree-ssa-loop.c.
2243 * tree-predcom.c (run_tree_predictive_commoning,
2244 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2245 Relocate here from tree-ssa-loop.c.
2246 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2247 ssa_name_values.release ().
2248 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2249 (ssa_name_values): Relocate from tree-flow.h.
2250 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2251 * tree-ssa-loop.c (run_tree_predictive_commoning,
2252 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2253 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2254 make_pass_graphite, pass_data_graphite_transforms,
2255 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2256 tree_parallelize_loops, pass_data_parallelize_loops,
2257 make_pass_parallelize_loops): Move to other files.
2258 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2260 * tree.h: Remove prototypes from tree-address.c.
2262 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2264 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2265 (struct int_tree_map): Move to tree-hasher.h
2266 (SCALE, LABEL, PERCENT): Move to gimple.h
2267 * tree-flow-inline.h: Delete. Move functions to other files.
2268 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2269 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2270 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2271 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2272 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2273 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2274 * tree-hasher.h: Don't include tree-flow.h.
2275 (struct int_tree_map): Relocate from tree-flow.h.
2276 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2277 tree-flow-inline.h and make static.
2278 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2280 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2281 tree-flow-inline.h and make static.
2282 * tree.h (is_global_var, may_be_aliased): Relocate from
2284 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2285 * value-prof.c: No longer include tree-flow-inline.h.
2286 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2288 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2290 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2291 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2292 (enum move_pos): Move to tree-ssa-loop-im.h
2293 * cfgloop.h: Move some prototypes.
2294 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2295 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2296 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2297 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2298 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2299 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2300 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2301 (enum move_pos): Relocate here.
2302 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2304 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2305 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2306 make_pass_lim): Relocate here from tree-ssa-loop.c.
2307 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2309 (loop_edge_to_cancel, unloop_loops): Make static.
2310 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2311 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2312 (tree_complete_unroll, gate_tree_complete_unroll,
2313 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2314 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2315 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2316 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2317 (stmt_invariant_in_loop_p): Remove unused function.
2318 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2319 * tree-ssa-loop-manip.h: New file. Add prototypes.
2320 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2321 (gcov_type_to_double_int): Move to cfgloop.h.
2322 (double_int_cmp, bound_index,
2323 estimate_numbers_of_iterations_loop): Make static.
2324 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2325 (max_loop_iterations): Factor out get_max_loop_iterations.
2326 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2328 * tree-ssa-loop-niter.h: New file. Add prototypes.
2329 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2330 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2331 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2332 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2333 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2334 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2335 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2336 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2337 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2338 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2339 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2340 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2341 pass_data_complete_unroll, make_pass_complete_unroll,
2342 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2343 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2344 tree-ssa-loop-ivcanon.c.
2345 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2346 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2347 tree-ssa-loop-prefetch.c.
2348 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2350 (get_lsm_tmp_name): Relocate and add suffix parameter.
2351 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2352 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2353 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2354 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2355 (get_estimated_loop_iterations): Factor out accessor from
2356 estimated_loop_iterations in tree-ssa-loop-niter.c.
2357 (get_max_loop_iterations): Factor out accessor from
2358 _max_loop_iterations in tree-ssa-niter.c.
2359 * loop-unroll.c (decide_unroll_constant_iterations,
2360 decide_unroll_runtime_iterations, decide_peel_simple,
2361 decide_unroll_stupid): Use new get_* accessors.
2363 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2365 PR tree-optimization/20318
2366 * doc/extend.texi (returns_nonnull): New function attribute.
2367 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2369 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2370 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2372 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2375 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2376 false if both components are bitfields.
2378 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2380 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2381 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2382 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2383 * config/aarch64/aarch64.h
2384 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2385 * config/aarch64/aarch64-simd-builtins.def
2386 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2388 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2390 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2392 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2394 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2396 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2398 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2399 (vneg_f64): New intrinsic.
2400 (vneg_s8): Asm replaced with C.
2401 (vneg_s16): Likewise.
2402 (vneg_s32): Likewise.
2403 (vneg_s64): New intrinsic.
2404 (vnegq_f32): Asm replaced with C.
2405 (vnegq_f64): Likewise.
2406 (vnegq_s8): Likewise.
2407 (vnegq_s16): Likewise.
2408 (vnegq_s32): Likewise.
2409 (vnegq_s64): Likewise.
2411 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2413 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2415 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2417 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2418 packed stack special handling.
2419 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2420 back to fixed stack slots for FPRs saved due to stdarg.
2422 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2424 * config/s390/s390.c (s390_frame_info): Restructure function.
2426 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2428 * config/s390/s390.c (struct s390_frame_layout): New field
2430 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2431 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2432 regs_ever_clobbered to char*.
2433 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2434 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2435 (s390_register_info): Enable FPR save slots. Move/Copy some
2436 functionality into ...
2437 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2438 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2440 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2441 not imply a stack frame.
2442 (s390_init_frame_layout): Remove variable clobbered_regs.
2443 (s390_update_register_info): Remove function.
2444 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2445 cannot be used for register renaming.
2446 (s390_hard_regno_scratch_ok): New function.
2447 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2448 (s390_initial_elimination_offset): Change offset calculation of
2449 the return address pointer.
2450 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2451 (restore_gprs): Set frame related flag.
2452 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2453 (s390_emit_prologue): Call s390_register_info instead of
2454 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2455 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2456 (s390_optimize_prologue): Call s390_optimize_register_info.
2457 Try to remove also FPR slot save/restore INSNs. Remove frame
2458 related flags from restore INSNs.
2460 2013-10-08 DJ Delorie <dj@redhat.com>
2462 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2465 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2466 avoid conflict with the MI use of %c.
2467 * config/rl78/rl78-real.md: change %c to %C throughout.
2468 * config/rl78/rl78-virt.md: Likewise.
2470 2013-10-08 Jan Hubicka <jh@suse.cz>
2472 * config/i386/i386.c (ix86_option_override_internal): Switch
2473 to SSE math for -ffast-math when target ISA supports SSE2.
2475 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2477 * tree-flow.h: Remove some prototypes.
2478 * tree.h: Remove some protypes, add a couple.
2479 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2480 using_eh_for_cleanups_p): Add interface routines for front ends.
2481 * tree-eh.h: New file. Add protoptyes.
2482 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2483 (add_stmt_to_eh_lp_fn): Make static.
2484 (lower_try_finally): Use new using_eh_for_cleanups_p.
2485 * emit-rtl.c: Include tree-eh.h.
2486 * gimple.h: Include tree-eh.h.
2488 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2490 PR tree-optimization/58480
2491 * tree-vrp.c (infer_nonnull_range): New function.
2492 (infer_value_range): Call infer_nonnull_range.
2494 2013-10-08 Dehao Chen <dehao@google.com>
2496 PR tree-optimization/58619
2497 * tree-inline.c (copy_phis_for_bb): Combine location data
2500 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2503 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2504 RTX_FRAME_RELATED_P on INSN.
2506 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2508 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2509 (altivec_expand_vec_perm_const): Call it.
2511 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2513 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2514 sequences for LE VSX loads and stores at expand time.
2515 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2517 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2518 (rs6000_gen_le_vsx_permute): New.
2519 (rs6000_gen_le_vsx_load): New.
2520 (rs6000_gen_le_vsx_store): New.
2521 (rs6000_gen_le_vsx_move): New.
2522 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2523 (*vsx_le_perm_load_v4si): New.
2524 (*vsx_le_perm_load_v8hi): New.
2525 (*vsx_le_perm_load_v16qi): New.
2526 (*vsx_le_perm_store_v2di): New.
2527 (*vsx_le_perm_store_v4si): New.
2528 (*vsx_le_perm_store_v8hi): New.
2529 (*vsx_le_perm_store_v16qi): New.
2530 (*vsx_xxpermdi2_le_<mode>): New.
2531 (*vsx_xxpermdi4_le_<mode>): New.
2532 (*vsx_xxpermdi8_le_V8HI): New.
2533 (*vsx_xxpermdi16_le_V16QI): New.
2534 (*vsx_lxvd2x2_le_<mode>): New.
2535 (*vsx_lxvd2x4_le_<mode>): New.
2536 (*vsx_lxvd2x8_le_V8HI): New.
2537 (*vsx_lxvd2x16_le_V16QI): New.
2538 (*vsx_stxvd2x2_le_<mode>): New.
2539 (*vsx_stxvd2x4_le_<mode>): New.
2540 (*vsx_stxvd2x8_le_V8HI): New.
2541 (*vsx_stxvd2x16_le_V16QI): New.
2543 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2545 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2547 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2549 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2550 loop to work also for 31bit ABI.
2551 Save the stack pointer for frame_size > 0.
2553 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2555 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2556 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2557 constraint letters from expanders.
2558 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2559 retry count to general_operand.
2560 ("tabort"): Give operand 0 a mode.
2561 ("tabort_1"): Add mode and constraint letter for operand 0.
2562 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2564 2013-10-04 Jeff Law <law@redhat.com>
2566 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2568 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2570 (thread_across_edge): Here. Call it.
2572 2013-10-04 Cary Coutant <ccoutant@google.com>
2574 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2575 discarding a location list expression (or a piece of one).
2577 2013-10-03 Jan Hubicka <jh@suse.cz>
2579 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2580 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2581 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2583 2013-10-03 Jan Hubicka <jh@suse.cz>
2585 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2586 accumulate-outgoing-args when producing unwind info.
2588 2013-10-03 Wei Mi <wmi@google.com>
2590 * lra-constraints.c (insert_move_for_subreg): New function
2591 extracted from simplify_operand_subreg.
2592 (simplify_operand_subreg): Add reload for paradoxical subreg.
2594 2013-10-03 Rong Xu <xur@google.com>
2596 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2597 the candidate of builtin_expect such that we should fix the
2598 size/time estimation.
2599 (estimate_function_body_sizes): Do the acutally size/time fix-up
2602 2013-10-03 Rong Xu <xur@google.com>
2604 * predict.c (tree_predict_by_opcode): Get the probability
2605 for builtin_expect from param builtin_expect_probability.
2606 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2607 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2608 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2610 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2613 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2614 * calls.c (alloca_call_p): Use get_callee_fndecl.
2615 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2616 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2618 (vrp_visit_stmt): Remove duplicated code.
2620 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2622 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2623 ceildf2, btruncdf2, instead of vsx_* name.
2625 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2626 iterators to only do V2DF and V4SF here. Move the DF code to
2627 rs6000.md where it is combined with SF mode. Replace <VSv> with
2628 just 'v' since only vector operations are handled with these insns
2629 after moving the DF support to rs6000.md.
2630 (vsx_sub<mode>3): Likewise.
2631 (vsx_mul<mode>3): Likewise.
2632 (vsx_div<mode>3): Likewise.
2633 (vsx_fre<mode>2): Likewise.
2634 (vsx_neg<mode>2): Likewise.
2635 (vsx_abs<mode>2): Likewise.
2636 (vsx_nabs<mode>2): Likewise.
2637 (vsx_smax<mode>3): Likewise.
2638 (vsx_smin<mode>3): Likewise.
2639 (vsx_sqrt<mode>2): Likewise.
2640 (vsx_rsqrte<mode>2): Likewise.
2641 (vsx_fms<mode>4): Likewise.
2642 (vsx_nfma<mode>4): Likewise.
2643 (vsx_copysign<mode>3): Likewise.
2644 (vsx_btrunc<mode>2): Likewise.
2645 (vsx_floor<mode>2): Likewise.
2646 (vsx_ceil<mode>2): Likewise.
2647 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2648 (vsx_sminsf3): Likewise.
2649 (vsx_fmadf4): Likewise.
2650 (vsx_fmsdf4): Likewise.
2651 (vsx_nfmadf4): Likewise.
2652 (vsx_nfmsdf4): Likewise.
2653 (vsx_cmpdf_internal1): Likewise.
2655 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2656 simpler to select whether a target has SPE or traditional floating
2657 point support in iterators.
2658 (TARGET_DF_SPE): Likewise.
2659 (TARGET_SF_FPR): Likewise.
2660 (TARGET_DF_FPR): Likewise.
2661 (TARGET_SF_INSN): Macros to say whether floating point support
2662 exists for a given operation for expanders.
2663 (TARGET_DF_INSN): Likewise.
2665 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2666 combining of SF/DF mode operations, using both traditional and VSX
2674 (abs<mode>2): Combine SF/DF modes using traditional floating point
2675 instructions. Add support for using the upper DF registers with
2676 VSX support, and SF registers with power8-vector support. Update
2677 expanders for operations supported by both the SPE and traditional
2678 floating point units.
2679 (abs<mode>2_fpr): Likewise.
2680 (nabs<mode>2): Likewise.
2681 (nabs<mode>2_fpr): Likewise.
2682 (neg<mode>2): Likewise.
2683 (neg<mode>2_fpr): Likewise.
2684 (add<mode>3): Likewise.
2685 (add<mode>3_fpr): Likewise.
2686 (sub<mode>3): Likewise.
2687 (sub<mode>3_fpr): Likewise.
2688 (mul<mode>3): Likewise.
2689 (mul<mode>3_fpr): Likewise.
2690 (div<mode>3): Likewise.
2691 (div<mode>3_fpr): Likewise.
2692 (sqrt<mode>3): Likewise.
2693 (sqrt<mode>3_fpr): Likewise.
2694 (fre<Fs>): Likewise.
2695 (rsqrt<mode>2): Likewise.
2696 (cmp<mode>_fpr): Likewise.
2697 (smax<mode>3): Likewise.
2698 (smin<mode>3): Likewise.
2699 (smax<mode>3_vsx): Likewise.
2700 (smin<mode>3_vsx): Likewise.
2701 (negsf2): Delete SF operations that are merged with DF.
2708 (fmasf4_fpr): Likewise.
2709 (fmssf4_fpr): Likewise.
2710 (nfmasf4_fpr): Likewise.
2711 (nfmssf4_fpr): Likewise.
2712 (sqrtsf2): Likewise.
2713 (rsqrtsf_internal1): Likewise.
2714 (smaxsf3): Likewise.
2715 (sminsf3): Likewise.
2716 (cmpsf_internal1): Likewise.
2717 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2718 (negdf2): Delete DF operations that are merged with SF.
2720 (nabsdf2): Likewise.
2726 (rsqrtdf_internal1): Likewise.
2727 (fmadf4_fpr): Likewise.
2728 (fmsdf4_fpr): Likewise.
2729 (nfmadf4_fpr): Likewise.
2730 (nfmsdf4_fpr): Likewise.
2731 (sqrtdf2): Likewise.
2732 (smaxdf3): Likewise.
2733 (smindf3): Likewise.
2734 (cmpdf_internal1): Likewise.
2735 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2736 (btrunc<mode>2): Delete separate expander, and combine with the
2737 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2738 (btrunc<mode>2_fpr): Likewise.
2739 (ceil<mode>2): Likewise.
2740 (ceil<mode>2_fpr): Likewise.
2741 (floor<mode>2): Likewise.
2742 (floor<mode>2_fpr): Likewise.
2743 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2744 Add support for using the upper registers with VSX and
2745 power8-vector. Move insns to be closer to the define_expands. On
2746 VSX systems, prefer the traditional form of FMA over the VSX
2747 version, since the traditional form allows the target not to
2748 overlap with the inputs.
2749 (fms<mode>4_fpr): Likewise.
2750 (nfma<mode>4_fpr): Likewise.
2751 (nfms<mode>4_fpr): Likewise.
2753 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2754 Richard Earnshaw <richard.earnshaw@arm.com>
2756 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2757 (struct mult_cost_table): Likewise.
2758 (struct mem_cost_table): Likewise.
2759 (struct fp_cost_table): Likewise.
2760 (struct vector_cost_table): Likewise.
2761 (cpu_cost_table): Likewise.
2762 * config/arm/arm.opt (mold-rts-costs): New option.
2763 (mnew-generic-costs): Likewise.
2764 * config/arm/arm.c (generic_extra_costs): New table.
2765 (cortexa15_extra_costs): Likewise.
2766 (arm_slowmul_tune): Use NULL as new costs.
2767 (arm_fastmul_tune): Likewise.
2768 (arm_strongarm_tune): Likewise.
2769 (arm_xscale_tune): Likewise.
2770 (arm_9e_tune): Likewise.
2771 (arm_v6t2_tune): Likewise.
2772 (arm_cortex_a5_tune): Likewise.
2773 (arm_cortex_a9_tune): Likewise.
2774 (arm_v6m_tune): Likewise.
2775 (arm_fa726te_tune): Likewise.
2776 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2777 (arm_cortex_tune): Use generict_extra_costs.
2778 (shifter_op_p): New function.
2779 (arm_unspec_cost): Likewise.
2780 (LIBCALL_COST): Define.
2781 (arm_new_rtx_costs): New function.
2782 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2783 table is available. Use old costs otherwise unless mnew-generic-costs
2785 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2786 (cpu_cost_table): Declare.
2788 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2791 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2792 (*subs_mul_imm_<mode>)
2793 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2794 (*sub_<shift>_<mode>)
2795 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2796 Remove k constraint.
2798 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2800 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2802 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2804 2013-10-02 Teresa Johnson <tejohnson@google.com>
2806 * predict.c (probably_never_executed): New function.
2807 (probably_never_executed_bb_p): Invoke probably_never_executed.
2808 (probably_never_executed_edge_p): Ditto.
2809 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2810 Treat profile insanities conservatively.
2812 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2814 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2816 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2818 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2819 use it. Use smaller increase for scratch. Don't increase reject
2820 for early clobber scratch.
2821 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2822 setting eliminated regs except setting fp from hfp.
2823 (lra_eliminate): Check lra_insn_recog_data on NULL.
2825 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2828 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2829 setting -mvsx-timode by default until the underlying problem is fixed.
2830 (RS6000_CPU, power7 defaults): Likewise.
2832 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2834 * config/x-linux (host-linux.o): Remove header dependencies.
2835 Use $(COMPILE) and $(POSTCOMPILE).
2836 * config/t-linux-android (linux-android.o): Ditto.
2838 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2840 * Makefile.in (expmed.o-warn): Remove.
2842 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2844 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2845 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2847 2013-10-02 Teresa Johnson <tejohnson@google.com>
2849 * dojump.c (do_jump_1): Divide probability between
2850 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2852 2013-10-02 Tom Tromey <tromey@redhat.com>
2854 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2856 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2858 * tree-flow.h: Remove some prototypes.
2859 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2860 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2861 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2862 mark_virtual_phi_result_for_renaming): Relocate here.
2863 * tree-into-ssa.h: Add prototypes.
2864 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2865 single_pred_before_succ_order.
2866 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2867 (nonfreeing_call_p) Move to gimple.c.
2868 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2870 * basic-block.h (single_pred_before_succ_order): Add prototype.
2871 * gimple.c (nonfreeing_call_p): Relocate here.
2872 * gimple.h: Add prototype.
2873 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2874 * tree-ssa-dom.h: New file. Relocate prototypes here.
2875 * tree-ssa.h: Include tree-ssa-dom.h.
2877 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2879 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2880 Use $(COMPILE) and $(POSTCOMPILE).
2882 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2884 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2886 * tree-flow.h: Remove some prototypes.
2887 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2888 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2889 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2890 propagate_tree_value*): Move from here to...
2891 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2892 propagate_tree_value*): Relocate here.
2893 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2894 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2895 * gimple-fold.c: Remove gimple-fold.h from include list.
2896 * tree-vrp.c: Remove gimple-fold.h from include list.
2897 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2898 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2899 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2900 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2901 * sese.c: Add tree-ssa-propagate.h to include list.
2903 2013-10-02 Richard Biener <rguenther@suse.de>
2905 * tree-loop-distribution.c: Include tree-vectorizer.h for
2907 (enum partition_kind): Remove PKIND_REDUCTION.
2908 (struct partition_s): Remove has_writes member, add reduction_p member.
2909 (partition_alloc): Adjust.
2910 (partition_builtin_p): Likewise.
2911 (partition_has_writes): Remove.
2912 (partition_reduction_p): New function.
2913 (partition_merge_into): Likewise.
2914 (generate_code_for_partition): Commonize builtin partition
2916 (rdg_cannot_recompute_vertex_p): Remove.
2917 (already_processed_vertex_p): Likewise.
2918 (rdg_flag_vertex): Do not set has_writes.
2919 (classify_partition): Adjust.
2920 (rdg_build_partitions): Do not set has_writes, treat all
2921 partitions as useful.
2922 (distribute_loop): Record number of library calls generated. Adjust.
2923 (tree_loop_distribution): Report number of loops and library
2924 calls generated as opt-info.
2926 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2928 * tree-flow.h: Include new .h files. Move prototypes.
2929 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2930 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2931 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2932 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2933 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2934 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2935 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2936 * tree.h (get_ref_base_and_extent): Move prototype out.
2937 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2939 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2940 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2941 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2942 * tree-scalar-evolution.c: Include tree.h.
2943 * sese.c: Include tree.h.
2944 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2945 * dwarf2out.c: Include tree-dfa.h.
2946 * tree-chrec.c: Include tree.h.
2947 * tree-data-ref.c: Include tree.h.
2949 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2951 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2954 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2956 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2957 Use $(COMPILE) and $(POSTCOMPILE).
2958 (sol2-cxx.o): Likewise.
2959 (sol2-stubs.o): Likewise.
2961 * config/x-solaris (host-solaris.o): Likewise.
2963 * config/sparc/t-sparc (sparc.o): Remove.
2964 (sparc-c.o): Remove header dependencies.
2965 Use $(COMPILE) and $(POSTCOMPILE).
2966 * config/sparc/x-sparc: Likewise.
2968 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2970 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2971 * config/arc/arc700.md: Likewise.
2972 * config/arc/arc-modes.def: Likewise.
2973 * config/arc/arc-simd.h: Likewise.
2974 * config/arc/t-arc-uClibc: Likewise.
2975 * config/arc/t-arc-newlib: Likewise.
2977 2013-10-02 Renlin Li <renlin.li@arm.com>
2979 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2981 (aarch64_expand_epilogue): Likewise.
2983 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2984 Yufeng Zhang <yufeng.zhang@arm.com>
2986 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2988 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2989 'base_in' represent a conversion and legal_cast_p_1 holds; set
2990 'base_in' with the returned value from get_unwidened.
2992 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2994 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2995 plus_constant is not used.
2997 2013-10-01 Wei Mi <wmi@google.com>
2999 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3000 * config/i386/i386.md: Add define_peephole2 to
3001 break partial reg stall for cvtss2sd/cvtsd2ss.
3003 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3005 * config/arc/arc.c (pass_arc_ifcvt::clone):
3006 Update for ctxt_ -> m_ctxt change.
3008 2013-10-01 Jeff Law <law@redhat.com>
3010 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3011 outgoing_edge and intermediate_edge fields. Instead store the path.
3012 (redirection_data::hash): Hash on the last edge's destination index.
3013 (redirection_data::equal): Check the entire thread path.
3014 (lookup_redirectio_data): Corresponding changes.
3015 (create_edge_and_update_destination_phis): Likewise.
3016 (thread_single_edge): Likewise.
3018 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3019 Diego Novillo <dnovillo@google.com>
3021 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3022 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3023 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3024 versions of these patterns.
3026 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3027 (__builtin_arc_aligned): Likewise.
3029 * config/arc/arc.md: Expand adc_0 comment stating the intended
3030 purpose and why it isn't ready.
3031 Replace commented out call_value_via_label_mixed with a
3032 plain comment about bl_s.
3034 * config/arc/arc.c (stdio.h): Don't include directly.
3035 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3036 Remove if (1) condition.
3037 (arc_encode_section_info): Fix comment.
3039 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3041 * config/arc/arc.c (arc_conditional_register_usage):
3042 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3043 Also set reg_alloc_order for DMA config regs.
3045 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3046 Jeremy Bennett <jeremy.bennett@embecosm.com>
3048 * doc/install.texi (--with-cpu): Mention ARC.
3049 (arc-*-elf32): New paragraph.
3050 (arc-linux-uclibc): Likewise.
3051 * doc/md.texi (Machine Constraints): Add ARC part.
3052 * doc/invoke.texi: (menu): Add ARC Options.
3053 (Machine Dependent Options) <ARC Options>: Add synopsis.
3054 (node ARC Options): Add.
3055 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3056 (ARC Built-in Functions): New section defining
3057 generic ARC built-in functions.
3058 (ARC SIMD Built-in Functions): New section defining SIMD specific
3060 (Declaring Attributes of Functions): Extended
3061 description of short_call and long_call attributes for ARC and
3062 added index entries.
3064 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3065 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3066 Joern Rennecke <joern.rennecke@embecosm.com>
3067 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3068 Brendan Kehoe <brendan@zen.org>
3069 Michael Eager <eager@eagercon.com>
3070 Simon Cook <simon.cook@embecosm.com>
3071 Jeremy Bennett <jeremy.bennett@embecosm.com>
3073 * config/arc, common/config/arc: New directories.
3075 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3076 Brendan Kehoe <brendan@zen.org>
3077 Simon Cook <simon.cook@embecosm.com>
3079 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3081 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3083 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3084 * tree-ssa-coalesce.h: New. Move prototype to here.
3085 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3086 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3087 (gimple_can_coalesce_p): Move to...
3088 * gimple.c (gimple_can_coalesce_p): Here.
3090 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3092 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3093 (dump_decl_set): Move to gimple.c.
3094 * gimple.h: Don't include tree-ssa-operands.h.
3095 (dump_decl_set): Add prototype.
3096 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3097 Move to gimple-ssa.h.
3098 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3099 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3100 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3101 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3102 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3104 (dump_decl_set): Relocate here.
3105 * gimple-ssa.h: New file.
3106 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3107 Relocate from gimple.h.
3108 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3109 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3110 swap_tree_operands and remove non-ssa path.
3111 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3112 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3114 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3115 vect_is_simple_reduction_1): Use swap_ssa_operands.
3116 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3117 (enum need_phi_state): Move to tree-into-ssa.c.
3118 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3119 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3120 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3121 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3122 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3123 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3124 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3125 num_imm_uses): Move to ssa-iterators.h.
3126 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3127 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3129 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3130 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3131 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3132 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3133 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3134 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3135 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3136 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3137 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3138 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3139 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3140 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3141 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3142 (set_phi_nodes): Move to tree-phinodes.h.
3143 * tree-ssa-operands.h (enum ssa_op_iter_type,
3144 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3145 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3146 (dump_decl_set): Remove prototype.
3147 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3148 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3149 (set_phi_nodes): Relocate from tree-flow-inline.h.
3150 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3152 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3153 include list. Temporarily add gimple.h to include list.
3154 * ssa-iterators.h: New file.
3155 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3156 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3157 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3158 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3159 Relocate from tree-ssa-operands.h.
3160 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3161 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3162 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3163 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3164 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3165 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3166 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3167 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3168 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3169 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3170 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3171 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3172 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3173 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3174 Relocate from tree-flow-inline.h.
3175 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3177 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3180 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3181 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3182 Insert '\t' to output template.
3183 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3184 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3185 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3186 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3188 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3190 * doc/install.texi (Host/target specific installation notes for GCC):
3191 Put @anchor before @heading.
3192 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3193 Use @email for email addresses.
3195 2013-10-01 Jeff Law <law@redhat.com>
3197 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3198 a vec. Only delete the path if we create one without successfully
3199 registering a jump thread.
3200 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3202 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3203 (paths): New vector of jump threading paths.
3204 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3205 (THREAD_PATH): New accessor macro for the entire thread path.
3206 (lookup_redirection_data): Get intermediate and final outgoing edge
3207 from the thread path.
3208 (create_edge_and_update_destination_phis): Copy the threading path.
3209 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3210 jump threading path.
3211 (ssa_redirect_edges): Get edges and block types from the jump threading
3212 path. Free the path vector.
3213 (thread_block): Get edges from the jump threading path. Look at the
3214 entire path to see if we thread to a loop exit. If we cancel a jump
3215 thread request, then free the path vector.
3216 (thread_single_edge): Get edges and block types from the jump threading
3217 path. Free the path vector.
3218 (thread_through_loop_header): Get edges and block types from the jump
3219 threading path. Free the path vector.
3220 (mark_threaded_blocks): Iterate over the vector of paths and store
3221 the path on the appropriate edge. Get edges and block types from the
3222 jump threading path.
3223 (mark_threaded_blocks): Get edges and block types from the jump
3224 threading path. Free the path vector.
3225 (thread_through_all_blocks): Use the vector of paths rather than
3226 a vector of 3-edge sets.
3227 (register_jump_thread): Accept pointer to a path vector rather
3228 than the path vector itself. Store the path vector for later use.
3231 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3232 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3235 * config/s390/s390.c (s390_split_branches): Modify check for table
3237 (s390_chunkify_start): Rearrange table jump insn check in order to
3238 deal with compare and branch insns correctly.
3240 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3244 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3245 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3246 define_insn_and_split.
3247 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3248 (shiftsi3_compare): New pattern.
3250 * config/arm/unspecs.md (UNSPEC_RRX): New.
3252 2013-10-01 Alan Modra <amodra@gmail.com>
3254 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3255 casing inout operands.
3257 2013-10-01 Richard Biener <rguenther@suse.de>
3259 PR tree-optimization/58553
3260 * tree-loop-distribution.c (struct partition_s): Add niter member.
3261 (classify_partition): Populate niter member for the partition
3262 and properly identify whether the relevant store happens before
3263 or after the loop exit.
3264 (generate_memset_builtin): Use niter member from the partition.
3265 (generate_memcpy_builtin): Likewise.
3267 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3269 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3270 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3272 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3274 * basic-block.h (edge_list): Prefix member names with "m_".
3275 * context.h (context): Likewise.
3276 * domwalk.h (dom_walker): Likewise.
3277 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3278 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3279 * hash-table.h (hash_table): Likewise.
3280 * machmode.h (bit_field_mode_iterator): Likewise.
3281 * pass_manager.h (pass_list): Likewise.
3282 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3283 * tree-pass.h (pass_data): Likewise.
3284 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3285 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3286 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3287 * asan.c (pass_data_asan): Update accordingly.
3288 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3289 (control_dependences::control_dependences): Likewise.
3290 (control_dependences::~control_dependences): Likewise.
3291 (control_dependences::~control_dependences): Likewise.
3292 (control_dependences::get_edges_dependent_on): Likewise.
3293 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3294 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3295 * context.c (gcc::context::context): Likewise.
3296 * cprop.c (pass_rtl_cprop::clone): Likewise.
3297 * domwalk.c (dom_walker::walk): Likewise.
3298 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3299 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3300 * mode-switching.c (pass_mode_switching::clone): Likewise.
3301 * passes.c (opt_pass::opt_pass): Likewise.
3302 (pass_manager::pass_manager): Likewise.
3303 * predict.c (pass_strip_predict_hints::clone): Likewise.
3304 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3305 (pass_split_all_insns::clone): Likewise.
3306 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3308 (bit_field_mode_iterator::next_mode): Likewise.
3309 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3310 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3311 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3312 * tree-complex.c (pass_lower_complex::clone): Likewise.
3313 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3314 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3315 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3316 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3317 (pass_fold_builtins::clone): Likewise.
3318 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3319 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3320 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3321 (pass_cd_dce::clone): Likewise.
3322 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3323 (pass_phi_only_cprop::clone): Likewise.
3324 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3325 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3326 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3327 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3328 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3329 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3330 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3331 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3332 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3333 * tree-vrp.c (pass_vrp::clone): Likewise.
3334 * tsan.c (pass_tsan::clone): Likewise.
3336 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3339 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3340 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3343 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3344 optimization, punt if sign_bit_p looked through any zero extension.
3346 2013-09-30 Teresa Johnson <tejohnson@google.com>
3348 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3349 Update redirected out edge count in joiner case.
3350 (ssa_redirect_edges): Common the joiner and non-joiner cases
3351 so that joiner case gets profile updates.
3353 2013-09-30 Richard Biener <rguenther@suse.de>
3355 PR tree-optimization/58554
3356 * tree-loop-distribution.c (classify_partition): Require
3357 unconditionally executed stores for memcpy and memset recognition.
3358 (tree_loop_distribution): Calculate dominance info.
3360 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3362 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3363 (NO_PROFILE_COUNTERS): Likewise.
3364 (PROFILE_HOOK): Likewise.
3365 (FUNCTION_PROFILER): Likewise.
3366 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3368 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3370 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3371 calls and defines in TARGET_MACHO conditional.
3372 (load_macho_picbase_di): Likewise.
3373 (reload_macho_picbase): Likewise.
3374 (reload_macho_picbase_si): Likewise.
3375 (reload_macho_picbase_di): Likewise.
3376 (nonlocal_goto_receiver): Likewise.
3378 2013-09-30 Nick Clifton <nickc@redhat.com>
3380 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3381 that use the MSP430X ISA.
3382 (msp430_option_override): Scan -mmcu command line option for any
3383 MCU name that supports the MSP430X ISA.
3384 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3385 -mmcu options which enable the MSP430X ISA.
3387 2013-09-30 Richard Biener <rguenther@suse.de>
3390 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3391 before looking for setjmp-like calls.
3393 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3396 * config/darwin-protos.h (machopic_get_function_picbase): New.
3397 * config/darwin.c (machopic_get_function_picbase): New.
3398 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3399 label for a new func. (load_macho_picbase_di): Likewise.
3400 (reload_macho_picbase): New expand.
3401 (reload_macho_picbase_si): New insn.
3402 (reload_macho_picbase_di): New insn.
3403 (nonlocal_goto_receiver): New define and split.
3404 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3405 (unspecv enum): Add UNSPECV_NLGR.
3407 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3409 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3410 that altivec registers are correctly sized on Darwin.
3412 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3414 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3415 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3416 * config/x-darwin (host-darwin.o): Likewise.
3417 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3418 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3419 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3421 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3423 * doc/invoke.texi: Fix usage of @tie{} command.
3425 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3427 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3429 2013-09-28 Jan Hubicka <jh@suse.cz>
3431 * config/i386/x86-tune.def: Add documentation for each of the options;
3434 2013-09-28 Jan Hubicka <jh@suse.cz>
3436 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3438 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3439 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3440 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3441 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3442 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3444 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3446 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3447 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3448 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3449 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3450 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3451 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3452 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3453 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3454 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3455 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3456 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3457 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3458 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3459 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3460 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3461 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3462 gimple.h, godump.c, graphite-clast-to-gimple.c,
3463 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3464 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3465 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3466 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3467 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3468 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3469 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3470 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3471 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3472 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3473 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3474 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3475 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3476 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3477 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3478 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3479 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3480 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3481 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3482 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3483 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3484 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3485 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3486 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3487 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3488 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3489 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3490 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3491 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3492 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3493 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3494 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3495 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3496 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3497 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3498 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3499 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3500 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3501 whitespace before "(".
3503 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3505 * expr.h (extract_bit_field): Remove packedp parameter.
3506 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3507 from forward declaration.
3508 (store_split_bit_field): Remove packedp arg from calls to
3509 extract_fixed_bit_field.
3510 (extract_bit_field_1): Remove packedp parameter and packedp
3511 argument from recursive calls and calls to extract_fixed_bit_field.
3512 (extract_bit_field): Remove packedp parameter and corresponding
3513 arg to extract_bit_field_1.
3514 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3516 (extract_split_bit_field): Remove packedp arg from call to
3517 extract_fixed_bit_field.
3518 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3519 (copy_blkmode_from_reg): Likewise.
3520 (copy_blkmode_to_reg): Likewise.
3521 (read_complex_part): Likewise.
3522 (store_field): Likewise.
3523 (expand_expr_real_1): Likewise.
3524 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3525 to extract_bit_field.
3526 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3527 call to extract_bit_field.
3528 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3529 call to extract_bit_field.
3530 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3531 and special packedp behavior from -fstrict-volatile-bitfields
3534 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3536 * lra-eliminations.c (init_elim_table): Guard value_p.
3538 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3540 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3541 DFmode, DImode, and SFmode in the upper VSX registers based on the
3542 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3543 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3544 of the VSX register handling.
3546 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3548 (zero_extendsidi2_lfiwzx): Add support for loading into the
3549 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3550 only do VSX memory options on Altivec registers.
3551 (extendsidi2_lfiwax): Likewise.
3552 (extendsfdf2_fpr): Likewise.
3553 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3554 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3555 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3556 (movdi_internal64): Likewise.
3558 2013-09-27 Xinliang David Li <davidxl@google.com>
3560 * opts.c (finish_options): Adjust parameters
3561 according to vect cost model.
3562 (common_handle_option): Set dynamic vect cost
3564 targhooks.c (default_add_stmt_cost): Compute stmt cost
3566 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3567 Use helper function.
3568 * tree-vectorizer.h (unlimited_cost_model): New function.
3569 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3570 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3572 (vect_enhance_data_refs_alignment): Ditto.
3573 * flag-types.h: New enum.
3574 * common/config/i386/i386-common.c (ix86_option_init_struct):
3575 No need to initialize vect_cost_model flag.
3576 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3579 2013-09-27 Diego Novillo <dnovillo@google.com>
3581 * gimple.h (enum ssa_mode): Remove.
3583 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3585 * cfgloop.h (number_of_loops): Fix typo in check for null.
3587 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3590 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3591 are children of outermost saved_cfun's loop, and set it up to
3592 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3593 if loop != loop0 and SESE region contains bbs that belong to loop0.
3595 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3597 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3598 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3599 (set_address_segment, set_address_base, set_address_index)
3600 (set_address_disp): Accept the argument unconditionally.
3601 (baseness): Remove must_be_base_p and must_be_index_p checks.
3602 (decompose_normal_address): Classify as much as possible in the
3605 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3607 * cse.c (count_reg_usage): Handle INT_LIST.
3608 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3609 * reginfo.c (reg_scan_mark_refs): Likewise.
3610 * reload1.c (eliminate_regs_1): Likewise.
3612 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3615 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3618 2013-09-27 Richard Biener <rguenther@suse.de>
3620 PR tree-optimization/58459
3621 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3622 restriction not propagating into loops.
3624 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3626 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3627 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3628 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3630 2013-09-26 Richard Biener <rguenther@suse.de>
3632 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3634 2013-09-26 Richard Biener <rguenther@suse.de>
3636 * alias.h (component_uses_parent_alias_set): Rename to ...
3637 (component_uses_parent_alias_set_from): ... this.
3638 * alias.c (component_uses_parent_alias_set): Rename to ...
3639 (component_uses_parent_alias_set_from): ... this and return
3641 (reference_alias_ptr_type_1): Use the result from
3642 component_uses_parent_alias_set_from instead of stripping
3643 components one at a time.
3644 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3646 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3648 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3649 Move prototypes to...
3650 * tree-ssa-ter.h: New File. Move prototypes here.
3651 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3652 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3653 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3654 from is_replaceable_p.
3655 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3656 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3657 (process_replaceable): Use ter_is_replaceable_p.
3658 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3659 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3660 newly refactored ssa_is_replaceable_p.
3661 * cfgexpand.c: Include tree-outof-ssa.h.
3662 * ssaexpand.h: Delete.
3664 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3666 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3667 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3669 (create_gimple_tmp): Delete.
3670 (get_expr_type, build_assign, build_type_cast): Move to...
3671 * gimple-builder.c: New File.
3672 (get_expr_type): Relocate from gimple.c.
3673 (build_assign, build_type_cast): Change to only create ssanames.
3674 * gimple.h: Move prototypes to...
3675 * gimple-builder.h: New File. Here.
3676 * tree-ssa.h: And here.
3677 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3678 count_uses_and_derefs): Relocate from gimple.c.
3679 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3680 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3681 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3682 gimple_replace_ssa_lhs.
3683 * asan.c: Include gimple-builder.h.
3684 * Makefile.in: Add gimple-builder.o.
3686 2013-09-26 Richard Biener <rguenther@suse.de>
3688 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3689 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3690 (loe_visit_block): Use gcc_checking_assert.
3691 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3692 gimple_assign_ssa_name_copy_p.
3693 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3696 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3698 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3699 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3701 2013-09-26 Richard Biener <rguenther@suse.de>
3703 PR tree-optimization/58539
3704 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3705 the fact that debug statements are not taking part in loop-closed
3708 2013-09-26 Nick Clifton <nickc@redhat.com>
3710 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3711 time warning message.
3712 (msp430_print_operand_raw): Delete unused letter parameter.
3713 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3714 (msp430_print_operand_address): New function.
3715 (msp430_print_operand): Move address printing code from here to
3717 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3719 (zero_extendpsisi2): Likewise.
3720 (extendpsisi2): New pattern.
3721 (andneghi3): New pattern.
3723 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3725 * config/aarch64/aarch64.opt (mlra): New option.
3726 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3727 (TARGET_LRA_P): Define.
3729 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3731 * expr.c (expand_assignment): Remove obsolete comment.
3733 2013-09-25 Jeff Law <law@redhat.com>
3735 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3736 tree-ssa-threadupdate.h.
3737 (register_jump_thread): Similarly.
3738 * tree-ssa-threadupdate.h: New header file.
3739 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3740 * tree-vrp.c: Likewise.
3741 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3742 (thread_around_empty_blocks): Change type of path vector argument to
3743 an edge,type pair from just an edge. Initialize both elements when
3744 appending to a jump threading path. Tweak references to elements
3746 (thread_across_edge): Similarly. Release memory for the elements
3748 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3749 (dump_jump_thread_path): New function broken out from
3750 register_jump_thread.
3751 (register_jump_thread): Use dump_jump_thread_path. Change type of
3752 path vector entries. Search the path for NULL edges and dump
3753 the path if one is found. Tweak the conversion of path to 3-edge
3754 form to use the block copy type information embedded in the path.
3756 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3758 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3760 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3761 Vladimir Makarov <vmakarov@redhat.com>
3763 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3764 from the least significant bit.
3765 (strip_address_mutations): Add bitfield operations handling.
3766 (must_be_index_p): Add shifting and rotate operations handling.
3767 (set_address_base): Use must_be_base_p predicate.
3768 (set_address_index): Use must_be_index_p predicate.
3770 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3771 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3772 Sergey Lega <sergey.s.lega@intel.com>
3773 Anna Tikhonova <anna.tikhonova@intel.com>
3774 Ilya Tocar <ilya.tocar@intel.com>
3775 Andrey Turetskiy <andrey.turetskiy@intel.com>
3776 Ilya Verbin <ilya.verbin@intel.com>
3777 Kirill Yukhin <kirill.yukhin@intel.com>
3778 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3780 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3782 (ix86_expand_vector_move_misalign): Support new unaligned load and
3783 stores and use new names.
3784 (CODE_FOR_sse2_storedqu): Rename to ...
3785 (CODE_FOR_sse2_storedquv16qi): ... this.
3786 (CODE_FOR_sse2_loaddqu): Rename to ...
3787 (CODE_FOR_sse2_loaddquv16qi): ... this.
3788 (CODE_FOR_avx_loaddqu256): Rename to ...
3789 (CODE_FOR_avx_loaddquv32qi): ... this.
3790 (CODE_FOR_avx_storedqu256): Rename to ...
3791 (CODE_FOR_avx_storedquv32qi): ... this.
3792 * config/i386/i386.md (fpint_logic): New.
3793 * config/i386/sse.md (VMOVE): Extend for AVX512.
3797 (VI_UNALIGNED_LOADSTORE): Ditto.
3798 (sse2_avx_avx512f): Ditto.
3799 (sse2_avx2): Extend for AVX512.
3800 (sse4_1_avx2): Ditto.
3801 (avx2_avx512f): New.
3802 (sse): Extend for AVX512.
3805 (avxsizesuffix): Ditto.
3806 (sseintvecmode): Ditto.
3808 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3809 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3810 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3811 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3812 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3813 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3814 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3815 (STORENT_MODE): Extend for AVX512.
3816 (*absneg<mode>2): Replace constraint "x" with "v".
3817 (*mul<mode>3): Ditto.
3818 (*ieee_smin<mode>3): Ditto.
3819 (*ieee_smax<mode>3): Ditto.
3820 (avx_cmp<mode>3): Replace VF with VF_128_256.
3821 (*<sse>_maskcmp<mode>3_comm): Ditto.
3822 (<sse>_maskcmp<mode>3): Ditto.
3823 (<sse>_andnot<mode>3): Extend for AVX512.
3824 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3825 (<code><mode>3, fpint_logic): New.
3826 (*<code><mode>3): Extend for AVX512.
3827 (avx512flogicsuff): New.
3828 (avx512f_<logic><mode>): Ditto.
3829 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3831 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3832 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3833 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3834 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3835 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3836 (xop_vpermil2<mode>3): Ditto.
3837 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3838 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3839 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3840 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3842 2013-09-25 Tom Tromey <tromey@redhat.com>
3844 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3845 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3846 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3847 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3848 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3849 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3850 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3851 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3852 (GRAPHITE_HTAB_H): Remove.
3854 2013-09-25 Tom Tromey <tromey@redhat.com>
3856 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3858 2013-09-25 Tom Tromey <tromey@redhat.com>
3860 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3862 2013-09-25 Tom Tromey <tromey@redhat.com>
3864 * config/i386/t-i386 (i386.o): Remove.
3865 (i386-c.o): Use COMPILE and POSTCOMPILE.
3867 2013-09-25 Tom Tromey <tromey@redhat.com>
3869 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3871 2013-09-25 Tom Tromey <tromey@redhat.com>
3873 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3874 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3875 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3876 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3877 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3878 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3879 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3880 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3881 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3882 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3883 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3884 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3885 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3886 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3887 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3888 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3889 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3890 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3891 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3892 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3893 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3894 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3895 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3896 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3897 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3898 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3899 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3900 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3901 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3902 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3903 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3904 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3905 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3906 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3907 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3908 (graphite-clast-to-gimple.o, graphite-dependences.o)
3909 (graphite-interchange.o, graphite-poly.o)
3910 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3911 (graphite-optimize-isl.o, tree-vect-loop.o)
3912 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3913 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3914 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3915 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3916 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3917 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3918 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3919 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3920 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3921 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3922 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3923 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3924 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3925 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3926 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3927 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3928 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3929 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3930 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3931 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3932 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3933 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3934 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3935 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3936 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3937 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3938 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3939 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3940 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3941 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3942 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3943 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3944 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3945 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3946 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3947 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3948 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3949 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3950 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3951 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3952 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3953 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3954 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3955 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3956 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3957 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3958 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3959 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3960 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3961 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3962 (gcov-dump.o): Remove.
3963 (default-c.o): Use COMPILE and POSTCOMPILE.
3964 (CFLAGS-gcc.o): New variable.
3965 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3967 2013-09-25 Tom Tromey <tromey@redhat.com>
3969 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3970 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3971 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3972 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3973 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3974 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3975 (c-family/array-notation-common.o, c-family/stub-objc.o)
3976 (c-family/c-ubsan.o): Remove.
3978 2013-09-25 Tom Tromey <tromey@redhat.com>
3980 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3982 2013-09-25 Tom Tromey <tromey@redhat.com>
3984 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3985 to add -DENABLE_SHARED_LIBGCC.
3986 (gcc.o): Don't use subshell.
3988 2013-09-25 Tom Tromey <tromey@redhat.com>
3990 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3991 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3992 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3993 * configure, config.in: Rebuild.
3995 2013-09-25 Tom Tromey <tromey@redhat.com>
3997 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3998 (COMPILE, POSTCOMPILE): New variables.
3999 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4000 (DEPFILES): New variable.
4001 Include ".Po" files.
4002 * configure.ac: Add checks for dependency checking.
4003 * configure, aclocal.m4: Regenerate.
4005 2013-09-25 Tom Tromey <tromey@redhat.com>
4007 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4008 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4010 2013-09-25 Tom Tromey <tromey@redhat.com>
4012 * Makefile.in (generated_files): Add options.h,
4013 target-hooks-def.h, insn-opinit.h,
4014 common/common-target-hooks-def.h, pass-instances.def,
4015 c-family/c-target-hooks-def.h.
4017 2013-09-25 Jeff Law <law@redhat.com>
4019 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4020 than foo[foo.length () - 1] to access last member in a vec.
4021 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4023 2013-09-25 Richard Biener <rguenther@suse.de>
4026 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4028 2013-09-25 Jan Hubicka <jh@suse.cz>
4030 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4033 2013-09-25 Marek Polacek <polacek@redhat.com>
4036 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4037 when determining the type name.
4039 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4041 * config/sh/sh.md: Fix formatting.
4043 2013-09-24 Xinliang David Li <davidxl@google.com>
4045 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4046 max peel iterations parameter.
4047 * param.def: New parameter.
4048 * doc/invoke.texi: Document New parameter.
4050 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4052 * gimple-pretty-print.c: Various whitespace tweaks.
4053 * tree-core.h: Likewise.
4054 * tree-pretty-print.c: Likewise.
4055 * tree-ssa-alias.c: Likewise.
4056 * tree-ssa-copy.c: Likewise.
4057 * tree-ssanames.c: Likewise.
4058 * tree-ssanames.h: Likewise.
4059 * tree-vrp.c: Likewise.
4061 2013-09-24 Alan Modra <amodra@gmail.com>
4065 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4066 for output operands that disallow regs. Don't use EXPAND_WRITE on
4069 2013-09-24 Richard Biener <rguenther@suse.de>
4072 * tree.c (reference_alias_ptr_type): Move ...
4073 * alias.c (reference_alias_ptr_type): ... here and implement
4074 in terms of the new reference_alias_ptr_type_1.
4075 (ref_all_alias_ptr_type_p): New helper.
4076 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4077 use ref_all_alias_ptr_type_p.
4078 (get_deref_alias_set): Add flag_strict_aliasing check here.
4079 (reference_alias_ptr_type_1): New function, split out from ...
4080 (get_alias_set): ... here.
4081 (alias_ptr_types_compatible_p): New function.
4082 * alias.h (reference_alias_ptr_type): Declare.
4083 (alias_ptr_types_compatible_p): Likewise.
4084 * tree.h (reference_alias_ptr_type): Remove.
4085 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4086 to compare MEM_REF alias types.
4088 2013-09-24 Richard Biener <rguenther@suse.de>
4090 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4092 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4094 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4095 reload helper function arrays into a single array reg_addr.
4096 (reload_fpr_gpr): Likewise.
4097 (reload_gpr_vsx): Likewise.
4098 (reload_vsx_gpr): Likewise.
4099 (struct rs6000_reg_addr): Likewise.
4100 (reg_addr): Likewise.
4101 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4102 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4103 (rs6000_init_hard_regno_mode_ok): Likewise.
4104 (rs6000_secondary_reload_direct_move): Likewise.
4105 (rs6000_secondary_reload): Likewise.
4107 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4108 constraints: wu, ww, and wy. Repurpose wv constraint added during
4109 power8 changes. Put wg constraint in alphabetical order.
4111 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4112 for future work to add ISA 2.07 VSX single precision support.
4113 (-mvsx-scalar-double): Change default from -1 to 1, update
4114 documentation comment.
4115 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4116 (-mupper-regs-df): New debug switch to control whether DF values
4117 can go in the traditional Altivec registers.
4118 (-mupper-regs-sf): New debug switch to control whether SF values
4119 can go in the traditional Altivec registers.
4121 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4123 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4124 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4125 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4126 (rs6000_debug_legitimate_address_p): Print if we are running
4127 before, during, or after reload.
4128 (rs6000_secondary_reload): Add a comment.
4129 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4131 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4132 constraints. Update documentation string.
4133 (wd constraint): Likewise.
4134 (wf constraint): Likewise.
4135 (wg constraint): Likewise.
4136 (wn constraint): Likewise.
4137 (ws constraint): Likewise.
4138 (wt constraint): Likewise.
4139 (wx constraint): Likewise.
4140 (wz constraint): Likewise.
4141 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4143 (ww constraint): Likewise.
4144 (wy constraint): Likewise.
4145 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4146 in the previous submissions.
4147 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4149 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4151 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4152 (REG_BR_PROB): Say that the probability is stored in an int_list.
4153 * reg-notes.def: Update commentary to mention INT_LIST.
4154 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4155 (INT_LIST): New rtx.
4156 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4157 * rtlanal.c (int_reg_note_p): New function.
4158 (alloc_reg_note): Assert that the note does not have an int argument.
4159 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4160 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4161 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4162 rather than an INSN_LIST. Handle INT_LIST.
4163 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4164 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4165 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4166 Manipulate them as ints rather than rtxes.
4167 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4168 * regmove.c (copy_src_to_dest): Likewise.
4169 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4171 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4172 into the cases that need it.
4173 * config/arm/arm.c (arm_unwind_emit): Likewise.
4175 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4176 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4177 * loop-doloop.c (add_test, doloop_modify): Likewise.
4178 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4179 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4180 * predict.c (combine_predictions_for_insn): Likewise.
4181 * print-rtl.c (print_rtx): Handle INT_LIST.
4182 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4183 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4184 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4185 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4186 (ix86_split_fp_branch, predict_jump): Likewise.
4187 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4188 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4189 * config/spu/spu.c (ea_load_store_inline): Likewise.
4191 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4192 value of a REG_BR_PROB note.
4193 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4194 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4195 * emit-rtl.c (try_split): Likewise.
4196 * predict.c (br_prob_note_reliable_p): Likewise.
4197 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4198 * reorg.c (mostly_true_jump): Likewise.
4199 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4200 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4201 * config/i386/i386.c (ix86_print_operand): Likewise.
4202 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4203 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4204 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4205 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4206 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4207 * config/sparc/sparc.c (output_cbranch): Likewise.
4208 * config/spu/spu.c (get_branch_target): Likewise.
4209 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4210 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4212 2013-09-23 Jan Hubicka <jh@suse.cz>
4214 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4216 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4217 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4219 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4220 (gimple_fold_call): Dump inconsistent devirtualizations; add
4221 sanity check for type based devirtualizations.
4222 * ipa-prop.c: Include ipa-utils.h
4223 (ipa_intraprocedural_devirtualization): Add sanity check.
4224 (try_make_edge_direct_virtual_call): Likewise.
4226 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4228 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4229 assignment statements.
4231 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4233 * gimple-pretty-print.c (dump_ssaname_info): New function.
4234 (dump_gimple_phi): Call it.
4235 (pp_gimple_stmt_1): Likewise.
4236 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4237 (range_info_def): Declare.
4238 * tree-pretty-print.c (pp_double_int): New function.
4239 (dump_generic_node): Call it.
4240 * tree-pretty-print.h (pp_double_int): Declare.
4241 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4242 * tree-ssanames.h (range_info_def): New structure.
4243 (value_range_type): Move definition here.
4244 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4246 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4248 (set_range_info): New function.
4249 (get_range_info): Likewise.
4250 (duplicate_ssa_name_range_info): Likewise.
4251 (duplicate_ssa_name_fn): Check pointer type and call
4252 duplicate_ssa_name_range_info.
4253 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4254 * tree-vrp.c (value_range_type): Remove definition, now in
4256 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4257 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4258 (SSA_NAME_RANGE_INFO): New macro.
4260 2013-09-23 Richard Biener <rguenther@suse.de>
4262 PR tree-optimization/58464
4263 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4264 (phi_trans_add): Change to add conditionally on being not present.
4265 (phi_translate_1): Remove recursion detection here.
4266 (phi_translate): Pre-seed the cache with NULL to catch
4267 recursion here in a more generic way.
4268 (bitmap_find_leader): Adjust comment.
4269 (get_representative_for): Dump value-numbers.
4270 (create_expression_by_pieces): Likewise.
4271 (insert_into_preds_of_block): Likewise.
4273 2013-09-23 Christian Bruel <christian.bruel@st.com>
4276 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4277 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4279 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4282 2013-09-20 Renlin Li <renlin.li@arm.com>
4284 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4286 (aarch64_expand_epilogue): Likewise.
4287 (aarch64_legitimize_reload_address): Likewise.
4289 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4291 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4292 NULL_TREE before pushing them onto the vector. Likewise for labels.
4294 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4296 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4297 comparison operators when -fno-trapping-math is in effect.
4298 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4299 comparison operators in TFmode and assert that unsupported operators
4301 (ia64_print_operand): Likewise.
4303 2013-09-21 Jan Hubicka <jh@suse.cz>
4305 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4306 (sse_typeless_stores): Enable for core
4307 (sse_load0_by_pxor): Likewise.
4308 (four_jump_limit): Disable for core.
4309 (pad_returns): Likewise.
4310 (avoid_vector_decode): Likewise.
4311 (fuse_cmp_and_branch): Enable for cores.
4312 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4314 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4317 * config/pa/pa.c (pa_option_override): Disable auto increment and
4318 decrement instructions until reload is completed.
4320 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4321 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4322 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4324 2013-09-20 DJ Delorie <dj@redhat.com>
4325 Nick Clifton <nickc@redhat.com>
4327 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4328 (need_to_save): Save bank 0 on interrupts.
4329 (characterize_address): Strip far address wrappers.
4330 (rl78_as_legitimate_address): Likewise.
4331 (transcode_memory_rtx): Likewise.
4332 (rl78_peep_movhi_p): Disable this peephole after devirt.
4333 (rl78_propogate_register_origins): Forget all origins when a
4335 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4336 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4338 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4339 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4340 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4342 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4344 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4346 * config/pa/pa.md: In "scc" insn patterns, change output template to
4347 handle const0_rtx in reg_or_0_operand operands.
4349 2013-09-20 Martin Husemann <martin@NetBSD.org>
4352 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4353 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4355 2013-09-20 Richard Biener <rguenther@suse.de>
4358 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4359 remembering SSA name version and block index.
4360 (new_scev_info_str): Adjust.
4361 (hash_scev_info): Likewise. Also hash the block index.
4362 (eq_scev_info): Adjust.
4363 (find_var_scev_info): Likewise.
4364 (struct instantiate_cache_entry): Remove.
4365 (struct instantiate_cache_type): Use a htab to map name, block
4367 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4368 (get_instantiated_value_entry): Likewise.
4369 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4370 (instantiate_scev_name): Adjust.
4372 2013-09-20 Jeff Law <law@redhat.com>
4374 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4376 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4378 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4379 Call aarch64_simd_expand_args to update op[argc].
4381 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4383 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4386 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4388 * gengtype.c (file_rules): Added rule for *.cc files.
4389 (get_output_file_with_visibility): Give fatal message when no
4392 2013-09-20 Renlin Li <renlin.li@arm.com>
4394 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4395 (aarch64_expand_epilogue): Likewise.
4396 (aarch64_legitimize_reload_address): Likewise.
4398 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4401 * expr.c (expand_assignment): Remove misalignp code path.
4403 2013-09-20 Marek Polacek <polacek@redhat.com>
4406 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4407 TYPE_PRECISION. Add asserts.
4409 2013-09-20 Richard Biener <rguenther@suse.de>
4411 PR tree-optimization/58453
4412 * tree-loop-distribution.c (distribute_loop): Apply the cost
4413 model for -ftree-loop-distribute-patterns, too.
4415 2013-09-20 Richard Biener <rguenther@suse.de>
4418 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4419 make type comparison less strict.
4421 2013-09-20 Alan Modra <amodra@gmail.com>
4423 * configure: Regenerate.
4424 * aclocal.m4: Regenerate.
4426 2013-09-20 Marek Polacek <polacek@redhat.com>
4429 * doc/extend.texi: Document that attribute used is meant to be used
4430 on variables with static storage duration.
4432 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4434 PR tree-optimization/58472
4435 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4436 simd_lane_access set inv_p = false.
4437 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4438 the simduid magic VAR_DECL.
4440 2013-09-19 Jan Hubicka <jh@suse.cz>
4442 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4444 2013-09-17 Jeff Law <law@redhat.com>
4446 * tree-ssa-dom.c (record_temporary_equivalences): New function
4447 split out of dom_opt_dom_walker::after_dom_children.
4448 (dom_opt_dom_walker::thread_across_edge): Move common code
4449 in here from dom_opt_dom_walker::after_dom_children.
4450 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4452 2013-09-19 Jan Hubicka <jh@suse.cz>
4454 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4455 (TARGET_GENERIC): Use PROCESOR_GENERIC
4456 (enum processor_type): Unify generic32 and 64.
4457 * i386.md (cpu): Likewise.
4458 * x86-tune.def (use_leave): Enable for generic32.
4459 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4460 * athlon.md: Change generic64 to generic in all occurences.
4461 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4462 (ix86_target_macros_internal): Likewise.
4463 * driver-i386.c (host_detect_local_cpu): Likewise.
4464 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4466 (generic_memcpy, generic_memset, generic_cost): This one.
4467 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4468 (m_GENERIC32, m_GENERIC64): Remove.
4469 (m_GENERIC): Turn into one flag.
4470 (processor_target): Unify generic tunnings.
4471 (ix86_option_override_internal): Replace generic32/64 by generic.
4472 (ix86_issue_rate): Likewise.
4473 (ix86_adjust_cost): Likewise.
4475 2013-09-19 Jan Hubicka <jh@suse.cz>
4477 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4478 of speculative flag.
4480 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4482 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4483 GOMP_sections_start, even if !exit_reachable.
4485 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4487 * lra-constraints.c (need_for_all_save_p): Use macro
4488 HARD_REGNO_CALL_PART_CLOBBERED.
4489 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4490 set up pseudo conflict hard regs.
4492 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4495 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4498 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4500 PR rtl-optimization/58438
4501 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4502 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4503 from previous subpasses.
4505 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4507 * arm.c (arm_get_frame_offsets): Validate architecture supports
4508 LDRD/STRD before accepting the tuning preference.
4509 (arm_expand_prologue): Likewise.
4510 (arm_expand_epilogue): Likewise.
4512 2013-09-18 Richard Biener <rguenther@suse.de>
4514 PR tree-optimization/58417
4515 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4516 have chrecs with symbols defined in the loop as operands.
4517 (chrec_fold_multiply): Likewise.
4518 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4519 parameters before folding binary operations.
4520 (struct instantiate_cache_entry_hasher): Remove.
4521 (struct instantiate_cache_type): Use a pointer-map.
4522 (instantiate_cache_type::instantiate_cache_type): New function.
4523 (instantiate_cache_type::get): Likewise.
4524 (instantiate_cache_type::set): Likewise.
4525 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4526 (get_instantiated_value_entry): Likewise.
4527 (global_cache): New global.
4528 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4529 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4530 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4531 (instantiate_scev_name): Adjust.
4532 (instantiate_scev): Construct global instead of local cache.
4533 (resolve_mixers): Likewise.
4535 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4536 Paolo Carlini <paolo.carlini@oracle.com>
4539 * doc/implement-cxx.texi: Fix references to the C++ standards.
4541 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4543 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4544 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4545 simd_lane_access drs, update also DR_ALIGNED_TO.
4547 2013-09-18 Marek Polacek <polacek@redhat.com>
4550 * doc/extend.texi: Document no_sanitize_undefined attribute.
4551 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4552 no_sanitize_undefined attribute.
4554 2013-09-18 Nick Clifton <nickc@redhat.com>
4556 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4557 (ASM_DECLARE_FUNCTION_NAME): Define.
4559 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4561 * compare-elim.c (find_comparison_dom_walker): New class
4562 (find_comparisons_in_bb): Rename to
4563 find_comparison_dom_walker::before_dom_children
4564 (find_comparisons): Adjust
4565 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4567 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4568 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4569 (init_walk_dominator_tree): Remove declaration.
4570 (fini_walk_dominator_tree): Remove declaration.
4571 * fwprop.c (single_def_use_dom_walker): New class
4572 (single_def_use_enter_block): Convert to
4573 single_def_use_dom_walker::before_dom_children.
4574 (single_def_use_leave_block): Convert to
4575 single_def_use_dom_walker::after_dom_children.
4576 (build_single_def_use_links): Adjust.
4577 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4579 (find_candidates_in_block): Convert to
4580 find_candidates_dom_walker::before_dom_children.
4581 (execute_strength_reduction): Adjust.
4582 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4583 (sese_dom_walker): New class.
4584 (sese_dom_walker::sese_dom_walker): New constructor.
4585 (sese_dom_walker::~sese_dom_walker): New destructor.
4586 (build_sese_conditions_before): Convert to
4587 sese_dom_walker::before_dom_children.
4588 (build_sese_conditions_after): Convert to
4589 sese_dom_walker::after_dom_children.
4590 (build_poly_scop): Adjust
4591 * tree-into-ssa.c (rewrite_dom_walker): New class
4592 (rewrite_enter_block): Convert to
4593 rewrite_dom_walker::before_dom_children.
4594 (rewrite_leave_block): Convert to
4595 rewrite_dom_walker::after_dom_children.
4596 (rewrite_update_dom_walker): New class.
4597 (rewrite_update_enter_block): Convert to
4598 rewrite_update_dom_walker::before_dom_children.
4599 (rewrite_update_leave_block): Convert to
4600 rewrite_update_dom_walker::after_dom_children.
4601 (rewrite_blocks, rewrite_into_ssa): Adjust.
4602 (mark_def_dom_walker): New class.
4603 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4604 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4605 (mark_def_sites_blocks): Convert to
4606 mark_def_dom_walker::before_dom_children.
4607 (mark_def_site_blocks): Remove.
4608 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4609 (tree_ssa_dominator_optimize): Adjust.
4610 (dom_thread_across_edge): Convert to method
4611 dom_opt_dom_walker::thread_across_edge.
4612 (dom_opt_enter_block): Convert to member function
4613 dom_opt_dom_walker::before_dom_children.
4614 (dom_opt_leave_block): Convert to member function
4615 dom_opt_dom_walker::after_dom_children.
4616 * tree-ssa-dse.c (dse_dom_walker): New class.
4617 (dse_enter_block): Convert to member function
4618 dse_dom_walker::before_dom_children.
4619 (tree_ssa_dse): Adjust.
4620 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4621 (determine_invariantness_stmt): Convert to method
4622 invariantness_dom_walker::before_dom_children.
4623 (determine_invariantness): Remove
4624 (move_computations_dom_walker): New class.
4625 (move_computations_stmt): Convert to method
4626 move_computations_dom_walker::before_dom_children.
4627 (move_computations, tree_ssa_lim): Adjust.
4628 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4629 (nt_init_block): Make method
4630 notrappping_dom_walker::before_dom_children.
4631 (nt_fini_block): Make
4632 method nontrapping_dom_walker::after_dom_children.
4633 (get_non_trapping): Adjust.
4634 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4635 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4636 (eliminate_leave_block): Make method.
4637 eliminate_dom_walker::after_dom_children.
4639 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4640 (strlen_enter_block): Make method
4641 strlen_dom_walker::before_dom_children.
4642 (strlen_leave_block): Make
4643 method strlen_dom_walker::after_dom_children.
4644 (tree_ssa_strlen): Adjust.
4645 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4646 (tree_ssa_uncprop): Adjust.
4647 (uncprop_leave_block): Make method
4648 uncprop_dom_walker::after_dom_children.
4649 (uncprop_leave_block): Make method
4650 uncprop_dom_walker::before_dom_children.
4652 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4654 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4655 branch in same basic block. Check both src and dest of the move insn.
4657 2013-09-17 Nick Clifton <nickc@redhat.com>
4659 * config/rl78/rl78-real.md (bf): New pattern.
4661 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4662 (rl78_print_operand): Do not put a # before a %B.
4663 * config/rl78/rl78.opt: Tweak doc strings.
4665 2013-09-17 DJ Delorie <dj@redhat.com>
4667 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4668 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4669 (rl78_option_override): Likewise, if -mallregs.
4670 (is_virtual_register): Likewise.
4671 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4672 (REGNO_OK_FOR_BASE_P): Likewise.
4673 * config/rl78/rl78.opt (-mallregs): New.
4675 2013-09-17 Nick Clifton <nickc@redhat.com>
4677 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4678 For interrupt functions: save all call clobbered registers if the
4679 interrupt handler is not a leaf function.
4680 (rl78_expand_prologue): Always recompute the frame information.
4681 For interrupt functions: only select bank 0 if one of the bank 0
4682 registers is going to be psuhed.
4684 2013-09-17 DJ Delorie <dj@redhat.com>
4686 * config/rl78/constraints.md: For each W* constraint, rename to C*
4687 and create a W* constraint that checks for an optional ES: prefix
4689 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4690 (es_addr): New. Used to wrap far addresses.
4691 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4692 (rl78_es_base): New.
4693 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4694 wrapped far addresses.
4695 (rl78_print_operand_1): Unwrap far addresses before processing.
4696 (rl78_lo16): Wrap far addresses in unspecs.
4697 (rl78_es_addr): New.
4698 (rl78_es_base): New.
4699 (insn_ok_now): Check for not-yet-wrapped far addresses.
4700 (transcode_memory_rtx): Properly re-wrap far addresses.
4702 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4704 * config/sparc/t-rtems: Add leon3 multilibs.
4706 2013-09-17 Cong Hou <congh@google.com>
4708 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4709 when checking the dot production pattern. The type of rhs operand
4710 of multiply is now checked correctly.
4712 2013-09-17 Jeff Law <law@redhat.com>
4714 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4715 edge implied equivalences into successor phis.
4716 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4717 here from tree-ssa-threadedge.c.
4718 (mark_threaded_blocks): When threading through a joiner, if both
4719 successors of the joiner's clone reach the same block, verify the
4720 PHI arguments are equal. If not, cancel the jump threading request.
4721 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4722 tree-ssa-threadupdate.c
4723 (thread_across_edge): Don't check PHI argument equality when
4724 threading through joiner block here.
4726 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4728 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4729 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4730 (warn_uninit, warn_uninitialized_vars,
4731 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4732 Move to tree-ssa-uninit.c.
4733 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4734 (has_undefined_value_p): New. Pass dependant parts of
4735 ssa_undefined_value_p.
4736 (uninit_undefined_value_p): Use has_undefined_value_p.
4737 (warn_uninit, warn_uninitialized_vars,
4738 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4739 Move from tree-ssa.c.
4740 * tree-ssa.h: Adjust prototypes.
4742 2013-09-17 Jan Hubicka <jh@suse.cz>
4745 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4746 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4747 FUNCTION_NOT_OPTIMIZED.
4748 * ipa-inline-analysis.c (compute_inline_parameters): Function
4749 not optimized is not inlinable unless it is alwaysinline.
4750 (inline_analyze_function): Force calls in not optimized
4751 function not inlinable.
4753 2013-09-17 Jan Hubicka <jh@suse.cz>
4756 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4758 * ipa.c (function_and_variable_visibility): Likewise.
4759 * ipa-profile.c (ipa_profile): Likewise.
4761 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4764 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4765 attribute, and don't inline the resolver in this case.
4767 2013-09-17 Teresa Johnson <tejohnson@google.com>
4769 * coverage.c (get_coverage_counts): Add missing newline.
4771 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4773 PR tree-optimization/58088
4774 * fold-const.c (mask_with_trailing_zeros): New function.
4775 (fold_binary_loc): Make sure we don't recurse infinitely
4776 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4777 Use mask_with_trailing_zeros where appropriate.
4779 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4781 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4782 of 'prev' var to get better distance estimation.
4784 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4786 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4787 * tree-inline.c (is_parameter_of): New predicate.
4788 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4789 a parameter has been remapped.
4790 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4791 (optimize_inline_calls): Initialize transform_parameter.
4792 (copy_gimple_seq_and_replace_locals): Likewise.
4793 (tree_function_versioning): Likewise.
4794 (maybe_inline_call_in_expr): Likewise.
4796 2013-09-17 Nick Clifton <nickc@redhat.com>
4798 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4799 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4801 (is_attr_func): New function.
4802 (msp430_is_interrupt_func): New function.
4803 (is_naked_func): New function.
4804 (is_reentrant_func): New function.
4805 (is_critical_func): New function.
4806 (msp430_start_function): Add annotations for function attributes.
4807 (msp430_attr): New function.
4808 (msp430_attribute_table): New.
4809 (msp430_function_section): New function.
4810 (TARGET_ASM_FUNCTION_SECTION): Define.
4811 (msp430_builtin): New enum.
4812 (msp430_init_builtins): New function.
4813 (msp430_builtin_devl): New function.
4814 (msp430_expand_builtin): New function.
4815 (TARGET_INIT_BUILTINS): Define.
4816 (TARGET_EXPAND_BUILTINS): Define.
4817 (TARGET_BUILTIN_DECL): Define.
4818 (msp430_expand_prologue): Add support for naked, interrupt,
4819 critical and reentrant functions.
4820 (msp430_expand_epilogue): Likewise.
4821 (msp430_print_operand): Handle 'O' character.
4822 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4824 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4825 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4826 (pushm): Use a 'n' rather than an 'i' constraint.
4827 (msp_return): Add generation of the interrupt return instruction.
4828 (disable_interrupts): New pattern.
4829 (enable_interrupts): New pattern.
4830 (push_intr_state): New pattern.
4831 (pop_intr_state): New pattern.
4832 (bic_SR): New pattern.
4833 (bis_SR): New pattern.
4834 * doc/extend.texi: Document MSP430 function attributes and builtin
4837 2013-09-17 Richard Biener <rguenther@suse.de>
4839 PR tree-optimization/58432
4840 * tree-loop-distribution.c (tree_loop_distribution): Also
4841 scan PHIs for outside loop uses and seed a partition from them.
4843 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4845 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4846 (restructure_reference): Call backtrace_base_for_ref.
4848 2013-09-17 Alan Modra <amodra@gmail.com>
4851 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4854 2013-09-16 DJ Delorie <dj@redhat.com>
4856 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4857 vregs location for RL78/G10.
4858 (rl78_expand_prologue): Avoid SEL on G10.
4859 (rl78_expand_epilogue): Likewise.
4860 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4861 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4862 __RL78_G10__ when appropriate.
4863 (ASM_SPEC): Pass -mg10 along to the assembler.
4864 * config/rl78/rl78.md (sel_rb): Disable for G10.
4865 * config/rl78/rl78.opt: Add -mg10 option.
4866 * config/rl78/t-rl78: Add -mg10 multilib.
4868 2013-09-16 Xinliang David Li <davidxl@google.com>
4870 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4871 * omp-low.c (omp_max_vf): Ditto.
4872 (expand_omp_simd): Ditto.
4873 * tree-vectorizer.c (vectorize_loops): Ditto.
4874 (gate_vect_slp): Ditto.
4875 (gate_increase_alignment): Ditto.
4876 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4877 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4878 (gate_tree_vectorize): Name change.
4879 (tree_vectorize): Ditto.
4880 (pass_vectorize::gate): Call new function.
4881 (pass_vectorize::execute): Ditto.
4882 * opts.c: O3 default setting change.
4883 (finish_options): Check new flag.
4884 * doc/invoke.texi: Document new flags.
4885 * common.opt: New flags.
4887 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4889 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4890 (Condition Code): Don't reference it.
4891 * doc/tm.texi: Regenerate.
4893 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4896 * lra-constraints.c (undo_optional_reloads): Consider all optional
4897 reload even if it did not get a hard reg.
4899 2013-09-16 Teresa Johnson <tejohnson@google.com>
4901 * dumpfile.c (dump_loc): Remove newline emission.
4902 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4903 emission to dump_printf_loc calls where missing.
4904 (vect_mark_for_runtime_alias_test): Ditto.
4905 (vect_analyze_data_ref_dependence): Ditto.
4906 (vect_analyze_data_ref_dependences): Ditto.
4907 (vect_slp_analyze_data_ref_dependence): Ditto.
4908 (vect_slp_analyze_data_ref_dependences): Ditto.
4909 (vect_compute_data_ref_alignment): Ditto.
4910 (vect_update_misalignment_for_peel): Ditto.
4911 (vect_verify_datarefs_alignment): Ditto.
4912 (vector_alignment_reachable_p): Ditto.
4913 (vect_get_data_access_cost): Ditto.
4914 (vect_enhance_data_refs_alignment): Ditto.
4915 (vect_find_same_alignment_drs): Ditto.
4916 (vect_analyze_data_refs_alignment): Ditto.
4917 (vect_analyze_group_access): Ditto.
4918 (vect_analyze_data_ref_access): Ditto.
4919 (vect_analyze_data_ref_accesses): Ditto.
4920 (vect_prune_runtime_alias_test_list): Ditto.
4921 (vect_analyze_data_refs): Ditto.
4922 (vect_create_addr_base_for_vector_ref): Ditto.
4923 (vect_create_data_ref_ptr): Ditto.
4924 (vect_grouped_store_supported): Ditto.
4925 (vect_grouped_load_supported): Ditto.
4926 * value-prof.c (check_counter): Ditto.
4927 (check_ic_target): Ditto.
4928 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4929 (vect_recog_widen_mult_pattern): Ditto.
4930 (vect_recog_widen_sum_pattern): Ditto.
4931 (vect_recog_over_widening_pattern): Ditto.
4932 (vect_recog_widen_shift_pattern): Ditto.
4933 (vect_recog_rotate_pattern): Ditto.
4934 (vect_recog_vector_vector_shift_pattern): Ditto.
4935 (vect_recog_divmod_pattern): Ditto.
4936 (vect_recog_mixed_size_cond_pattern): Ditto.
4937 (vect_recog_bool_pattern): Ditto.
4938 (vect_pattern_recog_1): Ditto.
4939 (vect_pattern_recog): Ditto.
4940 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4941 (vect_is_simple_iv_evolution): Ditto.
4942 (vect_analyze_scalar_cycles_1): Ditto.
4943 (vect_get_loop_niters): Ditto.
4944 (vect_analyze_loop_1): Ditto.
4945 (vect_analyze_loop_form): Ditto.
4946 (vect_analyze_loop_operations): Ditto.
4947 (vect_analyze_loop_2): Ditto.
4948 (vect_analyze_loop): Ditto.
4949 (report_vect_op): Ditto.
4950 (vect_is_slp_reduction): Ditto.
4951 (vect_is_simple_reduction_1): Ditto.
4952 (vect_get_known_peeling_cost): Ditto.
4953 (vect_estimate_min_profitable_iters): Ditto.
4954 (vect_model_reduction_cost): Ditto.
4955 (vect_model_induction_cost): Ditto.
4956 (get_initial_def_for_induction): Ditto.
4957 (vect_create_epilog_for_reduction): Ditto.
4958 (vectorizable_reduction): Ditto.
4959 (vectorizable_induction): Ditto.
4960 (vectorizable_live_operation): Ditto.
4961 (vect_loop_kill_debug_uses): Ditto.
4962 (vect_transform_loop): Ditto.
4963 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4964 (vect_stmt_relevant_p): Ditto.
4965 (process_use): Ditto.
4966 (vect_mark_stmts_to_be_vectorized): Ditto.
4967 (vect_model_simple_cost): Ditto.
4968 (vect_model_promotion_demotion_cost): Ditto.
4969 (vect_model_store_cost): Ditto.
4970 (vect_get_store_cost): Ditto.
4971 (vect_model_load_cost): Ditto.
4972 (vect_get_load_cost): Ditto.
4973 (vect_init_vector_1): Ditto.
4974 (vect_get_vec_def_for_operand): Ditto.
4975 (vect_finish_stmt_generation): Ditto.
4976 (vectorizable_call): Ditto.
4977 (vectorizable_conversion): Ditto.
4978 (vectorizable_assignment): Ditto.
4979 (vectorizable_shift): Ditto.
4980 (vectorizable_operation): Ditto.
4981 (vectorizable_store): Ditto.
4982 (vectorizable_load): Ditto.
4983 (vectorizable_condition): Ditto.
4984 (vect_analyze_stmt): Ditto.
4985 (vect_transform_stmt): Ditto.
4986 (vect_is_simple_use): Ditto.
4987 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4988 (vect_can_advance_ivs_p): Ditto.
4989 (vect_update_ivs_after_vectorizer): Ditto.
4990 (vect_do_peeling_for_loop_bound): Ditto.
4991 (vect_gen_niters_for_prolog_loop): Ditto.
4992 (vect_update_inits_of_drs): Ditto.
4993 (vect_create_cond_for_alias_checks): Ditto.
4994 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4995 (vect_build_slp_tree_1): Ditto.
4996 (vect_supported_load_permutation_p): Ditto.
4997 (vect_analyze_slp_instance): Ditto.
4998 (vect_analyze_slp): Ditto.
4999 (vect_make_slp_decision): Ditto.
5000 (vect_detect_hybrid_slp): Ditto.
5001 (vect_bb_vectorization_profitable_p): Ditto.
5002 (vect_slp_analyze_bb_1): Ditto.
5003 (vect_update_slp_costs_according_to_vf): Ditto.
5004 (vect_get_mask_element): Ditto.
5005 (vect_transform_slp_perm_load): Ditto.
5006 (vect_schedule_slp_instance): Ditto.
5007 (vect_schedule_slp): Ditto.
5008 (vect_slp_transform_bb): Ditto.
5009 * profile.c (read_profile_edge_counts): Ditto.
5010 (compute_branch_probabilities): Ditto.
5011 * coverage.c (get_coverage_counts): Ditto.
5013 2013-09-16 Diego Novillo <dnovillo@google.com>
5015 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5017 2013-09-16 Jan Hubicka <jh@suse.cz>
5019 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5020 abstract functions; for static functions check the presence of body.
5022 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5024 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5025 * config/aarch64/aarch64-simd.md
5026 (aarch64_mla_elt<mode>): New.
5027 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5028 (aarch64_mls_elt<mode>): Likewise.
5029 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5030 (aarch64_fma4_elt<mode>): Likewise.
5031 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5032 (aarch64_fma4_elt_to_128v2df): Likewise.
5033 (aarch64_fma4_elt_to_64df): Likewise.
5034 (fnma<mode>4): Likewise.
5035 (aarch64_fnma4_elt<mode>): Likewise.
5036 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5037 (aarch64_fnma4_elt_to_128v2df): Likewise.
5038 (aarch64_fnma4_elt_to_64df): Likewise.
5039 * config/aarch64/iterators.md (VDQSF): New.
5040 * config/aarch64/arm_neon.h
5041 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5042 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5044 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5046 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5047 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5048 (aarch64_mul3_elt_to_128df): Likewise.
5049 (aarch64_mul3_elt_to_64v2df): Likewise.
5050 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5052 (VMUL_CHANGE_NLANES) Likewise.
5055 * config/aarch64/arm_neon.h
5056 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5058 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5060 * config/aarch64/arm_neon.h
5061 (vcvtx_high_f32_f64): Fix parameters.
5063 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5064 Uros Bizjak <ubizjak@gmail.com>
5066 * config/alpha.c: Include tree-ssanames.h.
5068 2013-09-16 Richard Biener <rguenther@suse.de>
5070 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5071 (dot_rdg_1): Handle control_dd.
5072 (create_edge_for_control_dependence): New function.
5073 (create_rdg_edges): Add control dependences if asked for.
5074 (build_rdg): Likewise.
5075 (generate_loops_for_partition): If there are not necessary
5076 control stmts remove all their dependencies.
5077 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5078 (distribute_loop): Pass on control dependences.
5079 (tree_loop_distribution): Compute control dependences and remove
5080 restriction on number of loop nodes.
5082 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5084 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5087 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5089 * cse.c (try_const_anchors): Punt on CC modes.
5091 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5093 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5095 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5098 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5101 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5104 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5106 (darwin_objc1_section): Likewise.
5107 (darwin_file_end): Emit Image Info section when required.
5109 2013-09-14 Jan Hubicka <jh@suse.cz>
5111 * tree-into-ssa.c (gate_into_ssa): New.
5112 (pass_data_build_ssa): Use it.
5113 * cgraph.h (expand_thunk): Update prototype.
5114 * cgraphunit.c (analyze_function): Expand thunks early.
5115 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5116 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5117 set lowered flag; do not add new function.
5118 (assemble_thunks_and_aliases): Update.
5119 * tree-ssa.c (gate_init_datastructures): New gate.
5120 (pass_data_init_datastructures): Use it.
5122 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5125 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5126 xmm register set described in the psABI.
5128 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5130 * dwarf2out.c (should_emit_struct_debug): Add check
5131 for type_decl variable is not NULL.
5133 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5135 * config.gcc: Use new winnt-c.c target hooks
5136 * config/t-winnt: New file
5137 * config/winnt-c.c: New file
5138 * doc/tm.texi.in: Document new hook
5139 * doc/tm.texi: Regenerated
5141 2013-09-13 Jan Hubicka <jh@suse.cz>
5144 * ipa-inline.c (check_callers): New function.
5145 (check_caller_edge): Remove.
5146 (want_inline_function_to_all_callers_p): Also permit alises that are
5148 (inline_to_all_callers): Terminate the walk when devirtualization
5149 introduce new calls.
5151 2013-09-13 Jan Hubicka <jh@suse.cz>
5153 * ipa-inline-analysis.c (struct growth_data): Add node.
5154 (do_estimate_growth_1): Fix detection of recursion.
5156 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5158 PR tree-optimization/58392
5159 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5160 to avoid shadowing of outer loop variable. If
5161 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5162 replace_by_duplicate_decl simduid of loops that have it set and
5163 set dest_cfun->has_simduid_loops and/or
5164 dest_cfun->has_force_vect_loops.
5165 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5166 instead of maybe_lookup_decl.
5167 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5168 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5169 call. Copy over force_vect and copy and remap simduid. Set
5170 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5171 (copy_cfg_body): Remove blocks_to_copy argument. Use
5172 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5173 caller. Don't set cfun->has_simduid_loops and/or
5174 cfun->has_force_vect_loops here.
5175 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5177 (expand_call_inline, tree_function_versioning): Adjust copy_body
5180 2013-09-13 Martin Jambor <mjambor@suse.cz>
5183 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5184 the assert if the edge was a speculative one.
5186 2013-09-13 Richard Biener <rguenther@suse.de>
5188 * tree-data-ref.h (known_dependences_p): Move here ...
5189 * tree-loop-distribution.c (known_dependences_p): ... from here.
5190 (dump_rdg_component, debug_rdg_component): Remove.
5192 (generate_loops_for_partition): Use gimple_uid instead of
5193 relying on matching stmt visit order.
5194 (rdg_build_partitions): Take starting stmt vector.
5195 (ldist_gen): Merge into ...
5196 (distribute_loop): ... this function. Do not compute starting
5198 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5200 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5202 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5203 Set type attribute correctly. Set predicable_short_it attribute.
5204 (cmpsi_shiftsi): Remove %? from output template.
5206 2013-09-13 Richard Biener <rguenther@suse.de>
5208 * tree-loop-distribution.c (struct rdg_component,
5209 rdg_defs_used_in_other_loops_p, free_rdg_components,
5210 rdg_build_components): Remove.
5211 (stmts_from_loop): Do not record virtual PHIs.
5212 (generate_loops_for_partition): Skip virtual PHIs.
5213 (build_rdg_partition_for_component): Rename to ...
5214 (build_rdg_partition_for_vertex): ... this and adjust.
5215 (rdg_build_partitions): Take a vector of starting vertices
5216 instead of components. Remove unnecessary leftover handling.
5217 (ldist_gen): Do not build components or record other stores.
5218 (distribute_loop): Do not distribute loops containing stmts
5221 2013-09-13 Christian Bruel <christian.bruel@st.com>
5224 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5226 2013-09-13 Kai Tietz <ktietz@redhat.com>
5228 * config.gcc: Separate cases for mingw and cygwin targets,
5229 and add 64-bit cygwin target case.
5231 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5232 dll-export inline-functions.
5233 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5236 2013-09-13 Jeff Law <law@redhat.com>
5240 2013-09-06 Jeff Law <law@redhat.com>
5242 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5243 edge implied equivalences into successor phis.
5245 2013-09-12 DJ Delorie <dj@redhat.com>
5247 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5250 2013-09-12 Brooks Moses <bmoses@google.com>
5253 * Makefile.in: Do not install driver binaries in $(target)/bin.
5255 2013-09-12 DJ Delorie <dj@redhat.com>
5257 * config/rl78/rl78.opt (mrelax): New.
5258 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5259 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5261 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5263 (rl78_expand_epilogue): Likewise.
5265 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5268 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5269 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5270 pointer elimination with using remove_reg_equal_offset_note.
5272 2013-09-12 DJ Delorie <dj@redhat.com>
5274 * config/msp430/: New port.
5275 * config.gcc (msp430): Added.
5276 * doc/invoke.texi: Document MSP430 options.
5277 * doc/install.texi: Document msp430-elf
5278 * doc/md.texi: Document msp430-elf
5279 * doc/contrib.texi: Document msp430-elf
5281 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5284 2013-09-12 Martin Jambor <mjambor@suse.cz>
5287 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5288 reference descriptor is NULL.
5289 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5292 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5294 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5295 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5296 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5297 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5298 make_temp_ssa_name): move to tree-ssanames.h
5299 * tree-ssa-alias.h: Move prototype.
5300 * tree-ssa.h: Include tree-ssanames.h.
5301 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5302 * tree-ssanames.h: New. Move items from tree-flow*.h
5303 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5305 2013-09-12 Richard Biener <rguenther@suse.de>
5307 PR tree-optimization/58404
5308 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5309 propagate non-invariant addresses into dereferences wrapped
5310 in component references.
5312 2013-09-12 Richard Biener <rguenther@suse.de>
5314 PR tree-optimization/58402
5315 * passes.def: Move pass_late_warn_uninitialized later.
5317 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5319 * tree-ssa.h: New. Move content from tree-flow.h and
5321 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5322 Move prototypes belonging to tree-ssa.c.
5323 * tree-flow-inline.h (redirect_edge_var_map_def,
5324 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5326 * gimple.h: Adjust prototypes.
5327 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5329 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5330 * tree.h: Move prototype to tree-ssa.h.
5331 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5332 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5333 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5334 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5335 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5336 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5337 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5338 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5339 graphite-blocking.c, graphite-clast-to-gimple.c,
5340 graphite-dependences.c, graphite-interchange.c,
5341 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5342 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5343 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5344 ipa-reference.c, ipa-split.c, ipa-utils.c,
5345 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5346 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5347 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5348 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5349 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5350 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5351 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5352 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5353 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5354 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5355 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5356 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5357 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5358 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5359 value-prof.c, var-tracking.c,
5360 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5362 2013-09-12 Richard Biener <rguenther@suse.de>
5364 PR tree-optimization/58396
5365 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5366 (build_rdg): Take a loop-nest parameter, fix memleaks.
5367 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5369 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5371 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5374 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5376 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5377 The FMA instruction names should have a 'v' prefix.
5379 2013-09-12 Richard Biener <rguenther@suse.de>
5381 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5382 (dot_rdg): Use popen instead of system in optional code.
5383 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5384 (already_processed_vertex_p): Adjust.
5385 (has_anti_or_output_dependence, predecessor_has_mem_write,
5386 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5387 rdg_flag_uses): Remove.
5388 (rdg_flag_vertex): Simplify.
5389 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5391 (build_rdg_partition_for_component): Process the first vertex
5392 of a component only.
5393 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5395 2013-09-12 Alan Modra <amodra@gmail.com>
5397 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5399 2013-09-11 DJ Delorie <dj@redhat.com>
5400 Nick Clifton <nickc@redhat.com>
5402 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5403 (rl78_stack_based_mem): New.
5404 * config/rl78/constraints.md (Iv08): New.
5413 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5415 (movsi): Change from expand to insn-and-split.
5416 (ashrsi3): Clobber AX.
5420 * config/rl78/rl78.md (CC_REG): Fix.
5421 (addsi3): Allow memory and immediate operands.
5422 (addsi3_internal): Split into...
5423 (addsi3_internal_virt): ...new, and ...
5424 (addsi3_internal_real): ...new.
5426 (subsi3_internal_virt): New.
5427 (subsi3_internal_real): New.
5428 (mulsi3): Add memory operand.
5429 (mulsi3_rl78): Likewise.
5430 (mulsi3_g13): Likewise.
5431 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5432 (cbranchqi4_real): Add more constraint options.
5433 (cbranchhi4_real): Expand pattern.
5434 (cbranchhi4_real_signed): New.
5435 (cbranchhi4_real_inverted): New.
5436 (cbranchsi4_real_lt): New.
5437 (cbranchsi4_real_ge): New.
5438 (cbranchsi4_real_signed): New.
5439 (cbranchsi4_real): New.
5441 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5443 (lshrsi3_virt): Likewise.
5444 (ashlsi3_virt): Likewise.
5445 (cbranchqi4_virt_signed): New.
5446 (cbranchhi4_virt_signed): New.
5447 (cbranchsi4_virt): New.
5448 * config/rl78/rl78.c: Whitespace fixes throughout.
5449 (move_elim_pass): New.
5450 (pass_data_rl78_move_elim): New.
5451 (pass_rl78_move_elim): New.
5452 (make_pass_rl78_move_elim): New.
5453 (rl78_devirt_info): Run devirt earlier.
5454 (rl78_move_elim_info): New.
5455 (rl78_asm_file_start): Register it.
5456 (rl78_split_movsi): New.
5457 (rl78_as_legitimate_address): Allow virtual base registers when
5459 (rl78_addr_space_convert): Remove spurious debug stuff.
5460 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5461 (rl78_print_operand): More cases for not printing '#'.
5462 (rl78_expand_compare): Remove most of the logic.
5463 (content_memory): New.
5464 (clear_content_memory): New.
5465 (get_content_index): New.
5466 (get_content_name): New.
5467 (display_content_memory): New.
5468 (update_content): New.
5469 (record_content): New.
5470 (already_contains): New.
5471 (insn_ok_now): Re-recog insns with virtual registers.
5472 (add_postponed_content_update): New.
5473 (process_postponed_content_update): New.
5474 (gen_and_emit_move): New.
5475 (transcode_memory_rtx): Record new location content.
5476 Use gen_and_emit_move.
5477 (force_into_acc): New.
5478 (move_to_acc): Use gen_and_emit_move.
5479 (move_from_acc): Likewise.
5480 (move_acc_to_reg): Likewise.
5481 (move_to_x): Likewise.
5482 (move_to_hl): Likewise.
5483 (move_to_de): Likewise.
5484 (rl78_alloc_physical_registers_op1): Record location content.
5485 (has_constraint): New.
5486 (rl78_alloc_physical_registers_op2): Record location content.
5488 (rl78_alloc_physical_registers_ro1): Likewise.
5489 (rl78_alloc_physical_registers_cmp): Likewise.
5490 (rl78_alloc_physical_registers_umul): Likewise.
5491 (rl78_alloc_address_registers_macax): New.
5492 (rl78_alloc_physical_registers): Initialize and set location
5493 content memory as needed.
5494 (rl78_reorg): Make sure split2 is called.
5495 (rl78_rtx_costs): New.
5497 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5499 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5500 for (not (neg ...)) and (neg (not ...)) cases.
5502 2013-09-11 Richard Biener <rguenther@suse.de>
5505 * passes.def: Split critical edges before late uninit warning passes.
5506 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5508 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5510 PR tree-optimization/58385
5511 * fold-const.c (build_range_check): If both low and high are NULL,
5512 use omit_one_operand_loc to preserve exp side-effects.
5514 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5516 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5518 2013-09-11 Richard Biener <rguenther@suse.de>
5520 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5521 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5522 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5523 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5524 stmts_from_loop, known_dependences_p, build_empty_rdg,
5525 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5526 * tree-loop-distribution.c: ... here.
5527 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5528 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5529 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5530 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5531 * tree-loop-distribution.c: ... here.
5532 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5533 (struct partition_s): Add loops member.
5534 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5535 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5536 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5538 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5539 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5540 Sergey Lega <sergey.s.lega@intel.com>
5541 Anna Tikhonova <anna.tikhonova@intel.com>
5542 Ilya Tocar <ilya.tocar@intel.com>
5543 Andrey Turetskiy <andrey.turetskiy@intel.com>
5544 Ilya Verbin <ilya.verbin@intel.com>
5545 Kirill Yukhin <kirill.yukhin@intel.com>
5546 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5548 * config/i386/constraints.md (k): New.
5550 * config/i386/i386.c (const regclass_map): Add new mask registers.
5551 (dbx_register_map): Ditto.
5552 (dbx64_register_map): Ditto.
5553 (svr4_dbx_register_map): Ditto.
5554 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5556 (ix86_preferred_reload_class): Disable constants for mask registers.
5557 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5558 (ix86_hard_regno_mode_ok): Support new mask registers.
5559 (x86_order_regs_for_local_alloc): Ditto.
5560 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5561 (FIXED_REGISTERS): Add new mask registers.
5562 (CALL_USED_REGISTERS): Ditto.
5563 (REG_ALLOC_ORDER): Ditto.
5564 (VALID_MASK_REG_MODE): New.
5565 (FIRST_MASK_REG): Ditto.
5566 (LAST_MASK_REG): Ditto.
5567 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5568 (MAYBE_MASK_CLASS_P): New.
5569 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5570 (REG_CLASS_CONTENTS): Ditto.
5571 (MASK_REGNO_P): New.
5572 (ANY_MASK_REG_P): Ditto.
5573 (HI_REGISTER_NAMES): Add new mask registers.
5574 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5575 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5577 (attribute "type"): Add mskmov, msklog.
5578 (attribute "length_immediate"): Support them.
5579 (attribute "memory"): Ditto.
5580 (attribute "prefix_0f"): Ditto.
5581 (*movhi_internal): Support new mask registers.
5582 (*movqi_internal): Ditto.
5583 (define_split): Split out clobber pattern is a logic
5584 insn on mask registers.
5585 (*k<logic><mode>): New.
5586 (*andhi_1): Extend to support mask regs.
5587 (*andqi_1): Extend to support mask regs.
5589 (define_split): Split and-not to and and not if operands
5591 (*<code><mode>_1): Separate HI mode to new pattern...
5592 (*<code>hi_1): This.
5593 (*<code>qi_1): Extend to support mask regs.
5595 (kortestzhi): Ditto.
5596 (kortestchi): Ditto.
5598 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5599 (*one_cmplhi2_1): ...Here, now with mask registers support.
5600 (*one_cmplqi2_1): Support new mask registers.
5601 (HI/QImode arithmetics splitter): Don't split if mask registers
5603 (HI/QImode not splitter): Ditto.
5604 * config/i386/predicated.md (mask_reg_operand): New.
5605 (general_reg_operand): Ditto.
5607 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5609 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5610 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5612 2013-09-10 Jeff Law <law@redhat.com>
5614 PR tree-optimization/58380
5615 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5616 of threading through a buried loop header.
5618 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5619 return value for single successor case.
5621 2013-09-10 Jan Hubicka <jh@suse.cz>
5623 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5625 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5628 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5629 support conditional execution.
5630 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5632 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5634 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5636 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5637 Check destination too to check move insn.
5638 (undo_optional_reloads): Add check that the original peudo did not
5639 changed its allocation and the optional reload was inherited on last
5640 inheritance pass. Break loop after deciding to keep optional reload.
5641 (lra_undo_inheritance): Add check that inherited pseudo still in
5644 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5646 * config/aarch64/aarch64.md (generic_sched): New.
5647 * config/aarch64/aarch64-generic.md (load): Make conditional
5648 on generic_sched attribute.
5649 (nonload): Likewise.
5651 2013-09-10 Jan Hubicka <jh@suse.cz>
5653 * lto-cgraph.c: Include ipa-utils.h.
5654 (compute_ltrans_boundary): Also add possible targets into the boundary.
5656 2013-09-10 Jan Hubicka <jh@suse.cz>
5658 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5659 VAR_DECL of vtable rather than full expression.
5661 2013-09-10 Jan Hubicka <jh@suse.cz>
5662 Paolo Carlini <paolo.carlini@oracle.com>
5664 * cgraphunit.c (analyze_functions): Save input_location, set it
5665 to UNKNOWN_LOCATION and restore it at the end.
5667 2013-09-10 Martin Jambor <mjambor@suse.cz>
5669 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5670 represented by a thunk.
5672 2013-09-10 Jeff Law <law@redhat.com>
5674 PR tree-optimization/58343
5675 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5676 jump threading requests through loop headers buried in the middle
5677 of a jump threading path.
5679 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5680 in return value/type.
5682 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5684 PR rtl-optimization/58365
5685 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5686 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5689 2013-09-10 Richard Biener <rguenther@suse.de>
5691 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5692 * tree-data-ref.c (create_rdg_vertices): Collect all data
5693 references, signal failure to the caller, use data-ref API.
5694 (build_rdg): Compute data references only once. Maintain lifetime
5695 of data references and data dependences from within RDG.
5696 (free_rdg): Free dependence relations.
5697 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5698 inventing extra dependences.
5699 (distribute_loop): Update for RDG API changes.
5701 2013-09-10 Kai Tietz <ktietz@redhat.com>
5703 * doc/invoke.texi (fms-extensions): Document changed
5704 behavior for ms-abi targets.
5705 * config/i386/i386.c (ix86_option_override_internal):
5706 Set default value of option -fms-extension for ms-abi targets.
5708 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5710 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5712 2013-09-10 Alan Modra <amodra@gmail.com>
5715 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5717 2013-09-10 Alan Modra <amodra@gmail.com>
5719 * config/rs6000/predicates.md (add_cint_operand): New.
5720 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5721 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5722 using add_cint_operand.
5723 (largetoc_high_plus_aix): Likewise.
5725 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5727 PR tree-optimization/58364
5728 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5729 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5730 the current range can't be an unconditional true or false.
5732 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5734 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5736 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5738 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5740 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5743 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5745 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5747 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5748 NO_REGS for immediate that can't be moved directly into FP_REGS.
5750 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5752 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5753 comparison with negated operand.
5754 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5757 2013-09-09 Richard Biener <rguenther@suse.de>
5760 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5761 of a subloop record all its block as affecting loop-closed SSA form.
5763 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5765 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5766 of an rtx/bitpos pair.
5767 (store_fixed_bit_field): Update accordingly.
5769 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5771 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5773 * builtins.c (expand_errno_check): Likewise.
5774 * dwarf2cfi.c (init_return_column_size): Likewise.
5775 * except.c (sjlj_mark_call_sites): Likewise.
5776 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5777 * lra-constraints.c (emit_inc): Likewise.
5778 * ree.c (combine_set_extension): Likewise.
5779 * regmove.c (fixup_match_2): Likewise.
5780 * reload1.c (inc_for_reload): Likewise.
5782 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5784 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5785 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5786 Use gen_int_mode with the mode of the associated simplify_* call.
5787 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5788 * expmed.c (expand_shift_1): Likewise.
5789 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5790 * loop-iv.c (iv_number_of_iterations): Likewise.
5791 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5792 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5794 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5796 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5797 of the associated expand_* call.
5798 (asan_emit_stack_protection): Likewise.
5799 * builtins.c (round_trampoline_addr): Likewise.
5800 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5801 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5802 (emit_store_flag): Likewise.
5803 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5804 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5806 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5807 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5808 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5809 * modulo-sched.c (generate_prolog_epilog): Likewise.
5810 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5811 (expand_ctz, expand_ffs, expand_unop): Likewise.
5813 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5815 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5816 of the associated gen_rtx_* call.
5817 * caller-save.c (init_caller_save): Likewise.
5818 * combine.c (find_split_point, make_extraction): Likewise.
5819 (make_compound_operation): Likewise.
5820 * dwarf2out.c (mem_loc_descriptor): Likewise.
5821 * explow.c (plus_constant, probe_stack_range): Likewise.
5822 * expmed.c (expand_mult_const): Likewise.
5823 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5824 * reload1.c (init_reload): Likewise.
5825 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5826 * var-tracking.c (adjust_mems): Likewise.
5827 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5828 rather than gen_rtx_fmt_ee.
5830 2013-09-09 Jan Hubicka <jh@suse.cz>
5833 * value-prof.c (gimple_ic): Copy also abnormal edges.
5835 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5837 * asan.c (asan_shadow_cst): Use gen_int_mode.
5839 2013-09-08 Jan Hubicka <jh@suse.cz>
5841 * ipa-profile.c: Add toplevel comment.
5842 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5843 (contains_hot_call_p): New function.
5844 (ipa_propagate_frequency): Set frequencies based on counts when
5846 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5847 profile; do not tamper with profile after inlining if it is read.
5849 2013-09-08 Jan Hubicka <jh@suse.cz>
5851 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5854 2013-09-08 Jan Hubicka <jh@suse.cz>
5856 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5859 2013-09-08 Jeff Law <law@redhat.com>
5862 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5865 2013-09-08 Andi Kleen <ak@linux.intel.com>
5867 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5869 2013-09-08 Jan Hubicka <jh@suse.cz>
5871 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5873 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5875 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5876 for non-debug insns.
5877 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5878 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5879 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5882 2013-09-08 Jan Hubicka <jh@suse.cz>
5884 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5885 targets and devirtualize to BUILT_IN_UNREACHABLE.
5886 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5887 * ipa.c (walk_polymorphic_call_targets): New function.
5888 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5889 functions; use the new timevar.
5890 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5891 was removed from the program.
5892 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5893 not consider it in the walk when its vtable is dead.
5894 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5896 (devirt_variable_node_removal_hook): New function.
5897 (possible_polymorphic_call_targets): Also register
5898 devirt_variable_node_removal_hook.
5899 (ipa_devirt): Do not do non-speculative devirtualization.
5900 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5902 2013-09-08 Jan Hubicka <jh@suse.cz>
5904 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5905 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5906 varpool_remove_variable_insertion_hook): Declare.
5907 * varpool.c (varpool_node_hook_list): New structure.
5908 (first_varpool_node_removal_hook,
5909 first_varpool_variable_insertion_hook): New variables.
5910 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5911 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5912 varpool_remove_variable_insertion_hook,
5913 varpool_call_variable_insertion_hooks): New functions.
5914 (varpool_remove_node): Use it.
5916 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5919 * diagnostic.c (diagnostic_build_prefix): When s.file is
5920 "<built-in>" don't output line and column numbers.
5922 2013-09-06 Jan Hubicka <jh@suse.cz>
5924 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5925 * lto-streamer-out.c: Stream thunks, too.
5926 * lto-streamer-in.c (input_function): Pop cfun here
5927 (lto_read_body): Instead of here.
5929 2013-09-06 Caroline Tice <cmtice@google.com>
5931 * doc/install.texi: Add documentation for the --enable-vtable-verify
5932 and the --disable-libvtv configure options.
5934 2013-09-06 Jeff Law <law@redhat.com>
5936 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5937 edge implied equivalences into successor phis.
5939 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5941 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5943 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5945 * resource.c (mark_target_live_regs): Compute resources taking
5946 into account if a call is predicated or not.
5948 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5950 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5951 the compiler in the function names.
5953 2013-09-06 Jan Hubicka <jh@suse.cz>
5956 * ipa-inline.c (has_caller_p): New function.
5957 (want_inline_function_to_all_callers_p): Use it.
5958 (sum_callers, inline_to_all_callers): Break out from ...
5959 (ipa_inline): ... here.
5961 2013-09-06 Jan Hubicka <jh@suse.cz>
5963 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5964 only when AVX is enabled.
5966 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5968 * config/aarch64/aarch64.md
5969 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5970 is fpsimd_<load/store>2.
5971 (load_pair<mode>): Likewise.
5972 (store_pair<mode>): Likewise.
5974 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5976 * config/arm/types.md (type): Add "mrs" type.
5977 * config/aarch64/aarch64.md
5978 (aarch64_load_tp_hard): Make type "mrs".
5980 (load_tp_hard): Make type "mrs".
5981 * config/arm/cortex-a15.md: Update with new attributes.
5982 * config/arm/cortex-a5.md: Update with new attributes.
5983 * config/arm/cortex-a53.md: Update with new attributes.
5984 * config/arm/cortex-a7.md: Update with new attributes.
5985 * config/arm/cortex-a8.md: Update with new attributes.
5986 * config/arm/cortex-a9.md: Update with new attributes.
5987 * config/arm/cortex-m4.md: Update with new attributes.
5988 * config/arm/cortex-r4.md: Update with new attributes.
5989 * config/arm/fa526.md: Update with new attributes.
5990 * config/arm/fa606te.md: Update with new attributes.
5991 * config/arm/fa626te.md: Update with new attributes.
5992 * config/arm/fa726te.md: Update with new attributes.
5994 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5996 * config/aarch64/aarch64.md
5997 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5998 (*movtf_aarch64): Likewise.
6000 (thumb1_movdi_insn): Use "multiple" for type where more than one
6001 instruction is used for a move.
6002 (*arm32_movhf): Likewise.
6003 (*thumb_movdf_insn): Likewise.
6005 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6007 * config/arm/types.md (type): Rename fcpys to fmov.
6009 (*arm_movsi_vfp): Rename type fcpys as fmov.
6010 (*thumb2_movsi_vfp): Likewise
6011 (*movhf_vfp_neon): Likewise
6012 (*movhf_vfp): Likewise
6013 (*movsf_vfp): Likewise
6014 (*thumb2_movsf_vfp): Likewise
6015 (*movsfcc_vfp): Likewise
6016 (*thumb2_movsfcc_vfp): Likewise
6017 * config/aarch64/aarch64-simd.md
6018 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6019 * config/aarch64/aarch64.md
6020 (*movsi_aarch64): Replace type mov_reg with fmovs.
6021 (*movdi_aarch64): Likewise
6022 (*movsf_aarch64): Likewise
6023 (*movdf_aarch64): Likewise
6025 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6026 * config/arm/iwmmxt.md
6027 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6028 * config/arm/arm1020e.md: Update with new attributes.
6029 * config/arm/cortex-a15-neon.md: Update with new attributes.
6030 * config/arm/cortex-a5.md: Update with new attributes.
6031 * config/arm/cortex-a53.md: Update with new attributes.
6032 * config/arm/cortex-a7.md: Update with new attributes.
6033 * config/arm/cortex-a8-neon.md: Update with new attributes.
6034 * config/arm/cortex-a9.md: Update with new attributes.
6035 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6036 * config/arm/cortex-r4f.md: Update with new attributes.
6037 * config/arm/marvell-pj4.md: Update with new attributes.
6038 * config/arm/vfp11.md: Update with new attributes.
6040 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6042 * config/aarch64/aarch64.md
6043 (*madd<mode>): Fix type attribute.
6044 (*maddsi_uxtw): Likewise.
6045 (*msub<mode>): Likewise.
6046 (*msubsi_uxtw): Likewise.
6047 (<su_optab>maddsidi4): Likewise.
6048 (<su_optab>msubsidi4): Likewise.
6050 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6052 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6053 * config/arm/arm.md (core_cycles): Remove fdiv.
6054 * config/arm/vfp.md:
6055 (*sqrtsf2_vfp): Update for attribute changes.
6056 (*sqrtdf2_vfp): Likewise.
6057 * config/aarch64/aarch64.md:
6058 (sqrt<mode>2): Update for attribute changes.
6059 * config/arm/arm1020e.md: Update with new attributes.
6060 * config/arm/cortex-a15-neon.md: Update with new attributes.
6061 * config/arm/cortex-a5.md: Update with new attributes.
6062 * config/arm/cortex-a53.md: Update with new attributes.
6063 * config/arm/cortex-a7.md: Update with new attributes.
6064 * config/arm/cortex-a8-neon.md: Update with new attributes.
6065 * config/arm/cortex-a9.md: Update with new attributes.
6066 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6067 * config/arm/cortex-r4f.md: Update with new attributes.
6068 * config/arm/marvell-pj4.md: Update with new attributes.
6069 * config/arm/vfp11.md: Update with new attributes.
6071 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6073 * config/arm/types.md
6074 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6075 * config/aarch64/aarch64.md
6076 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6078 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6079 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6080 (float<GPI:mode><GPF:mode>2): Likewise.
6082 (*truncsisf2_vfp): Update with new attributes.
6083 (*truncsidf2_vfp): Likewise.
6084 (fixuns_truncsfsi2): Likewise.
6085 (fixuns_truncdfsi2): Likewise.
6086 (*floatsisf2_vfp): Likewise.
6087 (*floatsidf2_vfp): Likewise.
6088 (floatunssisf2): Likewise.
6089 (floatunssidf2): Likewise.
6090 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6091 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6092 * config/arm/arm1020e.md: Update with new attributes.
6093 * config/arm/cortex-a15-neon.md: Update with new attributes.
6094 * config/arm/cortex-a5.md: Update with new attributes.
6095 * config/arm/cortex-a53.md: Update with new attributes.
6096 * config/arm/cortex-a7.md: Update with new attributes.
6097 * config/arm/cortex-a8-neon.md: Update with new attributes.
6098 * config/arm/cortex-a9.md: Update with new attributes.
6099 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6100 * config/arm/cortex-r4f.md: Update with new attributes.
6101 * config/arm/marvell-pj4.md: Update with new attributes.
6102 * config/arm/vfp11.md: Update with new attributes.
6104 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6106 * config/aarch64/arm_neon.h
6107 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6108 (vqtbx<1,2,3,4><q>_s8): Likewise.
6110 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6112 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6113 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6114 (add<mode>3): Add type attribute.
6115 (add<mode>3): Likewise.
6116 (usadd<mode>3): Likewise.
6117 (ssadd<mode>3): Likewise.
6118 (sub<mode>3): Likewise.
6119 (sub<mode>3): Likewise.
6120 (ussub<mode>3): Likewise.
6121 (sssub<mode>3): Likewise.
6122 (ssmulsa3): Likewise.
6123 (usmulusa3): Likewise.
6124 (arm_usatsihi): Likewise.
6126 (*movdi_vfp): Add types for all instructions.
6127 (*movdi_vfp_cortexa8): Likewise.
6128 (*movhf_vfp_neon): Likewise.
6129 (*movhf_vfp): Likewise.
6130 (*movdf_vfp): Likewise.
6131 (*thumb2_movdf_vfp): Likewise.
6132 (*thumb2_movdfcc_vfp): Likewise.
6133 * config/arm/arm.md: Add type attribute to all insn patterns.
6134 (*thumb1_adddi3): Add type attribute.
6135 (*arm_adddi3): Likewise.
6136 (*adddi_sesidi_di): Likewise.
6137 (*adddi_zesidi_di): Likewise.
6138 (*thumb1_addsi3): Likewise.
6139 (addsi3_compare0): Likewise.
6140 (*addsi3_compare0_scratch): Likewise.
6141 (*compare_negsi_si): Likewise.
6142 (cmpsi2_addneg): Likewise.
6143 (*addsi3_carryin_<optab>): Likewise.
6144 (*addsi3_carryin_alt2_<optab>): Likewise.
6145 (*addsi3_carryin_clobercc_<optab>): Likewise.
6146 (*subsi3_carryin): Likewise.
6147 (*subsi3_carryin_const): Likewise.
6148 (*subsi3_carryin_compare): Likewise.
6149 (*subsi3_carryin_compare_const): Likewise.
6150 (*arm_subdi3): Likewise.
6151 (*thumb_subdi3): Likewise.
6152 (*subdi_di_zesidi): Likewise.
6153 (*subdi_di_sesidi): Likewise.
6154 (*subdi_zesidi_di): Likewise.
6155 (*subdi_sesidi_di): Likewise.
6156 (*subdi_zesidi_ze): Likewise.
6157 (thumb1_subsi3_insn): Likewise.
6158 (*arm_subsi3_insn): Likewise.
6159 (*anddi3_insn): Likewise.
6160 (*anddi_zesidi_di): Likewise.
6161 (*anddi_sesdi_di): Likewise.
6162 (*ne_zeroextracts): Likewise.
6163 (*ne_zeroextracts): Likewise.
6164 (*ite_ne_zeroextr): Likewise.
6165 (*ite_ne_zeroextr): Likewise.
6166 (*anddi_notdi_di): Likewise.
6167 (*anddi_notzesidi): Likewise.
6168 (*anddi_notsesidi): Likewise.
6169 (andsi_notsi_si): Likewise.
6170 (thumb1_bicsi3): Likewise.
6171 (*iordi3_insn): Likewise.
6172 (*iordi_zesidi_di): Likewise.
6173 (*iordi_sesidi_di): Likewise.
6174 (*thumb1_iorsi3_insn): Likewise.
6175 (*xordi3_insn): Likewise.
6176 (*xordi_zesidi_di): Likewise.
6177 (*xordi_sesidi_di): Likewise.
6178 (*arm_xorsi3): Likewise.
6179 (*andsi_iorsi3_no): Likewise.
6180 (*smax_0): Likewise.
6181 (*smax_m1): Likewise.
6182 (*arm_smax_insn): Likewise.
6183 (*smin_0): Likewise.
6184 (*arm_smin_insn): Likewise.
6185 (*arm_umaxsi3): Likewise.
6186 (*arm_uminsi3): Likewise.
6187 (*minmax_arithsi): Likewise.
6188 (*minmax_arithsi_): Likewise.
6189 (*satsi_<SAT:code>): Likewise.
6190 (arm_ashldi3_1bit): Likewise.
6191 (arm_ashrdi3_1bit): Likewise.
6192 (arm_lshrdi3_1bit): Likewise.
6193 (*arm_negdi2): Likewise.
6194 (*thumb1_negdi2): Likewise.
6195 (*arm_negsi2): Likewise.
6196 (*thumb1_negsi2): Likewise.
6197 (*negdi_extendsid): Likewise.
6198 (*negdi_zero_extend): Likewise.
6199 (*arm_abssi2): Likewise.
6200 (*thumb1_abssi2): Likewise.
6201 (*arm_neg_abssi2): Likewise.
6202 (*thumb1_neg_abss): Likewise.
6203 (one_cmpldi2): Likewise.
6204 (extend<mode>di2): Likewise.
6205 (*compareqi_eq0): Likewise.
6206 (*arm_extendhisi2addsi): Likewise.
6207 (*arm_movdi): Likewise.
6208 (*thumb1_movdi_insn): Likewise.
6209 (*arm_movt): Likewise.
6210 (*thumb1_movsi_insn): Likewise.
6211 (pic_add_dot_plus_four): Likewise.
6212 (pic_add_dot_plus_eight): Likewise.
6213 (tls_load_dot_plus_eight): Likewise.
6214 (*thumb1_movhi_insn): Likewise.
6215 (*thumb1_movsf_insn): Likewise.
6216 (*movdf_soft_insn): Likewise.
6217 (*thumb_movdf_insn): Likewise.
6218 (cbranchsi4_insn): Likewise.
6219 (cbranchsi4_scratch): Likewise.
6220 (*negated_cbranchsi4): Likewise.
6221 (*tbit_cbranch): Likewise.
6222 (*tlobits_cbranch): Likewise.
6223 (*tstsi3_cbranch): Likewise.
6224 (*cbranchne_decr1): Likewise.
6225 (*addsi3_cbranch): Likewise.
6226 (*addsi3_cbranch_scratch): Likewise.
6227 (*arm_cmpdi_insn): Likewise.
6228 (*arm_cmpdi_unsig): Likewise.
6229 (*arm_cmpdi_zero): Likewise.
6230 (*thumb_cmpdi_zero): Likewise.
6231 (*deleted_compare): Likewise.
6232 (*mov_scc): Likewise.
6233 (*mov_negscc): Likewise.
6234 (*mov_notscc): Likewise.
6235 (*cstoresi_eq0_thumb1_insn): Likewise.
6236 (cstoresi_nltu_thumb1): Likewise.
6237 (cstoresi_ltu_thu): Likewise.
6238 (thumb1_addsi3_addgeu): Likewise.
6239 (*arm_jump): Likewise.
6240 (*thumb_jump): Likewise.
6241 (*check_arch2): Likewise.
6242 (arm_casesi_internal): Likewise.
6243 (thumb1_casesi_dispatch): Likewise.
6244 (*arm_indirect_jump): Likewise.
6245 (*thumb1_indirect_jump): Likewise.
6247 (*and_scc): Likewise.
6248 (*ior_scc): Likewise.
6249 (*compare_scc): Likewise.
6250 (*cond_move): Likewise.
6251 (*cond_arith): Likewise.
6252 (*cond_sub): Likewise.
6253 (*cmp_ite0): Likewise.
6254 (*cmp_ite1): Likewise.
6255 (*cmp_and): Likewise.
6256 (*cmp_ior): Likewise.
6257 (*ior_scc_scc): Likewise.
6258 (*ior_scc_scc_cmp): Likewise.
6259 (*and_scc_scc): Likewise.
6260 (*and_scc_scc_cmp): Likewise.
6261 (*and_scc_scc_nod): Likewise.
6262 (*negscc): Likewise.
6263 (movcond_addsi): Likewise.
6264 (movcond): Likewise.
6265 (*ifcompare_plus_move): Likewise.
6266 (*if_plus_move): Likewise.
6267 (*ifcompare_move_plus): Likewise.
6268 (*if_move_plus): Likewise.
6269 (*ifcompare_arith_arith): Likewise.
6270 (*if_arith_arith): Likewise.
6271 (*ifcompare_arith_move): Likewise.
6272 (*if_arith_move): Likewise.
6273 (*ifcompare_move_arith): Likewise.
6274 (*if_move_arith): Likewise.
6275 (*ifcompare_move_not): Likewise.
6276 (*if_move_not): Likewise.
6277 (*ifcompare_not_move): Likewise.
6278 (*if_not_move): Likewise.
6279 (*ifcompare_shift_move): Likewise.
6280 (*if_shift_move): Likewise.
6281 (*ifcompare_move_shift): Likewise.
6282 (*if_move_shift): Likewise.
6283 (*ifcompare_shift_shift): Likewise.
6284 (*ifcompare_not_arith): Likewise.
6285 (*ifcompare_arith_not): Likewise.
6286 (*if_arith_not): Likewise.
6287 (*ifcompare_neg_move): Likewise.
6288 (*if_neg_move): Likewise.
6289 (*ifcompare_move_neg): Likewise.
6290 (*if_move_neg): Likewise.
6291 (prologue_thumb1_interwork): Likewise.
6292 (*cond_move_not): Likewise.
6293 (*sign_extract_onebit): Likewise.
6294 (*not_signextract_onebit): Likewise.
6295 (stack_tie): Likewise.
6296 (align_4): Likewise.
6297 (align_8): Likewise.
6298 (consttable_end): Likewise.
6299 (consttable_1): Likewise.
6300 (consttable_2): Likewise.
6301 (consttable_4): Likewise.
6302 (consttable_8): Likewise.
6303 (consttable_16): Likewise.
6304 (*thumb1_tablejump): Likewise.
6305 (prefetch): Likewise.
6306 (force_register_use): Likewise.
6307 (thumb_eh_return): Likewise.
6308 (load_tp_hard): Likewise.
6309 (load_tp_soft): Likewise.
6310 (tlscall): Likewise.
6311 (*arm_movtas_ze): Likewise.
6312 (*arm_rev): Likewise.
6313 (*arm_revsh): Likewise.
6314 (*arm_rev16): Likewise.
6315 * config/arm/thumb2.md
6316 (*thumb2_smaxsi3): Likewise.
6317 (*thumb2_sminsi3): Likewise.
6318 (*thumb32_umaxsi3): Likewise.
6319 (*thumb2_uminsi3): Likewise.
6320 (*thumb2_negdi2): Likewise.
6321 (*thumb2_abssi2): Likewise.
6322 (*thumb2_neg_abss): Likewise.
6323 (*thumb2_movsi_insn): Likewise.
6324 (tls_load_dot_plus_four): Likewise.
6325 (*thumb2_movhi_insn): Likewise.
6326 (*thumb2_mov_scc): Likewise.
6327 (*thumb2_mov_negs): Likewise.
6328 (*thumb2_mov_negs): Likewise.
6329 (*thumb2_mov_nots): Likewise.
6330 (*thumb2_mov_nots): Likewise.
6331 (*thumb2_movsicc_): Likewise.
6332 (*thumb2_movsfcc_soft_insn): Likewise.
6333 (*thumb2_indirect_jump): Likewise.
6334 (*thumb2_and_scc): Likewise.
6335 (*thumb2_ior_scc): Likewise.
6336 (*thumb2_ior_scc_strict_it): Likewise.
6337 (*thumb2_cond_move): Likewise.
6338 (*thumb2_cond_arith): Likewise.
6339 (*thumb2_cond_ari): Likewise.
6340 (*thumb2_cond_sub): Likewise.
6341 (*thumb2_negscc): Likewise.
6342 (*thumb2_movcond): Likewise.
6343 (thumb2_casesi_internal): Likewise.
6344 (thumb2_casesi_internal_pic): Likewise.
6345 (*thumb2_alusi3_short): Likewise.
6346 (*thumb2_mov<mode>_shortim): Likewise.
6347 (*thumb2_addsi_short): Likewise.
6348 (*thumb2_subsi_short): Likewise.
6349 (thumb2_addsi3_compare0): Likewise.
6350 (*thumb2_cbz): Likewise.
6351 (*thumb2_cbnz): Likewise.
6352 (*thumb2_one_cmplsi2_short): Likewise.
6353 (*thumb2_negsi2_short): Likewise.
6354 (*orsi_notsi_si): Likewise.
6355 * config/arm/arm1020e.md: Update with new attributes.
6356 * config/arm/arm1026ejs.md: Update with new attributes.
6357 * config/arm/arm1136jfs.md: Update with new attributes.
6358 * config/arm/arm926ejs.md: Update with new attributes.
6359 * config/arm/cortex-a15.md: Update with new attributes.
6360 * config/arm/cortex-a5.md: Update with new attributes.
6361 * config/arm/cortex-a53.md: Update with new attributes.
6362 * config/arm/cortex-a7.md: Update with new attributes.
6363 * config/arm/cortex-a8.md: Update with new attributes.
6364 * config/arm/cortex-a9.md: Update with new attributes.
6365 * config/arm/cortex-m4.md: Update with new attributes.
6366 * config/arm/cortex-r4.md: Update with new attributes.
6367 * config/arm/fa526.md: Update with new attributes.
6368 * config/arm/fa606te.md: Update with new attributes.
6369 * config/arm/fa626te.md: Update with new attributes.
6370 * config/arm/fa726te.md: Update with new attributes.
6372 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6374 * config/aarch64/aarch64-simd.md
6375 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6376 <vwx> iterator to ensure correct register choice.
6377 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6378 (aarch64_sqdmull_n<mode>): Likewise.
6379 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6380 * config/aarch64/arm_neon.h
6381 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6382 (vml<as><q>_n_<su>16): Likewise.
6383 (vml<as>l_high_lane<q>_<su>16): Likewise.
6384 (vml<as>l_high_n_<su>16): Likewise.
6385 (vml<as>l_lane<q>_<su>16): Likewise.
6386 (vml<as>l_n_<su>16): Likewise.
6387 (vmul<q>_lane<q>_<su>16): Likewise.
6388 (vmul<q>_n_<su>16): Likewise.
6389 (vmull_lane<q>_<su>16): Likewise.
6390 (vmull_n_<su>16): Likewise.
6391 (vmull_high_lane<q>_<su>16): Likewise.
6392 (vmull_high_n_<su>16): Likewise.
6393 (vqrdmulh<q>_n_s16): Likewise.
6395 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6397 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6398 have the correct lane parameter.
6400 2013-09-06 Richard Biener <rguenther@suse.de>
6402 * cfganal.c (control_dependences::~control_dependences):
6403 Properly free all of the vector.
6405 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6408 * config/i386/i386.c (ix86_conditional_register_usage):
6409 Proper initialize extended SSE registers.
6411 2013-09-06 Jan Hubicka <jh@suse.cz>
6413 PR tree-optimization/58311
6414 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6416 2013-09-06 Jan Hubicka <jh@suse.cz>
6418 * Makefile.in (tree-sra.o): Update dependencies.
6419 * tree-sra.c: Include ipa-utils.h
6420 (scan_function): Use recursive_call_p.
6421 (has_caller_p): New function.
6422 (cgraph_for_node_and_aliases): Count also callers of aliases.
6424 2013-09-06 Jan Hubicka <jh@suse.cz>
6427 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6428 * tree-tailcall.c: Include ipa-utils.h.
6429 (find_tail_calls): Use it.
6430 * ipa-pure-const.c (check_call): Likewise.
6431 * ipa-utils.c (recursive_call_p): New function.
6432 * ipa-utils.h (recursive_call_p): Dclare.
6433 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6434 (symtab_semantically_equivalent_p): New function.
6435 * Makefile.in (tree-tailcall.o): Update dependencies.
6437 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6439 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6442 2013-09-06 Richard Biener <rguenther@suse.de>
6444 * lto-streamer.h (lto_global_var_decls): Remove.
6445 * Makefile.in (OBJS): Remove lto-symtab.o.
6446 (lto-symtab.o): Remove.
6447 (GTFILES): Remove lto-symtab.c
6448 * lto-symtab.c: Move to lto/
6450 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6452 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6453 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6454 (UNSPEC_FPINT_RINT): New constant definitions.
6455 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6456 definition with 2 attributes.
6457 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6458 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6461 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6463 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6464 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6465 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6466 attribute to "z196_alone".
6467 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6470 2013-09-06 Richard Biener <rguenther@suse.de>
6472 * basic-block.h (class control_dependences): New.
6473 * tree-ssa-dce.c (control_dependence_map): Remove.
6475 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6476 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6477 find_pdom, find_control_dependence, find_all_control_dependences):
6479 (mark_control_dependent_edges_necessary,
6480 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6481 tree_dce_done, perform_tree_ssa_dce): Adjust.
6482 * cfganal.c (set_control_dependence_map_bit,
6483 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6484 find_all_control_dependences): Move from tree-ssa-dce.c and
6485 implement as methods of control_dependences class.
6486 (control_dependences::control_dependences): New.
6487 (control_dependences::~control_dependences): Likewise.
6488 (control_dependences::get_edges_dependent_on): Likewise.
6489 (control_dependences::get_edge): Likewise.
6491 2013-09-04 Jan Hubicka <jh@suse.cz>
6493 * tree.c (types_same_for_odr): Drop overactive check.
6494 * ipa-devirt.c (hash_type_name): Likewise.
6496 2013-09-04 Jan Hubicka <jh@suse.cz>
6498 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6499 (analyze_functions): ... here.
6501 2013-09-04 Jan Hubicka <jh@suse.cz>
6504 * cgraphunit.c (analyze_functions): Clear AUX fields
6505 after processing; initialize assembler name has.
6507 2013-09-05 Jeff Law <law@redhat.com>
6509 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6510 from thread_around_empty_block. Record threading path into PATH.
6511 Recurse if threading through the initial block is successful.
6512 (thread_across_edge): Corresponding changes to slightly simplify.
6514 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6516 * config/aarch64/aarch64.md
6517 (type): Remove frecpe, frecps, frecpx.
6518 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6519 fix to be a TARGET_SIMD instruction.
6520 (aarch64_frecps): Remove.
6521 * config/aarch64/aarch64-simd.md
6522 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6523 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6525 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6526 Sofiane Naci <sofiane.naci@arm.com>
6528 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6529 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6530 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6531 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6532 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6533 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6534 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6535 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6536 into "clz, "rbit". Rename "shift" to "shift_imm".
6537 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6538 changes. Update for attribute changes all occurrences of arlo_* and
6540 * config/arm/arm-fixed.md: Update for attribute changes
6541 all occurrences of arlo_* types.
6542 * config/arm/thumb2.md: Update for attribute changes all occurrences
6544 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6545 (cortexa7_older_only): Likewise.
6546 (cortexa7_younger): Likewise.
6547 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6548 (1020alu_shift_op): Likewise.
6549 (1020alu_shift_reg_op): Likewise.
6550 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6551 (alu_shift_op): Likewise.
6552 (alu_shift_reg_op): Likewise.
6553 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6554 (11_alu_shift_op): Likewise.
6555 (11_alu_shift_reg_op): Likewise.
6556 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6557 (9_alu_shift_reg_op): Likewise.
6558 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6560 (cortex_a15_alu_shift): Likewise.
6561 (cortex_a15_alu_shift_reg): Likewise.
6562 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6564 (cortex_a5_alu_shift): Likewise.
6565 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6567 (cortex_a53_alu_shift): Likewise.
6568 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6570 (cortex_a7_alu_reg): Likewise.
6571 (cortex_a7_alu_shift): Likewise.
6572 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6574 (cortex_a8_alu_shift): Likewise.
6575 (cortex_a8_alu_shift_reg): Likewise.
6576 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6577 (cortex_a9_dp_shift): Likewise.
6578 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6580 * config/arm/cortex-r4.md
6581 (cortex_r4_alu): Update for attribute changes.
6582 (cortex_r4_mov): Likewise.
6583 (cortex_r4_alu_shift_reg): Likewise.
6584 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6585 (526_alu_shift_op): Likewise.
6586 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6587 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6588 (626te_alu_shift_op): Likewise.
6589 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6590 (726te_alu_shift_op): Likewise.
6591 (726te_alu_shift_reg_op): Likewise.
6592 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6593 (mp626_alu_shift_op): Likewise.
6594 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6595 (pj4_alu_conds): Likewise.
6596 (pj4_shift): Likewise.
6597 (pj4_shift_conds): Likewise.
6598 (pj4_alu_shift): Likewise.
6599 (pj4_alu_shift_conds): Likewise.
6600 * config/aarch64/aarch64.md: Update for attribute change
6601 all occurrences of arlo_* and shift* types.
6603 2013-09-05 Mike Stump <mikestump@comcast.net>
6605 * tree.h: Move documentation for tree_function_decl to tree-core.h
6606 with the declaration.
6608 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6611 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6612 looking for widest mode.
6614 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6616 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6618 2013-09-05 Richard Biener <rguenther@suse.de>
6620 PR tree-optimization/58137
6621 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6622 Do not create vectors of pointers.
6623 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6624 types for the components of the vector initializer.
6625 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6626 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6628 2013-09-05 Martin Jambor <mjambor@suse.cz>
6630 * ipa-prop.c (remove_described_reference): Accept missing references,
6631 return false if that hppens, otherwise return true.
6632 (cgraph_node_for_jfunc): New function.
6633 (try_decrement_rdesc_refcount): Likewise.
6634 (try_make_edge_direct_simple_call): Use them.
6635 (ipa_edge_removal_hook): Remove references from rdescs.
6636 (ipa_edge_duplication_hook): Clone rdescs and their references
6637 when the new edge has the same caller as the old one.
6638 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6639 reference before removing any edges.
6641 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6643 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6645 * thumb2.md (thumb2_storewb_parisi): New pattern.
6647 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6649 * config/aarch64/aarch64-option-extensions.def: Add
6650 AARCH64_OPT_EXTENSION of 'crc'.
6651 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6652 (AARCH64_ISA_CRC): Ditto.
6653 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6654 description of the CRC extension.
6656 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6658 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6659 * config/rs6000/linux.h: Ditto.
6660 * alpha/linux.h: Ditto.
6661 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6662 no_c99_libc_has_function.
6663 * config/c6x/uclinux-elf.h: Ditto.
6664 * config/lm32/uclinux-elf.h: Ditto.
6665 * config/m68k/uclinux.h: Ditto.
6666 * config/moxie/uclinux.h: Ditto.
6667 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6668 (crisv32-*-linux*, cris-*-linux*): Ditto.
6669 * config/bfin/bfin.c: Include "tm_p.h".
6671 2013-09-05 Richard Biener <rguenther@suse.de>
6673 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6674 check for a definition without a basic-block.
6676 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6677 Sofiane Naci <sofiane.naci@arm.com>
6679 * config/aarch64/aarch64.md
6680 (*movti_aarch64): Rename r_2_f and f_2_r.
6681 (*movsf_aarch64): Likewise.
6682 (*movdf_aarch64): Likewise.
6683 (*movtf_aarch64): Likewise.
6684 (aarch64_movdi_<mode>low): Likewise.
6685 (aarch64_movdi_<mode>high): Likewise.
6686 (aarch64_mov<mode>high_di): Likewise.
6687 (aarch64_mov<mode>low_di): Likewise.
6688 (aarch64_movtilow_tilow): Likewise.
6689 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6690 values to config/arm/types.md
6691 (attribute "conds"): Update for attribute change.
6692 (anddi3_insn): Likewise.
6693 (iordi3_insn): Likewise.
6694 (xordi3_insn): Likewise.
6695 (one_cmpldi2): Likewise.
6696 * config/arm/types.md (type): Add Neon types.
6697 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6698 use "type" attribute.
6699 (movmisalign<mode>_neon_store): Likewise.
6700 (movmisalign<mode>_neon_load): Likewise.
6701 (vec_set<mode>_internal): Likewise.
6702 (vec_setv2di_internal): Likewise.
6703 (vec_extract<mode>): Likewise.
6704 (vec_extractv2di): Likewise.
6705 (add<mode>3_neon): Likewise.
6706 (adddi3_neon): Likewise.
6707 (sub<mode>3_neon): Likewise.
6708 (subdi3_neon): Likewise.
6709 (mul<mode>3_neon): Likewise.
6710 (mul<mode>3add<mode>_neon): Likewise.
6711 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6712 (fma<VCVTF:mode>4)): Likewise.
6713 (fma<VCVTF:mode>4_intrinsic): Likewise.
6714 (fmsub<VCVTF:mode>4)): Likewise.
6715 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6716 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6717 (ior<mode>3): Likewise.
6718 (and<mode>3): Likewise.
6719 (anddi3_neon): Likewise.
6720 (orn<mode>3_neon): Likewise.
6721 (orndi3_neon): Likewise.
6722 (bic<mode>3_neon): Likewise.
6723 (bicdi3_neon): Likewise.
6724 (xor<mode>3): Likewise.
6725 (one_cmpl<mode>2): Likewise.
6726 (abs<mode>2): Likewise.
6727 (neg<mode>2): Likewise.
6728 (umin<mode>3_neon): Likewise.
6729 (umax<mode>3_neon): Likewise.
6730 (smin<mode>3_neon): Likewise.
6731 (smax<mode>3_neon): Likewise.
6732 (vashl<mode>3): Likewise.
6733 (vashr<mode>3_imm): Likewise.
6734 (vlshr<mode>3_imm): Likewise.
6735 (ashl<mode>3_signed): Likewise.
6736 (ashl<mode>3_unsigned): Likewise.
6737 (neon_load_count): Likewise.
6738 (ashldi3_neon_noclobber): Likewise.
6739 (signed_shift_di3_neon): Likewise.
6740 (unsigned_shift_di3_neon): Likewise.
6741 (ashrdi3_neon_imm_noclobber): Likewise.
6742 (lshrdi3_neon_imm_noclobber): Likewise.
6743 (widen_ssum<mode>3): Likewise.
6744 (widen_usum<mode>3): Likewise.
6745 (quad_halves_<code>v4si): Likewise.
6746 (quad_halves_<code>v4sf): Likewise.
6747 (quad_halves_<code>v8hi): Likewise.
6748 (quad_halves_<code>v16qi): Likewise.
6749 (reduc_splus_v2di): Likewise.
6750 (neon_vpadd_internal<mode>): Likewise.
6751 (neon_vpsmin<mode>): Likewise.
6752 (neon_vpsmax<mode>): Likewise.
6753 (neon_vpumin<mode>): Likewise.
6754 (neon_vpumax<mode>): Likewise.
6755 (ss_add<mode>_neon): Likewise.
6756 (us_add<mode>_neon): Likewise.
6757 (ss_sub<mode>_neon): Likewise.
6758 (us_sub<mode>_neon): Likewise.
6759 (neon_vadd<mode>_unspec): Likewise.
6760 (neon_vaddl<mode>): Likewise.
6761 (neon_vaddw<mode>): Likewise.
6762 (neon_vhadd<mode>): Likewise.
6763 (neon_vqadd<mode>): Likewise.
6764 (neon_vaddhn<mode>): Likewise.
6765 (neon_vmul<mode>): Likewise.
6766 (neon_vmla<mode>): Likewise.
6767 (neon_vmlal<mode>): Likewise.
6768 (neon_vmls<mode>): Likewise.
6769 (neon_vmlsl<mode>): Likewise.
6770 (neon_vqdmulh<mode>): Likewise.
6771 (neon_vqdmlal<mode>): Likewise.
6772 (neon_vqdmlsl<mode>): Likewise.
6773 (neon_vmull<mode>): Likewise.
6774 (neon_vqdmull<mode>): Likewise.
6775 (neon_vsub<mode>_unspec): Likewise.
6776 (neon_vsubl<mode>): Likewise.
6777 (neon_vsubw<mode>): Likewise.
6778 (neon_vqsub<mode>): Likewise.
6779 (neon_vhsub<mode>): Likewise.
6780 (neon_vsubhn<mode>): Likewise.
6781 (neon_vceq<mode>): Likewise.
6782 (neon_vcge<mode>): Likewise.
6783 (neon_vcgeu<mode>): Likewise.
6784 (neon_vcgt<mode>): Likewise.
6785 (neon_vcgtu<mode>): Likewise.
6786 (neon_vcle<mode>): Likewise.
6787 (neon_vclt<mode>): Likewise.
6788 (neon_vcage<mode>): Likewise.
6789 (neon_vcagt<mode>): Likewise.
6790 (neon_vtst<mode>): Likewise.
6791 (neon_vabd<mode>): Likewise.
6792 (neon_vabdl<mode>): Likewise.
6793 (neon_vaba<mode>): Likewise.
6794 (neon_vabal<mode>): Likewise.
6795 (neon_vmax<mode>): Likewise.
6796 (neon_vmin<mode>): Likewise.
6797 (neon_vpaddl<mode>): Likewise.
6798 (neon_vpadal<mode>): Likewise.
6799 (neon_vpmax<mode>): Likewise.
6800 (neon_vpmin<mode>): Likewise.
6801 (neon_vrecps<mode>): Likewise.
6802 (neon_vrsqrts<mode>): Likewise.
6803 (neon_vqabs<mode>): Likewise.
6804 (neon_vqneg<mode>): Likewise.
6805 (neon_vcls<mode>): Likewise.
6806 (clz<mode>2): Likewise.
6807 (popcount<mode>2): Likewise.
6808 (neon_vrecpe): Likewise.
6809 (neon_vrsqrte): Likewise.
6810 (neon_vget_lane<mode>_sext_internal): Likewise.
6811 (neon_vget_lane<mode>_zext_internal): Likewise.
6812 (neon_vdup_n<mode>): Likewise.
6813 (neon_vdup_nv2di): Likewise.
6814 (neon_vdpu_lane<mode>_internal): Likewise.
6815 (neon_vswp<mode>): Likewise.
6816 (float<mode><V_cvtto>2): Likewise.
6817 (floatuns<mode><V_cvtto>2): Likewise.
6818 (fix_trunc<mode><V_cvtto>)2): Likewise
6819 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6820 (neon_vcvt<mode>): Likewise.
6821 (neon_vcvtv4sfv4hf): Likewise.
6822 (neon_vcvtv4hfv4sf): Likewise.
6823 (neon_vcvt_n<mode>): Likewise.
6824 (neon_vmovn<mode>): Likewise.
6825 (neon_vqmovn<mode>): Likewise.
6826 (neon_vqmovun<mode>): Likewise.
6827 (neon_vmovl<mode>): Likewise.
6828 (neon_vmul_lane<mode>): Likewise.
6829 (neon_vmull_lane<mode>): Likewise.
6830 (neon_vqdmull_lane<mode>): Likewise.
6831 (neon_vqdmulh_lane<mode>): Likewise.
6832 (neon_vmla_lane<mode>): Likewise.
6833 (neon_vmlal_lane<mode>): Likewise.
6834 (neon_vqdmlal_lane<mode>): Likewise.
6835 (neon_vmls_lane<mode>): Likewise.
6836 (neon_vmlsl_lane<mode>): Likewise.
6837 (neon_vqdmlsl_lane<mode>): Likewise.
6838 (neon_vext<mode>): Likewise.
6839 (neon_vrev64<mode>): Likewise.
6840 (neon_vrev32<mode>): Likewise.
6841 (neon_vrev16<mode>): Likewise.
6842 (neon_vbsl<mode>_internal): Likewise.
6843 (neon_vshl<mode>): Likewise.
6844 (neon_vqshl<mode>): Likewise.
6845 (neon_vshr_n<mode>): Likewise.
6846 (neon_vshrn_n<mode>): Likewise.
6847 (neon_vqshrn_n<mode>): Likewise.
6848 (neon_vqshrun_n<mode>): Likewise.
6849 (neon_vshl_n<mode>): Likewise.
6850 (neon_vqshl_n<mode>): Likewise.
6851 (neon_vqshlu_n<mode>): Likewise.
6852 (neon_vshll_n<mode>): Likewise.
6853 (neon_vsra_n<mode>): Likewise.
6854 (neon_vsri_n<mode>): Likewise.
6855 (neon_vsli_n<mode>): Likewise.
6856 (neon_vtbl1v8qi): Likewise.
6857 (neon_vtbl2v8qi): Likewise.
6858 (neon_vtbl3v8qi): Likewise.
6859 (neon_vtbl4v8qi): Likewise.
6860 (neon_vtbx1v8qi): Likewise.
6861 (neon_vtbx2v8qi): Likewise.
6862 (neon_vtbx3v8qi): Likewise.
6863 (neon_vtbx4v8qi): Likewise.
6864 (neon_vtrn<mode>_internal): Likewise.
6865 (neon_vzip<mode>_internal): Likewise.
6866 (neon_vuzp<mode>_internal): Likewise.
6867 (neon_vld1<mode>): Likewise.
6868 (neon_vld1_lane<mode>): Likewise.
6869 (neon_vld1_dup<mode>): Likewise.
6870 (neon_vld1_dupv2di): Likewise.
6871 (neon_vst1<mode>): Likewise.
6872 (neon_vst1_lane<mode>): Likewise.
6873 (neon_vld2<mode>): Likewise.
6874 (neon_vld2_lane<mode>): Likewise.
6875 (neon_vld2_dup<mode>): Likewise.
6876 (neon_vst2<mode>): Likewise.
6877 (neon_vst2_lane<mode>): Likewise.
6878 (neon_vld3<mode>): Likewise.
6879 (neon_vld3qa<mode>): Likewise.
6880 (neon_vld3qb<mode>): Likewise.
6881 (neon_vld3_lane<mode>): Likewise.
6882 (neon_vld3_dup<mode>): Likewise.
6883 (neon_vst3<mode>): Likewise.
6884 (neon_vst3qa<mode>): Likewise.
6885 (neon_vst3qb<mode>): Likewise.
6886 (neon_vst3_lane<mode>): Likewise.
6887 (neon_vld4<mode>): Likewise.
6888 (neon_vld4qa<mode>): Likewise.
6889 (neon_vld4qb<mode>): Likewise.
6890 (neon_vld4_lane<mode>): Likewise.
6891 (neon_vld4_dup<mode>): Likewise.
6892 (neon_vst4<mode>): Likewise.
6893 (neon_vst4qa<mode>): Likewise.
6894 (neon_vst4qb<mode>): Likewise.
6895 (neon_vst4_lane<mode>): Likewise.
6896 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6897 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6898 (neon_vec_<US>mult_lo_<mode>): Likewise.
6899 (neon_vec_<US>mult_hi_<mode>): Likewise.
6900 (neon_vec_<US>shiftl_<mode>): Likewise.
6901 (neon_unpack<US>_<mode>): Likewise.
6902 (neon_vec_<US>mult_<mode>): Likewise.
6903 (vec_pack_trunc_<mode>): Likewise.
6904 (neon_vec_pack_trunk_<mode>): Likewise.
6905 (neon_vabd<mode>_2): Likewise.
6906 (neon_vabd<mode>_3): Likewise.
6907 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6908 (thumb2_movsi_vfp): Likewise.
6909 (movdi_vfp): Likewise.
6910 (movdi_vfp_cortexa8): Likewise.
6911 (movhf_vfp_neon): Likewise.
6912 (movhf_vfp): Likewiwse.
6913 (movsf_vfp): Likewiwse.
6914 (thumb2_movsf_vfp): Likewiwse.
6915 (movdf_vfp): Likewise.
6916 (thumb2_movdf_vfp): Likewise.
6917 (movsfcc_vfp): Likewise.
6918 (thumb2_movsfcc_vfp): Likewise.
6919 (movdfcc_vfp): Likewise.
6920 (thumb2_movdfcc_vfp): Likewise.
6921 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6922 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6923 (v10_v2c): Likewise.
6924 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6926 (cortex_a15_neon_int_2): Likewise.
6927 (cortex_a15_neon_int_3): Likewise.
6928 (cortex_a15_neon_int_4): Likewise.
6929 (cortex_a15_neon_int_5): Likewise.
6930 (cortex_a15_neon_vqneg_vqabs): Likewise.
6931 (cortex_a15_neon_vmov): Likewise.
6932 (cortex_a15_neon_vaba): Likewise.
6933 (cortex_a15_neon_vaba_qqq): Likewise.
6934 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6935 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6936 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6937 scalar_64_32_long_scalar): Likewise.
6938 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6939 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6940 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6941 lotype_qdd_64_32_long): Likewise.
6942 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6943 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6944 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6945 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6946 (cortex_a15_neon_shift_1): Likewise.
6947 (cortex_a15_neon_shift_2): Likewise.
6948 (cortex_a15_neon_shift_3): Likewise.
6949 (cortex_a15_neon_vshl_ddd): Likewise.
6950 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6951 (cortex_a15_neon_vsra_vrsra): Likewise.
6952 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6953 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6954 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6955 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6956 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6957 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6958 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6959 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6960 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6961 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6962 (cortex_a15_neon_bp_simple): Likewise.
6963 (cortex_a15_neon_bp_2cycle): Likewise.
6964 (cortex_a15_neon_bp_3cycle): Likewise.
6965 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6966 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6967 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6968 (cortex_a15_neon_vld2_4_regs): Likewise.
6969 (cortex_a15_neon_vld3_vld4): Likewise.
6970 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6971 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6972 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6973 (cortex_a15_neon_vst3_vst4): Likewise.
6974 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6975 (cortex_a15_neon_vld3_vld4_lane" 10
6976 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6977 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6978 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6979 (cortex_a15_neon_ldm_2): Likewise.0
6980 (cortex_a15_neon_stm_2): Likewise.
6981 (cortex_a15_neon_mcr): Likewise.
6982 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6983 (cortex_a15_neon_mrc): Likewise.
6984 (cortex_a15_neon_mrrc): Likewise.
6985 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6987 (cortex_a15_alu_shift): Likewise.
6988 (cortex_a15_alu_shift_reg): Likewise.
6989 (cortex_a15_mult32): Likewise.
6990 (cortex_a15_mult64): Likewise.
6991 (cortex_a15_block): Likewise.
6992 (cortex_a15_branch): Likewise.
6993 (cortex_a15_load1): Likewise.
6994 (cortex_a15_load3): Likewise.
6995 (cortex_a15_store1): Likewise.
6996 (cortex_a15_store3): Likewise.
6997 (cortex_a15_call): Likewise.
6998 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6999 (cortex_a5_f2r): Likewise.
7000 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7002 (cortex_a53_f2r): Likewise.
7003 * config/arm/cortex-a7.md
7004 (cortex_a7_branch): Update for attribute change.
7005 (cortex_a7_call): Likewise.
7006 (cortex_a7_alu_imm): Likewise.
7007 (cortex_a7_alu_reg): Likewise.
7008 (cortex_a7_alu_shift): Likewise.
7009 (cortex_a7_mul): Likewise.
7010 (cortex_a7_load1): Likewise.
7011 (cortex_a7_store1): Likewise.
7012 (cortex_a7_load2): Likewise.
7013 (cortex_a7_store2): Likewise.
7014 (cortex_a7_load3): Likewise.
7015 (cortex_a7_store3): Likewise.
7016 (cortex_a7_load4): Likewise.
7017 (cortex_a7_store4): Likewise.
7018 (cortex_a7_fpalu): Likewise.
7019 (cortex_a7_fconst): Likewise.
7020 (cortex_a7_fpmuls): Likewise.
7021 (cortex_a7_neon_mul): Likewise.
7022 (cortex_a7_fpmacs): Likewise.
7023 (cortex_a7_neon_mla: Likewise.
7024 (cortex_a7_fpmuld: Likewise.
7025 (cortex_a7_fpmacd: Likewise.
7026 (cortex_a7_fpfmad: Likewise.
7027 (cortex_a7_fdivs: Likewise.
7028 (cortex_a7_fdivd: Likewise.
7029 (cortex_a7_r2f: Likewise.
7030 (cortex_a7_f2r: Likewise.
7031 (cortex_a7_f_flags: Likewise.
7032 (cortex_a7_f_loads: Likewise.
7033 (cortex_a7_f_loadd: Likewise.
7034 (cortex_a7_f_stores: Likewise.
7035 (cortex_a7_f_stored: Likewise.
7036 (cortex_a7_neon): Likewise.
7037 * config/arm/cortex-a8-neon.md
7038 (cortex_a8_neon_mrc): Update for attribute change.
7039 (cortex_a8_neon_mrrc): Likewise.
7040 (cortex_a8_neon_int_1): Likewise.
7041 (cortex_a8_neon_int_2): Likewise.
7042 (cortex_a8_neon_int_3): Likewise.
7043 (cortex_a8_neon_int_4): Likewise.
7044 (cortex_a8_neon_int_5): Likewise.
7045 (cortex_a8_neon_vqneg_vqabs): Likewise.
7046 (cortex_a8_neon_vmov): Likewise.
7047 (cortex_a8_neon_vaba): Likewise.
7048 (cortex_a8_neon_vaba_qqq): Likewise.
7049 (cortex_a8_neon_vsma): Likewise.
7050 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7051 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7052 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7053 long_scalar): Likewise.
7054 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7055 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7056 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7057 long_scalar_qdd_64_32_long): Likewise.
7058 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7059 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7060 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7061 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7062 (cortex_a8_neon_shift_1): Likewise.
7063 (cortex_a8_neon_shift_2): Likewise.
7064 (cortex_a8_neon_shift_3): Likewise.
7065 (cortex_a8_neon_vshl_ddd): Likewise.
7066 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7067 (cortex_a8_neon_vsra_vrsra): Likewise.
7068 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7069 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7070 (cortex_a8_neon_fp_vsum): Likewise.
7071 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7072 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7073 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7074 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7075 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7076 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7077 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7078 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7079 (cortex_a8_neon_bp_simple): Likewise.
7080 (cortex_a8_neon_bp_2cycle): Likewise.
7081 (cortex_a8_neon_bp_3cycle): Likewise.
7082 (cortex_a8_neon_ldr): Likewise.
7083 (cortex_a8_neon_str): Likewise.
7084 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7085 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7086 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7087 (cortex_a8_neon_vld2_4_regs): Likewise.
7088 (cortex_a8_neon_vld3_vld4): Likewise.
7089 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7090 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7091 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7092 (cortex_a8_neon_vst3_vst4): Likewise.
7093 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7094 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7095 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7096 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7097 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7098 (cortex_a8_neon_mcr): Likewise.
7099 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7100 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7101 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7103 (ca9_neon_mrrc): Likewise.
7104 (cortex_a9_neon_int_1): Likewise.
7105 (cortex_a9_neon_int_2): Likewise.
7106 (cortex_a9_neon_int_3): Likewise.
7107 (cortex_a9_neon_int_4): Likewise.
7108 (cortex_a9_neon_int_5): Likewise.
7109 (cortex_a9_neon_vqneg_vqabs): Likewise.
7110 (cortex_a9_neon_vmov): Likewise.
7111 (cortex_a9_neon_vaba): Likewise.
7112 (cortex_a9_neon_vaba_qqq): Likewise.
7113 (cortex_a9_neon_vsma): Likewise.
7114 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7115 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7116 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7117 long_scalar): Likewise.
7118 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7119 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7120 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7121 long_scalar_qdd_64_32_long): Likewise.
7122 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7123 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7124 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7125 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7126 (cortex_a9_neon_shift_1): Likewise.
7127 (cortex_a9_neon_shift_2): Likewise.
7128 (cortex_a9_neon_shift_3): Likewise.
7129 (cortex_a9_neon_vshl_ddd): Likewise.
7130 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7131 (cortex_a9_neon_vsra_vrsra): Likewise.
7132 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7133 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7134 (cortex_a9_neon_fp_vsum): Likewise.
7135 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7136 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7137 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7138 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7139 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7140 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7141 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7142 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7143 (cortex_a9_neon_bp_simple): Likewise.
7144 (cortex_a9_neon_bp_2cycle): Likewise.
7145 (cortex_a9_neon_bp_3cycle): Likewise.
7146 (cortex_a9_neon_ldr): Likewise.
7147 (cortex_a9_neon_str): Likewise.
7148 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7149 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7150 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7151 (cortex_a9_neon_vld2_4_regs): Likewise.
7152 (cortex_a9_neon_vld3_vld4): Likewise.
7153 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7154 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7155 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7156 (cortex_a9_neon_vst3_vst4): Likewise.
7157 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7158 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7159 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7160 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7161 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7162 (cortex_a9_neon_mcr): Likewise.
7163 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7164 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7165 (cortex_a9_fps): Likewise.
7166 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7168 (cortex_m4_fmuls): Likewise.
7169 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7171 (cortex_r4_mrc): Likewise.
7172 * config/arm/iterators.md: Update comment referring to neon_type.
7173 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7174 (iwmmxt_movsi_insn): Likewise.
7175 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7177 (pj4_core_to_vfp): Likewise.
7178 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7180 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7181 (vfp_fstore): Likewise.
7182 * doc/md.texi: Change references to neon_type to refer to type.
7184 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7186 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7188 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7190 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7193 2013-09-04 Easwaran Raman <eraman@google.com>
7196 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7197 (build_and_add_sum): Use it.
7198 (appears_later_in_bb): Simplify code.
7200 2013-09-04 Teresa Johnson <tejohnson@google.com>
7202 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7204 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7206 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7208 2013-09-04 Jan Hubicka <jh@suse.cz>
7210 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7211 * ipa-devirt.c: Include diganostic.h
7212 (odr_type_d): Add types and types_set.
7213 (hash_type_name): Work for types with vtables during LTO.
7214 (odr_hasher::remove): Fix comment; destroy types_set.
7215 (add_type_duplicate): New function,
7216 (get_odr_type): Use it.
7217 (dump_type_inheritance_graph): Dump type duplicates.
7218 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7220 * tree.c (types_same_for_odr): Give exact answers on types with
7223 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7225 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7226 explaining their differences.
7228 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7230 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7232 2013-09-03 Jeff Law <law@redhat.com>
7234 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7235 when not threading through a joiner block. Pass joiner/no joiner
7236 state to register_jump_thread.
7237 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7238 state from argument rather than implying on path length.
7239 Dump the entire jump thread path into debugging dump.
7240 * tree-flow.h (register_jump_thread): Update prototype.
7242 2013-08-29 Xinliang David Li <davidxl@google.com>
7244 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7245 Remove a trivial gcc_assert.
7247 2013-08-29 Xinliang David Li <davidxl@google.com>
7249 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7250 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7251 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7252 Delay base decl alignment adjustment.
7253 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7254 * tree-vectorizer.h: New data structure.
7255 (set_dr_misalignment): New function.
7256 (dr_misalignment): Ditto.
7257 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7258 (vectorizable_load): Ditto.
7259 (ensure_base_align): New function.
7260 (vectorize_loops): Add dbg_cnt support.
7261 (execute_vect_slp): Ditto.
7262 * dbgcnt.def: New debug counter.
7263 * Makefile: New dependency.
7265 2013-09-03 Meador Inge <meadori@codesourcery.com>
7269 2013-08-30 Meador Inge <meadori@codesourcery.com>
7271 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7273 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7275 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7276 function descriptor.
7278 2013-09-03 Richard Biener <rguenther@suse.de>
7280 * tree-affine.c (add_elt_to_tree): Fix association issue,
7281 avoid useless converts and make sure to always return a
7282 properly typed result.
7284 2013-09-03 Richard Biener <rguenther@suse.de>
7287 * fold-const.c (negate_expr_p): Fix division case.
7288 (negate_expr): Likewise.
7290 2013-09-03 Richard Biener <rguenther@suse.de>
7293 * tree-streamer-out.c: Include tm.h.
7294 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7296 2013-09-03 Jan Hubicka <jh@suse.cz>
7298 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7300 2013-09-03 Alan Modra <amodra@gmail.com>
7302 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7303 multilibs to big-endian target and vice versa.
7304 * config/rs6000/t-linux64: Use := assignment on all vars.
7305 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7306 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7307 * config/rs6000/t-linux64le: New file.
7308 * config/rs6000/t-linux64bele: New file.
7309 * config/rs6000/t-linux64lebe: New file.
7311 2013-09-02 Jan Hubicka <jh@suse.cz>
7313 * ipa-inline-transform.c (inline_transform): Do not
7314 optimize_inline_calls when not optimizing.
7316 2013-09-02 Jan Hubicka <jh@suse.cz>
7318 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7319 duplicated nodes for assembler names.
7320 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7321 hard registers from assembler name hash.
7323 2013-09-02 Jan Hubicka <jh@suse.cz>
7325 * ipa-split.c (execute_split_functions): Split externally visible
7326 functions called once.
7328 2013-09-02 Martin Jambor <mjambor@suse.cz>
7331 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7332 linked list. When finding the correct duplicate, also consider also
7333 the caller in additon to its inlined_to node.
7335 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7337 * config/aarch64/aarch64-simd-builtins.def
7338 (dup_lane_scalar): Remove.
7339 * config/aarch64/aarch64-simd.md
7340 (aarch64_simd_dup): Add 'w->w' alternative.
7341 (aarch64_dup_lane<mode>): Allow for VALL.
7342 (aarch64_dup_lane_scalar<mode>): Remove.
7343 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7344 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7345 (aarch64_get_lane_unsigned<mode>): Likewise.
7346 (aarch64_get_lane<mode>): Likewise.
7347 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7348 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7349 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7350 (VCON): Change container of V2SF.
7351 (vswap_width_name): Likewise.
7352 * config/aarch64/arm_neon.h
7353 (__aarch64_vdup_lane_any): New.
7354 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7355 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7356 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7358 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7361 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7362 the source or the destination is a single hard register.
7364 2013-09-02 Richard Biener <rguenther@suse.de>
7367 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7370 2013-09-02 Richard Biener <rguenther@suse.de>
7372 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7373 arithmetic to sizetype.
7375 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7377 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7378 Find auto-increment use both before and after candidate.
7380 2013-09-02 Marek Polacek <polacek@redhat.com>
7382 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7384 2013-09-01 Jan Hubicka <jh@suse.cz>
7386 * Makefile.in: Add ipa-profile.o
7387 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7388 * cgraph.c (struct cgraph_propagate_frequency_data,
7389 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7390 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7391 * cgraph.h (cgraph_propagate_frequency): Remove.
7392 * ipa-inline-analysis.c: Include ipa-utils.h;
7393 drop duplicated cfgloop.h.
7394 (inline_update_callee_summaries): Update.
7395 * ipa-profile.c: New file.
7396 * ipa-utils.h (ipa_propagate_frequency): Declare.
7397 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7398 data-streamer.h, value-prof.h.
7399 (symtab_remove_unreachable_nodes): Update profile.
7400 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7401 account_time_size, cmp_counts, dump_histogram,
7402 ipa_profile_generate_summary, ipa_profile_write_summary,
7403 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7404 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7405 Move to ipa-profile.c.
7407 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7409 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7411 2013-09-01 Jan Hubicka <jh@suse.cz>
7413 * common.opt (fdevirtualize-speculatively): New function.
7414 * invoke.texi (fdevirtualize-speculatively): Document.
7415 * ipa-devirt.c: Include ipa-inline.h
7416 (likely_target_p): New function.
7417 (ipa_devirt): New function.
7418 (gate_ipa_devirt): New function.
7419 (pass_data_ipa_devirt): New static var.
7420 (pass_ipa_devirt): Likewise.
7421 (make_pass_ipa_devirt): New function.
7422 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7423 (common_handle_option): Disable devirtualization when
7424 value range profiling is available.
7425 * passes.def (pass_ipa_devirt): Add.
7426 * timever.def (TV_IPA_DEVIRT): New timevar.
7427 * tree-pass.h (make_pass_ipa_devirt):
7429 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7431 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7432 include sanitize(undefined).
7434 2013-08-31 Diego Novillo <dnovillo@google.com>
7436 * Makefile.in (TREE_CORE_H): Define.
7438 (GTFILES): Add tree-core.h.
7439 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7441 * tree-core.h: New file.
7442 Move all data structures, enum, typedefs, global
7443 declarations and constants from ...
7446 2013-08-31 Jan Hubicka <jh@suse.cz>
7448 * bulitins.c (expand_builtin): Do not early exit for gcov
7449 instrumented functions.
7451 2013-08-31 Marek Polacek <polacek@redhat.com>
7453 * ubsan.c: Include tm_p.h.
7455 2013-08-31 Jan Hubicka <jh@suse.cz>
7457 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7460 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7461 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7462 * tree-cfg.c (verify_gimple_label): ... here.
7463 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7464 (ipa_merge_profiles): New function.
7465 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7466 (lto_input_function_body): Likewise.
7467 * ipa-utils.h (ipa_merge_profiles): Declare.
7468 * lto-streamer.h (lto_input_function_body): Update prototype.
7469 (emit_label_in_global_context_p): Remove.
7470 * lto-symtab.c: Include ipa-utils.h
7471 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7473 2013-08-31 Jan Hubicka <jh@suse.cz>
7475 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7477 2013-08-31 Jan Hubicka <jh@suse.cz>
7479 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7481 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7483 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7484 "cmp" RTX before signed_comparison_operator check to account
7487 2013-08-30 Jan Hubicka <jh@suse.cz>
7489 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7490 (detect_type_change_1): Rename to ...
7491 (detect_type_change): ... this one; early return on non-polymorphic
7493 (detect_type_change_ssa): Add comp_type parameter; update
7494 use of detect_type_change.
7495 (compute_complex_assign_jump_func): Add param_type parameter;
7496 update use of detect_type_change_ssa.
7497 (compute_complex_ancestor_jump_func): Likewise.
7498 (ipa_get_callee_param_type): New function.
7499 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7500 update calls to the jump function computation functions.
7502 2013-08-30 Teresa Johnson <tejohnson@google.com>
7503 Steven Bosscher <steven@gcc.gnu.org>
7505 * cfgrtl.c (fixup_new_cold_bb): New routine.
7506 (commit_edge_insertions): Invoke fixup_partitions.
7507 (find_partition_fixes): New routine.
7508 (fixup_partitions): Ditto.
7509 (verify_hot_cold_block_grouping): Update comments.
7510 (rtl_verify_edges): Invoke find_partition_fixes.
7511 (rtl_verify_bb_pointers): Update comments.
7512 (rtl_verify_bb_layout): Ditto.
7513 * basic-block.h (probably_never_executed_edge_p): Declare.
7514 (fixup_partitions): Ditto.
7515 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7516 * bb-reorder.c (sanitize_hot_paths): New function.
7517 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7519 * predict.c (probably_never_executed_edge_p): New routine.
7520 * cfg.c (check_bb_profile): Add partition insanity warnings.
7522 2013-08-30 Meador Inge <meadori@codesourcery.com>
7524 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7526 2013-08-30 Marek Polacek <polacek@redhat.com>
7528 * Makefile.in (ubsan.o): Add.
7529 (c-family/c-ubsan.o): Add.
7530 (builtins.o): Add ubsan.h dependency.
7531 * ubsan.h: New file.
7532 * ubsan.c: New file.
7533 * common.opt: Add -fsanitize=undefined option.
7534 (flag_sanitize): Add variable.
7535 (fsanitize=): Add option. Add Driver.
7536 (fsanitize=thread): Remove option.
7537 (fsanitize=address): Likewise.
7538 (static-libubsan): New option.
7539 * doc/invoke.texi: Document the new flag and -static-libubsan.
7540 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7541 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7542 * builtin-attrs.def (ATTR_COLD): Define.
7543 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7544 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7545 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7546 * flag-types.h (sanitize_code): New enum.
7547 * opts.c (common_handle_option): Parse command line arguments
7548 of -fsanitize=. Add -fsanitize=unreachable option.
7549 * varasm.c (get_variable_section): Adjust.
7550 (assemble_noswitch_variable): Likewise.
7551 (assemble_variable): Likewise.
7552 (output_constant_def_contents): Likewise.
7553 (categorize_decl_for_section): Likewise.
7554 (place_block_symbol): Likewise.
7555 (output_object_block): Likewise.
7556 * builtins.def: Likewise.
7557 * toplev.c (compile_file): Likewise.
7558 (process_options): Likewise.
7559 * cppbuiltin.c: Likewise.
7560 * tsan.c (tsan_pass): Likewise.
7561 (tsan_gate): Likewise.
7562 (tsan_gate_O0): Likewise.
7563 * cfgexpand.c (partition_stack_vars): Likewise.
7564 (expand_stack_vars): Likewise.
7565 (defer_stack_allocation): Likewise.
7566 (expand_used_vars): Likewise.
7567 * cfgcleanup.c (old_insns_match_p): Likewise.
7568 * asan.c (asan_finish_file): Likewise.
7569 (asan_instrument): Likewise.
7570 (gate_asan): Likewise.
7571 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7572 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7573 (asan_global_struct): Use pointer_sized_int_node instead
7574 calling build_nonstandard_integer_type.
7575 (initialize_sanitizer_builtins): Likewise.
7576 (asan_finish_file): Likewise.
7577 * gcc.c: Document %{%:function(args):X}.
7578 (static_spec_functions): Add sanitize.
7579 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7580 store funcval != NULL there.
7581 (do_spec_1): Adjust handle_spec_function caller.
7582 (handle_braces): Allow %:function(args) as condition.
7583 (sanitize_spec_function): New function.
7584 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7585 (LIBUBSAN_SPEC): Likewise.
7586 (LIBUBSAN_EARLY_SPEC): Likewise.
7587 (SANITIZER_SPEC): Handle libubsan.
7588 (SANITIZER_EARLY_SPEC): Likewise.
7589 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7590 instead of fsanitize=address.
7591 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7592 instead of fsanitize=address*.
7593 * builtins.c: Include ubsan.h.
7594 (fold_builtin_0): Instrument __builtin_unreachable.
7595 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7596 instead of flag_asan.
7597 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7598 (pointer_sized_int_node): Define.
7599 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7601 2013-08-30 Mike Stump <mikestump@comcast.net>
7603 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7606 2013-08-29 Jan Hubicka <jh@suse.cz>
7608 * cgraph.c (cgraph_function_body_availability): Handle weakref
7610 * passes.def: Remove pass_fixup_cfg.
7611 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7612 track when we need to remove functions.
7613 (gate_ipa_inline): Execute inlining always; add comment why.
7614 (pass_data_ipa_inline): Remove TODO_remove_functions.
7615 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7616 do not produce summaries.
7617 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7618 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7619 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7620 constant pool and vtable.
7622 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7624 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7625 New arm_neon.h's internal macros to specify 64-bit constants.
7626 Avoid using stdint.h's macros.
7628 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7630 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7631 before applying REG_P.
7633 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7635 PR tree-optimization/58277
7636 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7637 after seeing too many stmts with vdef in between dombb and current
7638 bb, invalidate everything.
7640 2013-08-30 Richard Biener <rguenther@suse.de>
7642 * fold-const.c (fold_single_bit_test): Fix overflow test.
7644 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7646 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7647 and which can live in a register, always retrieve the value on entry.
7648 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7649 passed by invisible reference specially.
7650 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7651 (vt_add_function_parameter): Correctly deal with a parameter passed by
7652 invisible reference.
7654 2013-08-30 Jan Hubicka <jh@suse.cz>
7656 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7658 2013-08-30 Richard Biener <rguenther@suse.de>
7660 PR tree-optimization/58228
7661 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7662 allow invariant loads in nested loop vectorization.
7664 2013-08-30 Richard Biener <rguenther@suse.de>
7666 PR tree-optimization/58223
7667 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7668 (has_anti_or_output_dependence): ... this and adjust to also
7669 look for output dependences.
7670 (mark_nodes_having_upstream_mem_writes): Adjust.
7671 (rdg_flag_uses): Likewise.
7673 2013-08-30 Richard Biener <rguenther@suse.de>
7675 PR tree-optimization/58010
7676 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7677 assert that we have a loop-closed PHI.
7679 2013-08-29 Jan Hubicka <jh@suse.cz>
7681 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7682 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7683 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7684 (lto_free_function_in_decl_state_for_node): New function.
7686 2013-08-29 Xinliang David Li <davidxl@google.com>
7688 * loop-unroll.c (report_unroll_peel): Minor message change.
7689 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7690 Emit alignment peeling message with default -fopt-info.
7691 (vect_loop_versioning): Emit loop version info message.
7692 * tree-vectorizer.c (vectorize_loops): Minor message change.
7693 (execute_vect_slp): Ditto.
7695 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7697 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7698 of the clone from the DECL_NAME of the original function.
7700 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7702 * passes.c (register_pass): Add overload.
7703 * tree-pass.h (register_pass): Forward declare it. Add comment.
7705 2013-08-29 Jan Hubicka <jh@suse.cz>
7707 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7708 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7709 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7710 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7711 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7712 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7713 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7714 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7716 2013-08-29 Teresa Johnson <tejohnson@google.com>
7718 * dumpfile.c (dump_loc): Output column number.
7719 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7720 * doc/invoke.texi: Document optall -fopt-info flag.
7721 * profile.c (read_profile_edge_counts): Use new dump framework.
7722 (compute_branch_probabilities): Ditto.
7723 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7724 when pass not in any opt group.
7725 * pass_manager.h (pass_manager::get_pass_profile): New method.
7726 * value-prof.c (check_counter): Use new dump framework.
7727 (check_ic_target): Ditto.
7728 * coverage.c (get_coverage_counts): Ditto.
7729 (coverage_init): Setup new dump framework.
7731 2013-08-29 Richard Biener <rguenther@suse.de>
7733 PR tree-optimization/58246
7734 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7735 handle the dominance check inside a basic-block.
7737 2013-08-29 Richard Biener <rguenther@suse.de>
7740 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7741 of default defs that appear in abnormal PHI nodes.
7743 2013-08-29 Richard Biener <rguenther@suse.de>
7745 PR tree-optimization/57685
7746 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7747 single-use operands to avoid exponential complexity.
7749 2013-08-28 Dehao Chen <dehao@google.com>
7751 * ipa-inline.c (edge_badness): Fix integer underflow.
7753 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7755 * gtm-builtins.def (_ITM_free): Declare leaf.
7757 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7760 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7761 (*tls_local_dynamic_base_64_largepic): Likewise.
7762 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7763 Remove predicate from call operand.
7764 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7765 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7767 2013-08-28 Jeff Law <law@redhat.com>
7769 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7770 checks for the number of predecessors and successors allowed.
7771 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7772 which require copying a joiner block if there is a request which
7773 is a subpath that requires no joiner block copying.
7775 2013-08-28 Jan Hubicka <jh@suse.cz>
7777 * lto-streamer-out.c (DFS_write_tree_body): Drop
7778 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7779 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7780 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7781 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7782 Do not read DECL_ERROR_ISSUED.
7783 (unpack_ts_decl_with_vis_value_fields): Do not read
7785 (lto_input_ts_binfo_tree_pointers): Do not read
7786 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7787 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7788 write DECL_ERROR_ISSUED..
7789 (pack_ts_decl_with_vis_value_fields): Do not write
7791 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7792 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7793 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7794 * tree.h (tree_decl_common): Update comment.
7795 (DECL_ERROR_ISSUED): Remove.
7797 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7800 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7802 2013-08-28 Jan Hubicka <jh@suse.cz>
7804 * builtins.def (free): Declare leaf.
7806 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7809 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7810 * configure: Regenerate.
7812 2013-08-27 Martin Jambor <mjambor@suse.cz>
7814 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7815 (ipa_ancestor_jf_data): Likewise.
7816 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7817 (ipa_get_jf_pass_through_type_preserved): New function.
7818 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7819 (ipa_get_jf_ancestor_type_preserved): New function.
7820 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7821 (ipa_get_jf_ancestor_result): Likewise.
7822 (propagate_vals_accross_pass_through): Use
7823 ipa_get_jf_pass_through_result to do all the value mappings.
7824 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7825 type_preserved flag.
7826 (ipa_set_jf_cst_copy): New function.
7827 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7828 (ipa_set_jf_arith_pass_through): Likewise.
7829 (ipa_set_ancestor_jf): Likewise.
7830 (compute_complex_assign_jump_func): Set type_preserved instead of
7832 (ipa_compute_jump_functions_for_edge): Likewise.
7833 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7834 (update_jump_functions_after_inlining): Update type_preserved.
7835 Explicitely create jump functions when combining one with pass_through.
7836 (ipa_write_jump_function): Stream the type_preserved flags.
7837 (ipa_read_jump_function): Likewise.
7839 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7840 Aldy Hernandez <aldyh@redhat.com>
7842 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7843 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7844 * function.h (struct function): Add has_force_vect_loops and
7846 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7847 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7849 * gimple.def: Update CLAUSES comments.
7850 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7851 (gimple_build_omp_for): Add argument to prototype.
7852 (gimple_omp_for_kind): New.
7853 (gimple_omp_for_set_kind): New.
7854 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7855 GOVD_DATA_SHARE_CLASS.
7856 (enum omp_region_type): Add ORT_SIMD.
7857 (gimple_add_tmp_var): Handle ORT_SIMD.
7858 (gimplify_var_or_parm_decl): Same.
7859 (is_gimple_stmt): Same.
7860 (omp_firstprivatize_variable): Same.
7861 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7862 (omp_notice_variable): Handle ORT_SIMD.
7863 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7864 (omp_check_private): Handle ORT_SIMD.
7865 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7867 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7868 Handle OMP_CLAUSE_LASTPRIVATE.
7869 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7871 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7872 (gimplify_expr): Handle OMP_SIMD.
7873 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7874 (expand_GOMP_SIMD_VF): New.
7875 (expand_GOMP_SIMD_LAST_LANE): New.
7876 * internal-fn.def (GOMP_SIMD_LANE): New.
7877 (GOMP_SIMD_VF): New.
7878 (GOMP_SIMD_LAST_LANE): New.
7879 * omp-low.c: Include target.h.
7880 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7882 (check_omp_nesting_restrictions): Same.
7884 (lower_rec_simd_input_clauses): New.
7885 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7887 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7888 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7889 (expand_omp_build_assign): New.
7890 (expand_omp_for_init_counts): New.
7891 (expand_omp_for_init_vars): New.
7892 (extract_omp_for_update_vars): New.
7893 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7894 and rewrite accordingly.
7895 (expand_omp_simd): New.
7896 (expand_omp_for): Use expand_omp_simd.
7897 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7898 (lower_omp_for): Do not lower the body.
7899 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7901 * tree-flow.h (find_omp_clause): Remove prototype.
7902 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7903 forcing vectorization of the loop, or if flag_tree_vectorize.
7904 (gate_tree_if_conversion): Similarly.
7905 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7906 gimple_build_omp_for.
7907 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7908 * tree-parloops (create_parallel_loop): Pass kind argument to
7909 gimple_build_omp_for.
7910 * tree-pretty-print.c (dump_omp_clause): Add cases for
7911 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7912 OMP_CLAUSE__SIMDUID_.
7913 (dump_generic_node): Handle OMP_SIMD.
7914 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7915 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7916 unroll OMP_SIMD loops here.
7917 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7918 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7920 (vect_analyze_data_refs): Handle simd loops.
7921 * tree-vect-loop.c (vectorizable_live_operation): Handle
7923 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7924 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7925 (vectorizable_load): Same.
7926 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7927 (struct simduid_to_vf): New.
7928 (simduid_to_vf::hash): New.
7929 (simduid_to-vf::equal): New.
7930 (struct simd_array_to_simduid): New.
7931 (simd_array_to_simduid::hash): New.
7932 (simd_array_to_simduid::equal): New.
7933 (adjust_simduid_builtins): New.
7934 (struct note_simd_array_uses_struct): New.
7935 (note_simd_array_uses_cb): New.
7936 (note_simd_array_uses): New.
7937 (vectorize_loops): Handle simd hints and adjust simd builtins
7939 * tree-vectorizer.h (struct _stmt_vec_info): Add
7940 simd_lane_access_p field.
7941 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7942 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7943 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7944 (omp_clause_code_name): Same.
7945 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7946 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7947 * tree.def (OMP_SIMD): New entry.
7948 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7949 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7950 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7951 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7952 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7953 (OMP_CLAUSE_LINEAR_STEP): New.
7954 (OMP_CLAUSE_SAFELEN_EXPR): New.
7955 (OMP_CLAUSE__SIMDUID__DECL): New.
7956 (find_omp_clause): New prototype.
7958 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7960 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7961 Haswell processor detection.
7963 2013-08-27 Christian Widmer <shadow@umbrox.de>
7966 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7967 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7968 AVX2 capable processors.
7970 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7972 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7973 of vget_low_* with implementations in terms of other intrinsics.
7975 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7978 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7979 values to -1, 0 and 1.
7981 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7983 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7984 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7985 (<optab><mode>3_insn): Remove.
7986 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7987 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7988 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7989 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7990 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7991 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7992 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7993 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7994 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7995 (ror<mode>3_insn): Likewise.
7996 * config/aarch64/predicates.md (aarch64_simd_register): New.
7998 2013-08-27 Richard Biener <rguenther@suse.de>
8000 PR tree-optimization/57521
8001 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8002 one edge is non-critical.
8003 (find_phi_replacement_condition): Make sure to use a non-critical
8004 edge. Cleanup and remove old bug workarounds.
8005 (bb_postdominates_preds): Remove.
8006 (if_convertible_loop_p_1): Do not compute post-dominators.
8007 (combine_blocks): Do not free post-dominators.
8008 (main_tree_if_conversion): Likewise.
8009 (pass_data_if_conversion): Add TODO_verify_ssa.
8011 2013-08-27 DJ Delorie <dj@redhat.com>
8013 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8015 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8017 * function.c (assign_parm_find_data_types): Set passed_mode and
8018 nominal_mode to the TYPE_MODE of nominal_type for the built
8019 pointer type in case of the struct-pass-by-reference.
8021 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8023 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8024 (UINT16_TYPE): Change default to "unsigned int".
8026 * config/avr/avr.opt (mfract-convert-truncate): New option.
8027 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8028 is set, round negative fractional integers according to n1169
8029 when converting to integer types.
8031 2013-08-26 Jan Hubicka <jh@suse.cz>
8033 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8034 methods can not be called indirectly when their address is not taken.
8036 2013-08-26 Jan Hubicka <jh@suse.cz>
8038 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8041 2013-08-26 Jan Hubicka <jh@suse.cz>
8043 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8046 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8048 * reload.c (find_valid_class): Allow classes that do not include
8049 FIRST_PSEUDO_REGISTER - 1.
8051 2013-08-26 Jan Hubicka <jh@suse.cz>
8053 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8054 fix edge count/frequency when speculation failed; fix type check
8055 for the direct call.
8057 2013-08-26 Jan Hubicka <jh@suse.cz>
8059 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8061 2013-08-26 Jan Hubicka <jh@suse.cz>
8063 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8064 to be changed by edge redirection.
8066 2013-08-26 Jan Hubicka <jh@suse.cz>
8068 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8069 formating; add sanity check.
8070 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8071 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8072 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8073 (resolve_noninline_speculation): Update callee keys, too.
8075 2013-08-26 Jan Hubicka <jh@suse.cz>
8077 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8078 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8080 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8082 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8085 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8087 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8088 (x86_64_elf_select_section): Ditto.
8089 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8091 (ix86_pass_by_reference): Ditto.
8092 (output_set_got): Ditto.
8093 (ix86_unary_operator_ok): Ditto.
8094 (ix86_expand_builtin): Ditto.
8096 2013-08-23 Jan Hubicka <jh@suse.cz>
8098 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8100 2013-08-23 Jan Hubicka <jh@suse.cz>
8102 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8103 (tree_decl_with_vis): Add FINAL field.
8105 2013-08-23 Jeff Law <law@redhat.com>
8107 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8108 the debugging dump when the expression is fully redundant.
8110 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8112 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8113 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8114 * pretty-print.c (pp_formatted_text_data): Likewise.
8115 (pp_write_text_to_stream): Likewise.
8116 (pp_write_text_as_dot_label_to_stream): Likewise.
8117 (pp_append_r): Likewise.
8118 (pp_format): Likewise.
8119 (pp_flush): Likewise.
8120 (pp_clear_output_area): Likewise.
8121 (pp_append_text): Likewise.
8122 (pp_formatted_text): Likewise.
8123 (pp_remaining_character_count_for_line): Likewise.
8124 (pp_newline): Likewise.
8125 (pp_character): Likewise.
8126 (output_buffer::~output_buffer): Define.
8127 (pretty_printer::~pretty_printer): Destruct output buffer.
8128 * pretty-print.h (output_buffer::~output_buffer): Declare.
8129 (pretty_printer::~pretty_printer): Declare virtual.
8131 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8134 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8135 HOST_WIDE_INT_M1U): New macros.
8136 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8137 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8138 unsigned -1 for lshift.
8139 * cse.c (cse_insn): Likewise.
8140 * double-int.c (rshift_double, lshift_double): Likewise.
8141 * builtins.c (fold_builtin_bitop): Likewise.
8142 * combine.c (force_to_mode): Likewise.
8143 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8144 * simplify-rtx.c (simplify_const_unary_operation,
8145 simplify_const_binary_operation): Likewise.
8146 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8147 check_va_list_escapes): Likewise.
8148 * rtlanal.c (nonzero_bits1): Likewise.
8149 * expmed.c (expand_smod_pow2): Likewise.
8150 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8152 2013-08-23 Jan Hubicka <jh@suse.cz>
8154 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8155 as having address taken.
8157 2013-08-23 Jan Hubicka <jh@suse.cz>
8159 * ipa-utils.h (method_class_type): Declare.
8160 * ipa-devirt.c (method_class_type): Export.
8162 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8163 do not walk base classes of anonymous types.
8165 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8167 PR rtl-optimization/58220
8169 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8170 handle SEQUENCE insns properly.
8172 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8174 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8176 (pp_newline_and_indent): Likewise.
8177 (pp_separate_with): Likewise.
8178 * pretty-print.c (pp_newline_and_flush): Define.
8179 (pp_newline_and_indent): Likewise.
8180 (pp_separate_with): Likewise.
8182 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8185 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8186 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8188 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8190 * config/i386/predicates.md (ext_sse_reg_operand): New.
8191 * config/i386/i386.md (*movti_internal): Use
8192 predicate to determine if EVEX is needed.
8193 (*movsi_internal): Ditto.
8194 (*movdf_internal): Ditto.
8195 (*movsf_internal): Ditto.
8196 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8198 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8200 PR tree-optimization/58209
8201 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8202 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8203 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8204 emit POINTER_PLUS_EXPR.
8205 (create_tailcall_accumulator): For pointer result type accumulate in
8208 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8210 * configure.ac: Add backslashes missing from the last change.
8211 * configure: Regenerate.
8213 2013-08-22 Jan Hubicka <jh@susue.cz>
8215 * ipa.c (function_and_variable_visibility): First remember function
8216 was global and then make it local.
8218 2013-08-22 Julian Brown <julian@codesourcery.com>
8220 * configure.ac: Add aarch64 to list of arches which use "nop" in
8222 * configure: Regenerate.
8224 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8226 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8227 gnu_libc_has_function.
8228 * config/s390/tpf.h: Likewise.
8230 2013-08-22 Jan Hubicka <jh@susue.cz>
8232 * timevar.c (validate_phases): Add cast.
8234 2013-08-22 Jan Hubicka <jh@susue.cz>
8236 * timevar.c (validate_phases): Use size_t for memory.
8237 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8239 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8241 * pretty-print.h (output_buffer::output_buffer): Declare.
8242 (pretty_printer::pretty_printer): Likewise.
8243 (pp_construct): Remove.
8244 * pretty-print.c (output_buffer::output_buffer): Define.
8245 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8246 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8247 (print_gimple_expr): Likewise.
8248 (print_gimple_seq): Likewise.
8249 (gimple_dump_bb): Likewise.
8250 * sched-vis.c (dump_value_slim): Likewise.
8251 (dump_insn_slim): Likewise.
8252 (dump_rtl_slim): Likewise.
8253 (str_pattern_slim): Likewise.
8254 * tree-mudflap.c (mf_varname_tree): Likewise.
8255 * graph.c (print_graph_cfg): Likewise.
8256 (start_graph_dump): Likewise.
8257 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8259 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8260 pretty printer initialization.
8261 * coretypes.h (diagnostic_context): Remove superflous type alias
8263 (pretty_printer): Likewise. Declare directly as a class.
8264 (pretty_print_info): Remove declaration as class.
8265 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8266 and pp_clear_output_area.
8267 (asan_add_global): Likewise.
8269 2013-08-22 Jan Hubicka <jh@suse.cz>
8271 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8272 * ipa-utils.h (update_type_inheritance_graph): Declare.
8273 (possible_polymorphic_call_target_p): Declare.
8274 (possible_polymorphic_call_target_p): New.
8275 * ipa-devirt.c: Update toplevel comments.
8276 (cached_polymorphic_call_targets): Move up.
8277 (odr_type_d): Move ID down.
8278 (polymorphic_type_binfo_p): Update comment.
8279 (odr_hasher::remove): Likewise;
8280 (get_odr_type): Set anonymous_namespace.
8281 (dump_odr_type): Dump it.
8282 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8283 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8284 (record_binfo): Add comment.
8285 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8287 (devirt_node_removal_hook): Do not iCE when cache is freed.
8288 (possible_polymorphic_call_target_p): New predicate.
8289 (update_type_inheritance_graph): New function.
8291 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8292 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8293 Sergey Lega <sergey.s.lega@intel.com>
8294 Anna Tikhonova <anna.tikhonova@intel.com>
8295 Ilya Tocar <ilya.tocar@intel.com>
8296 Andrey Turetskiy <andrey.turetskiy@intel.com>
8297 Ilya Verbin <ilya.verbin@intel.com>
8298 Kirill Yukhin <kirill.yukhin@intel.com>
8299 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8301 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8302 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8303 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8304 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8305 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8306 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8307 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8308 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8309 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8310 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8311 OPT_mavx512pf, OPT_mavx512er cases.
8312 * config/i386/constraints.md (v): New constraint.
8313 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8314 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8315 (bit_AVX512CD): New.
8316 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8317 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8318 * config/i386/i386-c.c (ix86_target_macros_internal):
8319 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8321 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8322 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8323 * config/i386/i386.c (regclass_map, dbx_register_map)
8324 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8325 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8326 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8328 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8329 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8330 -mavx512pf options. Fix formatting.
8331 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8332 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8333 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8334 -mavx512cd, -mavx512pf options.
8335 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8336 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8337 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8339 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8340 (ix86_set_reg_reg_cost): Ditto.
8341 (x86_order_regs_for_local_alloc): Ditto.
8342 (MAX_VECT_LEN): Extend to 64-byte.
8343 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8344 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8345 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8346 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8347 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8348 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8349 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8350 (SSE_REG_MODE_P): Support new modes.
8351 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8352 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8353 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8354 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8355 (REG_CLASS_CONTENTS): Add new registers.
8356 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8357 (EXT_REX_SSE_REGNO_P): New.
8358 (HI_REGISTER_NAMES): Add new registers.
8359 * config/i386/i386.md: Define constants for new registers.
8360 (mode): Add new 512-bit modes.
8361 (prefix): Support evex prefix.
8362 (isa): Support avx512f, noavx512f, fma_avx512f.
8363 (ssemodesuffix): Add new 512-bit modes.
8365 (*movxi_internal_avx512f): Ditto.
8366 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8368 (*movsi_internal): Likewise.
8369 (*movdf_internal): Likewise.
8370 (*movsf_internal): Likewise.
8371 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8372 (<code><mode>3): Likewise.
8373 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8375 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8376 with the new constraint "v".
8377 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8379 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8380 with the new constraint "v".
8381 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8382 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8383 (*<plusminus_insn><mode>3): Likewise.
8384 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8385 (*mul<mode>3): Likewise.
8386 (<sse>_vmmul<mode>3): Likewise.
8387 (<sse>_div<mode>3): Likewise.
8388 (<sse>_vmdiv<mode>3): Likewise.
8389 (<sse>_sqrt<mode>2): Likewise.
8390 (<sse>_vmsqrt<mode>2): Likewise.
8391 (*<code><mode>3_finite): Likewise.
8392 (*<code><mode>3) <smaxmin>: Likewise.
8393 (<sse>_vm<code><mode>3): Likewise.
8394 (*<code><mode>3) <any_logic>: Likewise.
8395 (*fma_fmadd_<mode>): Likewise.
8396 (*fma_fmsub_<mode>): Likewise.
8397 (*fma_fnmadd_<mode>): Likewise.
8398 (*fma_fnmsub_<mode>): Likewise.
8399 (*fma_fmaddsub_<mode>): Likewise.
8400 (*fma_fmsubadd_<mode>): Likewise.
8401 (*fmai_fmadd_<mode>): Likewise.
8402 (*fmai_fmsub_<mode>): Likewise.
8403 (*fmai_fnmadd_<mode>): Likewise.
8404 (*fmai_fnmsub_<mode>): Likewise.
8405 (sse_cvtsi2ss): Likewise.
8406 (sse_cvtsi2ssq): Likewise.
8407 (sse_cvtss2si): Likewise.
8408 (sse_cvtss2si_2): Likewise.
8409 (sse_cvtss2siq): Likewise.
8410 (sse_cvtss2siq_2): Likewise.
8411 (sse_cvttss2si): Likewise.
8412 (sse_cvtss2siq_2): Likewise.
8413 (float<sseintvecmodelower><mode>2): Likewise.
8414 (sse2_cvtsd2si_2): Likewise.
8415 (sse2_cvtsd2siq_2): Likewise.
8416 (*<plusminus_insn><mode>3): Likewise.
8417 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8418 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8419 (ashr<mode>3): Likewise.
8420 (<shift_insn><mode>3): Likewise.
8421 (avx2_<code><mode>3): Likewise.
8422 (*avx2_<code><mode>3): Likewise.
8423 (*andnot<mode>3): Likewise.
8424 (*<code><mode>3) <any_logic>: Likewise.
8425 (abs<mode>2): Likewise.
8426 (avx2_permvar<mode>): Likewise.
8427 (avx2_perm<mode>_1): Likewise.
8428 (*avx_vpermilp<mode>): Likewise.
8429 (avx_vpermilvar<mode>3): Likewise.
8430 (avx2_ashrv<mode>): Likewise.
8431 (avx2_<shift_insn>v<mode>): Likewise.
8432 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8434 * doc/rtl.texi: Document XImode.
8436 2013-08-21 Jeff Law <law@redhat.com>
8438 * tree-flow.h (register_jump_thread): Pass vector of edges
8439 instead of each important edge.
8440 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8441 thread path into a vector and pass that to register_jump_thread.
8442 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8443 passed in edge vector to the current 3-edge form.
8446 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8448 * dce.c (fini_dce): Call df_analyze again just in case
8449 delete_unmarked_insns removed anything.
8451 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8453 * reload.h (struct reg_equivs): Rename to ..
8454 (struct reg_equivs_s): .. this.
8456 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8458 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8460 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8462 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8464 2013-08-21 Jeff Law <law@redhat.com>
8466 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8467 simplify assignments too. If the RHS collapses to a singleton
8468 range, then return the value for the range.
8470 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8472 * config/i386/sse.md (V16): Rename to...
8474 (mov<mode>): Update iterator name.
8475 (*mov<mode>_internal): Ditto.
8476 (push<mode>1): Ditto.
8477 (movmisalign<mode>): Ditto.
8479 2013-08-20 Jan Hubicka <jh@suse.cz>
8482 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8483 entry for direct edges.
8484 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8486 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8488 Revert my last two changes, r201865 and r201864:
8491 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8493 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8494 instances can own GC refs.
8496 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8497 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8498 (gcc::context::gt_pch_nx): Likewise.
8499 (gcc::context::gt_pch_nx): Likewise.
8500 * ggc.h (gt_ggc_mx <T>): New.
8501 (gt_pch_nx_with_op <T>): New.
8502 (gt_pch_nx <T>): New.
8503 * passes.c (opt_pass::gt_ggc_mx): New.
8504 (opt_pass::gt_pch_nx): New.
8505 (opt_pass::gt_pch_nx_with_op): New.
8506 (pass_manager::gt_ggc_mx): New.
8507 (pass_manager::gt_pch_nx): New.
8508 (pass_manager::gt_pch_nx_with_op): New.
8509 (pass_manager::operator new): Use
8510 ggc_internal_cleared_alloc_stat rather than xcalloc.
8511 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8512 (pass_manager::gt_ggc_mx): New.
8513 (pass_manager::gt_pch_nx): New.
8514 (pass_manager::gt_pch_nx_with_op): New.
8515 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8516 (opt_pass::operator new): New.
8517 (opt_pass::gt_ggc_mx): New.
8518 (opt_pass::gt_pch_nx): New.
8519 (opt_pass::gt_pch_nx_with_op): New.
8522 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8524 * Makefile.in (GTFILES): Add context.h.
8525 * context.c (gcc::context::operator new): New.
8526 (gcc::context::gt_ggc_mx): New.
8527 (gcc::context::gt_pch_nx): New.
8528 (gcc::context::gt_pch_nx): New.
8529 * context.h (gcc::context): Add GTY((user)) marking.
8530 (gcc::context::operator new): New.
8531 (gcc::context::gt_ggc_mx): New.
8532 (gcc::context::gt_pch_nx): New.
8533 (gcc::context::gt_pch_nx): New.
8534 (g): Add GTY marking.
8535 (gt_ggc_mx (gcc::context *)): New.
8536 (gt_pch_nx (gcc::context *)): New.
8537 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8538 void *cookie)): New.
8539 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8541 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8543 * dce.c (fini_dce): Call df_analyze again just in case
8544 delete_unmarked_insns removed anything.
8546 2013-08-20 Teresa Johnson <tejohnson@google.com>
8548 PR rtl-optimizations/57451
8549 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8550 from crossing split section boundaries.
8552 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8554 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8555 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8556 with MULTLIB_DEFAULTS.
8558 2013-08-20 Nick Clifton <nickc@redhat.com>
8560 * target.def (narrow_volatile_bitfield): Note that the default
8561 value is false, not !TARGET_STRICT_ALIGN.
8562 * doc/tm.texi: Regenerate.
8564 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8566 Fix LIB_SPEC for systems without libpthread.
8568 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8569 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8571 * config/i386/linux-common.h: Likewise.
8572 * config/mips/linux-common.h: Likewise.
8574 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8576 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8579 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8581 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8582 instances can own GC refs.
8584 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8585 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8586 (gcc::context::gt_pch_nx): Likewise.
8587 (gcc::context::gt_pch_nx): Likewise.
8588 * ggc.h (gt_ggc_mx <T>): New.
8589 (gt_pch_nx_with_op <T>): New.
8590 (gt_pch_nx <T>): New.
8591 * passes.c (opt_pass::gt_ggc_mx): New.
8592 (opt_pass::gt_pch_nx): New.
8593 (opt_pass::gt_pch_nx_with_op): New.
8594 (pass_manager::gt_ggc_mx): New.
8595 (pass_manager::gt_pch_nx): New.
8596 (pass_manager::gt_pch_nx_with_op): New.
8597 (pass_manager::operator new): Use
8598 ggc_internal_cleared_alloc_stat rather than xcalloc.
8599 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8600 (pass_manager::gt_ggc_mx): New.
8601 (pass_manager::gt_pch_nx): New.
8602 (pass_manager::gt_pch_nx_with_op): New.
8603 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8604 (opt_pass::operator new): New.
8605 (opt_pass::gt_ggc_mx): New.
8606 (opt_pass::gt_pch_nx): New.
8607 (opt_pass::gt_pch_nx_with_op): New.
8609 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8611 * Makefile.in (GTFILES): Add context.h.
8612 * context.c (gcc::context::operator new): New.
8613 (gcc::context::gt_ggc_mx): New.
8614 (gcc::context::gt_pch_nx): New.
8615 (gcc::context::gt_pch_nx): New.
8616 * context.h (gcc::context): Add GTY((user)) marking.
8617 (gcc::context::operator new): New.
8618 (gcc::context::gt_ggc_mx): New.
8619 (gcc::context::gt_pch_nx): New.
8620 (gcc::context::gt_pch_nx): New.
8621 (g): Add GTY marking.
8622 (gt_ggc_mx (gcc::context *)): New.
8623 (gt_pch_nx (gcc::context *)): New.
8624 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8625 void *cookie)): New.
8626 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8628 2013-08-20 Alan Modra <amodra@gmail.com>
8631 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8632 (rs6000_emit_epilogue): Likewise.
8634 2013-08-19 Dehao Chen <dehao@google.com>
8636 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8638 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8639 Jakub Jelinek <jakub@redhat.com>
8641 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8642 (BUILT_IN_FABSD64): Likewise.
8643 (BUILT_IN_FABSD128): Likewise.
8644 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8645 (fold_builtin_1): Likewise.
8646 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8647 destination and source operands.
8648 (*abstd2_fpr): Likewise.
8649 (*nabstd2_fpr): Likewise.
8651 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8653 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8656 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8658 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8661 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8663 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8664 * builtins.c (default_libc_has_function): New.
8665 (gnu_libc_has_function): Ditto.
8666 (no_c99_libc_has_function): Ditto.
8667 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8668 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8669 (fold_builtin_sincos): Likewise.
8670 (fold_builtin_cexp): Likewise.
8671 * builtins.def (DEF_C94_BUILTIN): Likewise.
8672 (DEF_C99_BUILTIN): Likewise.
8673 (DEF_C99_C90RES_BUILTIN): Likewise.
8674 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8675 definitions to using this define.
8676 * config/darwin-protos.h (darwin_libc_has_function): New.
8677 * config/darwin.c (darwin_libc_has_function): Ditto.
8678 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8679 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8680 * config/darwin.h: Ditto.
8681 * config/elfos.h: Ditto.
8682 * config/freebsd.h: Ditto.
8683 * config/i386/cygming.h: Ditto.
8684 * config/i386/djgpp.h: Ditto.
8685 * config/i386/i386-interix.h: Ditto.
8686 * config/microblaze/microblaze.h: Ditto.
8687 * config/mmix/mmix.h: Ditto.
8688 * config/gnu-user.h: Ditto.
8689 * config/ia64/hpux.h: Ditto.
8690 * config/pa/pa-hpux.h: Ditto.
8691 * config/pdp11/pdp11.h: Ditto.
8692 * config/picochip/picochip.h: Ditto.
8693 * config/linux.h: Ditto.
8694 * config/netbsd.h: Ditto.
8695 * config/openbsd.h: Ditto.
8696 * config/rs6000/aix43.h: Ditto.
8697 * config/rs6000/aix51.h: Ditto.
8698 * config/rs6000/aix52.h: Ditto.
8699 * config/rs6000/aix53.h: Ditto.
8700 * config/rs6000/aix61.h: Ditto.
8701 * config/rs6000/darwin.h: Ditto.
8702 * config/rs6000/linux.h: Ditto.
8703 * config/rs6000/linux64.h: Ditto.
8704 * config/s390/tpf.h: Ditto.
8705 * config/sol2-10.h: Ditto.
8706 * config/sol2.h: Ditto.
8707 * config/vms/vms.h: Ditto.
8708 * config/vxworks.h: Ditto.
8709 * config/linux-android.c (linux_android_libc_has_function):
8710 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8711 * config/linux-protos.h (linux_android_libc_has_function):
8713 * config/i386/i386.c (ix86_libc_has_function): New.
8714 * config/i386/i386-protos.h
8715 (ix86_libc_has_function): New declaration.
8716 * config/i386/i386.md
8717 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8718 ("isinf<mode>2): Likewise.
8719 * convert.c (convert_to_integer): Using new target hook
8720 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8721 TARGET_C99_FUNCTIONS.
8722 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8723 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8724 * coretypes.h (function_class): New enum for different
8725 classes of functions.
8726 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8727 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8728 (TARGET_HAS_SINCOS): Likewise.
8729 (TARGET_LIBC_HAS_FUNCTION): New.
8730 * doc/tm.texi: Regenerated.
8731 * targhooks.h (default_libc_has_function): New declaration.
8732 (no_c99_libc_has_function): Ditto.
8733 (gnu_libc_has_function): Ditto.
8734 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8735 and TARGET_HAS_SINCOS.
8737 2013-08-18 Jan Hubicka <jh@suse.cz>
8739 * Makeifle-in (ipa-devirt.o): New.
8740 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8741 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8742 (analyze_functions): Look into possible targets of polymorphic call.
8743 * dumpfile.c (dump_files): Add type-inheritance dump.
8744 * dumpfile.h (TDI_inheritance): New.
8745 * ipa-devirt.c: New file.
8746 * ipa-utils.h (odr_type_d): Forward declare.
8747 (odr_type): New type.
8748 (build_type_inheritance_graph): Declare.
8749 (possible_polymorphic_call_targets): Declare and introduce inline
8750 variant when only edge is pased.
8751 (dump_possible_polymorphic_call_targets): Likewise.
8752 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8753 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8754 (types_same_for_odr): ... here.
8755 * tree.h (type_in_anonymous_namespace_p): Declare.
8757 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8759 PR tree-optimization/58006
8760 * tree-parloops.c (take_address_of): Don't ICE if get_name
8762 (eliminate_local_variables_stmt): Remove clobber stmts.
8764 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8766 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8767 error message is issued for an alias to undefined symbol.
8769 2013-08-18 Jan Hubicka <jh@suse.cz>
8771 * cgraph.c (cgraph_create_indirect_edge): Discover
8772 polymorphic calls and record basic info into indirect_info.
8773 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8774 devirtualization, ignore objc function calls.
8775 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8776 call with no parm index info.
8777 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8778 * tree.c (virtual_method_call_p): New function.
8779 * tree.h (virtual_method_call_p): Declare.
8781 2013-08-16 Jan Hubicka <jh@suse.cz>
8784 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8786 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8788 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8791 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8793 * gengtype.c (type_for_name): Add special-case support for
8794 locating types within the "gcc::" namespace.
8795 (open_base_files): Emit a "using namespace gcc" directive.
8797 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8800 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8801 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8803 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8804 array instead of each individual operand as a separate argument.
8805 (emit_fusion_gpr_load): Likewise.
8806 (expand_fusion_gpr_load): Add new function declaration.
8808 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8809 signature to have the operands passed as an array, instead of as
8810 separate arguments. Allow ZERO_EXTEND to be in the memory
8811 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8812 depend on the register live/dead flags when peepholes are run.
8813 (expand_fusion_gpr_load): New function to be called from the
8814 peephole2 pass, to change the register that addis sets to be the
8816 (emit_fusion_gpr_load): Change the calling signature to have the
8817 operands passed as an array, instead of as separate arguments.
8818 Allow ZERO_EXTEND to be in the memory address, and also
8819 SIGN_EXTEND if -mpower8-fusion-sign.
8821 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8823 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8824 adjust the register addis loads up in the peephole2 pass. Do not
8825 depend on the register live/dead state when the peephole pass is done.
8827 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8829 * gengtype.c (create_user_defined_type): Ensure that the kind
8830 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8831 declaration is seen before the GTY((user)) marking.
8833 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8836 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8838 2013-08-16 Jan Hubicka <jh@suse.cz>
8840 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8841 arugment expected_type.
8842 (gimple_fold_call): Use it.
8843 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8844 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8845 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8846 (try_make_edge_direct_virtual_call): Likewise.
8847 * tree.c (obj_type_ref_class): New.
8848 * tree.h (obj_type_ref_class): Use it.
8850 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8852 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8853 (rtl_slim_pp): Likewise.
8854 (init_rtl_slim_pretty_print): Likewise.
8855 (dump_value_slim): Don't call it. Use local pretty printer.
8856 (dump_insn_slim): Likewise.
8857 (dump_rtl_slim): Likewise.
8858 (str_pattern_slim): Likewise.
8859 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8862 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8864 PR tree-optimization/58164
8865 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8866 walk gimple_goto_dest of GIMPLE_GOTO.
8868 PR tree-optimization/58165
8869 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8870 bi_call must be the last stmt in a bb, don't split_block, instead
8871 use fallthru edge from it and give up if there is none.
8872 Release conds vector when returning early.
8874 2013-08-14 Xinliang David Li <davidxl@google.com>
8876 * config/i386/i386.c (ix86_option_override_internal):
8877 Remove unused variable and field.
8879 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8882 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8883 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8884 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8885 and Linux, correct BLKmode alignment when 128-bit alignment is
8886 required and compatibility flag is not set.
8887 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8888 for zero-size arguments when compatibility flag is not set.
8890 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8892 PR tree-optimization/58145
8893 * tree-sra.c (build_ref_for_offset): If prev_base has
8894 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8896 2013-08-14 Xinliang David Li <davidxl@google.com>
8898 * config/i386/i386.c (ix86_option_override_internal):
8899 Fix uninitialized variable error.
8901 2013-08-14 Xinliang David Li <davidxl@google.com>
8903 * config/i386/i386.opt: Define two new options.
8904 * config/i386/x86-tune.def: Add arch selector field in macros.
8905 * config/i386/i386.h: Adjust macro definition.
8906 * config/i386/i386.c (ix86_option_override_internal):
8908 (parse_mtune_ctrl_str): New function.
8909 (set_ix86_tune_features): New function.
8910 (ix86_function_specific_restore): Call the new helper function.
8912 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8914 PR rtl-optimization/57662
8915 * sel-sched.c (code_motion_process_successors): When the current insn
8916 is removed after the recursive traversal, break from the loop.
8917 Add comments and debug printouts.
8919 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8920 Alexandre Oliva <aoliva@redhat.com>
8923 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8924 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8925 there also UNSPEC_PLTOFF.
8927 2013-08-14 Marek Polacek <polacek@redhat.com>
8929 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8930 NUM_CONDITIONS bit positions.
8932 2013-08-13 Cary Coutant <ccoutant@google.com>
8934 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8935 (attr_checksum): Hash vector contents instead of pointer.
8936 (attr_checksum_ordered): Likewise.
8938 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8940 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8941 when Pmode != word_mode. Add length_address attribute.
8942 (sse3_monitor_<mode>): Merge from sse3_monitor and
8943 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8944 Pmode != word_mode. Update insn length attribute.
8945 * config/i386/i386.c (ix86_option_override_internal): Update
8946 ix86_gen_monitor selection for merged sse3_monitor insn.
8948 2013-08-13 Julian Brown <julian@codesourcery.com>
8950 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8951 perform invalid legitimization on greater-than-word-size modes for
8954 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8956 * ira.c (setup_class_translate_array): Use aclass instead of cl
8957 for classes not fully covered by allocno classes.
8959 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8961 PR tree-optimization/57661
8962 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8963 * tree-inline.c (tree_function_versioning): Initialize it.
8964 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8965 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8966 that is not being copied.
8969 * asan.c (instrument_strlen_call): Fix typo in comment.
8970 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8972 2013-08-13 Steve Ellcey <sellcey@mips.com>
8974 * config/mips/mips.md (prefetch): Use lw instead of ld on
8975 loongson in 32bit mode.
8977 2013-08-13 Nick Clifton <nickc@redhat.com>
8979 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8981 2013-08-13 Jan Hubicka <jh@suse.cz>
8983 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8984 introduced edge; fix typo in sanity check.
8985 (cgraph_resolve_speculation): Export; improve diagnostic.
8986 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8987 speculation at type mismatch.
8988 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8989 (cgraph_resolve_speculation): Declare.
8990 (symtab_can_be_discarded): New function.
8991 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8992 * ipa-inline-transform.c (speculation_removed): New global var.
8993 (clone_inlined_nodes): See if speculation can be removed.
8994 (inline_call): If speculations was removed, we growths may not match.
8995 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8996 (speculation_useful_p): New function.
8997 (resolve_noninline_speculation): New function.
8998 (inline_small_functions): Resolve useless speculations.
8999 * ipa-inline.h (speculation_useful_p): Declare
9000 * ipa.c (can_replace_by_local_alias): Simplify.
9001 (ipa_profile): Produce speculative calls in non-lto, too;
9002 add simple cost model; produce local aliases.
9004 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9006 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9009 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9011 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9013 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9015 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9016 iterator instead of X87MODEF.
9018 2013-08-12 Perez Read <netfirewall@gmail.com>
9021 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9022 operand 0 for intel asm alternative.
9023 (*movabs<mode>_2): Ditto for operand 1.
9025 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9027 * config/aarch64/arm_none.h
9028 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9030 2013-08-12 Nick Clifton <nickc@redhat.com>
9032 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9034 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9036 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9037 expand for QI/HImode operand to produce more effictive code for
9038 unsigned char(short) --> float(double) conversion.
9040 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9042 * doc/invoke.texi: Mention that -ftls-model does not force the final
9045 2013-08-12 Marek Polacek <polacek@redhat.com>
9046 Marc Glisse <marc.glisse@inria.fr>
9048 PR tree-optimization/57980
9049 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9050 when creating -1 constant.
9052 2013-08-10 Jan Hubicka <jh@suse.cz>
9054 Workaround binutils PR14342.
9055 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9056 (gimple_init_edge_profiler): Likewise.
9057 (gimple_gen_ic_func_profiler): Likewise.
9059 2013-08-09 Jan Hubicka <jh@suse.cz>
9061 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9063 2013-08-09 Xinliang David Li <davidxl@google.com>
9065 * config/i386/stringop.def: New file.
9066 * config/i386/stringop.opt: New file.
9067 * config/i386/i386-opts.h: Include stringopt.def.
9068 * config/i386/i386.opt: Include stringopt.opt.
9069 * config/i386/i386.c (ix86_option_override_internal):
9070 Override default size based stringop inline strategies with options.
9071 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9074 2013-08-09 Jan Hubicka <jh@suse.cz>
9076 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9078 2013-08-09 Jan Hubicka <jh@suse.cz>
9080 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9082 (dump_cgraph_node): Dump profile-id.
9083 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9084 and common_target_probability.
9085 * lto-cgraph.c (lto_output_edge): Stream common targets.
9086 (lto_output_node): Stream profile ids.
9087 (input_node): Stream profile ids.
9088 (input_edge): Stream common targets.
9089 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9090 * ipa.c: Include value-prof.h
9091 (ipa_profile_generate_summary): Turn indirect call statement histograms
9092 into common targets.
9093 (ipa_profile): Turn common targets into speculative edges.
9095 2013-08-09 Jan Hubicka <jh@suse.cz>
9097 * cgraph.h (cgraph_node): Add profile_id.
9098 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9099 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9100 (del_node_map): Update.
9101 (find_func_by_funcdef_no): Replace by ...
9102 (find_func_by_profile_id): ... this one.
9103 (gimple_ic_transform): Do not remove useful histograms when
9104 speculation is not done; dump info when indirect call removal
9106 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9107 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9108 (__gcov_indirect_call_profiler_v2): .. this one.
9109 * profile.h (init_node_map): Update.
9110 * coverage.c (coverage_compute_profile_id): New function.
9111 * coverage.h (coverage_compute_profile_id): Declare.
9112 * tree-profile.c (init_ic_make_global_vars): Make
9113 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9114 (gimple_init_edge_profiler): Update prototype of
9115 __gcov_indirect_call_profiler.
9116 (gimple_gen_ic_func_profiler): Simplify.
9117 (tree_profiling): Use init_node_map
9119 2013-08-09 Jan Hubicka <jh@suse.cz>
9121 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9122 non-speculative refs.
9123 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9124 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9125 (cgraph_set_call_stmt): Likewise.
9126 (cgraph_create_edge_1): Fix release checking compilatoin;
9128 (cgraph_free_edge): Free indirect info.
9129 (cgraph_turn_edge_to_speculative): New function.
9130 (cgraph_speculative_call_info): New function.
9131 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9132 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9133 (dump_cgraph_node): Dump speculation.
9134 (verify_edge_count_and_frequency): Accept speculative edges.
9135 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9136 (verify_cgraph_node): Handle speculation.
9137 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9138 (cgraph_set_call_stmt): Update prototype.
9139 (cgraph_make_edge_direct): Update prototype.
9140 (cgraph_speculative_call_info): Declare.
9141 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9142 to change; update call of ipa_find_references.
9143 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9144 lto_stmt_uid and speculative flags.
9145 (ipa_dump_references): Dump speculation.
9146 (ipa_clone_references): Clone speculative flag.
9147 (ipa_clone_referring): Likewise.
9148 (ipa_clone_ref): New function.
9149 (ipa_find_reference): Look into lto_stmt_uids
9150 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9151 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9152 (ipa_find_reference): Update declaration.
9153 (ipa_clone_ref): Declare.
9154 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9155 stream speculative flag.
9156 (lto_output_ref): Stream statements uids and speculation.
9157 (input_ref): Likewise.
9158 (input_edge): Stream speuclation.
9159 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9160 (cgraph_set_call_stmt_including_clones): Handle speculation.
9161 * ipa-inline.c (heap_edge_removal_hook): New function.
9162 (inline_small_functions): Register it.
9163 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9164 also initialize refs.
9165 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9167 (try_make_edge_direct_simple_call): Likewise.
9168 (try_make_edge_direct_simple_call): Likewise.
9169 (update_indirect_edges_after_inlining): Likewise.
9170 (remove_described_reference): Look proper lto_stmt_uid.
9171 (propagate_controlled_uses): Likewise.
9172 (propagate_controlled_uses): Liekwise.
9173 * tree-inline.c (copy_bb): Copy speculative edges.
9174 (redirect_all_calls): New function.
9175 (copy_cfg_body): Do redirection after loop info is updated.
9176 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9178 2013-08-09 Jan Hubicka <jh@suse.cz>
9180 * lto-streamer-out.c (output_function): Renumber PHIs.
9181 * lto-streamer-in.c (input_function): Likewise.
9183 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9185 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9186 (get_lane_unsigned): Likewise.
9187 (dup_lane_scalar): Likewise.
9188 (get_lane): enable for VALL.
9189 * config/aarch64/aarch64-simd.md
9190 (aarch64_dup_lane_scalar<mode>): Remove.
9191 (aarch64_get_lane_signed<mode>): Likewise.
9192 (aarch64_get_lane_unsigned<mode>): Likewise.
9193 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9194 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9195 (aarch64_get_lane<mode>): Enable for all vector modes.
9196 (aarch64_get_lanedi): Remove misleading constraints.
9197 * config/aarch64/arm_neon.h
9198 (__aarch64_vget_lane_any): Define.
9199 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9200 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9201 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9202 * config/aarch64/iterators.md (VDQQH): New.
9206 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9208 * configure.ac: Add GAS check for LEON instructions on SPARC.
9209 * configure: Regenerate.
9210 * config.in: Likewise.
9211 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9213 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9214 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9216 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9217 (AS_LEON_FLAG): New macro.
9218 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9219 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9220 Deal with LEON and LEON3 for the memory model.
9221 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9222 (atomic_compare_and_swap<mode>_1): Likewise.
9223 (*atomic_compare_and_swap<mode>_1): Likewise.
9225 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9227 * config/arm/neon.md (vcond): Fix floating-point vector
9228 comparisons against 0.
9230 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9232 * lra-constraints.c (emit_spill_move): Remove assert.
9233 (process_alt_operands): Add more debugging
9234 output. Increase reject for spilling into memory. Decrease
9235 reject for reloading scratch.
9236 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9238 2013-08-08 Steve Ellcey <sellcey@mips.com>
9240 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9241 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9242 micromips incompatible. Add nan2008.
9243 (MULTILIB_DIRNAMES): Add nan2008.
9244 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9245 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9246 and micromips incompatible. Add nan2008.
9247 (MULTILIB_DIRNAMES): Add nan2008.
9248 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9250 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9252 PR rtl-optimization/58079
9253 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9254 simplify_comparison has widened a comparison with an integer.
9256 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9258 * config/arm/neon.md (movmisalign<mode>): Disable when we
9259 don't allow unaligned accesses.
9260 (*movmisalign<mode>_neon_store): Likewise.
9261 (*movmisalign<mode>_neon_load): Likewise.
9262 (*movmisalign<mode>_neon_store): Likewise.
9263 (*movmisalign<mode>_neon_load): Likewise.
9265 2013-08-08 Jan Hubicka <jh@suse.cz>
9267 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9268 (make_pass_rebuild_cgraph_edges): Also clear references.
9269 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9270 * ipa-inline-transform.c (inline_transform): Remove all references
9272 * cgraphunit.c (expand_function): Remove all references after
9274 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9275 (ipa_find_reference): Rewrite to iterator.
9276 (remove_stmt_references): Likewise.
9277 (ipa_clear_stmts_in_references): New function.
9278 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9279 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9281 * ipa-split.c (split_function): Remove references in split function.
9283 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9286 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9287 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9289 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9292 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9293 suggested mode for the assignment isn't compatible with the
9296 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9299 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9301 2013-08-07 Xinliang David Li <davidxl@google.com>
9303 * config/i386/i386.opt: New option -mtune-ctrl=.
9304 * config/i386/x86-tune.def: New file.
9305 * config/i386/i386.h: include x86-tune.def.
9306 * config/i386/i386.c (ix86_option_override_internal):
9307 Parsing -mtune-ctrl= option and set tune features.
9309 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9312 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9313 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9316 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9318 * rtl.h (update_alignments): Declare.
9319 * final.c (grow_label_align): New function extracted from...
9320 (shorten_branches): ...here. Call it.
9321 (update_alignments): New function.
9322 * reorg.c (sibling_labels): New variable.
9323 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9324 the new label along with it onto the sibling_labels vector.
9325 (fill_simple_delay_slots): Adjust call to get_label_before.
9326 (fill_slots_from_thread): Likewise.
9327 (relax_delay_slots): Likewise.
9328 (make_return_insns): Likewise.
9329 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9331 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9333 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9334 document its semantics.
9335 (diagnostic_report_diagnostic): Adjust accordingly.
9337 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9339 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9340 (sparc_option_override): ...and port to new C++ pass API.
9341 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9343 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9345 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9347 2013-08-06 Caroline Tice <cmtice@google.com>
9349 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9350 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9351 * tree-pass.h: Add pass_vtable_verify.
9352 * varasm.c (assemble_variable): Add code to properly set the comdat
9353 section and name for the .vtable_map_vars section.
9354 (assemble_vtyv_preinit_initializer): New function.
9355 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9357 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9358 * vtable-verify.c: New file.
9359 * vtable-verify.h: New file.
9360 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9361 initialiation levels.
9362 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9363 * passes.def: Insert pass_vtable_verify.
9364 * aclocal.m4: Reorder includes.
9365 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9366 -fvtv-counts options.
9367 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9368 as appropriate, if -fvtable-verify=... is used.
9369 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9370 -fvtable-verify=... is used.
9371 * Makefile.in (OBJS): Add vtable-verify.o to list.
9372 (vtable-verify.o): Add new build rule.
9373 (GTFILES): Add vtable-verify.c to list.
9374 * common.opt (fvtable-verify=): New flag.
9375 (vtv_priority): Values for fvtable-verify= flag.
9376 (fvtv-counts): New flag.
9377 (fvtv-debug): New flag.
9378 * tree.h (save_vtable_map_decl): New extern function decl.
9380 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9382 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9383 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9384 (pass_data_rl78_devirt): ...new pass_data instance and...
9385 (make_pass_rl78_devirt): ...new function.
9386 (rl78_asm_file_start): Port pass registration to new C++ API.
9388 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9390 * coretypes.h (rtl_opt_pass): Add.
9391 (gcc::context): Add.
9392 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9393 (epiphany_init): Port to new C++ pass API.
9394 (epiphany_optimize_mode_switching): Likewise.
9395 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9396 (pass_manager::get_pass_mode_switching): New.
9397 (pass_manager::get_pass_peephole2): New.
9398 * mode-switching.c (pass_mode_switching): Add clone method.
9399 * recog.c (pass_peephole2): Add clone method.
9400 (pass_split_all_insns): Add clone method.
9402 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9404 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9405 (mips_option_override): ...here, porting to new C++ API for passes.
9407 2013-08-06 Jan Hubicka <jh@suse.cz>
9409 * cgraph.c (cgraph_get_body): New function based on lto.c
9411 * cgraph.h (cgraph_get_body): Declare.
9412 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9414 * cgraphunit.c (expand_function): Get body prior expanding.
9415 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9416 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9418 * passes.c (do_per_function_toporder): Get body.
9419 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9420 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9422 2013-08-06 Martin Jambor <mjambor@suse.cz>
9425 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9426 is not re-finalized. Rename second parameter to no_collect.
9428 2013-08-06 Martin Jambor <mjambor@suse.cz>
9431 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9432 MEM_REF has proper alignment information.
9434 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9437 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9439 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9440 rtx (*) (rtx, ...) with insn_gen_fn.
9441 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9442 insn_gen_fn::stored_funcptr. Add initializer braces.
9444 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9446 Rewrite how instances of passes are cloned to remove assumptions
9447 about their sizes (thus allowing pass subclasses to have
9448 additional data fields, albeit non-GC-managed ones at this point).
9450 * passes.c (make_pass_instance): Now that passes have clone
9451 methods, rewrite this function to eliminate XNEW and memcpy
9452 calls that used hardcoded sizes. Since this function no longer
9453 creates pass instances, rename it to...
9454 (add_pass_instance): ...this. Document the old way that passes were
9455 numbered and flagged, and rework this function to continue using it.
9456 (next_pass_1): Add an initial_pass argument for use by
9458 (position_pass): When adding multiple instances of a pass, use
9459 the pass's clone method, rather than relying on the XNEW/memcpy
9460 within the former make_pass_instance (now add_pass_instance).
9461 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9462 the initial instance of the current pass within the pass manager.
9464 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9466 This is the automated part of the conversion of passes from C
9467 structs to C++ classes.
9469 Patch autogenerated by refactor_passes.py from
9470 https://github.com/davidmalcolm/gcc-refactoring-scripts
9471 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9473 * asan.c (pass_asan): Convert from a global struct to a subclass of
9474 gimple_opt_pass along with...
9475 (pass_data_asan): ...new pass_data instance and...
9476 (make_pass_asan): ...new function.
9477 (pass_asan_O0): Convert from a global struct to a subclass of
9478 gimple_opt_pass along with...
9479 (pass_data_asan_O0): ...new pass_data instance and...
9480 (make_pass_asan_O0): ...new function.
9481 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9482 subclass of rtl_opt_pass along with...
9483 (pass_data_inc_dec): ...new pass_data instance and...
9484 (make_pass_inc_dec): ...new function.
9485 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9486 a subclass of rtl_opt_pass along with...
9487 (pass_data_reorder_blocks): ...new pass_data instance and...
9488 (make_pass_reorder_blocks): ...new function.
9489 (pass_duplicate_computed_gotos): Convert from a global struct to a
9490 subclass of rtl_opt_pass along with...
9491 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9492 (make_pass_duplicate_computed_gotos): ...new function.
9493 (pass_partition_blocks): Convert from a global struct to a subclass of
9494 rtl_opt_pass along with...
9495 (pass_data_partition_blocks): ...new pass_data instance and...
9496 (make_pass_partition_blocks): ...new function.
9497 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9498 struct to a subclass of rtl_opt_pass along with...
9499 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9501 (make_pass_branch_target_load_optimize1): ...new function.
9502 (pass_branch_target_load_optimize2): Convert from a global struct to a
9503 subclass of rtl_opt_pass along with...
9504 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9506 (make_pass_branch_target_load_optimize2): ...new function.
9507 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9508 of rtl_opt_pass along with...
9509 (pass_data_jump): ...new pass_data instance and...
9510 (make_pass_jump): ...new function.
9511 (pass_jump2): Convert from a global struct to a subclass of
9512 rtl_opt_pass along with...
9513 (pass_data_jump2): ...new pass_data instance and...
9514 (make_pass_jump2): ...new function.
9515 * cfgexpand.c (pass_expand): Convert from a global struct to a
9516 subclass of rtl_opt_pass along with...
9517 (pass_data_expand): ...new pass_data instance and...
9518 (make_pass_expand): ...new function.
9519 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9520 of rtl_opt_pass along with...
9521 (pass_data_free_cfg): ...new pass_data instance and...
9522 (make_pass_free_cfg): ...new function.
9523 (pass_into_cfg_layout_mode): Convert from a global struct to a
9524 subclass of rtl_opt_pass along with...
9525 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9526 (make_pass_into_cfg_layout_mode): ...new function.
9527 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9528 subclass of rtl_opt_pass along with...
9529 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9530 (make_pass_outof_cfg_layout_mode): ...new function.
9531 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9532 struct to a subclass of gimple_opt_pass along with...
9533 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9534 (make_pass_build_cgraph_edges): ...new function.
9535 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9536 subclass of gimple_opt_pass along with...
9537 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9538 (make_pass_rebuild_cgraph_edges): ...new function.
9539 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9540 subclass of gimple_opt_pass along with...
9541 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9543 (make_pass_remove_cgraph_callee_edges): ...new function.
9544 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9545 struct to a subclass of rtl_opt_pass along with...
9546 (pass_data_stack_adjustments): ...new pass_data instance and...
9547 (make_pass_stack_adjustments): ...new function.
9548 * combine.c (pass_combine): Convert from a global struct to a subclass
9549 of rtl_opt_pass along with...
9550 (pass_data_combine): ...new pass_data instance and...
9551 (make_pass_combine): ...new function.
9552 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9553 global struct to a subclass of rtl_opt_pass along with...
9554 (pass_data_compare_elim_after_reload): ...new pass_data instance
9556 (make_pass_compare_elim_after_reload): ...new function.
9557 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9558 of rtl_opt_pass along with...
9559 (pass_data_rtl_cprop): ...new pass_data instance and...
9560 (make_pass_rtl_cprop): ...new function.
9561 * cse.c (pass_cse): Convert from a global struct to a subclass of
9562 rtl_opt_pass along with...
9563 (pass_data_cse): ...new pass_data instance and...
9564 (make_pass_cse): ...new function.
9565 (pass_cse2): Convert from a global struct to a subclass of
9566 rtl_opt_pass along with...
9567 (pass_data_cse2): ...new pass_data instance and...
9568 (make_pass_cse2): ...new function.
9569 (pass_cse_after_global_opts): Convert from a global struct to a
9570 subclass of rtl_opt_pass along with...
9571 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9572 (make_pass_cse_after_global_opts): ...new function.
9573 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9574 of rtl_opt_pass along with...
9575 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9576 (make_pass_ud_rtl_dce): ...new function.
9577 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9578 rtl_opt_pass along with...
9579 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9580 (make_pass_fast_rtl_dce): ...new function.
9581 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9582 a subclass of rtl_opt_pass along with...
9583 (pass_data_df_initialize_opt): ...new pass_data instance and...
9584 (make_pass_df_initialize_opt): ...new function.
9585 (pass_df_initialize_no_opt): Convert from a global struct to a
9586 subclass of rtl_opt_pass along with...
9587 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9588 (make_pass_df_initialize_no_opt): ...new function.
9589 (pass_df_finish): Convert from a global struct to a subclass of
9590 rtl_opt_pass along with...
9591 (pass_data_df_finish): ...new pass_data instance and...
9592 (make_pass_df_finish): ...new function.
9593 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9594 rtl_opt_pass along with...
9595 (pass_data_rtl_dse1): ...new pass_data instance and...
9596 (make_pass_rtl_dse1): ...new function.
9597 (pass_rtl_dse2): Convert from a global struct to a subclass of
9598 rtl_opt_pass along with...
9599 (pass_data_rtl_dse2): ...new pass_data instance and...
9600 (make_pass_rtl_dse2): ...new function.
9601 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9602 subclass of rtl_opt_pass along with...
9603 (pass_data_dwarf2_frame): ...new pass_data instance and...
9604 (make_pass_dwarf2_frame): ...new function.
9605 * except.c (pass_set_nothrow_function_flags): Convert from a global
9606 struct to a subclass of rtl_opt_pass along with...
9607 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9609 (make_pass_set_nothrow_function_flags): ...new function.
9610 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9611 subclass of rtl_opt_pass along with...
9612 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9614 (make_pass_convert_to_eh_region_ranges): ...new function.
9615 * final.c (pass_compute_alignments): Convert from a global struct to a
9616 subclass of rtl_opt_pass along with...
9617 (pass_data_compute_alignments): ...new pass_data instance and...
9618 (make_pass_compute_alignments): ...new function.
9619 (pass_final): Convert from a global struct to a subclass of
9620 rtl_opt_pass along with...
9621 (pass_data_final): ...new pass_data instance and...
9622 (make_pass_final): ...new function.
9623 (pass_shorten_branches): Convert from a global struct to a subclass of
9624 rtl_opt_pass along with...
9625 (pass_data_shorten_branches): ...new pass_data instance and...
9626 (make_pass_shorten_branches): ...new function.
9627 (pass_clean_state): Convert from a global struct to a subclass of
9628 rtl_opt_pass along with...
9629 (pass_data_clean_state): ...new pass_data instance and...
9630 (make_pass_clean_state): ...new function.
9631 * function.c (pass_instantiate_virtual_regs): Convert from a global
9632 struct to a subclass of rtl_opt_pass along with...
9633 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9634 (make_pass_instantiate_virtual_regs): ...new function.
9635 (pass_leaf_regs): Convert from a global struct to a subclass of
9636 rtl_opt_pass along with...
9637 (pass_data_leaf_regs): ...new pass_data instance and...
9638 (make_pass_leaf_regs): ...new function.
9639 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9640 subclass of rtl_opt_pass along with...
9641 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9643 (make_pass_thread_prologue_and_epilogue): ...new function.
9644 (pass_match_asm_constraints): Convert from a global struct to a
9645 subclass of rtl_opt_pass along with...
9646 (pass_data_match_asm_constraints): ...new pass_data instance and...
9647 (make_pass_match_asm_constraints): ...new function.
9648 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9649 subclass of rtl_opt_pass along with...
9650 (pass_data_rtl_fwprop): ...new pass_data instance and...
9651 (make_pass_rtl_fwprop): ...new function.
9652 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9653 rtl_opt_pass along with...
9654 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9655 (make_pass_rtl_fwprop_addr): ...new function.
9656 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9657 rtl_opt_pass along with...
9658 (pass_data_rtl_pre): ...new pass_data instance and...
9659 (make_pass_rtl_pre): ...new function.
9660 (pass_rtl_hoist): Convert from a global struct to a subclass of
9661 rtl_opt_pass along with...
9662 (pass_data_rtl_hoist): ...new pass_data instance and...
9663 (make_pass_rtl_hoist): ...new function.
9664 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9665 subclass of gimple_opt_pass along with...
9666 (pass_data_lower_cf): ...new pass_data instance and...
9667 (make_pass_lower_cf): ...new function.
9668 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9669 from a global struct to a subclass of gimple_opt_pass along with...
9670 (pass_data_strength_reduction): ...new pass_data instance and...
9671 (make_pass_strength_reduction): ...new function.
9672 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9673 of rtl_opt_pass along with...
9674 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9675 (make_pass_rtl_ifcvt): ...new function.
9676 (pass_if_after_combine): Convert from a global struct to a subclass of
9677 rtl_opt_pass along with...
9678 (pass_data_if_after_combine): ...new pass_data instance and...
9679 (make_pass_if_after_combine): ...new function.
9680 (pass_if_after_reload): Convert from a global struct to a subclass of
9681 rtl_opt_pass along with...
9682 (pass_data_if_after_reload): ...new pass_data instance and...
9683 (make_pass_if_after_reload): ...new function.
9684 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9685 a subclass of rtl_opt_pass along with...
9686 (pass_data_initialize_regs): ...new pass_data instance and...
9687 (make_pass_initialize_regs): ...new function.
9688 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9689 of ipa_opt_pass_d along with...
9690 (pass_data_ipa_cp): ...new pass_data instance and...
9691 (make_pass_ipa_cp): ...new function.
9692 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9693 global struct to a subclass of gimple_opt_pass along with...
9694 (pass_data_inline_parameters): ...new pass_data instance and...
9695 (make_pass_inline_parameters): ...new function.
9696 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9697 subclass of gimple_opt_pass along with...
9698 (pass_data_early_inline): ...new pass_data instance and...
9699 (make_pass_early_inline): ...new function.
9700 (pass_ipa_inline): Convert from a global struct to a subclass of
9701 ipa_opt_pass_d along with...
9702 (pass_data_ipa_inline): ...new pass_data instance and...
9703 (make_pass_ipa_inline): ...new function.
9704 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9705 struct to a subclass of gimple_opt_pass along with...
9706 (pass_data_local_pure_const): ...new pass_data instance and...
9707 (make_pass_local_pure_const): ...new function.
9708 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9709 ipa_opt_pass_d along with...
9710 (pass_data_ipa_pure_const): ...new pass_data instance and...
9711 (make_pass_ipa_pure_const): ...new function.
9712 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9713 to a subclass of ipa_opt_pass_d along with...
9714 (pass_data_ipa_reference): ...new pass_data instance and...
9715 (make_pass_ipa_reference): ...new function.
9716 * ipa-split.c (pass_split_functions): Convert from a global struct to
9717 a subclass of gimple_opt_pass along with...
9718 (pass_data_split_functions): ...new pass_data instance and...
9719 (make_pass_split_functions): ...new function.
9720 (pass_feedback_split_functions): Convert from a global struct to a
9721 subclass of gimple_opt_pass along with...
9722 (pass_data_feedback_split_functions): ...new pass_data instance and...
9723 (make_pass_feedback_split_functions): ...new function.
9724 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9725 global struct to a subclass of simple_ipa_opt_pass along with...
9726 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9728 (make_pass_ipa_function_and_variable_visibility): ...new function.
9729 (pass_ipa_free_inline_summary): Convert from a global struct to a
9730 subclass of simple_ipa_opt_pass along with...
9731 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9732 (make_pass_ipa_free_inline_summary): ...new function.
9733 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9734 subclass of ipa_opt_pass_d along with...
9735 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9737 (make_pass_ipa_whole_program_visibility): ...new function.
9738 (pass_ipa_profile): Convert from a global struct to a subclass of
9739 ipa_opt_pass_d along with...
9740 (pass_data_ipa_profile): ...new pass_data instance and...
9741 (make_pass_ipa_profile): ...new function.
9742 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9743 ipa_opt_pass_d along with...
9744 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9745 (make_pass_ipa_cdtor_merge): ...new function.
9746 * ira.c (pass_ira): Convert from a global struct to a subclass of
9747 rtl_opt_pass along with...
9748 (pass_data_ira): ...new pass_data instance and...
9749 (make_pass_ira): ...new function.
9750 (pass_reload): Convert from a global struct to a subclass of
9751 rtl_opt_pass along with...
9752 (pass_data_reload): ...new pass_data instance and...
9753 (make_pass_reload): ...new function.
9754 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9755 subclass of rtl_opt_pass along with...
9756 (pass_data_cleanup_barriers): ...new pass_data instance and...
9757 (make_pass_cleanup_barriers): ...new function.
9758 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9759 of rtl_opt_pass along with...
9760 (pass_data_loop2): ...new pass_data instance and...
9761 (make_pass_loop2): ...new function.
9762 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9763 rtl_opt_pass along with...
9764 (pass_data_rtl_loop_init): ...new pass_data instance and...
9765 (make_pass_rtl_loop_init): ...new function.
9766 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9767 rtl_opt_pass along with...
9768 (pass_data_rtl_loop_done): ...new pass_data instance and...
9769 (make_pass_rtl_loop_done): ...new function.
9770 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9771 subclass of rtl_opt_pass along with...
9772 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9773 (make_pass_rtl_move_loop_invariants): ...new function.
9774 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9775 rtl_opt_pass along with...
9776 (pass_data_rtl_unswitch): ...new pass_data instance and...
9777 (make_pass_rtl_unswitch): ...new function.
9778 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9779 subclass of rtl_opt_pass along with...
9780 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9782 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9783 (pass_rtl_doloop): Convert from a global struct to a subclass of
9784 rtl_opt_pass along with...
9785 (pass_data_rtl_doloop): ...new pass_data instance and...
9786 (make_pass_rtl_doloop): ...new function.
9787 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9788 a subclass of rtl_opt_pass along with...
9789 (pass_data_lower_subreg): ...new pass_data instance and...
9790 (make_pass_lower_subreg): ...new function.
9791 (pass_lower_subreg2): Convert from a global struct to a subclass of
9792 rtl_opt_pass along with...
9793 (pass_data_lower_subreg2): ...new pass_data instance and...
9794 (make_pass_lower_subreg2): ...new function.
9795 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9796 struct to a subclass of ipa_opt_pass_d along with...
9797 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9798 (make_pass_ipa_lto_gimple_out): ...new function.
9799 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9800 of ipa_opt_pass_d along with...
9801 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9802 (make_pass_ipa_lto_finish_out): ...new function.
9803 * mode-switching.c (pass_mode_switching): Convert from a global struct
9804 to a subclass of rtl_opt_pass along with...
9805 (pass_data_mode_switching): ...new pass_data instance and...
9806 (make_pass_mode_switching): ...new function.
9807 * modulo-sched.c (pass_sms): Convert from a global struct to a
9808 subclass of rtl_opt_pass along with...
9809 (pass_data_sms): ...new pass_data instance and...
9810 (make_pass_sms): ...new function.
9811 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9812 subclass of gimple_opt_pass along with...
9813 (pass_data_expand_omp): ...new pass_data instance and...
9814 (make_pass_expand_omp): ...new function.
9815 (pass_lower_omp): Convert from a global struct to a subclass of
9816 gimple_opt_pass along with...
9817 (pass_data_lower_omp): ...new pass_data instance and...
9818 (make_pass_lower_omp): ...new function.
9819 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9820 of gimple_opt_pass along with...
9821 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9822 (make_pass_diagnose_omp_blocks): ...new function.
9823 * passes.c (pass_early_local_passes): Convert from a global struct to
9824 a subclass of simple_ipa_opt_pass along with...
9825 (pass_data_early_local_passes): ...new pass_data instance and...
9826 (make_pass_early_local_passes): ...new function.
9827 (pass_all_early_optimizations): Convert from a global struct to a
9828 subclass of gimple_opt_pass along with...
9829 (pass_data_all_early_optimizations): ...new pass_data instance and...
9830 (make_pass_all_early_optimizations): ...new function.
9831 (pass_all_optimizations): Convert from a global struct to a subclass
9832 of gimple_opt_pass along with...
9833 (pass_data_all_optimizations): ...new pass_data instance and...
9834 (make_pass_all_optimizations): ...new function.
9835 (pass_all_optimizations_g): Convert from a global struct to a subclass
9836 of gimple_opt_pass along with...
9837 (pass_data_all_optimizations_g): ...new pass_data instance and...
9838 (make_pass_all_optimizations_g): ...new function.
9839 (pass_rest_of_compilation): Convert from a global struct to a subclass
9840 of rtl_opt_pass along with...
9841 (pass_data_rest_of_compilation): ...new pass_data instance and...
9842 (make_pass_rest_of_compilation): ...new function.
9843 (pass_postreload): Convert from a global struct to a subclass of
9844 rtl_opt_pass along with...
9845 (pass_data_postreload): ...new pass_data instance and...
9846 (make_pass_postreload): ...new function.
9847 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9848 subclass of rtl_opt_pass along with...
9849 (pass_data_gcse2): ...new pass_data instance and...
9850 (make_pass_gcse2): ...new function.
9851 * postreload.c (pass_postreload_cse): Convert from a global struct to
9852 a subclass of rtl_opt_pass along with...
9853 (pass_data_postreload_cse): ...new pass_data instance and...
9854 (make_pass_postreload_cse): ...new function.
9855 * predict.c (pass_profile): Convert from a global struct to a subclass
9856 of gimple_opt_pass along with...
9857 (pass_data_profile): ...new pass_data instance and...
9858 (make_pass_profile): ...new function.
9859 (pass_strip_predict_hints): Convert from a global struct to a subclass
9860 of gimple_opt_pass along with...
9861 (pass_data_strip_predict_hints): ...new pass_data instance and...
9862 (make_pass_strip_predict_hints): ...new function.
9863 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9864 of rtl_opt_pass along with...
9865 (pass_data_peephole2): ...new pass_data instance and...
9866 (make_pass_peephole2): ...new function.
9867 (pass_split_all_insns): Convert from a global struct to a subclass of
9868 rtl_opt_pass along with...
9869 (pass_data_split_all_insns): ...new pass_data instance and...
9870 (make_pass_split_all_insns): ...new function.
9871 (pass_split_after_reload): Convert from a global struct to a subclass
9872 of rtl_opt_pass along with...
9873 (pass_data_split_after_reload): ...new pass_data instance and...
9874 (make_pass_split_after_reload): ...new function.
9875 (pass_split_before_regstack): Convert from a global struct to a
9876 subclass of rtl_opt_pass along with...
9877 (pass_data_split_before_regstack): ...new pass_data instance and...
9878 (make_pass_split_before_regstack): ...new function.
9879 (pass_split_before_sched2): Convert from a global struct to a subclass
9880 of rtl_opt_pass along with...
9881 (pass_data_split_before_sched2): ...new pass_data instance and...
9882 (make_pass_split_before_sched2): ...new function.
9883 (pass_split_for_shorten_branches): Convert from a global struct to a
9884 subclass of rtl_opt_pass along with...
9885 (pass_data_split_for_shorten_branches): ...new pass_data instance
9887 (make_pass_split_for_shorten_branches): ...new function.
9888 * ree.c (pass_ree): Convert from a global struct to a subclass of
9889 rtl_opt_pass along with...
9890 (pass_data_ree): ...new pass_data instance and...
9891 (make_pass_ree): ...new function.
9892 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9893 subclass of rtl_opt_pass along with...
9894 (pass_data_stack_regs): ...new pass_data instance and...
9895 (make_pass_stack_regs): ...new function.
9896 (pass_stack_regs_run): Convert from a global struct to a subclass of
9897 rtl_opt_pass along with...
9898 (pass_data_stack_regs_run): ...new pass_data instance and...
9899 (make_pass_stack_regs_run): ...new function.
9900 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9901 subclass of rtl_opt_pass along with...
9902 (pass_data_cprop_hardreg): ...new pass_data instance and...
9903 (make_pass_cprop_hardreg): ...new function.
9904 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9905 subclass of rtl_opt_pass along with...
9906 (pass_data_reginfo_init): ...new pass_data instance and...
9907 (make_pass_reginfo_init): ...new function.
9908 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9909 of rtl_opt_pass along with...
9910 (pass_data_regmove): ...new pass_data instance and...
9911 (make_pass_regmove): ...new function.
9912 * regrename.c (pass_regrename): Convert from a global struct to a
9913 subclass of rtl_opt_pass along with...
9914 (pass_data_regrename): ...new pass_data instance and...
9915 (make_pass_regrename): ...new function.
9916 * reorg.c (pass_delay_slots): Convert from a global struct to a
9917 subclass of rtl_opt_pass along with...
9918 (pass_data_delay_slots): ...new pass_data instance and...
9919 (make_pass_delay_slots): ...new function.
9920 (pass_machine_reorg): Convert from a global struct to a subclass of
9921 rtl_opt_pass along with...
9922 (pass_data_machine_reorg): ...new pass_data instance and...
9923 (make_pass_machine_reorg): ...new function.
9924 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9925 of rtl_opt_pass along with...
9926 (pass_data_sched): ...new pass_data instance and...
9927 (make_pass_sched): ...new function.
9928 (pass_sched2): Convert from a global struct to a subclass of
9929 rtl_opt_pass along with...
9930 (pass_data_sched2): ...new pass_data instance and...
9931 (make_pass_sched2): ...new function.
9932 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9933 to a subclass of rtl_opt_pass along with...
9934 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9935 (make_pass_stack_ptr_mod): ...new function.
9936 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9937 to a subclass of rtl_opt_pass along with...
9938 (pass_data_rtl_store_motion): ...new pass_data instance and...
9939 (make_pass_rtl_store_motion): ...new function.
9940 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9941 of gimple_opt_pass along with...
9942 (pass_data_tracer): ...new pass_data instance and...
9943 (make_pass_tracer): ...new function.
9944 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9945 to a subclass of gimple_opt_pass along with...
9946 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9947 (make_pass_diagnose_tm_blocks): ...new function.
9948 (pass_lower_tm): Convert from a global struct to a subclass of
9949 gimple_opt_pass along with...
9950 (pass_data_lower_tm): ...new pass_data instance and...
9951 (make_pass_lower_tm): ...new function.
9952 (pass_tm_init): Convert from a global struct to a subclass of
9953 gimple_opt_pass along with...
9954 (pass_data_tm_init): ...new pass_data instance and...
9955 (make_pass_tm_init): ...new function.
9956 (pass_tm_mark): Convert from a global struct to a subclass of
9957 gimple_opt_pass along with...
9958 (pass_data_tm_mark): ...new pass_data instance and...
9959 (make_pass_tm_mark): ...new function.
9960 (pass_tm_edges): Convert from a global struct to a subclass of
9961 gimple_opt_pass along with...
9962 (pass_data_tm_edges): ...new pass_data instance and...
9963 (make_pass_tm_edges): ...new function.
9964 (pass_tm_memopt): Convert from a global struct to a subclass of
9965 gimple_opt_pass along with...
9966 (pass_data_tm_memopt): ...new pass_data instance and...
9967 (make_pass_tm_memopt): ...new function.
9968 (pass_ipa_tm): Convert from a global struct to a subclass of
9969 simple_ipa_opt_pass along with...
9970 (pass_data_ipa_tm): ...new pass_data instance and...
9971 (make_pass_ipa_tm): ...new function.
9972 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9973 subclass of gimple_opt_pass along with...
9974 (pass_data_call_cdce): ...new pass_data instance and...
9975 (make_pass_call_cdce): ...new function.
9976 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9977 subclass of gimple_opt_pass along with...
9978 (pass_data_build_cfg): ...new pass_data instance and...
9979 (make_pass_build_cfg): ...new function.
9980 (pass_split_crit_edges): Convert from a global struct to a subclass of
9981 gimple_opt_pass along with...
9982 (pass_data_split_crit_edges): ...new pass_data instance and...
9983 (make_pass_split_crit_edges): ...new function.
9984 (pass_warn_function_return): Convert from a global struct to a
9985 subclass of gimple_opt_pass along with...
9986 (pass_data_warn_function_return): ...new pass_data instance and...
9987 (make_pass_warn_function_return): ...new function.
9988 (pass_warn_function_noreturn): Convert from a global struct to a
9989 subclass of gimple_opt_pass along with...
9990 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9991 (make_pass_warn_function_noreturn): ...new function.
9992 (pass_warn_unused_result): Convert from a global struct to a subclass
9993 of gimple_opt_pass along with...
9994 (pass_data_warn_unused_result): ...new pass_data instance and...
9995 (make_pass_warn_unused_result): ...new function.
9996 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9997 a subclass of gimple_opt_pass along with...
9998 (pass_data_merge_phi): ...new pass_data instance and...
9999 (make_pass_merge_phi): ...new function.
10000 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10001 a subclass of gimple_opt_pass along with...
10002 (pass_data_lower_complex): ...new pass_data instance and...
10003 (make_pass_lower_complex): ...new function.
10004 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10005 gimple_opt_pass along with...
10006 (pass_data_lower_complex_O0): ...new pass_data instance and...
10007 (make_pass_lower_complex_O0): ...new function.
10008 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10009 subclass of gimple_opt_pass along with...
10010 (pass_data_lower_eh): ...new pass_data instance and...
10011 (make_pass_lower_eh): ...new function.
10012 (pass_refactor_eh): Convert from a global struct to a subclass of
10013 gimple_opt_pass along with...
10014 (pass_data_refactor_eh): ...new pass_data instance and...
10015 (make_pass_refactor_eh): ...new function.
10016 (pass_lower_resx): Convert from a global struct to a subclass of
10017 gimple_opt_pass along with...
10018 (pass_data_lower_resx): ...new pass_data instance and...
10019 (make_pass_lower_resx): ...new function.
10020 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10021 of gimple_opt_pass along with...
10022 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10023 (make_pass_lower_eh_dispatch): ...new function.
10024 (pass_cleanup_eh): Convert from a global struct to a subclass of
10025 gimple_opt_pass along with...
10026 (pass_data_cleanup_eh): ...new pass_data instance and...
10027 (make_pass_cleanup_eh): ...new function.
10028 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10029 to a subclass of simple_ipa_opt_pass along with...
10030 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10031 (make_pass_ipa_lower_emutls): ...new function.
10032 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10033 a subclass of gimple_opt_pass along with...
10034 (pass_data_if_conversion): ...new pass_data instance and...
10035 (make_pass_if_conversion): ...new function.
10036 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10037 subclass of gimple_opt_pass along with...
10038 (pass_data_build_ssa): ...new pass_data instance and...
10039 (make_pass_build_ssa): ...new function.
10040 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10041 global struct to a subclass of gimple_opt_pass along with...
10042 (pass_data_loop_distribution): ...new pass_data instance and...
10043 (make_pass_loop_distribution): ...new function.
10044 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10045 subclass of gimple_opt_pass along with...
10046 (pass_data_mudflap_1): ...new pass_data instance and...
10047 (make_pass_mudflap_1): ...new function.
10048 (pass_mudflap_2): Convert from a global struct to a subclass of
10049 gimple_opt_pass along with...
10050 (pass_data_mudflap_2): ...new pass_data instance and...
10051 (make_pass_mudflap_2): ...new function.
10052 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10053 subclass of gimple_opt_pass along with...
10054 (pass_data_mudflap_1): ...new pass_data instance and...
10055 (make_pass_mudflap_1): ...new function.
10056 (pass_mudflap_2): Convert from a global struct to a subclass of
10057 gimple_opt_pass along with...
10058 (pass_data_mudflap_2): ...new pass_data instance and...
10059 (make_pass_mudflap_2): ...new function.
10060 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10061 gimple_opt_pass along with...
10062 (pass_data_nrv): ...new pass_data instance and...
10063 (make_pass_nrv): ...new function.
10064 (pass_return_slot): Convert from a global struct to a subclass of
10065 gimple_opt_pass along with...
10066 (pass_data_return_slot): ...new pass_data instance and...
10067 (make_pass_return_slot): ...new function.
10068 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10069 to a subclass of gimple_opt_pass along with...
10070 (pass_data_object_sizes): ...new pass_data instance and...
10071 (make_pass_object_sizes): ...new function.
10072 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10073 global struct to a subclass of gimple_opt_pass along with...
10074 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10076 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10077 (pass_fixup_cfg): Convert from a global struct to a subclass of
10078 gimple_opt_pass along with...
10079 (pass_data_fixup_cfg): ...new pass_data instance and...
10080 (make_pass_fixup_cfg): ...new function.
10081 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10082 (make_pass_mudflap_1): ...new function.
10083 (pass_mudflap_2): Replace declaration with that of...
10084 (make_pass_mudflap_2): ...new function.
10085 (pass_asan): Replace declaration with that of...
10086 (make_pass_asan): ...new function.
10087 (pass_asan_O0): Replace declaration with that of...
10088 (make_pass_asan_O0): ...new function.
10089 (pass_tsan): Replace declaration with that of...
10090 (make_pass_tsan): ...new function.
10091 (pass_tsan_O0): Replace declaration with that of...
10092 (make_pass_tsan_O0): ...new function.
10093 (pass_lower_cf): Replace declaration with that of...
10094 (make_pass_lower_cf): ...new function.
10095 (pass_refactor_eh): Replace declaration with that of...
10096 (make_pass_refactor_eh): ...new function.
10097 (pass_lower_eh): Replace declaration with that of...
10098 (make_pass_lower_eh): ...new function.
10099 (pass_lower_eh_dispatch): Replace declaration with that of...
10100 (make_pass_lower_eh_dispatch): ...new function.
10101 (pass_lower_resx): Replace declaration with that of...
10102 (make_pass_lower_resx): ...new function.
10103 (pass_build_cfg): Replace declaration with that of...
10104 (make_pass_build_cfg): ...new function.
10105 (pass_early_tree_profile): Replace declaration with that of...
10106 (make_pass_early_tree_profile): ...new function.
10107 (pass_cleanup_eh): Replace declaration with that of...
10108 (make_pass_cleanup_eh): ...new function.
10109 (pass_sra): Replace declaration with that of...
10110 (make_pass_sra): ...new function.
10111 (pass_sra_early): Replace declaration with that of...
10112 (make_pass_sra_early): ...new function.
10113 (pass_early_ipa_sra): Replace declaration with that of...
10114 (make_pass_early_ipa_sra): ...new function.
10115 (pass_tail_recursion): Replace declaration with that of...
10116 (make_pass_tail_recursion): ...new function.
10117 (pass_tail_calls): Replace declaration with that of...
10118 (make_pass_tail_calls): ...new function.
10119 (pass_tree_loop): Replace declaration with that of...
10120 (make_pass_tree_loop): ...new function.
10121 (pass_tree_loop_init): Replace declaration with that of...
10122 (make_pass_tree_loop_init): ...new function.
10123 (pass_lim): Replace declaration with that of...
10124 (make_pass_lim): ...new function.
10125 (pass_tree_unswitch): Replace declaration with that of...
10126 (make_pass_tree_unswitch): ...new function.
10127 (pass_predcom): Replace declaration with that of...
10128 (make_pass_predcom): ...new function.
10129 (pass_iv_canon): Replace declaration with that of...
10130 (make_pass_iv_canon): ...new function.
10131 (pass_scev_cprop): Replace declaration with that of...
10132 (make_pass_scev_cprop): ...new function.
10133 (pass_empty_loop): Replace declaration with that of...
10134 (make_pass_empty_loop): ...new function.
10135 (pass_record_bounds): Replace declaration with that of...
10136 (make_pass_record_bounds): ...new function.
10137 (pass_graphite): Replace declaration with that of...
10138 (make_pass_graphite): ...new function.
10139 (pass_graphite_transforms): Replace declaration with that of...
10140 (make_pass_graphite_transforms): ...new function.
10141 (pass_if_conversion): Replace declaration with that of...
10142 (make_pass_if_conversion): ...new function.
10143 (pass_loop_distribution): Replace declaration with that of...
10144 (make_pass_loop_distribution): ...new function.
10145 (pass_vectorize): Replace declaration with that of...
10146 (make_pass_vectorize): ...new function.
10147 (pass_slp_vectorize): Replace declaration with that of...
10148 (make_pass_slp_vectorize): ...new function.
10149 (pass_complete_unroll): Replace declaration with that of...
10150 (make_pass_complete_unroll): ...new function.
10151 (pass_complete_unrolli): Replace declaration with that of...
10152 (make_pass_complete_unrolli): ...new function.
10153 (pass_parallelize_loops): Replace declaration with that of...
10154 (make_pass_parallelize_loops): ...new function.
10155 (pass_loop_prefetch): Replace declaration with that of...
10156 (make_pass_loop_prefetch): ...new function.
10157 (pass_iv_optimize): Replace declaration with that of...
10158 (make_pass_iv_optimize): ...new function.
10159 (pass_tree_loop_done): Replace declaration with that of...
10160 (make_pass_tree_loop_done): ...new function.
10161 (pass_ch): Replace declaration with that of...
10162 (make_pass_ch): ...new function.
10163 (pass_ccp): Replace declaration with that of...
10164 (make_pass_ccp): ...new function.
10165 (pass_phi_only_cprop): Replace declaration with that of...
10166 (make_pass_phi_only_cprop): ...new function.
10167 (pass_build_ssa): Replace declaration with that of...
10168 (make_pass_build_ssa): ...new function.
10169 (pass_build_alias): Replace declaration with that of...
10170 (make_pass_build_alias): ...new function.
10171 (pass_build_ealias): Replace declaration with that of...
10172 (make_pass_build_ealias): ...new function.
10173 (pass_dominator): Replace declaration with that of...
10174 (make_pass_dominator): ...new function.
10175 (pass_dce): Replace declaration with that of...
10176 (make_pass_dce): ...new function.
10177 (pass_dce_loop): Replace declaration with that of...
10178 (make_pass_dce_loop): ...new function.
10179 (pass_cd_dce): Replace declaration with that of...
10180 (make_pass_cd_dce): ...new function.
10181 (pass_call_cdce): Replace declaration with that of...
10182 (make_pass_call_cdce): ...new function.
10183 (pass_merge_phi): Replace declaration with that of...
10184 (make_pass_merge_phi): ...new function.
10185 (pass_split_crit_edges): Replace declaration with that of...
10186 (make_pass_split_crit_edges): ...new function.
10187 (pass_pre): Replace declaration with that of...
10188 (make_pass_pre): ...new function.
10189 (pass_profile): Replace declaration with that of...
10190 (make_pass_profile): ...new function.
10191 (pass_strip_predict_hints): Replace declaration with that of...
10192 (make_pass_strip_predict_hints): ...new function.
10193 (pass_lower_complex_O0): Replace declaration with that of...
10194 (make_pass_lower_complex_O0): ...new function.
10195 (pass_lower_complex): Replace declaration with that of...
10196 (make_pass_lower_complex): ...new function.
10197 (pass_lower_vector): Replace declaration with that of...
10198 (make_pass_lower_vector): ...new function.
10199 (pass_lower_vector_ssa): Replace declaration with that of...
10200 (make_pass_lower_vector_ssa): ...new function.
10201 (pass_lower_omp): Replace declaration with that of...
10202 (make_pass_lower_omp): ...new function.
10203 (pass_diagnose_omp_blocks): Replace declaration with that of...
10204 (make_pass_diagnose_omp_blocks): ...new function.
10205 (pass_expand_omp): Replace declaration with that of...
10206 (make_pass_expand_omp): ...new function.
10207 (pass_expand_omp_ssa): Replace declaration with that of...
10208 (make_pass_expand_omp_ssa): ...new function.
10209 (pass_object_sizes): Replace declaration with that of...
10210 (make_pass_object_sizes): ...new function.
10211 (pass_strlen): Replace declaration with that of...
10212 (make_pass_strlen): ...new function.
10213 (pass_fold_builtins): Replace declaration with that of...
10214 (make_pass_fold_builtins): ...new function.
10215 (pass_stdarg): Replace declaration with that of...
10216 (make_pass_stdarg): ...new function.
10217 (pass_early_warn_uninitialized): Replace declaration with that of...
10218 (make_pass_early_warn_uninitialized): ...new function.
10219 (pass_late_warn_uninitialized): Replace declaration with that of...
10220 (make_pass_late_warn_uninitialized): ...new function.
10221 (pass_cse_reciprocals): Replace declaration with that of...
10222 (make_pass_cse_reciprocals): ...new function.
10223 (pass_cse_sincos): Replace declaration with that of...
10224 (make_pass_cse_sincos): ...new function.
10225 (pass_optimize_bswap): Replace declaration with that of...
10226 (make_pass_optimize_bswap): ...new function.
10227 (pass_optimize_widening_mul): Replace declaration with that of...
10228 (make_pass_optimize_widening_mul): ...new function.
10229 (pass_warn_function_return): Replace declaration with that of...
10230 (make_pass_warn_function_return): ...new function.
10231 (pass_warn_function_noreturn): Replace declaration with that of...
10232 (make_pass_warn_function_noreturn): ...new function.
10233 (pass_cselim): Replace declaration with that of...
10234 (make_pass_cselim): ...new function.
10235 (pass_phiopt): Replace declaration with that of...
10236 (make_pass_phiopt): ...new function.
10237 (pass_forwprop): Replace declaration with that of...
10238 (make_pass_forwprop): ...new function.
10239 (pass_phiprop): Replace declaration with that of...
10240 (make_pass_phiprop): ...new function.
10241 (pass_tree_ifcombine): Replace declaration with that of...
10242 (make_pass_tree_ifcombine): ...new function.
10243 (pass_dse): Replace declaration with that of...
10244 (make_pass_dse): ...new function.
10245 (pass_nrv): Replace declaration with that of...
10246 (make_pass_nrv): ...new function.
10247 (pass_rename_ssa_copies): Replace declaration with that of...
10248 (make_pass_rename_ssa_copies): ...new function.
10249 (pass_sink_code): Replace declaration with that of...
10250 (make_pass_sink_code): ...new function.
10251 (pass_fre): Replace declaration with that of...
10252 (make_pass_fre): ...new function.
10253 (pass_check_data_deps): Replace declaration with that of...
10254 (make_pass_check_data_deps): ...new function.
10255 (pass_copy_prop): Replace declaration with that of...
10256 (make_pass_copy_prop): ...new function.
10257 (pass_vrp): Replace declaration with that of...
10258 (make_pass_vrp): ...new function.
10259 (pass_uncprop): Replace declaration with that of...
10260 (make_pass_uncprop): ...new function.
10261 (pass_return_slot): Replace declaration with that of...
10262 (make_pass_return_slot): ...new function.
10263 (pass_reassoc): Replace declaration with that of...
10264 (make_pass_reassoc): ...new function.
10265 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10266 (make_pass_rebuild_cgraph_edges): ...new function.
10267 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10268 (make_pass_remove_cgraph_callee_edges): ...new function.
10269 (pass_build_cgraph_edges): Replace declaration with that of...
10270 (make_pass_build_cgraph_edges): ...new function.
10271 (pass_local_pure_const): Replace declaration with that of...
10272 (make_pass_local_pure_const): ...new function.
10273 (pass_tracer): Replace declaration with that of...
10274 (make_pass_tracer): ...new function.
10275 (pass_warn_unused_result): Replace declaration with that of...
10276 (make_pass_warn_unused_result): ...new function.
10277 (pass_diagnose_tm_blocks): Replace declaration with that of...
10278 (make_pass_diagnose_tm_blocks): ...new function.
10279 (pass_lower_tm): Replace declaration with that of...
10280 (make_pass_lower_tm): ...new function.
10281 (pass_tm_init): Replace declaration with that of...
10282 (make_pass_tm_init): ...new function.
10283 (pass_tm_mark): Replace declaration with that of...
10284 (make_pass_tm_mark): ...new function.
10285 (pass_tm_memopt): Replace declaration with that of...
10286 (make_pass_tm_memopt): ...new function.
10287 (pass_tm_edges): Replace declaration with that of...
10288 (make_pass_tm_edges): ...new function.
10289 (pass_split_functions): Replace declaration with that of...
10290 (make_pass_split_functions): ...new function.
10291 (pass_feedback_split_functions): Replace declaration with that of...
10292 (make_pass_feedback_split_functions): ...new function.
10293 (pass_strength_reduction): Replace declaration with that of...
10294 (make_pass_strength_reduction): ...new function.
10295 (pass_ipa_lower_emutls): Replace declaration with that of...
10296 (make_pass_ipa_lower_emutls): ...new function.
10297 (pass_ipa_function_and_variable_visibility): Replace declaration with
10299 (make_pass_ipa_function_and_variable_visibility): ...new function.
10300 (pass_ipa_tree_profile): Replace declaration with that of...
10301 (make_pass_ipa_tree_profile): ...new function.
10302 (pass_early_local_passes): Replace declaration with that of...
10303 (make_pass_early_local_passes): ...new function.
10304 (pass_ipa_whole_program_visibility): Replace declaration with that
10306 (make_pass_ipa_whole_program_visibility): ...new function.
10307 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10308 (make_pass_ipa_lto_gimple_out): ...new function.
10309 (pass_ipa_increase_alignment): Replace declaration with that of...
10310 (make_pass_ipa_increase_alignment): ...new function.
10311 (pass_ipa_inline): Replace declaration with that of...
10312 (make_pass_ipa_inline): ...new function.
10313 (pass_ipa_free_lang_data): Replace declaration with that of...
10314 (make_pass_ipa_free_lang_data): ...new function.
10315 (pass_ipa_free_inline_summary): Replace declaration with that of...
10316 (make_pass_ipa_free_inline_summary): ...new function.
10317 (pass_ipa_cp): Replace declaration with that of...
10318 (make_pass_ipa_cp): ...new function.
10319 (pass_ipa_reference): Replace declaration with that of...
10320 (make_pass_ipa_reference): ...new function.
10321 (pass_ipa_pure_const): Replace declaration with that of...
10322 (make_pass_ipa_pure_const): ...new function.
10323 (pass_ipa_pta): Replace declaration with that of...
10324 (make_pass_ipa_pta): ...new function.
10325 (pass_ipa_lto_finish_out): Replace declaration with that of...
10326 (make_pass_ipa_lto_finish_out): ...new function.
10327 (pass_ipa_tm): Replace declaration with that of...
10328 (make_pass_ipa_tm): ...new function.
10329 (pass_ipa_profile): Replace declaration with that of...
10330 (make_pass_ipa_profile): ...new function.
10331 (pass_ipa_cdtor_merge): Replace declaration with that of...
10332 (make_pass_ipa_cdtor_merge): ...new function.
10333 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10335 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10336 (pass_init_datastructures): Replace declaration with that of...
10337 (make_pass_init_datastructures): ...new function.
10338 (pass_fixup_cfg): Replace declaration with that of...
10339 (make_pass_fixup_cfg): ...new function.
10340 (pass_expand): Replace declaration with that of...
10341 (make_pass_expand): ...new function.
10342 (pass_instantiate_virtual_regs): Replace declaration with that of...
10343 (make_pass_instantiate_virtual_regs): ...new function.
10344 (pass_rtl_fwprop): Replace declaration with that of...
10345 (make_pass_rtl_fwprop): ...new function.
10346 (pass_rtl_fwprop_addr): Replace declaration with that of...
10347 (make_pass_rtl_fwprop_addr): ...new function.
10348 (pass_jump): Replace declaration with that of...
10349 (make_pass_jump): ...new function.
10350 (pass_jump2): Replace declaration with that of...
10351 (make_pass_jump2): ...new function.
10352 (pass_lower_subreg): Replace declaration with that of...
10353 (make_pass_lower_subreg): ...new function.
10354 (pass_cse): Replace declaration with that of...
10355 (make_pass_cse): ...new function.
10356 (pass_fast_rtl_dce): Replace declaration with that of...
10357 (make_pass_fast_rtl_dce): ...new function.
10358 (pass_ud_rtl_dce): Replace declaration with that of...
10359 (make_pass_ud_rtl_dce): ...new function.
10360 (pass_rtl_dce): Replace declaration with that of...
10361 (make_pass_rtl_dce): ...new function.
10362 (pass_rtl_dse1): Replace declaration with that of...
10363 (make_pass_rtl_dse1): ...new function.
10364 (pass_rtl_dse2): Replace declaration with that of...
10365 (make_pass_rtl_dse2): ...new function.
10366 (pass_rtl_dse3): Replace declaration with that of...
10367 (make_pass_rtl_dse3): ...new function.
10368 (pass_rtl_cprop): Replace declaration with that of...
10369 (make_pass_rtl_cprop): ...new function.
10370 (pass_rtl_pre): Replace declaration with that of...
10371 (make_pass_rtl_pre): ...new function.
10372 (pass_rtl_hoist): Replace declaration with that of...
10373 (make_pass_rtl_hoist): ...new function.
10374 (pass_rtl_store_motion): Replace declaration with that of...
10375 (make_pass_rtl_store_motion): ...new function.
10376 (pass_cse_after_global_opts): Replace declaration with that of...
10377 (make_pass_cse_after_global_opts): ...new function.
10378 (pass_rtl_ifcvt): Replace declaration with that of...
10379 (make_pass_rtl_ifcvt): ...new function.
10380 (pass_into_cfg_layout_mode): Replace declaration with that of...
10381 (make_pass_into_cfg_layout_mode): ...new function.
10382 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10383 (make_pass_outof_cfg_layout_mode): ...new function.
10384 (pass_loop2): Replace declaration with that of...
10385 (make_pass_loop2): ...new function.
10386 (pass_rtl_loop_init): Replace declaration with that of...
10387 (make_pass_rtl_loop_init): ...new function.
10388 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10389 (make_pass_rtl_move_loop_invariants): ...new function.
10390 (pass_rtl_unswitch): Replace declaration with that of...
10391 (make_pass_rtl_unswitch): ...new function.
10392 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10393 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10394 (pass_rtl_doloop): Replace declaration with that of...
10395 (make_pass_rtl_doloop): ...new function.
10396 (pass_rtl_loop_done): Replace declaration with that of...
10397 (make_pass_rtl_loop_done): ...new function.
10398 (pass_web): Replace declaration with that of...
10399 (make_pass_web): ...new function.
10400 (pass_cse2): Replace declaration with that of...
10401 (make_pass_cse2): ...new function.
10402 (pass_df_initialize_opt): Replace declaration with that of...
10403 (make_pass_df_initialize_opt): ...new function.
10404 (pass_df_initialize_no_opt): Replace declaration with that of...
10405 (make_pass_df_initialize_no_opt): ...new function.
10406 (pass_reginfo_init): Replace declaration with that of...
10407 (make_pass_reginfo_init): ...new function.
10408 (pass_inc_dec): Replace declaration with that of...
10409 (make_pass_inc_dec): ...new function.
10410 (pass_stack_ptr_mod): Replace declaration with that of...
10411 (make_pass_stack_ptr_mod): ...new function.
10412 (pass_initialize_regs): Replace declaration with that of...
10413 (make_pass_initialize_regs): ...new function.
10414 (pass_combine): Replace declaration with that of...
10415 (make_pass_combine): ...new function.
10416 (pass_if_after_combine): Replace declaration with that of...
10417 (make_pass_if_after_combine): ...new function.
10418 (pass_ree): Replace declaration with that of...
10419 (make_pass_ree): ...new function.
10420 (pass_partition_blocks): Replace declaration with that of...
10421 (make_pass_partition_blocks): ...new function.
10422 (pass_match_asm_constraints): Replace declaration with that of...
10423 (make_pass_match_asm_constraints): ...new function.
10424 (pass_regmove): Replace declaration with that of...
10425 (make_pass_regmove): ...new function.
10426 (pass_split_all_insns): Replace declaration with that of...
10427 (make_pass_split_all_insns): ...new function.
10428 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10429 (make_pass_fast_rtl_byte_dce): ...new function.
10430 (pass_lower_subreg2): Replace declaration with that of...
10431 (make_pass_lower_subreg2): ...new function.
10432 (pass_mode_switching): Replace declaration with that of...
10433 (make_pass_mode_switching): ...new function.
10434 (pass_sms): Replace declaration with that of...
10435 (make_pass_sms): ...new function.
10436 (pass_sched): Replace declaration with that of...
10437 (make_pass_sched): ...new function.
10438 (pass_ira): Replace declaration with that of...
10439 (make_pass_ira): ...new function.
10440 (pass_reload): Replace declaration with that of...
10441 (make_pass_reload): ...new function.
10442 (pass_clean_state): Replace declaration with that of...
10443 (make_pass_clean_state): ...new function.
10444 (pass_branch_prob): Replace declaration with that of...
10445 (make_pass_branch_prob): ...new function.
10446 (pass_value_profile_transformations): Replace declaration with that
10448 (make_pass_value_profile_transformations): ...new function.
10449 (pass_postreload_cse): Replace declaration with that of...
10450 (make_pass_postreload_cse): ...new function.
10451 (pass_gcse2): Replace declaration with that of...
10452 (make_pass_gcse2): ...new function.
10453 (pass_split_after_reload): Replace declaration with that of...
10454 (make_pass_split_after_reload): ...new function.
10455 (pass_branch_target_load_optimize1): Replace declaration with that
10457 (make_pass_branch_target_load_optimize1): ...new function.
10458 (pass_thread_prologue_and_epilogue): Replace declaration with that
10460 (make_pass_thread_prologue_and_epilogue): ...new function.
10461 (pass_stack_adjustments): Replace declaration with that of...
10462 (make_pass_stack_adjustments): ...new function.
10463 (pass_peephole2): Replace declaration with that of...
10464 (make_pass_peephole2): ...new function.
10465 (pass_if_after_reload): Replace declaration with that of...
10466 (make_pass_if_after_reload): ...new function.
10467 (pass_regrename): Replace declaration with that of...
10468 (make_pass_regrename): ...new function.
10469 (pass_cprop_hardreg): Replace declaration with that of...
10470 (make_pass_cprop_hardreg): ...new function.
10471 (pass_reorder_blocks): Replace declaration with that of...
10472 (make_pass_reorder_blocks): ...new function.
10473 (pass_branch_target_load_optimize2): Replace declaration with that
10475 (make_pass_branch_target_load_optimize2): ...new function.
10476 (pass_leaf_regs): Replace declaration with that of...
10477 (make_pass_leaf_regs): ...new function.
10478 (pass_split_before_sched2): Replace declaration with that of...
10479 (make_pass_split_before_sched2): ...new function.
10480 (pass_compare_elim_after_reload): Replace declaration with that of...
10481 (make_pass_compare_elim_after_reload): ...new function.
10482 (pass_sched2): Replace declaration with that of...
10483 (make_pass_sched2): ...new function.
10484 (pass_stack_regs): Replace declaration with that of...
10485 (make_pass_stack_regs): ...new function.
10486 (pass_stack_regs_run): Replace declaration with that of...
10487 (make_pass_stack_regs_run): ...new function.
10488 (pass_df_finish): Replace declaration with that of...
10489 (make_pass_df_finish): ...new function.
10490 (pass_compute_alignments): Replace declaration with that of...
10491 (make_pass_compute_alignments): ...new function.
10492 (pass_duplicate_computed_gotos): Replace declaration with that of...
10493 (make_pass_duplicate_computed_gotos): ...new function.
10494 (pass_variable_tracking): Replace declaration with that of...
10495 (make_pass_variable_tracking): ...new function.
10496 (pass_free_cfg): Replace declaration with that of...
10497 (make_pass_free_cfg): ...new function.
10498 (pass_machine_reorg): Replace declaration with that of...
10499 (make_pass_machine_reorg): ...new function.
10500 (pass_cleanup_barriers): Replace declaration with that of...
10501 (make_pass_cleanup_barriers): ...new function.
10502 (pass_delay_slots): Replace declaration with that of...
10503 (make_pass_delay_slots): ...new function.
10504 (pass_split_for_shorten_branches): Replace declaration with that of...
10505 (make_pass_split_for_shorten_branches): ...new function.
10506 (pass_split_before_regstack): Replace declaration with that of...
10507 (make_pass_split_before_regstack): ...new function.
10508 (pass_convert_to_eh_region_ranges): Replace declaration with that
10510 (make_pass_convert_to_eh_region_ranges): ...new function.
10511 (pass_shorten_branches): Replace declaration with that of...
10512 (make_pass_shorten_branches): ...new function.
10513 (pass_set_nothrow_function_flags): Replace declaration with that of...
10514 (make_pass_set_nothrow_function_flags): ...new function.
10515 (pass_dwarf2_frame): Replace declaration with that of...
10516 (make_pass_dwarf2_frame): ...new function.
10517 (pass_final): Replace declaration with that of...
10518 (make_pass_final): ...new function.
10519 (pass_rtl_seqabstr): Replace declaration with that of...
10520 (make_pass_rtl_seqabstr): ...new function.
10521 (pass_release_ssa_names): Replace declaration with that of...
10522 (make_pass_release_ssa_names): ...new function.
10523 (pass_early_inline): Replace declaration with that of...
10524 (make_pass_early_inline): ...new function.
10525 (pass_inline_parameters): Replace declaration with that of...
10526 (make_pass_inline_parameters): ...new function.
10527 (pass_update_address_taken): Replace declaration with that of...
10528 (make_pass_update_address_taken): ...new function.
10529 (pass_convert_switch): Replace declaration with that of...
10530 (make_pass_convert_switch): ...new function.
10531 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10532 to a subclass of simple_ipa_opt_pass along with...
10533 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10534 (make_pass_ipa_tree_profile): ...new function.
10535 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10536 subclass of gimple_opt_pass along with...
10537 (pass_data_sra_early): ...new pass_data instance and...
10538 (make_pass_sra_early): ...new function.
10539 (pass_sra): Convert from a global struct to a subclass of
10540 gimple_opt_pass along with...
10541 (pass_data_sra): ...new pass_data instance and...
10542 (make_pass_sra): ...new function.
10543 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10544 gimple_opt_pass along with...
10545 (pass_data_early_ipa_sra): ...new pass_data instance and...
10546 (make_pass_early_ipa_sra): ...new function.
10547 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10548 subclass of gimple_opt_pass along with...
10549 (pass_data_ccp): ...new pass_data instance and...
10550 (make_pass_ccp): ...new function.
10551 (pass_fold_builtins): Convert from a global struct to a subclass of
10552 gimple_opt_pass along with...
10553 (pass_data_fold_builtins): ...new pass_data instance and...
10554 (make_pass_fold_builtins): ...new function.
10555 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10556 subclass of gimple_opt_pass along with...
10557 (pass_data_copy_prop): ...new pass_data instance and...
10558 (make_pass_copy_prop): ...new function.
10559 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10560 global struct to a subclass of gimple_opt_pass along with...
10561 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10562 (make_pass_rename_ssa_copies): ...new function.
10563 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10564 subclass of gimple_opt_pass along with...
10565 (pass_data_dce): ...new pass_data instance and...
10566 (make_pass_dce): ...new function.
10567 (pass_dce_loop): Convert from a global struct to a subclass of
10568 gimple_opt_pass along with...
10569 (pass_data_dce_loop): ...new pass_data instance and...
10570 (make_pass_dce_loop): ...new function.
10571 (pass_cd_dce): Convert from a global struct to a subclass of
10572 gimple_opt_pass along with...
10573 (pass_data_cd_dce): ...new pass_data instance and...
10574 (make_pass_cd_dce): ...new function.
10575 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10576 subclass of gimple_opt_pass along with...
10577 (pass_data_dominator): ...new pass_data instance and...
10578 (make_pass_dominator): ...new function.
10579 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10580 gimple_opt_pass along with...
10581 (pass_data_phi_only_cprop): ...new pass_data instance and...
10582 (make_pass_phi_only_cprop): ...new function.
10583 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10584 subclass of gimple_opt_pass along with...
10585 (pass_data_dse): ...new pass_data instance and...
10586 (make_pass_dse): ...new function.
10587 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10588 a subclass of gimple_opt_pass along with...
10589 (pass_data_forwprop): ...new pass_data instance and...
10590 (make_pass_forwprop): ...new function.
10591 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10592 struct to a subclass of gimple_opt_pass along with...
10593 (pass_data_tree_ifcombine): ...new pass_data instance and...
10594 (make_pass_tree_ifcombine): ...new function.
10595 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10596 subclass of gimple_opt_pass along with...
10597 (pass_data_ch): ...new pass_data instance and...
10598 (make_pass_ch): ...new function.
10599 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10600 subclass of gimple_opt_pass along with...
10601 (pass_data_tree_loop): ...new pass_data instance and...
10602 (make_pass_tree_loop): ...new function.
10603 (pass_tree_loop_init): Convert from a global struct to a subclass of
10604 gimple_opt_pass along with...
10605 (pass_data_tree_loop_init): ...new pass_data instance and...
10606 (make_pass_tree_loop_init): ...new function.
10607 (pass_lim): Convert from a global struct to a subclass of
10608 gimple_opt_pass along with...
10609 (pass_data_lim): ...new pass_data instance and...
10610 (make_pass_lim): ...new function.
10611 (pass_tree_unswitch): Convert from a global struct to a subclass of
10612 gimple_opt_pass along with...
10613 (pass_data_tree_unswitch): ...new pass_data instance and...
10614 (make_pass_tree_unswitch): ...new function.
10615 (pass_predcom): Convert from a global struct to a subclass of
10616 gimple_opt_pass along with...
10617 (pass_data_predcom): ...new pass_data instance and...
10618 (make_pass_predcom): ...new function.
10619 (pass_vectorize): Convert from a global struct to a subclass of
10620 gimple_opt_pass along with...
10621 (pass_data_vectorize): ...new pass_data instance and...
10622 (make_pass_vectorize): ...new function.
10623 (pass_graphite): Convert from a global struct to a subclass of
10624 gimple_opt_pass along with...
10625 (pass_data_graphite): ...new pass_data instance and...
10626 (make_pass_graphite): ...new function.
10627 (pass_graphite_transforms): Convert from a global struct to a subclass
10628 of gimple_opt_pass along with...
10629 (pass_data_graphite_transforms): ...new pass_data instance and...
10630 (make_pass_graphite_transforms): ...new function.
10631 (pass_check_data_deps): Convert from a global struct to a subclass of
10632 gimple_opt_pass along with...
10633 (pass_data_check_data_deps): ...new pass_data instance and...
10634 (make_pass_check_data_deps): ...new function.
10635 (pass_iv_canon): Convert from a global struct to a subclass of
10636 gimple_opt_pass along with...
10637 (pass_data_iv_canon): ...new pass_data instance and...
10638 (make_pass_iv_canon): ...new function.
10639 (pass_scev_cprop): Convert from a global struct to a subclass of
10640 gimple_opt_pass along with...
10641 (pass_data_scev_cprop): ...new pass_data instance and...
10642 (make_pass_scev_cprop): ...new function.
10643 (pass_record_bounds): Convert from a global struct to a subclass of
10644 gimple_opt_pass along with...
10645 (pass_data_record_bounds): ...new pass_data instance and...
10646 (make_pass_record_bounds): ...new function.
10647 (pass_complete_unroll): Convert from a global struct to a subclass of
10648 gimple_opt_pass along with...
10649 (pass_data_complete_unroll): ...new pass_data instance and...
10650 (make_pass_complete_unroll): ...new function.
10651 (pass_complete_unrolli): Convert from a global struct to a subclass of
10652 gimple_opt_pass along with...
10653 (pass_data_complete_unrolli): ...new pass_data instance and...
10654 (make_pass_complete_unrolli): ...new function.
10655 (pass_parallelize_loops): Convert from a global struct to a subclass
10656 of gimple_opt_pass along with...
10657 (pass_data_parallelize_loops): ...new pass_data instance and...
10658 (make_pass_parallelize_loops): ...new function.
10659 (pass_loop_prefetch): Convert from a global struct to a subclass of
10660 gimple_opt_pass along with...
10661 (pass_data_loop_prefetch): ...new pass_data instance and...
10662 (make_pass_loop_prefetch): ...new function.
10663 (pass_iv_optimize): Convert from a global struct to a subclass of
10664 gimple_opt_pass along with...
10665 (pass_data_iv_optimize): ...new pass_data instance and...
10666 (make_pass_iv_optimize): ...new function.
10667 (pass_tree_loop_done): Convert from a global struct to a subclass of
10668 gimple_opt_pass along with...
10669 (pass_data_tree_loop_done): ...new pass_data instance and...
10670 (make_pass_tree_loop_done): ...new function.
10671 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10672 struct to a subclass of gimple_opt_pass along with...
10673 (pass_data_cse_reciprocals): ...new pass_data instance and...
10674 (make_pass_cse_reciprocals): ...new function.
10675 (pass_cse_sincos): Convert from a global struct to a subclass of
10676 gimple_opt_pass along with...
10677 (pass_data_cse_sincos): ...new pass_data instance and...
10678 (make_pass_cse_sincos): ...new function.
10679 (pass_optimize_bswap): Convert from a global struct to a subclass of
10680 gimple_opt_pass along with...
10681 (pass_data_optimize_bswap): ...new pass_data instance and...
10682 (make_pass_optimize_bswap): ...new function.
10683 (pass_optimize_widening_mul): Convert from a global struct to a
10684 subclass of gimple_opt_pass along with...
10685 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10686 (make_pass_optimize_widening_mul): ...new function.
10687 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10688 subclass of gimple_opt_pass along with...
10689 (pass_data_phiopt): ...new pass_data instance and...
10690 (make_pass_phiopt): ...new function.
10691 (pass_cselim): Convert from a global struct to a subclass of
10692 gimple_opt_pass along with...
10693 (pass_data_cselim): ...new pass_data instance and...
10694 (make_pass_cselim): ...new function.
10695 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10696 subclass of gimple_opt_pass along with...
10697 (pass_data_phiprop): ...new pass_data instance and...
10698 (make_pass_phiprop): ...new function.
10699 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10700 subclass of gimple_opt_pass along with...
10701 (pass_data_pre): ...new pass_data instance and...
10702 (make_pass_pre): ...new function.
10703 (pass_fre): Convert from a global struct to a subclass of
10704 gimple_opt_pass along with...
10705 (pass_data_fre): ...new pass_data instance and...
10706 (make_pass_fre): ...new function.
10707 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10708 subclass of gimple_opt_pass along with...
10709 (pass_data_reassoc): ...new pass_data instance and...
10710 (make_pass_reassoc): ...new function.
10711 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10712 subclass of gimple_opt_pass along with...
10713 (pass_data_sink_code): ...new pass_data instance and...
10714 (make_pass_sink_code): ...new function.
10715 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10716 subclass of gimple_opt_pass along with...
10717 (pass_data_strlen): ...new pass_data instance and...
10718 (make_pass_strlen): ...new function.
10719 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10720 struct to a subclass of gimple_opt_pass along with...
10721 (pass_data_build_alias): ...new pass_data instance and...
10722 (make_pass_build_alias): ...new function.
10723 (pass_build_ealias): Convert from a global struct to a subclass of
10724 gimple_opt_pass along with...
10725 (pass_data_build_ealias): ...new pass_data instance and...
10726 (make_pass_build_ealias): ...new function.
10727 (pass_ipa_pta): Convert from a global struct to a subclass of
10728 simple_ipa_opt_pass along with...
10729 (pass_data_ipa_pta): ...new pass_data instance and...
10730 (make_pass_ipa_pta): ...new function.
10731 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10732 subclass of gimple_opt_pass along with...
10733 (pass_data_uncprop): ...new pass_data instance and...
10734 (make_pass_uncprop): ...new function.
10735 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10736 global struct to a subclass of gimple_opt_pass along with...
10737 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10738 (make_pass_late_warn_uninitialized): ...new function.
10739 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10740 to a subclass of gimple_opt_pass along with...
10741 (pass_data_init_datastructures): ...new pass_data instance and...
10742 (make_pass_init_datastructures): ...new function.
10743 (pass_early_warn_uninitialized): Convert from a global struct to a
10744 subclass of gimple_opt_pass along with...
10745 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10746 (make_pass_early_warn_uninitialized): ...new function.
10747 (pass_update_address_taken): Convert from a global struct to a
10748 subclass of gimple_opt_pass along with...
10749 (pass_data_update_address_taken): ...new pass_data instance and...
10750 (make_pass_update_address_taken): ...new function.
10751 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10752 struct to a subclass of gimple_opt_pass along with...
10753 (pass_data_release_ssa_names): ...new pass_data instance and...
10754 (make_pass_release_ssa_names): ...new function.
10755 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10756 subclass of gimple_opt_pass along with...
10757 (pass_data_stdarg): ...new pass_data instance and...
10758 (make_pass_stdarg): ...new function.
10759 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10760 global struct to a subclass of gimple_opt_pass along with...
10761 (pass_data_convert_switch): ...new pass_data instance and...
10762 (make_pass_convert_switch): ...new function.
10763 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10764 to a subclass of gimple_opt_pass along with...
10765 (pass_data_tail_recursion): ...new pass_data instance and...
10766 (make_pass_tail_recursion): ...new function.
10767 (pass_tail_calls): Convert from a global struct to a subclass of
10768 gimple_opt_pass along with...
10769 (pass_data_tail_calls): ...new pass_data instance and...
10770 (make_pass_tail_calls): ...new function.
10771 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10772 struct to a subclass of gimple_opt_pass along with...
10773 (pass_data_lower_vector): ...new pass_data instance and...
10774 (make_pass_lower_vector): ...new function.
10775 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10776 gimple_opt_pass along with...
10777 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10778 (make_pass_lower_vector_ssa): ...new function.
10779 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10780 to a subclass of gimple_opt_pass along with...
10781 (pass_data_slp_vectorize): ...new pass_data instance and...
10782 (make_pass_slp_vectorize): ...new function.
10783 (pass_ipa_increase_alignment): Convert from a global struct to a
10784 subclass of simple_ipa_opt_pass along with...
10785 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10786 (make_pass_ipa_increase_alignment): ...new function.
10787 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10788 gimple_opt_pass along with...
10789 (pass_data_vrp): ...new pass_data instance and...
10790 (make_pass_vrp): ...new function.
10791 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10792 subclass of simple_ipa_opt_pass along with...
10793 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10794 (make_pass_ipa_free_lang_data): ...new function.
10795 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10796 gimple_opt_pass along with...
10797 (pass_data_tsan): ...new pass_data instance and...
10798 (make_pass_tsan): ...new function.
10799 (pass_tsan_O0): Convert from a global struct to a subclass of
10800 gimple_opt_pass along with...
10801 (pass_data_tsan_O0): ...new pass_data instance and...
10802 (make_pass_tsan_O0): ...new function.
10803 * var-tracking.c (pass_variable_tracking): Convert from a global
10804 struct to a subclass of rtl_opt_pass along with...
10805 (pass_data_variable_tracking): ...new pass_data instance and...
10806 (make_pass_variable_tracking): ...new function.
10807 * web.c (pass_web): Convert from a global struct to a subclass of
10808 rtl_opt_pass along with...
10809 (pass_data_web): ...new pass_data instance and...
10810 (make_pass_web): ...new function.
10811 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10812 declaration with that of...
10813 (make_pass_mode_switch_use): ...new function.
10814 (pass_resolve_sw_modes): Replace declaration with that of...
10815 (make_pass_resolve_sw_modes): ...new function.
10816 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10817 from a global struct to a subclass of rtl_opt_pass along with...
10818 (pass_data_mode_switch_use): ...new pass_data instance and...
10819 (make_pass_mode_switch_use): ...new function.
10820 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10821 from a global struct to a subclass of rtl_opt_pass along with...
10822 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10823 (make_pass_resolve_sw_modes): ...new function.
10824 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10825 struct to a subclass of rtl_opt_pass along with...
10826 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10827 (make_pass_insert_vzeroupper): ...new function.
10828 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10829 global struct to a subclass of rtl_opt_pass along with...
10830 (pass_data_work_around_errata): ...new pass_data instance and...
10831 (make_pass_work_around_errata): ...new function.
10832 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10833 struct to a subclass of rtl_opt_pass along with...
10834 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10835 (make_pass_mips_machine_reorg2): ...new function.
10837 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10839 * passes.c (pass_manager::operator new): New.
10841 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10843 Handwritten part of conversion of passes to C++ classes.
10845 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10846 (toplev.o): Add dep on PASS_MANAGER_H.
10847 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10848 of early local pases to reflect this moving from a global to a
10849 member of gcc::pass_manager.
10850 (cgraph_add_new_function): Likewise.
10851 * lto-cgraph.c (lto_output_node): Update for conversion of
10852 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10853 * passes.c (opt_pass::clone): New.
10854 (opt_pass::gate): New.
10855 (opt_pass::execute): New.
10856 (opt_pass::opt_pass): New.
10857 (pass_manager::execute_early_local_passes): New.
10858 (pass_manager::execute_pass_mode_switching): new.
10859 (finish_optimization_passes): Convert to...
10860 (pass_manager::finish_optimization_passes): ...this.
10861 (finish_optimization_passes): Update for conversion of passes to
10863 (register_dump_files_1): Use has_gate since we cannot portably
10864 check a vtable entry against NULL.
10865 (dump_one_pass): Likewise.
10866 (ipa_write_summaries_2): Likewise.
10867 (ipa_write_optimization_summaries_1): Likewise.
10868 (ipa_read_summaries_1): Likewise.
10869 (ipa_read_optimization_summaries_1): Likewise.
10870 (execute_ipa_stmt_fixups): Likewise.
10871 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10872 pass-creation functions rather than wiring up globals, and
10873 storing the results in fields of pass_manager generated using
10874 pass-instances.def.
10875 (pass_manager::dump_profile_report): Update for conversion of
10876 passes to C++ classes.
10877 (pass_manager::execute_ipa_summary_passes): Likewise.
10878 (execute_one_ipa_transform_pass): Likewise.
10879 (execute_one_pass): Use has_gate and has_execute since we cannot
10880 portably check a vtable entry against NULL.
10881 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10882 (pass_manager): Use pass-instances.def to add fields for the
10883 various pass instances.
10884 * toplev.c (finalize): Update for move of
10885 finish_optimization_passes to a method of gcc::pass_manager.
10886 * toplev.h (finish_optimization_passes): Move to method of class
10888 * tree-pass.h (struct pass_data): New.
10889 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10890 (opt_pass::gate): Convert to virtual function.
10891 (opt_pass::~opt_pass): New.
10892 (opt_pass::clone): New.
10893 (opt_pass::execute): Convert to virtual function.
10894 (opt_pass::opt_pass): New.
10895 (opt_pass::ctxt_): new.
10896 (gimple_opt_pass): Convert to subclass of opt_pass.
10897 (gimple_opt_pass::gimple_opt_pass): New.
10898 (rtl_opt_pass): Convert to subclass of opt_pass.
10899 (rtl_opt_pass::rtl_opt_pass): New.
10900 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10901 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10902 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10903 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10904 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10905 invocation of pass_mode_switching to reflect this moving from a
10906 global to a member of gcc::pass_manager.
10907 (ix86_option_override): Rework how pass_insert_vzeroupper is
10908 added to the pass_manager to reflect autogenerated changes.
10909 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10911 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10913 PR rtl-optimization/57708
10914 * recog.c (peep2_find_free_register): Validate all regs in a
10917 2013-08-05 Jan Hubicka <jh@suse.cz>
10920 * cgraph.c (verify_cgraph_node): Accept local flags from other
10922 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10923 (function_and_variable_visibility): Likewise.
10924 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10926 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10928 * graph.c (init_graph_slim_pretty_print): Remove.
10929 (print_graph_cfg): Do not call it. Use local pretty printer.
10930 (start_graph_dump): Likewise.
10932 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10934 * gimple-pretty-print.c (buffer): Remove.
10935 (initialized): Likewise.
10936 (maybe_init_pretty_print): Likewise.
10937 (print_gimple_stmt): Do not call it. Use non-static local
10938 pretty_printer variable.
10939 (print_gimple_expr): Likewise.
10940 (print_gimple_seq): Likewise.
10941 (gimple_dump_bb): Likewise.
10943 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10945 * asan.c (asan_pp): Remove.
10946 (asan_pp_initialized): Likewise.
10947 (asan_pp_initialize): Likewise.
10948 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10949 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10950 (asan_add_global): Likewise.
10952 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10954 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10955 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10956 * pretty-print.c (pp_formatted_text_data): Likewise.
10957 (pp_indent): Rename from pp_base_indent.
10958 (pp_format): Rename from pp_base_format.
10959 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10960 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10961 (pp_flush): Rename from pp_base_flush.
10962 (pp_set_line_maximum_length): Rename from
10963 pp_base_set_line_maximum_length.
10964 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10965 (pp_set_prefix): Rename from pp_base_set_prefix.
10966 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10967 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10968 (pp_append_text): Rename from pp_base_append_text.
10969 (pp_formatted_text): Rename from pp_base_formatted_text.
10970 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10971 (pp_remaining_character_count_for_line): Rename from
10972 pp_base_remaining_character_count_for_line.
10973 (pp_newline): Rename from pp_base_newline.
10974 (pp_character): Rename from pp_base_character.
10975 (pp_string): Rename from pp_base_string.
10976 (pp_maybe_space): Rename from pp_base_maybe_space.
10977 * asan.c (asan_pp_string): Adjust.
10978 (asan_emit_stack_protection): Likewise.
10979 (asan_add_global): Likewise.
10980 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10981 * tree-mudflap.c (mf_varname_tree): Likewise.
10982 * tree-pretty-print.c (pp_tree_identifier): Rename from
10983 pp_base_tree_identifier.
10984 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10985 Declare as function.
10987 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10989 * pretty-print.h (pp_bar_bar): New.
10990 (pp_ampersand_ampersand): Likewise.
10991 (pp_less_equal): Likewise.
10992 (pp_greater_equal): Likewise.
10993 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10994 printer functions instead of pp_string or operators and punctuators.
10995 (dump_gimple_call): Likewise.
10996 (dump_gimple_omp_for): Likewise.
10997 (dump_gimple_transaction): Likewise.
10998 (dump_gimple_phi): Likewise.
10999 (pp_gimple_stmt_1): Likewise.
11000 * sched-vis.c (print_insn): Likewise.
11001 * tree-mudflap.c (mf_varname_tree): Likewise.
11002 * tree-pretty-print.c (dump_block_node): Likewise.
11003 (dump_generic_node): Likewise.
11005 2013-08-02 Jan Hubicka <jh@suse.cz>
11007 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11009 * lto-streamer-out.c (tree_is_indexable): Results decls and
11010 parm decls are not indexable.
11011 (DFS_write_tree_body): Do not follow args and results.
11012 (hash_tree): Likewise.
11013 (output_functions): Rearrange so struct function is needed
11014 only when real body is output; be able to also ouptut abstract
11015 functions; output DECL_ARGUMENTS and DECL_RESULT.
11016 (lto_output): When not in WPA, ale store abstract functions.
11017 (write_symbol): Do not care about RESULT_DECL.
11018 (output_symbol_p): Handle correctly sbtract decls.
11019 * lto-streamer-in.c (input_function): Rearrange so struct
11020 function can be NULL at entry; allow streaming of
11021 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11022 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11023 sanity check during LTO.
11024 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11025 RESULT_DECl and DECL_ARGUMENTS.
11026 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11029 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11031 * pretty-print.h (pp_underscore): New.
11033 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11034 printer functions instead of pp_character.
11035 (dump_binary_rhs): Likewise.
11036 (dump_ternary_rhs): Likewise.
11037 (dump_gimple_call_args): Likewise.
11038 (pp_points_to_solution): Likewise.
11039 (dump_gimple_call): Likewise.
11040 (dump_gimple_switch): Likewise.
11041 (dump_gimple_cond): Likewise.
11042 (dump_gimple_bind): Likewise.
11043 (dump_gimple_try): Likewise.
11044 (dump_gimple_omp_for): Likewise.
11045 (dump_gimple_omp_continue): Likewise.
11046 (dump_gimple_omp_single): Likewise.
11047 (dump_gimple_omp_sections): Likewise.
11048 (dump_gimple_omp_block): Likewise.
11049 (dump_gimple_omp_critical): Likewise.
11050 (dump_gimple_transaction): Likewise.
11051 (dump_gimple_asm): Likewise.
11052 (dump_gimple_phi): Likewise.
11053 (dump_gimple_omp_parallel): Likewise.
11054 (dump_gimple_omp_task): Likewise.
11055 (dump_gimple_omp_atomic_load): Likewise.
11056 (dump_gimple_omp_atomic_store): Likewise.
11057 (dump_gimple_mem_ops): Likewise.
11058 (pp_gimple_stmt_1): Likewise.
11059 (pp_cfg_jump): Likewise.
11060 (dump_implicit_edges): Likewise.
11061 (gimple_dump_bb_for_graph): Likewise.
11062 * graph.c (draw_cfg_node): Likewise.
11063 * langhooks.c (lhd_print_error_function): Likewise.
11064 * sched-vis.c (print_exp): Likewise.
11065 (print_value): Likewise.
11066 (print_pattern): Likewise.
11067 (print_insn): Likewise.
11068 (rtl_dump_bb_for_graph): Likewise.
11069 * tree-pretty-print.c (dump_function_declaration): Likewise.
11070 (dump_array_domain): Likewise.
11071 (dump_omp_clause): Likewise.
11072 (dump_location): Likewise.
11073 (dump_generic_node): Likewise.
11074 (print_struct_decl): Likewise.
11075 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11077 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11079 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11080 candidate table when replacing a candidate statement.
11081 (replace_rhs_if_not_dup): Likewise.
11082 (replace_one_candidate): Likewise.
11084 2013-08-02 Jan Hubicka <jh@suse.cz>
11085 Martin Liska <marxin.liska@gmail.com>
11087 * cgraphunit.c (add_new_function): Fix logic when adding from
11089 (assemble_thunk): Rename to ...
11090 (expand_thunk); .. this one; export; get it working with
11091 general functions; make produced gimple valid.
11092 * cgraph.h (expand_thunk): Declare.
11094 2013-08-02 Jan Hubicka <jh@suse.cz>
11096 * ipa-cp.c (gather_context_independent_values): Use
11097 ipa_get_param_move_cost.
11098 (get_replacement_map): Remove PARAM; move parameter folding
11100 (create_specialized_node): Update.
11101 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11102 assert that we have gimple body; update move_cost.
11103 (count_formal_params): Assert that we have gimple body.
11104 (ipa_dump_param): New function.
11105 (ipa_alloc_node_params): Break out from ...
11106 (ipa_initialize_node_params): ... here.
11107 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11108 (ipa_write_node_info): Stream move costs.
11109 (ipa_read_node_info): Read move costs.
11110 (ipa_update_after_lto_read): Do not recompute node params.
11111 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11112 (ipa_get_param): Check we are not in WPA.
11113 (ipa_get_param_move_cost): New.
11114 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11115 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11116 parm numbers to be present.
11118 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11120 PR rtl-optimization/58048
11121 * lra-constraints.c (process_alt_operands): Don't check asm
11122 operand on register.
11124 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11126 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11127 the implied StoreLoad barrier for atomic operations if before.
11129 2013-08-02 Jan Hubicka <jh@suse.cz>
11130 Martin Liska <marxin.liska@gmail.com>
11132 * cgraph.c (cgraph_function_body_availability): Do not check
11134 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11135 symtab_node_availability): Declare.
11136 * ipa.c (can_replace_by_local_alias): New.
11137 (function_and_variable_visibility): Use it.
11138 * symtab.c (symtab_for_node_and_aliases,
11139 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11141 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11143 PR rtl-optimization/57963
11144 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11145 (lra_constraints): Use them.
11147 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11149 * config/arm/types.md (define_attr "type"): Add "load_acq"
11151 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11153 (cortex_a53_store1): Likewise.
11155 2013-08-01 Jan Hubicka <jh@suse.cz>
11157 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11158 partitions are not needed.
11160 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11162 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11163 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11164 MAYBE_NON_Q_CLASS_P where appropriate.
11166 2013-08-01 Jan Hubicka <jh@suse.cz>
11168 * cgraph.h (release_function_body): Declare.
11169 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11170 of unused delcarations.
11172 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11174 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11175 RTL form when subtracting a constant.
11177 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11179 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11180 Generate canonical plus rtx with negated immediate instead of minus
11182 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11184 2013-08-01 Jan Hubicka <jh@suse.cz>
11186 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11187 (cgraph_release_function_body): Likewise.
11188 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11189 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11190 to used_as_abstract_origin.
11191 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11192 symbols used as abstract origins.
11193 * cgraphunit.c (analyze_functions): Update.
11194 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11195 used_as_abstract_origin.
11196 * tree-inline.c (tree_function_versioning): Update
11197 used_as_abstract_origin; be ready for DECL_RESULT and
11198 DECL_ARGUMENTS to be NULL.
11200 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11201 for abstract functions.
11202 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11205 2013-08-01 Jan Hubicka <jh@suse.cz>
11207 * profile.c (compute_value_histograms): Fix thinko.
11209 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11211 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11212 aarch-common-protos.h to extra_headers.
11213 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11214 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11215 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11217 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11219 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11220 Include "../arm/types.md". Define "type" attribute for all patterns.
11221 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11224 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11226 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11227 to support power8 load fusion.
11228 (fusion_gpr_mem_load): Likewise.
11230 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11232 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11233 declarations for power8 load fusion.
11234 (emit_fusion_gpr_load): Likewise.
11236 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11237 tuning for power8, turn on fusion mode by default. Turn on sign
11238 extending fusion mode if normal fusion mode is on, and we are at
11240 (fusion_gpr_load_p): New function, return true if we can fuse an
11241 addis instruction with a dependent load to a GPR.
11242 (emit_fusion_gpr_load): Emit the instructions for power8 load
11245 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11246 (VSX load fusion peepholes): New peepholes to fuse together an
11247 addi instruction with a VSX load instruction.
11249 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11250 peepholes to fuse an addis instruction with a load to a GPR base
11251 register. If we are supporting sign extending fusions, convert
11252 sign extending loads to zero extending loads and add an explicit
11255 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11257 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11258 aarch-common-protos.h to extra_headers.
11259 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11260 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11261 (arm_early_store_addr_dep): Likewise.
11262 (arm_no_early_alu_shift_dep): Likewise.
11263 (arm_no_early_alu_shift_value_dep): Likewise.
11264 (arm_no_early_mul_dep): Likewise.
11265 (arm_no_early_store_addr_dep): Likewise.
11266 (arm_mac_accumulator_is_mul_result): Likewise.
11267 (arm_mac_accumulator_is_result): Likewise.
11268 * config/arm/aarch-common.c: ... here. New file.
11269 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11271 (arm_early_store_addr_dep): Likewise.
11272 (arm_no_early_alu_shift_dep): Likewise.
11273 (arm_no_early_alu_shift_value_dep): Likewise.
11274 (arm_no_early_mul_dep): Likewise.
11275 (arm_no_early_store_addr_dep): Likewise.
11276 (arm_mac_accumulator_is_mul_result): Likewise.
11277 (arm_mac_accumulator_is_result): Likewise.
11278 * config/arm/aarch-common-protos.h: ... here. New file.
11279 * config/arm/t-arm (aarch-common.o): Define.
11281 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11283 * config/arm/arm.md: Include new file "types.md".
11284 (define_attr "type"): Move from here to ...
11285 (define_attr "mul32"): Likewise.
11286 (define_attr "mul64"): Likewise.
11287 * config/arm/types.md: ... here. New file.
11289 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11291 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11292 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11294 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11296 * gen-pass-instances.awk: Fix offset of substr().
11298 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11300 * Makefile.in (pass-instances.def): New.
11301 (passes.o): Replace dependency on passes.def with one on
11304 * gen-pass-instances.awk: New.
11306 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11307 rather than passes.def, updating local definition of NEXT_PASS
11308 macro to add an extra NUM parameter (currently unused).
11310 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11312 * Makefile.in (PASS_MANAGER_H): New.
11313 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11314 (passes.o): Likewise.
11315 (statistics.o): Likewise.
11316 (cgraphunit.o): Likewise.
11317 (context.o): Depend on PASS_MANAGER_H.
11319 * pass_manager.h: New.
11321 * cgraphunit.c (cgraph_add_new_function): Update for moves
11322 of globals to fields of pass_manager.
11323 (analyze_function): Likewise.
11324 (expand_function): Likewise.
11325 (ipa_passes): Likewise.
11326 (compile): Likewise.
11328 * context.c (context::context): New.
11329 * context.h (context::context): New.
11330 (context::get_passes): New.
11331 (context::passes_): New.
11333 * lto-cgraph.c (input_node): Update for moves of globals to
11334 fields of pass_manager.
11336 * passes.c (all_passes): Remove, in favor of a field of the
11337 same name within the new class pass_manager.
11338 (all_small_ipa_passes): Likewise.
11339 (all_lowering_passes): Likewise.
11340 (all_regular_ipa_passes): Likewise.
11341 (all_late_ipa_passes): Likewise.
11342 (all_lto_gen_passes): Likewise.
11343 (passes_by_id): Likewise.
11344 (passes_by_id_size): Likewise.
11345 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11346 the new class pass_manager.
11347 (set_pass_for_id): Convert to...
11348 (pass_manager::set_pass_for_id): ...method.
11349 (get_pass_for_id): Convert to...
11350 (pass_manager::get_pass_for_id): ...method.
11351 (register_one_dump_file): Move body of implementation into...
11352 (pass_manager::register_one_dump_file): ...here.
11353 (register_dump_files_1): Convert to...
11354 (pass_manager::register_dump_files_1): ...method.
11355 (register_dump_files): Convert to...
11356 (pass_manager::register_dump_files): ...method.
11357 (create_pass_tab): Update for moves of globals to fields of
11359 (dump_passes): Move body of implementation into...
11360 (pass_manager::dump_passes): ...here.
11361 (register_pass): Move body of implementation into...
11362 (pass_manager::register_pass): ...here.
11363 (init_optimization_passes): Convert into...
11364 (pass_manager::pass_manager): ...constructor for new
11365 pass_manager class, and initialize the pass_lists array.
11366 (check_profile_consistency): Update for moves of globals to
11367 fields of pass_manager.
11368 (dump_profile_report): Move body of implementation into...
11369 (pass_manager::dump_profile_report): ...here.
11370 (ipa_write_summaries_1): Update for moves of pass lists from
11371 being globals to fields of pass_manager.
11372 (ipa_write_optimization_summaries): Likewise.
11373 (ipa_read_summaries): Likewise.
11374 (ipa_read_optimization_summaries): Likewise.
11375 (execute_all_ipa_stmt_fixups): Likewise.
11377 * statistics.c (statistics_fini): Update for moves of globals to
11378 fields of pass_manager.
11380 * toplev.c (general_init): Replace call to
11381 init_optimization_passes with construction of the pass_manager
11384 * tree-pass.h (all_passes): Remove, in favor of a field of the
11385 same name within the new class pass_manager.
11386 (all_small_ipa_passes): Likewise.
11387 (all_lowering_passes): Likewise.
11388 (all_regular_ipa_passes): Likewise.
11389 (all_lto_gen_passes): Likewise.
11390 (all_late_ipa_passes): Likewise.
11391 (passes_by_id): Likewise.
11392 (passes_by_id_size): Likewise.
11393 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11394 the new class pass_manager.
11395 (get_pass_for_id): Remove.
11397 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11399 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11402 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11404 * arm.md (mulhi3): New expand pattern.
11406 2013-07-30 Jan Hubicka <jh@suse.cz>
11407 Martin Liska <marxin.liska@gmail.com>
11409 * profile.c (compute_value_histograms): Do not ICE when
11410 there is mismatch only on some counters.
11412 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11414 PR rtl-optimization/57637
11415 * function.c (move_insn_for_shrink_wrap): Also check the
11416 GEN set of the LIVE problem for the liveness analysis
11417 if it exists, otherwise give up.
11419 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11421 PR tree-optimization/57993
11422 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11423 replaced statement in the candidate table.
11424 (phi_add_costs): Return infinite cost when the hidden basis does
11425 not dominate all phis on which the candidate is dependent.
11426 (replace_one_candidate): Record replaced statement in the
11429 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11431 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11432 (ashlv2si3): New expander.
11433 (*ashlv2si3_i): New define_insn_and_split.
11434 * predicates.md (float_operation): Allow patterns with three
11435 basic sub-patterns.
11437 PR rtl-optimization/58021
11438 * mode-switching.c (create_pre_exit): Always split off preceding
11439 insns if we are not at the basic block head.
11441 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11443 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11444 (UCLIBC_DYNAMIC_LINKER): New macro.
11445 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11447 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11448 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11450 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11451 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11452 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11453 for SF and DF modes. Use ieee_quad_format for TF mode.
11454 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11455 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11456 (mips_option_override): Handle `-mnan=legacy'.
11457 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11458 `-mabs=2008' and `-mnan=2008'.
11459 (OPTION_DEFAULT_SPECS): Add "nan" default.
11460 (ASM_SPEC): Handle `-mnan='.
11461 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11462 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11463 comment accordingly.
11464 (neg<mode>2): Likewise.
11465 * config/mips/mips.opt (mabs, mnan): New options.
11466 * doc/install.texi (Configuration): Document `--with-nan=' option.
11467 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11469 (MIPS Options): Document them.
11470 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11471 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11472 * configure: Regenerate.
11473 * config.in: Regenerate.
11475 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11477 * config/i386/i386.md (float post-reload splitters): Do not check
11478 for subregs of SSE registers.
11480 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11481 H.J. Lu <hongjiu.lu@intel.com>
11485 * config/i386/i386.md (post-reload splitter
11486 to avoid partial SSE reg dependency stalls): New pattern.
11488 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11490 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11492 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11494 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11495 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11497 2013-07-26 Cary Coutant <ccoutant@google.com>
11499 * dwarf2out.c (die_checksum_ordered): Don't include template
11500 instantiations in signature.
11501 (is_template_parameter): New function.
11502 (is_template_instantiation): New function.
11503 (generate_skeleton_bottom_up): Don't include template instantiations
11505 (generate_skeleton): Likewise.
11506 (break_out_comdat_types): Move recursive call to break out nested
11508 (prune_unused_types_mark_generic_parms_dies): Call
11509 is_template_parameter.
11511 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11513 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11514 uses vector registers.
11515 * config/aarch64/iterators.md: Add attributes rtn and vas.
11517 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11518 Richard Earnshaw <richard.earnshaw@arm.com>
11520 * combine.c (simplify_comparison): Re-canonicalize operands
11522 * config/arm/arm.md (movcond_addsi): New splitter.
11524 2013-07-25 Sterling Augustine <saugustine@google.com>
11526 * dwarf2out.c (size_of_pubnames): Move code to...
11527 (include_pubname_in_output): ...here. New.
11528 (want_pubnames): Rearrange.
11529 (output_pubnames): Call include_pubname_in_output. Move assertion.
11531 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11533 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11535 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11538 * doc/extend.texi: Remove obsolete builtins. Fix
11539 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11541 2013-07-25 Jan Hubicka <jh@suse.cz>
11543 * cgraph.c (release_function_body): Break out from ...
11544 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11545 and DECL_ARGUMENTS.
11546 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11547 old_tree in the map.
11548 (create_specialized_node): Update.
11549 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11551 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11552 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11553 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11554 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11557 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11559 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11560 addsi3_carryin_alt2_<optab>): Correct output template.
11562 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11564 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11565 Adjust for arm_restrict_it.
11566 Remove trailing whitespace.
11568 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11570 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11571 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11573 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11575 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11577 PR rtl-optimization/57960
11578 * lra-constraints.c (process_alt_operands): Use the right mode
11579 when checking strict_low.
11581 2013-07-25 Jan Hubicka <jh@suse.cz>
11583 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11584 * cgraph.c (cgraph_remove_node): Do not release function body
11585 when in cgraph streaming.
11586 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11587 in other partitions are not considered reachable; fix handling of
11590 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11592 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11594 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11599 * config/arm/arm.md ("*sibcall_insn): Replace use of
11600 Ss with US. Adjust output for v5 and v4t.
11601 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11603 * config/arm/constraints.md ("Ss"): Rename to US.
11605 2013-07-25 Terry Guo <terry.guo@arm.com>
11607 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11608 shift_add/shift_sub0/shift_sub1 RTXs.
11610 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11611 Anton Blanchard <anton@au1.ibm.com>
11613 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11614 (altivec_vpks<VI_char>ss): Likewise.
11615 (altivec_vpks<VI_char>us): Likewise.
11616 (altivec_vpku<VI_char>us): Likewise.
11617 (altivec_vpku<VI_char>um): Likewise.
11619 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11621 Introduce context class.
11623 * Makefile.in (CONTEXT_H): New.
11624 (OBJS): Add context.o.
11625 (toplev.o): Add CONTEXT_H to dependencies.
11628 * toplev.c (general_init): Create the singleton gcc::context instance.
11634 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11636 PR rtl-optimization/57968
11637 * mode-switching.c (create_pre_exit): Allow instructions that
11638 don't set a return register to need a non-exit mode.
11640 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11641 Anton Blanchard <anton@au1.ibm.com>
11643 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11644 operands to vperm for little endian.
11645 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11646 of lvsl to create the control mask for a vperm for little endian.
11648 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11649 Anton Blanchard <anton@au1.ibm.com>
11651 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11652 two operands for little-endian.
11654 2013-07-23 Steve Ellcey <sellcey@mips.com>
11656 * config/mips/mips.c (mips_case_values_threshold): New.
11657 (TARGET_CASE_VALUES_THRESHOLD): Define.
11659 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11660 Anton Blanchard <anton@au1.ibm.com>
11662 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11663 selection of field for vector splat in little endian mode.
11665 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11667 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11668 expanders to rs6000.md.
11669 (ior<mode>3): Likewise.
11670 (and<mode>3): Likewise.
11671 (one_cmpl<mode>2): Likewise.
11672 (nor<mode>3): Likewise.
11673 (andc<mode>3): Likewise.
11674 (eqv<mode>3): Likewise.
11675 (nand<mode>3): Likewise.
11676 (orc<mode>3): Likewise.
11678 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11681 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11682 to split multi-word logical operations.
11683 (rs6000_split_logical_di): Likewise.
11684 (rs6000_split_logical): Likewise.
11686 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11687 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11688 and allow TImode operations in 32-bit.
11689 (vsx_and<mode>3_64bit): Likewise.
11690 (vsx_ior<mode>3_32bit): Likewise.
11691 (vsx_ior<mode>3_64bit): Likewise.
11692 (vsx_xor<mode>3_32bit): Likewise.
11693 (vsx_xor<mode>3_64bit): Likewise.
11694 (vsx_one_cmpl<mode>2_32bit): Likewise.
11695 (vsx_one_cmpl<mode>2_64bit): Likewise.
11696 (vsx_nor<mode>3_32bit): Likewise.
11697 (vsx_nor<mode>3_64bit): Likewise.
11698 (vsx_andc<mode>3_32bit): Likewise.
11699 (vsx_andc<mode>3_64bit): Likewise.
11700 (vsx_eqv<mode>3_32bit): Likewise.
11701 (vsx_eqv<mode>3_64bit): Likewise.
11702 (vsx_nand<mode>3_32bit): Likewise.
11703 (vsx_nand<mode>3_64bit): Likewise.
11704 (vsx_orc<mode>3_32bit): Likewise.
11705 (vsx_orc<mode>3_64bit): Likewise.
11707 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11708 logical types in GPRs.
11710 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11711 logical insns to rs6000.md, and allow TImode operations in
11713 (altivec_ior<mode>3): Likewise.
11714 (altivec_xor<mode>3): Likewise.
11715 (altivec_one_cmpl<mode>2): Likewise.
11716 (altivec_nor<mode>3): Likewise.
11717 (altivec_andc<mode>3): Likewise.
11719 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11720 attributes for moving the 128-bit logical operations into
11722 (BOOL_REGS_OUTPUT): Likewise.
11723 (BOOL_REGS_OP1): Likewise.
11724 (BOOL_REGS_OP2): Likewise.
11725 (BOOL_REGS_UNARY): Likewise.
11726 (BOOL_REGS_AND_CR0): Likewise.
11727 (one_cmpl<mode>2): Add support for DI logical operations on
11728 32-bit, splitting the operations to 32-bit.
11729 (anddi3): Likewise.
11730 (iordi3): Likewise.
11731 (xordi3): Likewise.
11732 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11733 changes to combine the 32/64-bit code, allow logical operations on
11734 TI mode in 32-bit, and to use similar match_operator patterns like
11735 scalar mode uses. Combine the Altivec and VSX code for logical
11736 operations, and move it here.
11737 (ior<mode>3, 128-bit types): Likewise.
11738 (xor<mode>3, 128-bit types): Likewise.
11739 (one_cmpl<mode>3, 128-bit types): Likewise.
11740 (nor<mode>3, 128-bit types): Likewise.
11741 (andc<mode>3, 128-bit types): Likewise.
11742 (eqv<mode>3, 128-bit types): Likewise.
11743 (nand<mode>3, 128-bit types): Likewise.
11744 (orc<mode>3, 128-bit types): Likewise.
11745 (and<mode>3_internal): Likewise.
11746 (bool<mode>3_internal): Likewise.
11747 (boolc<mode>3_internal1): Likewise.
11748 (boolc<mode>3_internal2): Likewise.
11749 (boolcc<mode>3_internal1): Likewise.
11750 (boolcc<mode>3_internal2): Likewise.
11751 (eqv<mode>3_internal1): Likewise.
11752 (eqv<mode>3_internal2): Likewise.
11753 (one_cmpl1<mode>3_internal): Likewise.
11755 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11757 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11758 Rename flag_stack_usage to flag_stack_usage_info.
11760 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11762 * config/microblaze/sync.md: New file.
11763 * config/microblaze/microblaze.md: Include sync.md
11764 * config/microblaze/microblaze.c: Add print_operand 'y'.
11765 * config/microblaze/constraints.md: Add memory_contraint
11766 'Q' which is a single register.
11768 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11770 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11772 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11774 * reload.c (find_reloads): Exit loop once we find this operand
11775 cannot be reloaded somehow for this alternative.
11777 * reload.c (find_reloads): Exit loop once we find a hard register.
11779 * rtlanal.c (computed_jump_p): Exit loop once we find label
11782 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11784 * cfgloopmanip.c (remove_path): Exit loop after setting
11787 * gensupport.c (subst_dup): Avoid loop if code is not
11788 MATCH_DUP nor MATCH_OP_DUP.
11790 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11792 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11794 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11796 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11797 true for SP_REGNUM if mode == ptr_mode.
11798 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11799 with value R0_REGNUM + 31.
11801 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11803 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11804 pad pointer-typed argument downward.
11806 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11808 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11809 and __ILP32__ when the ILP32 model is in use.
11811 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11813 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11814 (aarch64_load_symref_appropriately): In the case of
11815 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11816 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11817 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11818 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11819 if 'mode' doesn't equal to 'ptr_mode'.
11820 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11821 'vcall_offset'; change to call aarch64_emit_move differently depending
11822 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11823 to calculate the upper bound of 'vcall_offset'.
11824 (aarch64_cannot_force_const_mem): Change to also return true if
11826 (aarch64_legitimize_reload_address): In the case of large
11827 displacements, add new local variable 'xmode' and an assertion
11828 based on it; change to use 'xmode' to generate the new rtx and
11830 (aarch64_asm_trampoline_template): Change to generate the template
11831 differently depending on TARGET_ILP32 or not; change to use
11832 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11833 (aarch64_trampoline_size): Removed.
11834 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11835 and replace immediate literals with it. Change to use 'ptr_mode'
11836 instead of 'DImode' and call convert_memory_address if the mode
11837 of 'fnaddr' doesn't equal to 'ptr_mode'.
11838 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11840 (aarch64_elf_asm_destructor): Likewise.
11841 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11842 on TARGET_ILP32 instead of aarch64_trampoline_size.
11843 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11844 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11845 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11846 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11847 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11848 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11849 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11850 depending on the value of 'mode'.
11851 (add_losym_<mode>): New.
11852 (ldr_got_small_<mode>): New, based on ldr_got_small.
11853 (ldr_got_small): Remove.
11854 (ldr_got_small_sidi): New.
11855 * config/aarch64/iterators.md (P): New.
11856 (PTR): Change to 'ptr_mode' in the condition.
11858 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11860 * config.gcc (aarch64*-*-*): Support --with-abi.
11861 (aarch64*-*-elf): Support --with-multilib-list.
11862 (aarch64*-*-linux*): Likewise.
11863 (supported_defaults): Add abi to aarch64*-*-*.
11864 * configure.ac: Mention AArch64 for --with-multilib-list.
11865 * configure: Re-generated.
11866 * config/aarch64/biarchilp32.h: New file.
11867 * config/aarch64/biarchlp64.h: New file.
11868 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11870 (MULTILIB_DEFAULTS): Ditto.
11871 (DRIVER_SELF_SPECS): Ditto.
11872 (ASM_SPEC): Update to also substitute -mabi.
11873 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11874 file whose name depends on -mabi= and -mbig-endian.
11875 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11877 (POINTER_SIZE): New define.
11878 (POINTERS_EXTEND_UNSIGNED): Ditto.
11879 (enum aarch64_abi_type): New enumeration tag.
11880 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11881 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11882 (TARGET_ILP32): New define.
11883 * config/aarch64/aarch64.opt (mabi): New.
11884 (aarch64_abi): New.
11885 (ilp32, lp64): New values for -mabi.
11886 * config/aarch64/t-aarch64 (comma): New define.
11887 (MULTILIB_OPTIONS): Ditto.
11888 (MULTILIB_DIRNAMES): Ditto.
11889 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11890 * doc/invoke.texi: Document -mabi for AArch64.
11892 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11894 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11896 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11897 Anton Blanchard <anton@au1.ibm.com>
11899 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11900 endianness when selecting field to splat.
11902 2013-07-22 Eric Christopher <echristo@gmail.com>
11904 * dwarf2out.c (die_odr_checksum): New function to use
11905 CHECKSUM_ macros and ULEB128 for DIE tag.
11906 (generate_type_signature): Use.
11908 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11910 * config.gcc (sparc*-*-*): Accept leon3 processor.
11911 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11912 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11913 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11914 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11915 (mfix-ut699): Adjust comment.
11916 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11917 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11918 (CPP_CPU_SPEC): Likewise.
11919 (ASM_CPU_SPEC): Likewise.
11920 * config/sparc/sparc.c (leon3_cost): New constant.
11921 (sparc_option_override): Add leon3 support.
11922 (mem_ref): New function.
11923 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11924 (sparc_do_work_around_errata): Look into the instruction in the delay
11925 slot and adjust accordingly. Add fix for the data cache nullify issues
11926 of the UT699. Change insertion position for the NOP.
11927 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11928 (leon3_load): New reservation.
11929 (leon_store): Bump latency to 2.
11930 (grfpu): New automaton.
11931 (grfpu_alu): New unit.
11932 (grfpu_ds): Likewise.
11933 (leon_fp_alu): Adjust.
11934 (leon_fp_mult): Delete.
11935 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11936 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11937 * config/sparc/sparc.md (cpu): Add leon3.
11938 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11939 (swapsi): Likewise.
11940 (atomic_test_and_set): Likewise.
11941 (ldstub): Likewise.
11943 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11945 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11946 default for R5900 targets.
11947 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11948 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11949 * config/mips/mips.c (mips_option_override): Report an error for
11950 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11951 for -march=r5900 -mhard-float.
11953 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11955 * df-problems.c (can_move_insns_across): Exit loop once we
11956 find a non-fixed, non-global register.
11958 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11961 * omega.c (omega_eliminate_red): Break after setting red_found.
11962 (omega_problem_has_red_equations): Similarly after setting found.
11963 (omega_query_variable): Similarly after setting coupled.
11965 2013-07-22 Marek Polacek <polacek@redhat.com>
11967 * gimplify.c: Don't include gimple.h twice.
11969 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11971 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11972 instead of TARGET_THUMB1.
11973 (Pz): New constraint.
11974 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11976 (compare_negsi_si): Likewise.
11977 (compare_addsi2_op0): Likewise.
11978 (compare_addsi2_op1): Likewise.
11979 (addsi3_carryin_<optab>): Likewise.
11980 (addsi3_carryin_alt2_<optab>): Likewise.
11981 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11982 for arm_restrict_it.
11983 (subsi3_carryin): Likewise.
11984 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11985 (minmax_arithsi): Disable for arm_restrict_it.
11986 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11987 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11988 (satsi_<SAT:code>_shift): Likewise.
11989 (arm_shiftsi3): Add alternative for 16-bit encoding.
11990 (arm32_movhf): Disable for arm_restrict_it.
11991 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11992 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11994 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11996 * config/arm/arm.md (attribute "insn"): Delete.
11997 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11998 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11999 (not_shiftsi): Update for attribute change.
12000 (not_shiftsi_compare0): Likewise.
12001 (not_shiftsi_compare0_scratch): Likewise.
12002 (arm_one_cmplsi2): Likewise.
12003 (thumb1_one_cmplsi2): Likewise.
12004 (notsi_compare0): Likewise.
12005 (notsi_compare0_scratch): Likewise.
12006 (thumb1_movdi_insn): Likewise.
12007 (arm_movsi_insn): Likewise.
12008 (movhi_insn_arch4): Likewise.
12009 (movhi_bytes): Likewise.
12010 (arm_movqi_insn): Likewise.
12011 (thumb1_movqi_insn): Likewise.
12012 (arm32_movhf): Likewise.
12013 (thumb1_movhf): Likewise.
12014 (arm_movsf_soft_insn): Likewise.
12015 (thumb1_movsf_insn): Likewise.
12016 (thumb_movdf_insn): Likewise.
12017 (movsicc_insn): Likewise.
12018 (movsfcc_soft_insn): Likewise.
12019 (and_scc): Likewise.
12020 (cond_move): Likewise.
12021 (if_move_not): Likewise.
12022 (if_not_move): Likewise.
12023 (if_shift_move): Likewise.
12024 (if_move_shift): Likewise.
12025 (if_shift_shift): Likewise.
12026 (if_not_arith): Likewise.
12027 (if_arith_not): Likewise.
12028 (cond_move_not): Likewise.
12029 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12030 (neon_mov<mode>): Likewise.
12031 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12032 (thumb2_movsi_vfp): Likewise.
12033 (movsf_vfp): Likewise.
12034 (thumb2_movsf_vfp): Likewise.
12035 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12037 (cortexa7_older_only): Likewise.
12038 (cortexa7_younger): Likewise.
12039 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12040 (1020alu_shift_op): Likewise.
12041 (1020alu_shift_reg_op): Likewise.
12042 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12043 (alu_shift_op): Likewise.
12044 (alu_shift_reg_op): Likewise.
12045 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12046 (11_alu_shift_op): Likewise.
12047 (11_alu_shift_reg_op): Likewise.
12048 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12049 (9_alu_shift_reg_op): Likewise.
12050 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12052 (cortex_a15_alu_shift): Likewise.
12053 (cortex_a15_alu_shift_reg): Likewise.
12054 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12055 (cortex_a5_alu_shift): Likewise.
12056 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12058 (cortex_a53_alu_shift): Likewise.
12059 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12061 (cortex_a7_alu_reg): Likewise.
12062 (cortex_a7_alu_shift): Likewise.
12063 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12064 (cortex_a8_alu_shift): Likewise.
12065 (cortex_a8_alu_shift_reg): Likewise.
12066 (cortex_a8_mov): Likewise.
12067 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12068 (cortex_a9_dp_shift): Likewise.
12069 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12070 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12071 (cortex_r4_mov): Likewise.
12072 (cortex_r4_alu_shift): Likewise.
12073 (cortex_r4_alu_shift_reg): Likewise.
12074 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12075 (526_alu_shift_op): Likewise.
12076 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12077 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12078 (626te_alu_shift_op): Likewise.
12079 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12080 (726te_alu_op): Likewise.
12081 (726te_alu_shift_op): Likewise.
12082 (726te_alu_shift_reg_op): Likewise.
12083 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12084 (mp626_alu_shift_op): Likewise.
12085 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12086 (pj4_alu_e1_conds): Likewise.
12087 (pj4_alu): Likewise.
12088 (pj4_alu_conds): Likewise.
12089 (pj4_shift): Likewise.
12090 (pj4_shift_conds): Likewise.
12091 (pj4_alu_shift): Likewise.
12092 (pj4_alu_shift_conds): Likewise.
12094 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12096 * config/arm/predicates.md (shiftable_operator_strict_it):
12098 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12099 Disable cond_exec version for arm_restrict_it.
12100 (thumb2_smaxsi3): Convert to generate cond_exec.
12101 (thumb2_sminsi3): Likewise.
12102 (thumb32_umaxsi3): Likewise.
12103 (thumb2_uminsi3): Likewise.
12104 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12105 (thumb2_neg_abssi2): Likewise.
12106 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12107 (thumb2_movsicc_insn): Adjust alternatives.
12108 (thumb2_mov_negscc): Disable for arm_restrict_it.
12109 (thumb2_mov_negscc_strict_it): New pattern.
12110 (thumb2_mov_notscc_strict_it): New pattern.
12111 (thumb2_mov_notscc): Disable for arm_restrict_it.
12112 (thumb2_ior_scc): Likewise.
12113 (thumb2_ior_scc_strict_it): New pattern.
12114 (thumb2_cond_move): Adjust for arm_restrict_it.
12115 (thumb2_cond_arith): Disable for arm_restrict_it.
12116 (thumb2_cond_arith_strict_it): New pattern.
12117 (thumb2_cond_sub): Adjust for arm_restrict_it.
12118 (thumb2_movcond): Likewise.
12119 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12120 (thumb2_zero_extendhisi2_v6): Likewise.
12121 (thumb2_zero_extendqisi2_v6): Likewise.
12122 (orsi_notsi_si): Likewise.
12123 (orsi_not_shiftsi_si): Likewise.
12125 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12127 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12128 instruction sequence is 1 byte shorter.
12130 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12132 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12133 it is not needed after split.
12135 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12138 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12139 second label for nonlocal goto receivers. Don't output pic base labels
12140 unless we're producing PIC; mark that action unreachable().
12141 (ix86_save_reg): If the function contains a nonlocal label, save the
12143 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12144 * config/darwin.c (emitted_pic_label_num): New GTY.
12145 (update_pic_label_number_if_needed): New.
12146 (machopic_output_function_base_name): Adjust for nonlocal receiver
12148 (machopic_should_output_picbase_label): New.
12149 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12150 (nonlocal_goto_receiver): New insn and split.
12152 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12154 * config/aarch64/aarch64-builtins.c
12155 (aarch64_fold_builtin): Fold abs in all modes.
12156 * config/aarch64/aarch64-simd-builtins.def
12157 (abs): Enable for all modes.
12158 * config/aarch64/arm_neon.h
12159 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12160 (vabs_f64): Add missing intrinsic.
12162 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12164 * config/aarch64/arm_neon.h (vabs_s64): New function
12166 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12169 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12170 * config/avr/avr.md (adjust_len): Add `round'.
12171 * config/avr/avr-protos.h (avr_out_round): New prototype.
12172 (avr_out_plus): Add `out_label' argument.
12173 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12174 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12175 Handle the case where `insn' is just a pattern.
12176 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12177 (avr_out_round): New function.
12178 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12180 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12182 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12183 Add check for flag_stack_usage to handle -fstack-usage support
12185 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12187 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12188 interaction for new Power8 flags and VSX.
12190 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12192 PR middle-end/57698
12193 * tree-inline.c (expand_call_inline): Emit errors during
12194 early_inlining only if optimization is not turned on.
12196 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12200 * passes.c (init_optimization_passes): Move the construction of
12201 the pass hierarchy into a new passes.def file.
12203 * Makefile.in (passes.o): Add dependency on passes.def.
12205 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12207 * passes.c (init_optimization_passes): Introduce macros for
12208 constructing the tree of passes (INSERT_PASSES_AFTER,
12209 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12210 TERMINATE_PASS_LIST).
12212 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12213 Wei Mi <wmi@google.com>
12215 PR rtl-optimization/57878
12216 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12218 (reload_pseudo_compare_func): Check nregs first for reload
12221 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12223 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12225 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12227 * read-rtl.c (validate_const_int): Once an invalid character is
12228 seen, quit the loop.
12230 * gengtype.c (write_roots): Similarly once we find the "deletable"
12231 or "if_marked" option.
12233 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12235 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12236 "xtab" and "sat". Move value "clz" from here to ...
12237 (attriubte "type"): ... here.
12238 (satsi_<SAT:code>): Delete "insn" attribute.
12239 (satsi_<SAT:code>_shift): Likewise.
12240 (arm_zero_extendqisi2addsi): Likewise.
12241 (arm_extendqisi2addsi): Likewise.
12242 (clzsi2): Update for attribute changes.
12243 (rbitsi2): Likewise.
12244 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12246 (arm_usatsihi): Likewise.
12247 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12249 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12251 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12252 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12253 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12254 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12255 in alphabetical order.
12256 (attribute "core_cycles"): Update for attribute changes.
12257 (arm_addsi3): Likewise.
12258 (addsi3_compare0): Likewise.
12259 (addsi3_compare0_scratch): Likewise.
12260 (addsi3_compare_op1): Likewise.
12261 (addsi3_compare_op2): Likewise.
12262 (compare_addsi2_op0): Likewise.
12263 (compare_addsi2_op1): Likewise.
12264 (addsi3_carryin_shift_<optab>): Likewise.
12265 (subsi3_carryin_shift): Likewise.
12266 (rsbsi3_carryin_shift): Likewise.
12267 (arm_subsi3_insn): Likewise.
12268 (subsi3_compare0): Likewise.
12269 (subsi3_compare): Likewise.
12270 (arm_andsi3_insn): Likewise.
12271 (thumb1_andsi3_insn): Likewise.
12272 (andsi3_compare0): Likewise.
12273 (andsi3_compare0_scratch): Likewise.
12274 (zeroextractsi_compare0_scratch
12275 (andsi_not_shiftsi_si): Likewise.
12276 (iorsi3_insn): Likewise.
12277 (iorsi3_compare0): Likewise.
12278 (iorsi3_compare0_scratch): Likewise.
12279 (arm_xorsi3): Likewise.
12280 (thumb1_xorsi3_insn): Likewise.
12281 (xorsi3_compare0): Likewise.
12282 (xorsi3_compare0_scratch): Likewise.
12283 (satsi_<SAT:code>_shift): Likewise.
12285 (arm_shiftsi3): Likewise.
12286 (shiftsi3_compare0): Likewise.
12287 (not_shiftsi): Likewise.
12288 (not_shiftsi_compare0): Likewise.
12289 (not_shiftsi_compare0_scratch): Likewise.
12290 (arm_one_cmplsi2): Likewise.
12291 (thumb_one_complsi2): Likewise.
12292 (notsi_compare0): Likewise.
12293 (notsi_compare0_scratch): Likewise.
12294 (thumb1_zero_extendhisi2): Likewise.
12295 (arm_zero_extendhisi2): Likewise.
12296 (arm_zero_extendhisi2_v6): Likewise.
12297 (arm_zero_extendhisi2addsi): Likewise.
12298 (thumb1_zero_extendqisi2): Likewise.
12299 (thumb1_zero_extendqisi2_v6): Likewise.
12300 (arm_zero_extendqisi2): Likewise.
12301 (arm_zero_extendqisi2_v6): Likewise.
12302 (arm_zero_extendqisi2addsi): Likewise.
12303 (thumb1_extendhisi2): Likewise.
12304 (arm_extendhisi2): Likewise.
12305 (arm_extendhisi2_v6): Likewise.
12306 (arm_extendqisi): Likewise.
12307 (arm_extendqisi_v6): Likewise.
12308 (arm_extendqisi2addsi): Likewise.
12309 (thumb1_extendqisi2): Likewise.
12310 (thumb1_movdi_insn): Likewise.
12311 (arm_movsi_insn): Likewise.
12312 (movsi_compare0): Likewise.
12313 (movhi_insn_arch4): Likewise.
12314 (movhi_bytes): Likewise.
12315 (arm_movqi_insn): Likewise.
12316 (thumb1_movqi_insn): Likewise.
12317 (arm32_movhf): Likewise.
12318 (thumb1_movhf): Likewise.
12319 (arm_movsf_soft_insn): Likewise.
12320 (thumb1_movsf_insn): Likewise.
12321 (movdf_soft_insn): Likewise.
12322 (thumb_movdf_insn): Likewise.
12323 (arm_cmpsi_insn): Likewise.
12324 (cmpsi_shiftsi): Likewise.
12325 (cmpsi_shiftsi_swp): Likewise.
12326 (arm_cmpsi_negshiftsi_si): Likewise.
12327 (movsicc_insn): Likewise.
12328 (movsfcc_soft_insn): Likewise.
12329 (arith_shiftsi): Likewise.
12330 (arith_shiftsi_compare0
12331 (arith_shiftsi_compare0_scratch
12332 (sub_shiftsi): Likewise.
12333 (sub_shiftsi_compare0
12334 (sub_shiftsi_compare0_scratch
12335 (and_scc): Likewise.
12336 (cond_move): Likewise.
12337 (if_plus_move): Likewise.
12338 (if_move_plus): Likewise.
12339 (if_move_not): Likewise.
12340 (if_not_move): Likewise.
12341 (if_shift_move): Likewise.
12342 (if_move_shift): Likewise.
12343 (if_shift_shift): Likewise.
12344 (if_not_arith): Likewise.
12345 (if_arith_not): Likewise.
12346 (cond_move_not): Likewise.
12347 (thumb1_ashlsi3): Set type attribute.
12348 (thumb1_ashrsi3): Likewise.
12349 (thumb1_lshrsi3): Likewise.
12350 (thumb1_rotrsi3): Likewise.
12351 (shiftsi3_compare0_scratch): Likewise.
12352 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12353 (neon_mov<mode>): Likewise.
12354 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12356 (thumb2_movsi_insn): Likewise.
12357 (thumb2_cmpsi_neg_shiftsi): Likewise.
12358 (thumb2_extendqisi_v6): Likewise.
12359 (thumb2_zero_extendhisi2_v6): Likewise.
12360 (thumb2_zero_extendqisi2_v6): Likewise.
12361 (thumb2_shiftsi3_short): Likewise.
12362 (thumb2_addsi3_compare0_scratch): Likewise.
12363 (orsi_not_shiftsi_si): Likewise.
12364 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12365 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12367 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12368 (1020alu_shift_op): Likewise.
12369 (1020alu_shift_reg_op): Likewise.
12370 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12371 (alu_shift_op): Likewise.
12372 (alu_shift_reg_op): Likewise.
12373 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12374 (11_alu_shift_op): Likewise.
12375 (11_alu_shift_reg_op): Likewise.
12376 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12377 (9_alu_shift_reg_op): Likewise.
12378 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12380 (cortex_a15_alu_shift): Likewise.
12381 (cortex_a15_alu_shift_reg): Likewise.
12382 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12384 (cortex_a5_alu_shift): Likewise.
12385 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12387 (cortex_a53_alu_shift): Likewise.
12388 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12390 (cortex_a7_alu_reg): Likewise.
12391 (cortex_a7_alu_shift): Likewise.
12392 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12394 (cortex_a8_alu_shift): Likewise.
12395 (cortex_a8_alu_shift_reg): Likewise.
12396 (cortex_a8_mov): Likewise.
12397 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12398 (cortex_a9_dp_shift): Likewise.
12399 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12401 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12403 (cortex_r4_mov): Likewise.
12404 (cortex_r4_alu_shift): Likewise.
12405 (cortex_r4_alu_shift_reg): Likewise.
12406 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12407 (526_alu_shift_op): Likewise.
12408 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12409 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12410 (626te_alu_shift_op): Likewise.
12411 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12412 (726te_alu_op): Likewise.
12413 (726te_alu_shift_op): Likewise.
12414 (726te_alu_shift_reg_op): Likewise.
12415 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12416 (mp626_alu_shift_op): Likewise.
12417 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12418 (pj4_alu_e1_conds): Likewise.
12419 (pj4_alu): Likewise.
12420 (pj4_alu_conds): Likewise.
12421 (pj4_shift): Likewise.
12422 (pj4_shift_conds): Likewise.
12423 (pj4_alu_shift): Likewise.
12424 (pj4_alu_shift_conds): Likewise.
12425 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12427 (cortexa7_older_only): Likewise.
12428 (cortexa7_younger): Likewise.
12430 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12432 * ipa-pure-const.c (generate_summary): Rename to...
12433 (pure_const_generate_summary): ... this.
12435 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12437 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12439 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12442 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12445 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12447 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12448 enabled without -march=zEC12.
12449 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12452 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12454 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12455 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12456 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12457 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12458 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12459 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12460 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12462 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12463 and ISA_HAS_NMADD3_NMSUB3.
12464 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12465 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12466 (nmsub4<mode>, nmsub3<mode>): Likewise.
12467 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12469 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12471 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12472 TARGET_MIPS5400 checking.
12474 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12475 Peter Bergner <bergner@vnet.ibm.com>
12477 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12478 registers in the comment.
12479 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12480 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12481 rather than FIRST_PSEUDO_REGISTERS.
12483 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12485 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12486 enable extra ISA flags with TARGET_HTM.
12488 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12490 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12493 2013-07-15 Cong Hou <congh@google.com>
12495 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12496 in compare function for sorting.
12498 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12500 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12501 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12502 * config/rs6000/rs6000.opt: Add -mhtm option.
12503 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12504 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12505 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12506 __HTM__ if the HTM instructions are available.
12507 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12508 htm_spr_reg_operand): New define_predicates.
12509 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12510 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12512 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12513 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12514 HTM builtin functions.
12515 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12516 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12517 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12518 (rs6000_builtin_mask_calculate): Likewise.
12519 (rs6000_option_override_internal): Likewise.
12520 (bdesc_htm): Add new HTM builtin support.
12521 (htm_spr_num): New function.
12522 (htm_spr_regno): Likewise.
12523 (rs6000_htm_spr_icode): Likewise.
12524 (htm_expand_builtin): Likewise.
12525 (htm_init_builtins): Likewise.
12526 (rs6000_expand_builtin): Add support for HTM builtin functions.
12527 (rs6000_init_builtins): Likewise.
12528 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12530 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12531 (TARGET_HTM, MASK_HTM): Define macros.
12532 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12533 (FIXED_REGISTERS): Likewise.
12534 (CALL_USED_REGISTERS): Likewise.
12535 (CALL_REALLY_USED_REGISTERS): Likewise.
12536 (REG_ALLOC_ORDER): Likewise.
12537 (enum reg_class): Likewise.
12538 (REG_CLASS_NAMES): Likewise.
12539 (REG_CLASS_CONTENTS): Likewise.
12540 (REGISTER_NAMES): Likewise.
12541 (ADDITIONAL_REGISTER_NAMES): Likewise.
12542 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12543 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12544 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12545 * config/rs6000/htm.md: New file.
12546 * config/rs6000/htmintrin.h: New file.
12547 * config/rs6000/htmxlintrin.h: New file.
12549 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12551 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12552 Define SYMBOL_TINY_GOT, update comment.
12553 * config/aarch64/aarch64.c
12554 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12555 (aarch64_expand_mov_immediate): Likewise.
12556 (aarch64_print_operand): Likewise.
12557 (aarch64_classify_symbol): Likewise.
12558 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12559 (ldr_got_tiny): Define.
12561 2013-07-13 Tobias Grosser <tobias@grosser.es>
12563 PR tree-optimization/54094
12564 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12565 scheduling dimension for the parallelism check from the polyhedral
12566 information in the AST.
12567 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12570 2013-07-13 Jason Merrill <jason@redhat.com>
12572 * print-tree.c (debug_vec_tree): Use debug_raw.
12573 (debug_raw (vec<tree, va_gc> &)): New.
12574 (debug_raw (vec<tree, va_gc> *)): New.
12575 * tree.h: Declare them.
12577 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12579 * ifcvt.c (ifcvt_after_combine): New static variable.
12580 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12582 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12583 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12584 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12586 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12588 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12590 2013-07-12 Michael Matz <matz@suse.de>
12592 PR middle-end/55771
12593 * convert.c (convert_to_real): Reject non-float inner types.
12595 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12597 * config/aarch64/aarch64-protos.h
12598 (aarch64_simd_immediate_valid_for_move): Remove.
12599 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12600 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12602 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12604 2013-07-11 Steve Ellcey <sellcey@mips.com>
12606 * config/mips/mips.c (mips_conditional_register_usage): Do not
12607 use t[0-7] registers in MIPS16 mode when optimizing for size.
12609 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12611 * config/i386/i386.c (dispatch_function_versions): Fix array
12612 indexing of function_version_info to match actual_versions.
12614 2013-07-11 Teresa Johnson <tejohnson@google.com>
12616 * vec.h (struct va_gc): Move release out-of-line.
12617 (va_gc::release): Call ggc_free on released vec.
12619 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12621 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12622 Require GOT register as additional operand in UNSPEC.
12623 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12624 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12625 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12626 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12627 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12628 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12629 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12631 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12634 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12635 name seen by assembler/linker rather if available.
12637 2013-07-11 Andreas Schwab <schwab@suse.de>
12639 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12641 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12643 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12645 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12647 * doc/tm.texi.in: Move hook documentation to ....
12648 * target.def: ... here.
12650 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12651 text on @hook line.
12652 * doc/tm.texi: Regenerate.
12654 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12657 * doc/invoke.texi: Document -Wconditionally-supported.
12659 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12662 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12665 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12668 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12669 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12670 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12671 Remove duplicate devices.
12672 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12673 * config/avr/t-multilib: Regenerate.
12674 * config/avr/avr-tables.opt: Regenerate.
12675 * doc/avr-mmcu.texi: Regenerate.
12677 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12680 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12682 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12684 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12685 the cost of MULT when optimizing for size.
12687 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12689 * config/cr16/cr16-protos.h: Don't include target.h.
12691 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12693 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12694 adjust register size for TDmode and TFmode for VSX registers.
12696 2013-07-08 Kai Tietz <ktietz@redhat.com>
12699 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12700 hook_bool_const_tree_true.
12702 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12704 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12705 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12706 * config/s390/s390.md: Define FPR*_REGNUM constants.
12707 Fix FPR2_REGNUM constant (18 -> 17).
12708 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12709 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12710 ("trunc<BFP:mode><DFP_ALL:mode>2")
12711 ("trunc<DFP_ALL:mode><BFP:mode>2")
12712 ("*extend<BFP:mode><DFP_ALL:mode>2")
12713 ("*extend<DFP_ALL:mode><BFP:mode>2")
12714 ("extend<BFP:mode><DFP_ALL:mode>2")
12715 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12718 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12720 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12722 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12724 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12725 and cfun_fpr_bit_p to cfun_fpr_save_p.
12726 (s390_frame_area, s390_register_info, s390_frame_info)
12727 (s390_emit_prologue, s390_emit_epilogue)
12728 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12730 * config/s390/s390.h: Define *_REGNUM macros for floating point
12733 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12735 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12737 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12739 PR rtl-optimization/57786
12740 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12741 and break out of the loop when it is set to false.
12743 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12746 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12747 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12748 (const_int 63)) 0)).
12749 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12750 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12751 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12753 PR rtl-optimization/57829
12754 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12755 mask bits outside of mode are just sign-extension from mode to HWI.
12757 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12759 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12760 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12761 adjust_address instead of change_address to keep info about alignment.
12762 (emit_strmov): Remove.
12763 (emit_memmov): New function.
12764 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12765 (expand_movmem_epilogue): Likewise and return updated rtx for
12767 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12768 destination and source.
12769 (decide_alignment): Refactor, handle vector_loop.
12770 (ix86_expand_movmem): Likewise.
12771 (ix86_expand_setmem): Likewise.
12772 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12774 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12776 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12777 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12779 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12781 * config/i386/sse.md (sse_movlhps): Change alternative 3
12782 of operand 2 to "m".
12784 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12787 * config/i386/sse.md (iptr): New mode attribute.
12788 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12789 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12790 (<sse>_vmmul<mode>3): Ditto.
12791 (<sse>_vmdiv<mode>3): Ditto.
12792 (sse_vmrcpv4sf2): Ditto.
12793 (<sse>_vmsqrt<mode>2): Ditto.
12794 (sse_vmrsqrtv4sf2): Ditto.
12795 (<sse>_vm<code><mode>3): Ditto.
12796 (avx_vmcmp<mode>3): Ditto.
12797 (<sse>_vmmaskcmp<mode>3): Ditto.
12798 (<sse>_comi): Ditto.
12799 (<sse>_ucomi): Ditto.
12800 (*xop_vmfrcz_<mode>): Ditto.
12801 (*fmai_fmadd_<mode>): Ditto.
12802 (*fmai_fmsub_<mode>): Ditto.
12803 (*fmai_fnmadd_<mode>): Ditto.
12804 (*fmai_fnmsub_<mode>): Ditto.
12805 (*fma4i_vmfmadd_<mode>): Ditto.
12806 (*fma4i_vmfmsub_<mode>): Ditto.
12807 (*fma4i_vmfnmadd_<mode>): Ditto.
12808 (*fma4i_vmfnmsub_<mode>): Ditto.
12809 (*xop_vmfrcz_<mode>): Ditto.
12810 (sse_cvtps2pi): Ditto.
12811 (sse_cvttps2pi): Ditto.
12812 (sse_cvtss2si): Ditto.
12813 (sse_cvtss2si_2): Ditto.
12814 (sse_cvtss2siq_2): Ditto.
12815 (sse_cvttss2si): Ditto.
12816 (sse_cvttss2siq): Ditto.
12817 (sse_cvtsd2si): Ditto.
12818 (sse_cvtsd2si_2): Ditto.
12819 (sse_cvtsd2siq_2): Ditto.
12820 (sse_cvttsd2si): Ditto.
12821 (sse_cvttsd2siq): Ditto.
12822 (sse_cvtsd2ss): Ditto.
12823 (sse_cvtss2sd): Ditto.
12824 (avx2_pbroadcast<mode>): Ditto.
12825 (avx2_pbroadcast<mode>_1): Ditto.
12826 (*avx_vperm_broadcast_v4sf): Ditto.
12828 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12829 (sse_movlhps): Ditto.
12830 (sse_storehps): Ditto.
12831 (sse_loadhps): Ditto.
12832 (sse_storelps): Ditto.
12833 (sse_loadlps): Ditto.
12834 (*vec_concatv4sf): Ditto.
12835 (*vec_interleave_highv2df): Ditto.
12836 (*vec_interleave_lowv2df): Ditto.
12837 (*vec_extractv2df_1_sse): Ditto.
12838 (*vec_extractv2df_0_sse): Ditto.
12839 (sse2_storelpd): Ditto.
12840 (sse2_loadlpd): Ditto.
12841 (sse2_movsd): Ditto.
12842 (*vec_concatv4si): Ditto.
12843 (vec_concatv2di): Ditto.
12845 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12846 for Intel asm dialect.
12847 (mmx_punpcklwd): Ditto.
12848 (mmx_punpckldq): Ditto.
12850 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12851 for intel assembler dialect.
12853 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12856 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12857 for BUILT_IN_C{LZ,LRSB}*.
12858 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12859 * tree-vrp.c (extract_range_basic): Handle
12860 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12861 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12862 fall thru to code calling set_value*.
12863 * builtins.c (expand_builtin): Remove *IMAX cases.
12864 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12865 if width is bigger than 2*HWI.
12867 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12869 PR rtl-optimization/55342
12870 * lra-int.h (lra_subreg_reload_pseudos): New.
12871 * lra.c: Add undoing optional reloads to the block diagram.
12872 (lra_subreg_reload_pseudos): New.
12873 (lra_optional_reload_pseudos): Change comments.
12874 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12875 lra_optional_reload_pseudos after undo transformations.
12876 * lra-assigns.c (pseudo_prefix_title): New.
12877 (lra_setup_reg_renumber): Use it.
12878 (spill_for): Ditto. Check subreg reload pseudos too.
12879 (assign_by_spills): Consider subreg reload pseudos too.
12880 * lra-constraints.c (simplify_operand_subreg): Use
12881 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12882 (curr_insn_transform): Recognize and do optional reloads.
12883 (undo_optional_reloads): New.
12884 (lra_undo_inheritance): Call undo_optional_reloads.
12886 2013-07-05 Thomas Quinot <quinot@adacore.com>
12888 * tree-complex.c (expand_complex_operations_1): Fix typo.
12890 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12892 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12893 (tune_params): New member 'const vec_costs'.
12894 * config/aarch64/aarch64.c (generic_vector_cost): New.
12895 (generic_tunings): New member 'generic_vector_cost'.
12896 (aarch64_builtin_vectorization_cost): New.
12897 (aarch64_add_stmt_cost): New.
12898 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12899 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12901 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12904 * config/i386/predicates.md (vsib_address_operand): Disallow
12905 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12907 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12909 PR middle-end/55030
12910 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12911 expand_builtin_setjmp_receiver.
12912 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12913 with NULL for the label parameter.
12914 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12915 the frame-pointer. Adjust comments.
12916 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12917 only if LABEL is non-NULL.
12919 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12921 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12922 (ARM_ABI_AAPCS64): Ditto.
12924 (ARM_DEFAULT_ABI): Ditto.
12926 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12928 * config/aarch64/aarch64-builtins.c
12929 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12930 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12932 * config/aarch64/aarch64-simd.md
12933 (aarch64_ld1<VALL:mode>): New.
12934 (aarch64_st1<VALL:mode>): Likewise.
12935 * config/aarch64/arm_neon.h
12936 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12938 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12940 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12942 (ix86_option_override_internal):Turn on all -mavx target flags by
12943 default as they are dependent on AVX anyway.
12945 2013-07-02 Cary Coutant <ccoutant@google.com>
12947 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12948 deterministic hash.
12949 (loc_checksum_ordered): Likewise.
12950 (hash_loc_operands): Remove inline keyword.
12952 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12954 PR tree-optimization/57741
12955 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12956 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12957 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12958 Allow REAL_CST step_exprs if flag_associative_math.
12959 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12961 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12963 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12965 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12967 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12969 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12971 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12973 (iorsi3_insn): Likewise.
12974 (arm_xorsi3): Likewise.
12976 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12978 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12980 (attribute "type"): ... here, and prefix with "wmmx_".
12981 (attribute "core_cycles"): Update for attribute changes.
12982 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12983 (tbcstv4hi): Likewise.
12984 (tbcstv2si): Likewise.
12985 (iwmmxt_iordi3): Likewise.
12986 (iwmmxt_xordi3): Likewise.
12987 (iwmmxt_anddi3): Likewise.
12988 (iwmmxt_nanddi3): Likewise.
12989 (iwmmxt_arm_movdi): Likewise.
12990 (iwmmxt_movsi_insn): Likewise.
12991 (mov<mode>_internal): Likewise.
12992 (and<mode>3_iwmmxt): Likewise.
12993 (ior<mode>3_iwmmxt): Likewise.
12994 (xor<mode>3_iwmmxt): Likewise.
12995 (add<mode>3_iwmmxt): Likewise.
12996 (ssaddv8qi3): Likewise.
12997 (ssaddv4hi3): Likewise.
12998 (ssaddv2si3): Likewise.
12999 (usaddv8qi3): Likewise.
13000 (usaddv4hi3): Likewise.
13001 (usaddv2si3): Likewise.
13002 (sub<mode>3_iwmmxt): Likewise.
13003 (sssubv8qi3): Likewise.
13004 (sssubv4hi3): Likewise.
13005 (sssubv2si3): Likewise.
13006 (ussubv8qi3): Likewise.
13007 (ussubv4hi3): Likewise.
13008 (ussubv2si3): Likewise.
13009 (mulv4hi3_iwmmxt): Likewise.
13010 (smulv4hi3_highpart): Likewise.
13011 (umulv4hi3_highpart): Likewise.
13012 (iwmmxt_wmacs): Likewise.
13013 (iwmmxt_wmacsz): Likewise.
13014 (iwmmxt_wmacu): Likewise.
13015 (iwmmxt_wmacuz): Likewise.
13016 (iwmmxt_clrdi): Likewise.
13017 (iwmmxt_clrv8qi): Likewise.
13018 (iwmmxt_clr4hi): Likewise.
13019 (iwmmxt_clr2si): Likewise.
13020 (iwmmxt_uavgrndv8qi3): Likewise.
13021 (iwmmxt_uavgrndv4hi3): Likewise.
13022 (iwmmxt_uavgv8qi3): Likewise.
13023 (iwmmxt_uavgv4hi3): Likewise.
13024 (iwmmxt_tinsrb): Likewise.
13025 (iwmmxt_tinsrh): Likewise.
13026 (iwmmxt_tinsrw): Likewise.
13027 (iwmmxt_textrmub): Likewise.
13028 (iwmmxt_textrmsb): Likewise.
13029 (iwmmxt_textrmuh): Likewise.
13030 (iwmmxt_textrmsh): Likewise.
13031 (iwmmxt_textrmw): Likewise.
13032 (iwmxxt_wshufh): Likewise.
13033 (eqv8qi3): Likewise.
13034 (eqv4hi3): Likewise.
13035 (eqv2si3): Likewise.
13036 (gtuv8qi3): Likewise.
13037 (gtuv4hi3): Likewise.
13038 (gtuv2si3): Likewise.
13039 (gtv8qi3): Likewise.
13040 (gtv4hi3): Likewise.
13041 (gtv2si3): Likewise.
13042 (smax<mode>3_iwmmxt): Likewise.
13043 (umax<mode>3_iwmmxt): Likewise.
13044 (smin<mode>3_iwmmxt): Likewise.
13045 (umin<mode>3_iwmmxt): Likewise.
13046 (iwmmxt_wpackhss): Likewise.
13047 (iwmmxt_wpackwss): Likewise.
13048 (iwmmxt_wpackdss): Likewise.
13049 (iwmmxt_wpackhus): Likewise.
13050 (iwmmxt_wpackwus): Likewise.
13051 (iwmmxt_wpackdus): Likewise.
13052 (iwmmxt_wunpckihb): Likewise.
13053 (iwmmxt_wunpckihh): Likewise.
13054 (iwmmxt_wunpckihw): Likewise.
13055 (iwmmxt_wunpckilb): Likewise.
13056 (iwmmxt_wunpckilh): Likewise.
13057 (iwmmxt_wunpckilw): Likewise.
13058 (iwmmxt_wunpckehub): Likewise.
13059 (iwmmxt_wunpckehuh): Likewise.
13060 (iwmmxt_wunpckehuw): Likewise.
13061 (iwmmxt_wunpckehsb): Likewise.
13062 (iwmmxt_wunpckehsh): Likewise.
13063 (iwmmxt_wunpckehsw): Likewise.
13064 (iwmmxt_wunpckelub): Likewise.
13065 (iwmmxt_wunpckeluh): Likewise.
13066 (iwmmxt_wunpckeluw): Likewise.
13067 (iwmmxt_wunpckelsb): Likewise.
13068 (iwmmxt_wunpckelsh): Likewise.
13069 (iwmmxt_wunpckelsw): Likewise.
13070 (ror<mode>3): Likewise.
13071 (ashr<mode>3_iwmmxt): Likewise.
13072 (lshr<mode>3_iwmmxt): Likewise.
13073 (ashl<mode>3_iwmmxt): Likewise.
13074 (ror<mode>3_di): Likewise.
13075 (ashr<mode>3_di): Likewise.
13076 (lshr<mode>3_di): Likewise.
13077 (ashl<mode>3_di): Likewise.
13078 (iwmmxt_wmadds): Likewise.
13079 (iwmmxt_wmaddu): Likewise.
13080 (iwmmxt_tmia): Likewise.
13081 (iwmmxt_tmiaph): Likewise.
13082 (iwmmxt_tmiabb): Likewise.
13083 (iwmmxt_tmiatb): Likewise.
13084 (iwmmxt_tmiabt): Likewise.
13085 (iwmmxt_tmiatt): Likewise.
13086 (iwmmxt_tmovmskb): Likewise.
13087 (iwmmxt_tmovmskh): Likewise.
13088 (iwmmxt_tmovmskw): Likewise.
13089 (iwmmxt_waccb): Likewise.
13090 (iwmmxt_wacch): Likewise.
13091 (iwmmxt_waccw): Likewise.
13092 (iwmmxt_waligni): Likewise.
13093 (iwmmxt_walignr): Likewise.
13094 (iwmmxt_walignr0): Likewise.
13095 (iwmmxt_walignr1): Likewise.
13096 (iwmmxt_walignr2): Likewise.
13097 (iwmmxt_walignr3): Likewise.
13098 (iwmmxt_wsadb): Likewise.
13099 (iwmmxt_wsadh): Likewise.
13100 (iwmmxt_wsadbz): Likewise.
13101 (iwmmxt_wsadhz): Likewise.
13102 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13103 (iwmmxt_wabsdiffb): Likewise.
13104 (iwmmxt_wabsdiffh): Likewise.
13105 (iwmmxt_wabsdiffw): Likewise.
13106 (iwmmxt_waddsubhx): Likewise
13107 (iwmmxt_wsubaddhx): Likewise.
13108 (addc<mode>3): Likewise.
13109 (iwmmxt_avg4): Likewise.
13110 (iwmmxt_avg4r): Likewise.
13111 (iwmmxt_wmaddsx): Likewise.
13112 (iwmmxt_wmaddux): Likewise.
13113 (iwmmxt_wmaddsn): Likewise.
13114 (iwmmxt_wmaddun): Likewise.
13115 (iwmmxt_wmulwsm): Likewise.
13116 (iwmmxt_wmulwum): Likewise.
13117 (iwmmxt_wmulsmr): Likewise.
13118 (iwmmxt_wmulumr): Likewise.
13119 (iwmmxt_wmulwsmr): Likewise.
13120 (iwmmxt_wmulwumr): Likewise.
13121 (iwmmxt_wmulwl): Likewise.
13122 (iwmmxt_wqmulm): Likewise.
13123 (iwmmxt_wqmulwm): Likewise.
13124 (iwmmxt_wqmulmr): Likewise.
13125 (iwmmxt_wqmulwmr): Likewise.
13126 (iwmmxt_waddbhusm): Likewise.
13127 (iwmmxt_waddbhusl): Likewise.
13128 (iwmmxt_wqmiabb): Likewise.
13129 (iwmmxt_wqmiabt): Likewise.
13130 (iwmmxt_wqmiatb): Likewise.
13131 (iwmmxt_wqmiatt): Likewise.
13132 (iwmmxt_wqmiabbn): Likewise.
13133 (iwmmxt_wqmiabtn): Likewise.
13134 (iwmmxt_wqmiatbn): Likewise.
13135 (iwmmxt_wqmiattn): Likewise.
13136 (iwmmxt_wmiabb): Likewise.
13137 (iwmmxt_wmiabt): Likewise.
13138 (iwmmxt_wmiatb): Likewise.
13139 (iwmmxt_wmiatt): Likewise.
13140 (iwmmxt_wmiabbn): Likewise.
13141 (iwmmxt_wmiabtn): Likewise.
13142 (iwmmxt_wmiatbn): Likewise.
13143 (iwmmxt_wmiattn): Likewise.
13144 (iwmmxt_wmiawbb): Likewise.
13145 (iwmmxt_wmiawbt): Likewise.
13146 (iwmmxt_wmiawtb): Likewise.
13147 (iwmmxt_wmiawtt): Likewise.
13148 (iwmmxt_wmiawbbn): Likewise.
13149 (iwmmxt_wmiawbtn): Likewise.
13150 (iwmmxt_wmiawtbn): Likewise.
13151 (iwmmxt_wmiawttn): Likewise.
13152 (iwmmxt_wmerge): Likewise.
13153 (iwmmxt_tandc<mode>3): Likewise.
13154 (iwmmxt_torc<mode>3): Likewise.
13155 (iwmmxt_torvsc<mode>3): Likewise.
13156 (iwmmxt_textrc<mode>3): Likewise.
13157 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13158 (wmmxt_pack): Likewise.
13159 (wmmxt_mult_c1): Likewise.
13160 (wmmxt_mult_c2): Likewise.
13161 (wmmxt_alu_c1): Likewise.
13162 (wmmxt_alu_c2): Likewise.
13163 (wmmxt_alu_c3): Likewise.
13164 (wmmxt_transfer_c1): Likewise.
13165 (wmmxt_transfer_c2): Likewise.
13166 (wmmxt_transfer_c3): Likewise.
13167 (marvell_f_iwmmxt_wstr): Likewise.
13168 (marvell_f_iwmmxt_wldr): Likewise.
13170 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13172 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13174 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13177 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13178 * lra-constraints.c (need_for_split_p): Check call used hard regs
13179 living through calls.
13181 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13182 call used regs for call insn.
13184 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13187 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13188 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13190 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13192 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13194 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13196 * lra-constraints.c (need_for_split_p): Check call used hard regs
13197 living through calls.
13199 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13202 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13203 to tie with any other modes. Eliminate Altivec vector mode tests,
13204 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13205 code, to return 0 if testing MODE2 for a condition, if we've
13206 already tested MODE1 for the same condition.
13208 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13210 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13213 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13215 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13216 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13218 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13220 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13222 (aarch64_symbolic_constant_p): Remove.
13223 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13224 static. Fix line length and white space.
13225 (aarch64_symbolic_constant_p): Remove.
13226 * config/aarch64/predicates.md (aarch64_valid_symref):
13227 Use aarch64_classify_symbol_expression.
13229 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13231 * config/arm/constraints.md (Ts): New constraint.
13232 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13234 (compare_scc): Use "Ts" constraint for operand 0.
13235 (ior_scc_scc): Likewise.
13236 (and_scc_scc): Likewise.
13237 (and_scc_scc_nodom): Likewise.
13238 (ior_scc_scc_cmp): Likewise for operand 7.
13239 (and_scc_scc_cmp): Likewise.
13240 * config/arm/thumb2.md (thumb2_movsi_insn):
13241 Add alternatives for 16-bit encodings.
13242 (thumb2_movhi_insn): Likewise.
13243 (thumb2_movsicc_insn): Likewise.
13244 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13245 (thumb2_negscc): Use "Ts" constraint.
13246 Move mvn instruction outside cond_exec block.
13247 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13248 for 16-bit encodings.
13250 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13252 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13254 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13255 (mulsi3subsi): Likewise.
13256 (mulsidi3adddi): Likewise.
13257 (mulsidi3_v6): Likewise.
13258 (umulsidi3_v6): Likewise.
13259 (umulsidi3adddi_v6): Likewise.
13260 (smulsi3_highpart_v6): Likewise.
13261 (umulsi3_highpart_v6): Likewise.
13262 (mulhisi3tb): Likewise.
13263 (mulhisi3bt): Likewise.
13264 (mulhisi3tt): Likewise.
13265 (maddhisi4): Likewise.
13266 (maddhisi4tb): Likewise.
13267 (maddhisi4tt): Likewise.
13268 (maddhidi4): Likewise.
13269 (maddhidi4tb): Likewise.
13270 (maddhidi4tt): Likewise.
13271 (zeroextractsi_compare0_scratch): Likewise.
13272 (insv_zero): Likewise.
13273 (insv_t2): Likewise.
13274 (anddi_notzesidi_di): Likewise.
13275 (anddi_notsesidi_di): Likewise.
13276 (andsi_notsi_si): Likewise.
13277 (iordi_zesidi_di): Likewise.
13278 (xordi_zesidi_di): Likewise.
13279 (andsi_iorsi3_notsi): Likewise.
13280 (smax_0): Likewise.
13281 (smax_m1): Likewise.
13282 (smin_0): Likewise.
13283 (not_shiftsi): Likewise.
13284 (unaligned_loadsi): Likewise.
13285 (unaligned_loadhis): Likewise.
13286 (unaligned_loadhiu): Likewise.
13287 (unaligned_storesi): Likewise.
13288 (unaligned_storehi): Likewise.
13289 (extv_reg): Likewise.
13290 (extzv_t2): Likewise.
13291 (divsi3): Likewise.
13292 (udivsi3): Likewise.
13293 (arm_zero_extendhisi2addsi): Likewise.
13294 (arm_zero_extendqisi2addsi): Likewise.
13295 (compareqi_eq0): Likewise.
13296 (arm_extendhisi2_v6): Likewise.
13297 (arm_extendqisi2addsi): Likewise.
13298 (arm_movt): Likewise.
13299 (thumb2_ldrd): Likewise.
13300 (thumb2_ldrd_base): Likewise.
13301 (thumb2_ldrd_base_neg): Likewise.
13302 (thumb2_strd): Likewise.
13303 (thumb2_strd_base): Likewise.
13304 (thumb2_strd_base_neg): Likewise.
13305 (arm_negsi2): Add alternative for 16-bit encoding.
13306 (arm_one_cmplsi2): Likewise.
13308 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13310 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13311 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13312 (movdfcc): Likewise.
13313 * config/arm/vfp.md (*thumb2_movsf_vfp):
13314 Disable predication for arm_restrict_it.
13315 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13316 (*thumb2_movdfcc_vfp): Likewise.
13317 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13318 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13319 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13320 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13321 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13322 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13323 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13324 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13325 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13326 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13327 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13328 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13329 Disable predication for arm_restrict_it.
13331 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13333 * config/i386/bmiintrin.h (_bextr_u32): New.
13334 (_bextr_u64): Ditto.
13336 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13338 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13339 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13340 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13341 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13342 * config/mips/n32-elf.h: ...this new file.
13344 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13347 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13348 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13350 2013-06-27 Catherine Moore <clm@codesourcery.com>
13352 * config/mips/mips-tables.opt: Regenerate.
13353 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13354 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13355 * doc/invoke.texi: Add -m14kc.
13357 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13360 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13361 constraints of operand 1 and 2.
13364 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13365 to match RTL canonicalization. Swap predicates and
13366 constraints of operand 1 and 2.
13368 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13370 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13371 Process OP_INOUT regs for splitting too.
13373 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13375 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13376 decl before the loop, initialize to NULL.
13377 (vectorizable_load): Initialize ptr_incr to NULL.
13379 2013-06-27 Martin Jambor <mjambor@suse.cz>
13382 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13383 * ipa-ref.c (ipa_maybe_record_reference): New function.
13384 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13385 * ipa-cp.c (create_specialized_node): Record potential references from
13387 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13389 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13391 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13392 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13393 'mode' to force_reg.
13394 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13395 (aarch64_expand_mov_immediate): Likewise.
13397 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13399 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13400 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13402 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13404 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13405 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13406 (struct machine_function): Add tbegin_p.
13407 (s390_canonicalize_comparison): Fold CC mode compares to
13408 conditional jump if possible.
13409 (s390_emit_jump): Return the emitted jump.
13410 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13411 Handle CCRAWmode compares.
13412 (s390_option_override): Default to -mhtm if available.
13413 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13414 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13415 FPRs instead of df_regs_ever_live_p.
13416 (s390_optimize_nonescaping_tx): New function.
13417 (s390_init_frame_layout): Extend clobbered_regs array to cover
13419 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13420 (s390_expand_tbegin): New function.
13421 (enum s390_builtin): New enum definition.
13422 (code_for_builtin): New array definition.
13423 (s390_init_builtins): New function.
13424 (s390_expand_builtin): New function.
13425 (TARGET_INIT_BUILTINS): Define.
13426 (TARGET_EXPAND_BUILTIN): Define.
13427 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13428 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13429 (s390_alc_comparison): Likewise.
13430 * config/s390/s390-modes.def: Add CCRAWmode.
13431 * config/s390/s390.h (processor_flags): Add PF_TX.
13432 (TARGET_CPU_HTM): Define macro.
13433 (TARGET_HTM): Define macro.
13434 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13435 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13436 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13437 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13438 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13439 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13440 ("*cc_to_int"): Move up.
13441 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13442 constants other than 0.
13443 ("*ccraw_to_int"): New insn and splitter definition.
13444 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13445 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13446 ("tx_assist"): New expander.
13447 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13448 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13449 * config/s390/s390.opt: Add -mhtm option.
13450 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13451 * config/s390/htmxlintrin.h: New file.
13452 * config/s390/htmintrin.h: New file.
13453 * config/s390/s390intrin.h: New file.
13454 * doc/extend.texi: Document htm builtins.
13455 * config.gcc: Add the new header files to extra_headers.
13457 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13459 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13460 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13462 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13463 Pat Haugen <pthaugen@us.ibm.com>
13464 Peter Bergner <bergner@vnet.ibm.com>
13466 * config/rs6000/power8.md: New.
13467 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13468 setting for power8 entry.
13469 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13470 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13471 test for Power4/Power5 only.
13472 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13474 (force_new_group): Adjust comment.
13475 * config/rs6000/rs6000.md: Include power8.md.
13477 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13479 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13480 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13482 (tune_params): Update comment.
13483 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13484 (arm_max_conditional_execute): New function.
13485 (thumb2_final_prescan_insn): Use max_insn_skipped and
13486 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13488 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13490 PR tree-optimization/57705
13491 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13492 SSA_NAME step, provided that it is not defined inside the loop.
13493 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13494 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13496 2013-06-25 Martin Jambor <mjambor@suse.cz>
13498 PR middle-end/57670
13499 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13500 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13502 (ipa_note_param_call): Initialize member_ptr flag.
13503 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13504 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13505 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13506 (ipa_read_indirect_edge_info): Likewise.
13508 2013-06-25 Richard Biener <rguenther@suse.de>
13510 PR middle-end/56977
13511 * passes.c (init_optimization_passes): Move pass_fold_builtins
13512 and pass_dce earlier with -Og.
13514 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13516 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13517 <BIT_FIELD_REF>: Remove trailing TAB.
13518 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13521 2013-06-24 Martin Jambor <mjambor@suse.cz>
13523 PR tree-optimization/57358
13524 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13525 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13526 (ipa_analyze_params_uses): Generate pessimistic info when true.
13528 2013-06-24 Martin Jambor <mjambor@suse.cz>
13530 PR tree-optimization/57539
13531 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13532 global.inlined_to of the new node to it. All callers changed.
13533 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13534 inlining_into, pass it to cgraph_clone_node.
13535 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13536 ipa_free_edge_args_substructures.
13537 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13538 rdesc linked list. Do not assert rdesc edges have inlined caller.
13539 Assert we have found an rdesc in the rdesc list.
13541 2013-06-24 Richard Biener <rguenther@suse.de>
13543 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13544 (pointer_set_lookup): Declare.
13545 (class pointer_map): New template class implementing a
13546 generic pointer to T map.
13547 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13548 pointer_map<T>::contains, pointer_map<T>::insert,
13549 pointer_map<T>::traverse): New functions.
13550 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13551 (pointer_set_lookup): New function.
13552 (pointer_set_contains): Use pointer_set_lookup.
13553 (pointer_set_insert): Likewise.
13554 (insert_aux): Remove.
13555 (struct pointer_map_t): Embed a pointer_set_t.
13556 (pointer_map_create): Adjust.
13557 (pointer_map_destroy): Likewise.
13558 (pointer_map_contains): Likewise.
13559 (pointer_map_insert): Likewise.
13560 (pointer_map_traverse): Likewise.
13561 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13562 pointer_map<unsigned> instead of a pointer_map_t.
13563 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13564 (streamer_tree_cache_lookup): Likewise.
13565 (streamer_tree_cache_create): Likewise.
13566 (streamer_tree_cache_delete): Likewise.
13567 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13568 pointer_map<unsigned> instead of a pointer_map_t.
13569 (lto_init_tree_ref_encoder): Adjust.
13570 (lto_destroy_tree_ref_encoder): Likewise.
13571 * lto-section-out.c (lto_output_decl_index): Likewise.
13572 (lto_record_function_out_decl_state): Likewise.
13573 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13575 2013-06-24 Richard Biener <rguenther@suse.de>
13577 PR tree-optimization/57488
13578 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13580 2013-06-24 Alan Modra <amodra@gmail.com>
13582 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13583 (gen_easy_altivec_constant): Likewise.
13584 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13585 easy_vector_constant_msb): Likewise.
13587 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13590 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13591 add missing return true.
13593 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13596 * config/sh/predicates.md (general_extend_operand): Invoke
13597 general_movsrc_operand for memory operands.
13598 (general_movsrc_operand): Allow reg+reg addressing, do not use
13599 general_operand for memory operands.
13601 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13603 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13604 when current target options does not apply.
13605 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13606 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13607 * config/i386/bmiintrin.h: Pass appropriate target
13608 attributes to header.
13609 * config/i386/mmintrin.h: Ditto.
13610 * config/i386/nmmintrin.h: Ditto.
13611 * config/i386/avx2intrin.h: Ditto.
13612 * config/i386/fxsrintrin.h: Ditto.
13613 * config/i386/tbmintrin.h: Ditto.
13614 * config/i386/xsaveintrin.h: Ditto.
13615 * config/i386/f16cintrin.h: Ditto.
13616 * config/i386/xtestintrin.h: Ditto.
13617 * config/i386/xsaveoptintrin.h: Ditto.
13618 * config/i386/bmi2intrin.h: Ditto.
13619 * config/i386/lzcntintrin.h: Ditto.
13620 * config/i386/smmintrin.h: Ditto.
13621 * config/i386/wmmintrin.h: Ditto.
13622 * config/i386/x86intrin.h: Remove all header include guards.
13623 * config/i386/prfchwintrin.h: Ditto.
13624 * config/i386/pmmintrin.h: Ditto.
13625 * config/i386/tmmintrin.h: Ditto.
13626 * config/i386/xmmintrin.h: Ditto.
13627 * config/i386/popcntintrin.h: Ditto.
13628 * config/i386/rdseedintrin.h: Ditto.
13629 * config/i386/ammintrin.h: Ditto.
13630 * config/i386/emmintrin.h: Ditto.
13631 * config/i386/immintrin.h: Remove all header include guards.
13632 * config/i386/fma4intrin.h: Ditto.
13633 * config/i386/lwpintrin.h: Ditto.
13634 * config/i386/xopintrin.h: Ditto.
13635 * config/i386/ia32intrin.h: Ditto.
13636 * config/i386/avxintrin.h: Ditto.
13637 * config/i386/rtmintrin.h: Ditto.
13638 * config/i386/fmaintrin.h: Ditto.
13639 * config/i386/mm3dnow.h: Ditto.
13641 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13643 * common/config/i386/i386-common.c: Handle LZCNT.
13645 2013-06-22 Andi Kleen <ak@linux.intel.com>
13647 * doc/extend.texi: Use __atomic_store_n instead of
13648 __atomic_store in HLE example.
13650 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13652 * config/sh/sh.c: Remove <cstdlib> workaround.
13654 2013-06-21 Andi Kleen <ak@linux.intel.com>
13656 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13658 2013-06-21 Andi Kleen <ak@linux.intel.com>
13660 * doc/extend.texi: Document that __atomic_clear and
13661 __atomic_test_and_set should only be used with bool.
13663 2013-06-20 Jan Hubicka <jh@suse.cz>
13665 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13666 types_same_for_odr.
13667 * tree.c (decls_same_for_odr): New function.
13668 (same_for_edr): New function.
13669 (types_same_for_odr): New function.
13670 (get_binfo_at_offset): Use it.
13671 * tree.h (types_same_for_odr): Declare.
13673 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13674 Jason Merrill <jason@redhat.com>
13676 * system.h: Include <cstdlib> as well as <stdlib.h>.
13678 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13681 * config/i386/i386.c (construct_container): Report error if
13682 long double is used with disabled x87 float returns.
13684 2013-06-20 Jan Hubicka <jh@suse.cz>
13686 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13688 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13690 PR rtl-optimization/57425
13691 PR rtl-optimization/57569
13692 * alias.c (write_dependence_p): Remove parameters mem_mode and
13693 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13694 Changed all callers.
13695 (canon_anti_dependence): Get comments and semantics in sync.
13696 Add parameter mem_canonicalized. Changed all callers.
13697 * rtl.h (canon_anti_dependence): Update prototype.
13699 2013-06-20 Richard Biener <rguenther@suse.de>
13701 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13702 byte case, inline streamer_read_uchar and defer section
13705 2013-06-20 Richard Biener <rguenther@suse.de>
13707 PR tree-optimization/57584
13708 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13709 SSA names into the expanded expression that take part in
13710 abnormal coalescing.
13712 2013-06-19 Sharad Singhai <singhai@google.com>
13714 * gcov.c (print_usage): Handle new option.
13715 (process_args): Ditto.
13716 (get_gcov_intermediate_filename): New function.
13717 (output_intermediate_file): New function.
13718 (output_gcov_file): New function
13719 (generate_results): Handle new option.
13720 (release_function): Relase demangled name.
13721 (read_graph_file): Handle demangled name.
13722 (output_lines): Ditto.
13723 * doc/gcov.texi: Document gcov intermediate format.
13725 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13728 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13729 (lra_emit_add): Use the functions. Add comment about Y as an
13732 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13735 * collect2.c (collect_atexit): New.
13736 (collect_exit): Delete.
13737 (main): Register collect_atexit with atexit.
13738 (collect_wait): Change collect_exit to exit.
13740 * collect2.h (collect_exit): Delete.
13741 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13743 2013-06-19 Wei Mi <wmi@google.com>
13745 PR rtl-optimization/57518
13746 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13747 if regno is used in paradoxical subreg.
13748 (update_equiv_regs): Check pdx_subregs[regno] before
13749 set a reg to be equivalent with a mem.
13751 2013-06-19 Matthias Klose <doko@ubuntu.com>
13754 * file-find.h (find_a_file): Add a mode parameter.
13755 * file-find.c (find_a_file): Likewise.
13756 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13757 with X_OK for the executables.
13758 * collect2.c (main): Call find_a_file with X_OK.
13760 2013-06-19 Steve Ellcey <sellcey@mips.com>
13763 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13764 Use NEXT_INSN instead of next_real_insn.
13766 2013-06-19 Jan Hubicka <jh@suse.cz>
13768 * cgraph.h (const_value_known_p): Replace by ...
13769 (ctor_for_folding): .. this one.
13770 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13771 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13772 * expr.c (expand_expr_real_1): Likewise.
13773 (string_constant): Likewise.
13774 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13775 * ipa.c (process_references): Likewise.
13776 (symtab_remove_unreachable_nodes): Likewise.
13777 * ipa-inline-analysis.c (param_change_prob): Likewise.
13778 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13779 (get_base_constructor): Likwise.
13780 * varpool.c (varpool_remove_node): Likewise.
13781 (varpool_remove_initializer): LIkewise.
13782 (dump_varpool_node): LIkwise.
13783 (const_value_known_p): Rewrite to ...
13784 (ctor_for_folding): ... this one.
13786 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13789 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13790 PERSONALITY in $PATH derived prefixes.
13792 2013-06-19 Jeff Law <law@redhat.com>
13794 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13797 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13798 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13801 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13803 * config/arm/vfp.md: Move VFP instruction classification documentation
13805 * config/arm/arm.md: ... here. Update instruction classification
13808 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13810 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13811 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13812 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13814 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13817 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13818 with NEXT_INSN. Use tablejump_p to check for jump table data
13821 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13824 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13825 that now in C++ the value is correct per the C++ standards.
13827 2013-06-19 Richard Biener <rguenther@suse.de>
13829 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13830 for global context.
13832 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13835 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13838 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13839 with next_active_insn.
13841 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13843 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13844 functions are inlined during failures to flag an error.
13845 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13846 in early inline pass.
13848 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13850 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13853 2013-06-18 Julian Brown <julian@codesourcery.com>
13855 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13856 Permit virtual register pre-reload if !strict.
13857 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13859 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13861 * config/arm/neon.md (movmisalign<mode>): Use
13862 neon_perm_struct_or_reg_operand instead of
13863 neon_struct_or_register_operand.
13864 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13865 neon_permissive_struct_operand instead of neon_struct_operand.
13866 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13867 neon_vector_mem_operand.
13868 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13869 neon_vector_mem_operand.
13870 (neon_permissive_struct_operand): New.
13871 (neon_struct_or_register_operand): Rename to...
13872 (neon_perm_struct_or_reg_operand): This. Adjust call to
13873 neon_vector_mem_operand.
13875 2013-06-18 Richard Biener <rguenther@suse.de>
13877 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13878 * lto-streamer.h: Include pointer-set.h.
13879 (struct lto_decl_slot): Remove.
13880 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13881 Remove next_index entry.
13882 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13883 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13884 (lto_init_tree_ref_encoder): Adjust.
13885 (lto_destroy_tree_ref_encoder): Likewise.
13886 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13887 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13888 (lto_output_decl_index): Adjust.
13889 (lto_new_out_decl_state): Likewise.
13890 (lto_record_function_out_decl_state): Likewise.
13891 * lto-streamer-out.c (copy_function): Likewise.
13893 2013-06-18 Richard Biener <rguenther@suse.de>
13895 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13896 * cgraphunit.c: Include cfgloop.h.
13897 (init_lowered_empty_function): Initialize the loop tree.
13898 (assemble_thunk): Insert new BBs into loops.
13900 2013-06-18 Richard Biener <rguenther@suse.de>
13902 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13903 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13904 the map from cache entry to cache index optional.
13905 (streamer_tree_cache_replace_tree): Adjust accordingly.
13906 (streamer_tree_cache_append): Likewise.
13907 (streamer_tree_cache_delete): Likewise.
13908 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13909 streamer cache map from cache entry to cache index.
13910 * lto-streamer-out.c (create_output_block): Adjust.
13912 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13914 * config/arm/arm.md (attribute "insn"): Move multiplication and
13915 division attributes to...
13916 (attribute "type"): ... here. Remove mult.
13917 (attribute "mul32"): New attribute.
13918 (attribute "mul64"): Add umaal.
13919 (*arm_mulsi3): Update attributes.
13920 (*arm_mulsi3_v6): Likewise.
13921 (*thumb_mulsi3): Likewise.
13922 (*thumb_mulsi3_v6): Likewise.
13923 (*mulsi3_compare0): Likewise.
13924 (*mulsi3_compare0_v6): Likewise.
13925 (*mulsi_compare0_scratch): Likewise.
13926 (*mulsi_compare0_scratch_v6): Likewise.
13927 (*mulsi3addsi): Likewise.
13928 (*mulsi3addsi_v6): Likewise.
13929 (*mulsi3addsi_compare0): Likewise.
13930 (*mulsi3addsi_compare0_v6): Likewise.
13931 (*mulsi3addsi_compare0_scratch): Likewise.
13932 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13933 (*mulsi3subsi): Likewise.
13934 (*mulsidi3adddi): Likewise.
13935 (*mulsi3addsi_v6): Likewise.
13936 (*mulsidi3adddi_v6): Likewise.
13937 (*mulsidi3_nov6): Likewise.
13938 (*mulsidi3_v6): Likewise.
13939 (*umulsidi3_nov6): Likewise.
13940 (*umulsidi3_v6): Likewise.
13941 (*umulsidi3adddi): Likewise.
13942 (*umulsidi3adddi_v6): Likewise.
13943 (*smulsi3_highpart_nov6): Likewise.
13944 (*smulsi3_highpart_v6): Likewise.
13945 (*umulsi3_highpart_nov6): Likewise.
13946 (*umulsi3_highpart_v6): Likewise.
13947 (mulhisi3): Likewise.
13948 (*mulhisi3tb): Likewise.
13949 (*mulhisi3bt): Likewise.
13950 (*mulhisi3tt): Likewise.
13951 (maddhisi4): Likewise.
13952 (*maddhisi4tb): Likewise.
13953 (*maddhisi4tt): Likewise.
13954 (maddhidi4): Likewise.
13955 (*maddhidi4tb): Likewise.
13956 (*maddhidi4tt): Likewise.
13957 (divsi3): Likewise.
13958 (udivsi3): Likewise.
13959 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13960 (thumb2_mulsi_short_compare0): Likewise.
13961 (thumb2_mulsi_short_compare0_scratch): Likewise.
13962 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13963 (1020mult2): Likewise.
13964 (1020mult3): Likewise.
13965 (1020mult4): Likewise.
13966 (1020mult5): Likewise.
13967 (1020mult6): Likewise.
13968 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13970 (cortex_a15_mult64): Likewise.
13971 (cortex_a15_sdiv): Likewise.
13972 (cortex_a15_udiv): Likewise.
13973 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13979 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13980 (pj4_ir_div): Likewise.
13981 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13982 (11_mult2): Likewise.
13983 (11_mult3): Likewise.
13984 (11_mult4): Likewise.
13985 (11_mult5): Likewise.
13986 (11_mult6): Likewise.
13987 (11_mult7): Likewise.
13988 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13989 (cortex_a8_mla): Likewise.
13990 (cortex_a8_mull): Likewise.
13991 (cortex_a8_smulwy): Likewise.
13992 (cortex_a8_smlald): Likewise.
13993 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13994 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13995 (cortex_r4_mul_3): Likewise.
13996 (cortex_r4_mla_4): Likewise.
13997 (cortex_r4_mla_3): Likewise.
13998 (cortex_r4_smlald): Likewise.
13999 (cortex_r4_mull): Likewise.
14000 (cortex_r4_sdiv): Likewise.
14001 (cortex_r4_udiv): Likewise.
14002 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14003 (cortex_a7_idiv): Likewise.
14004 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14005 (9_mult2): Likewise.
14006 (9_mult3): Likewise.
14007 (9_mult4): Likewise.
14008 (9_mult5): Likewise.
14009 (9_mult6): Likewise.
14010 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14011 (cortex_a53_sdiv): Likewise.
14012 (cortex_a53_udiv): Likewise.
14013 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14014 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14015 (mp626_mult2): Likewise.
14016 (mp626_mult3): Likewise.
14017 (mp626_mult4): Likewise.
14018 * config/arm/fa526.md (526_mult1): Update attribute change.
14019 (526_mult2): Likewise.
14020 * config/arm/arm-generic.md (mult): Update attribute change.
14021 (mult_ldsched_strongarm): Likewise.
14022 (mult_ldsched): Likewise.
14023 (multi_cycle): Likewise.
14024 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14025 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14026 (606te_mult2): Likewise.
14027 (606te_mult3): Likewise.
14028 (606te_mult4): Likewise.
14029 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14030 (cortex_a9_mac16): Likewise.
14031 (cortex_a9_multiply): Likewise.
14032 (cortex_a9_mac): Likewise.
14033 (cortex_a9_multiply_long): Likewise.
14034 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14035 (626te_mult2): Likewise.
14036 (626te_mult3): Likewise.
14037 (626te_mult4): Likewise.
14039 2013-06-18 Richard Biener <rguenther@suse.de>
14042 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14044 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14047 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14048 with next_active_insn.
14050 2013-06-18 Alan Modra <amodra@gmail.com>
14052 * config/rs6000/rs6000.h (enum data_align): New.
14053 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14054 (DATA_ABI_ALIGNMENT): Define.
14055 (CONSTANT_ALIGNMENT): Correct comment.
14056 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14057 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14059 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14061 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14062 ATTRIBUTE_UNUSED marking.
14064 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14066 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14067 alternative and update.
14068 (aarch64_dup_lanedi): Delete.
14069 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14070 * config/aarch64/aarch64-simd-builtins.def: Update.
14072 2013-06-17 Richard Biener <rguenther@suse.de>
14074 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14075 (lto_input_scc): Declare.
14076 (lto_input_tree_1): Likewise.
14077 (struct lto_stats_d): Add num_tree_bodies_output and
14078 num_pickle_refs_output.
14079 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14080 (lto_read_tree_1): Split out from ...
14081 (lto_read_tree): ... this.
14082 (lto_input_scc): New function.
14083 (lto_input_tree_1): Split out from ...
14084 (lto_input_tree): ... this. Handle LTO_tree_scc.
14085 (lto_data_in_create): Create the streamer cache without hashes.
14086 * lto-streamer-out.c (create_output_block): Create the streamer
14087 cache with hashes when not doing WPA.
14088 (lto_write_tree_1): Split out from ...
14089 (lto_write_tree): ... this.
14090 (get_symbol_initial_value): New function.
14091 (lto_output_tree_1): Split out from ...
14092 (lto_output_tree): ... this. Write trees as series of SCCs
14093 using a DFS walk via DFS_write_tree.
14094 (struct sccs, struct scc_entry): New types.
14095 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14096 (DFS_write_tree_body): New function.
14097 (DFS_write_tree): Likewise.
14098 (hash_tree): Likewise.
14099 (scc_entry_compare): Likewise.
14100 (hash_scc): Likewise.
14101 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14102 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14103 TREE_CHAIN as regular reference.
14104 (streamer_read_integer_cst): Remove.
14105 (streamer_get_pickled_tree): Adjust.
14106 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14107 of DECL_EXTERNALs in BLOCK_VARS for now.
14108 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14110 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14111 Add hash value argument and record that if hashes are recorded
14113 (streamer_tree_cache_insert_1): Adjust.
14114 (streamer_tree_cache_insert): Likewise.
14115 (streamer_tree_cache_insert_at): Rename to ...
14116 (streamer_tree_cache_replace_tree): ... this and adjust.
14117 (streamer_tree_cache_append): Adjust.
14118 (record_common_node): Likewise.
14119 (streamer_tree_cache_create): Add argument whether to
14120 record hash values together with trees.
14121 (streamer_tree_cache_delete): Adjust.
14122 * tree-streamer.h (struct streamer_tree_cache_d): Add
14124 (streamer_read_integer_cst): Remove.
14125 (streamer_tree_cache_insert): Adjust.
14126 (streamer_tree_cache_append): Likewise.
14127 (streamer_tree_cache_insert_at): Rename to ...
14128 (streamer_tree_cache_replace_tree): ... this and adjust.
14129 (streamer_tree_cache_create): Add argument whether to record hashes.
14130 (streamer_tree_cache_get): Rename to ...
14131 (streamer_tree_cache_get_tree): ... this.
14132 (streamer_tree_cache_get_hash): New function.
14133 * tree.c (cache_integer_cst): New function.
14134 * tree.h (cache_integer_cst): Declare.
14135 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14136 * lto-symtab.c (lto_varpool_replace_node): Only release
14137 DECL_INITIAL of non-prevailing decls.
14138 * varpool.c (varpool_remove_initializer): Do not release
14139 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14141 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14143 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14144 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14145 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14146 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14147 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14148 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14149 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14150 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14151 instead of TARGET_64BIT.
14152 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14153 Require ISA_HAS_<D>DIV.
14155 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14157 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14158 (mips*-*-linux*): Move default with_llsc setting to where other
14160 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14162 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14163 Likewise. Remove default with_tune setting. Move default float
14164 setting to its own block. Handle with_llsc in the same block as above.
14166 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14168 PR rtl-optimization/57425
14169 PR rtl-optimization/57569
14170 * alias.c (write_dependence_p): Add new parameters mem_mode,
14171 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14172 Changed all callers.
14173 (canon_anti_dependence): New function.
14174 * cse.c (check_dependence): Use canon_anti_dependence.
14175 * cselib.c (cselib_invalidate_mem): Likewise.
14176 * rtl.h (canon_anti_dependence): Declare.
14178 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14180 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14181 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14182 ".set mips3" for 64-bit targets.
14184 2013-06-15 Dehao Chen <dehao@google.com>
14186 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14187 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14188 (gimple_check_call_args): Likewise.
14189 * value-prof.c (check_ic_target): Likewise.
14190 * ipa-inline.c (early_inliner): Likewise.
14191 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14192 * cgraph.c (cgraph_create_edge_1): Likewise.
14193 (cgraph_make_edge_direct): Likewise.
14195 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14198 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14199 rs6000_output_move_128bit to handle emitting quad memory
14200 operations. Set attribute length to 8 bytes.
14202 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14204 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14206 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14207 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14208 (aarch64_<su>mlsl<mode>): Likewise.
14210 2013-06-14 Mike Stump <mikestump@comcast.net>
14212 * Makefile.in (TARGET_H): Add insn-codes.h.
14214 2013-06-14 Alan Modra <amodra@gmail.com>
14216 PR middle-end/57134
14217 PR middle-end/57586
14218 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14219 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14220 bitfield expansion when EXPAND_MEMORY.
14221 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14223 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14225 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14226 test for clearing quad memory on 32-bit later.
14228 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14230 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14231 (fold_negate_expr): Likewise.
14232 (fold_real_zero_addition_p): Handle vectors.
14233 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14235 2013-06-14 Alan Modra <amodra@gmail.com>
14237 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14239 2013-06-13 Jan Hubicka <jh@suse.cz>
14241 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14242 Local comdats are not externally visible.
14243 * symtab.c (dump_symtab_base): Dump externally visible.
14244 (verify_symtab_base): Verify back links in the symtab hash.
14246 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14248 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14249 CONVERT_EXPR as equal nodes.
14251 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14253 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14255 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14257 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14258 Generalize to complex and vector.
14259 * tree.c (build_all_ones_cst): New function.
14260 * tree.h (build_all_ones_cst): Declare it.
14262 2013-06-13 Alan Modra <amodra@gmail.com>
14264 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14265 * config/rs6000/rs6000.md (signbittf2): New insn.
14266 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14267 (abstf2_internal, cmptf_internal2): Likewise.
14268 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14270 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14271 Pat Haugen <pthaugen@us.ibm.com>
14272 Peter Bergner <bergner@vnet.ibm.com>
14274 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14275 power8 byte, half-word, and quad-word atomic instructions.
14276 (emit_store_conditional): Likewise.
14277 (rs6000_expand_atomic_compare_and_swap): Likewise.
14278 (rs6000_expand_atomic_op): Likewise.
14280 * config/rs6000/sync.md (larx): Add new modes for power8.
14282 (AINT): New mode iterator to include TImode as well as normal
14283 integer modes on power8.
14284 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14285 that VSX registers are not considered. Use AINT mode iterator
14286 instead of INT1 to allow inclusion of quad word atomic operations
14288 (load_locked<mode>): Likewise.
14289 (store_conditional<mode>): Likewise.
14290 (atomic_compare_and_swap<mode>): Likewise.
14291 (atomic_exchange<mode>): Likewise.
14292 (atomic_nand<mode>): Likewise.
14293 (atomic_fetch_<fetchop_name><mode>): Likewise.
14294 (atomic_nand_fetch<mode>): Likewise.
14295 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14297 (ATOMIC): On power8, add QImode, HImode modes.
14298 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14299 modes that promote to SImode.
14300 (load_lockedti): Convert TImode arguments to PTImode, so that we
14301 get a guaranteed even/odd register pair.
14302 (load_lockedpti): Likewise.
14303 (store_conditionalti): Likewise.
14304 (store_conditionalpti): Likewise.
14306 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14307 atomic load/store instructions.
14310 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14312 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14314 (insn_count): New attribute, with most cases extracted from...
14315 (length): ...here. Redefine most cases in terms of insn_count.
14316 (single_insn): Delete.
14317 (can_delay): Use insn_count to check for single instructions.
14318 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14319 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14320 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14321 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14322 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14323 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14324 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14325 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14326 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14327 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14328 rather than "length".
14329 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14330 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14331 Use "insn_count" rather than "length".
14332 * config/mips/mips-dsp.md
14333 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14334 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14337 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14339 PR tree-optimization/57361
14340 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14342 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14344 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14346 (aarch64_simd_combine<mode>): New instruction expansion.
14347 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14348 function prototype.
14349 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14350 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14352 2013-06-12 Jan Hubicka <jh@suse.cz>
14354 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14355 decl has when in streaming stage.
14356 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14357 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14359 2013-06-12 Roland Stigge <stigge@antcom.de>
14362 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14364 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14366 PR tree-optimization/57537
14367 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14368 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14370 2013-06-12 Richard Biener <rguenther@suse.de>
14372 * data-streamer.h (streamer_write_char_stream): CSE
14373 obs->current_pointer.
14374 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14375 streamer_write_char_stream manually and optimize the resulting loop.
14376 (streamer_write_hwi_stream): Likewise.
14378 2013-06-12 Jan Hubicka <jh@suse.cz>
14380 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14381 * cgraph.h (varpool_create_empty_node): Declare.
14382 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14384 * symtab.c (symtab_unregister_node): Be lax about missin entries
14386 (symtab_get_node): Update comment.
14387 * varpool.c (varpool_create_empty_node): Break out from ...
14388 (varpool_node_for_decl): ... here.
14389 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14391 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14393 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14394 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14395 part. Use straight-line flow at the end.
14396 <COMPONENT_REF>: Remove superfluous else.
14397 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14399 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14402 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14403 target hook even for !TREE_PUBLIC decls. If no resolution info
14404 is available, return false for common and external decls.
14406 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14408 * config/rl78/constraints.md (U): New constraint.
14409 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14412 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14415 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14416 to allow returning address to AT_PLATFORM name.
14418 2013-06-11 Jan Hubicka <jh@suse.cz>
14420 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14421 * cgraph.h (symtab_node_base): Add weakref flag.
14422 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14423 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14424 (output_weakrefs): Use weakref flag.
14425 * fold-const.c (simple_operand_p): Handle WEAK.
14426 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14427 * ipa.c (varpool_externally_visible_p): Drop weakref.
14428 (function_and_variable_visibility): Update comment; fix weakref
14429 sanity checks; do not clear DECL_WEAK on them.
14430 * lto-cgraph.c (lto_output_node): update.
14431 (lto_output_varpool_node): Update.
14432 (input_overwrite_node): Update.
14433 (input_node): Update.
14434 (input_varpool_node): Update.
14435 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14436 (lto_symtab_merge_symbols): Add sanity check.
14437 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14438 * passes.c (rest_of_decl_compilation): Set static flag, too.
14439 * symtab.c (dump_symtab_base): Dump weakref.
14440 (verify_symtab_base): Sanity check weakrefs.
14441 (symtab_make_decl_local): Remove duplicated code.
14442 (symtab_alias_ultimate_target): Simplify.
14443 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14445 2013-06-11 Tom de Vries <tom@codesourcery.com>
14447 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14448 sequence_vect == NULL.
14450 2013-06-11 DJ Delorie <dj@redhat.com>
14452 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14453 (rl78_unwind_word_mode): New.
14455 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14457 * final.c (debug_prefix_maps): Make static.
14459 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14461 * function.c (initial_trampoline): Remove stray copy.
14463 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14465 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14467 2013-06-11 Martin Jambor <mjambor@suse.cz>
14469 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14470 within bounds at the beginning of the function.
14472 2013-06-11 Alan Modra <amodra@gmail.com>
14474 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14476 (get_named_section): Don't NULL !DECL_P decl.
14478 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14480 * doc/invoke.texi (core-avx2): Document.
14482 (atom): Updated with MOVBE.
14484 2013-06-11 Richard Biener <rguenther@suse.de>
14486 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14488 2013-06-11 Anton Blanchard <anton@samba.org>
14490 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14491 correct shift value in little-endian mode.
14493 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14496 * varasm.c (get_variable_align): Move #endif to the right place.
14498 2013-06-10 Cary Coutant <ccoutant@google.com>
14500 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14501 for hash so that hash table traversal order is deterministic.
14503 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14504 Pat Haugen <pthaugen@us.ibm.com>
14505 Peter Bergner <bergner@vnet.ibm.com>
14507 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14508 of vectors in GPRS if they are direct moves or quad word load or
14511 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14513 (direct_move_p): Likewise.
14514 (quad_load_store_p): Likewise.
14516 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14517 classes into bins based on the physical register type.
14518 (reg_class_to_reg_type): Likewise.
14519 (IS_STD_REG_TYPE): Likewise.
14520 (IS_FP_VECT_REG_TYPE): Likewise.
14521 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14522 use direct move instructions.
14523 (reload_gpr_vsx): Likewise.
14524 (reload_vsx_gpr): Likewise.
14525 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14526 information that is a simplification of register classes. Also
14527 precalculate direct move reload helpers.
14528 (direct_move_p): New function to return true if the operation can
14529 be done as a direct move instruciton.
14530 (quad_load_store_p): New function to return true if the operation
14531 is a quad memory operation.
14532 (rs6000_legitimize_address): If quad memory, only allow register
14533 indirect for TImode addresses.
14534 (rs6000_legitimate_address_p): Likewise.
14535 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14536 (rs6000_reload_register_type): Likewise.
14537 (register_to_reg_type): Return register type.
14538 (rs6000_secondary_reload_simple_move): New helper function for
14539 secondary reload and secondary memory needed to identify anything
14540 that is a simple move, and does not need reloading.
14541 (rs6000_secondary_reload_direct_move): New helper function for
14542 secondary reload to identify cases that can be done with several
14543 instructions via the direct move instructions.
14544 (rs6000_secondary_reload_move): New helper function for secondary
14545 reload to identify moves between register types that can be done.
14546 (rs6000_secondary_reload): Add support for quad memory operations
14547 and for direct move.
14548 (rs6000_secondary_memory_needed): Likewise.
14549 (rs6000_debug_secondary_memory_needed): Change argument names.
14550 (rs6000_output_move_128bit): New function to return the move to
14551 use for 128-bit moves, including knowing about the various
14552 limitations of quad memory operations.
14554 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14555 memory operations. call rs6000_output_move_128bit for the actual
14556 instruciton(s) to generate.
14557 (vsx_movti_64bit): Likewise.
14559 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14560 (UNSPEC_P8V_MTVSRWZ): Likewise.
14561 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14562 (UNSPEC_P8V_MTVSRD): Likewise.
14563 (UNSPEC_P8V_XXPERMDI): Likewise.
14564 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14565 (UNSPEC_FUSION_GPR): Likewise.
14566 (FMOVE128_GPR): New iterator for direct move.
14567 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14568 (f32_sv): Likewise.
14569 (f32_dm): Likewise.
14570 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14571 loads and direct move instructions.
14572 (zero_extendsidi2_lfiwzx): Likewise.
14573 (extendsidi2_lfiwax): Likewise.
14574 (extendsidi2_nocell): Likewise.
14575 (floatsi<mode>2_lfiwax): Likewise.
14576 (lfiwax): Likewise.
14577 (floatunssi<mode>2_lfiwzx): Likewise.
14578 (lfiwzx): Likewise.
14579 (fix_trunc<mode>_stfiwx): Likewise.
14580 (fixuns_trunc<mode>_stfiwx): Likewise.
14581 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14582 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14583 (parity<mode>2_cmpb): Set length/type attr.
14584 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14585 for 'mr.' to fast_compare.
14586 (bpermd_<mode>): Change type attr to popcnt.
14587 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14588 (p8_mtvsrwz_1): Likewise.
14589 (p8_mtvsrwz_2): Likewise.
14590 (reload_fpr_from_gpr<mode>): Likewise.
14591 (p8_mtvsrd_1): Likewise.
14592 (p8_mtvsrd_2): Likewise.
14593 (p8_xxpermdi_<mode>): Likewise.
14594 (reload_vsx_from_gpr<mode>): Likewise.
14595 (reload_vsx_from_gprsf): Likewise.
14596 (p8_mfvsrd_3_<mode>): LIkewise.
14597 (reload_gpr_from_vsx<mode>): Likewise.
14598 (reload_gpr_from_vsxsf): Likewise.
14599 (p8_mfvsrd_4_disf): Likewise.
14600 (multi-word GPR splits): Do not split direct moves or quad memory
14603 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14605 * tree-into-ssa.c (interesting_blocks): Make static.
14607 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14610 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14611 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14612 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14613 (get_variable_align): New function.
14614 (get_variable_section, emit_bss, emit_common,
14615 assemble_variable_contents, place_block_symbol): Use
14616 get_variable_align instead of DECL_ALIGN.
14617 (assemble_noswitch_variable): Add align argument, use it
14618 instead of DECL_ALIGN.
14619 (assemble_variable): Adjust caller. Use get_variable_align
14620 instead of DECL_ALIGN.
14621 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14623 (DATA_ABI_ALIGNMENT): Define.
14624 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14625 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14626 opt is false, only return the psABI mandated alignment increase.
14627 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14628 (DATA_ABI_ALIGNMENT): ... this.
14629 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14630 (DATA_ABI_ALIGNMENT): ... this.
14631 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14632 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14633 (DATA_ABI_ALIGNMENT): ... this.
14634 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14635 * doc/tm.texi: Regenerated.
14637 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14639 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14640 cmp_code to construct REG_EQUAL note.
14642 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14645 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14646 that operands[2] doesn't overlap with operands[0].
14648 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14649 Jan Hubicka <jh@suse.cz>
14651 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14652 hack to mark symbols as used.
14654 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14656 PR rtl-optimization/57559
14657 * lra-constraints.c (process_alt_operands): Don't discourage
14658 memory with known offset for offsetable memory constraint.
14659 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14661 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14663 * varasm.c (struct oc_local_state): Reorder fields.
14664 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14665 and adjust accordingly.
14666 (output_constructor): Reorder initialization code and adjust call to
14667 output_constructor_bitfield.
14669 2013-06-07 Jan Hubicka <jh@suse.cz>
14671 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14673 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14675 * tree-object-size.c (unknown): Make const.
14677 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14679 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14680 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14681 for last alternative in the cpu_facility attribute.
14683 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14686 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14687 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14688 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14689 * config/arm/constraints.md (Dg): New constraint.
14690 * config/arm/neon.md (xordi3_neon): Remove.
14691 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14692 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14694 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14696 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14697 Clean up alternatives.
14699 2013-06-07 Alan Modra <amodra@gmail.com>
14701 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14704 2013-06-07 Alan Modra <amodra@gmail.com>
14706 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14708 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14710 * config/arm/constraints.md (Df): New constraint.
14711 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14712 Correct length attribute for last two alternatives.
14714 2013-06-07 Alan Modra <amodra@gmail.com>
14716 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14717 override user -mfp-in-toc.
14718 (offsettable_ok_by_alignment): Consider just the current access
14719 rather than the whole object, unless BLKmode. Handle
14720 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14721 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14722 for -mcmodel=medium.
14723 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14724 override user -mfp-in-toc or -msum-in-toc. Default to
14725 -mno-fp-in-toc for -mcmodel=medium.
14727 2013-06-06 DJ Delorie <dj@redhat.com>
14729 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14730 TARGET_VALID_POINTER_MODE.
14732 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14733 Pat Haugen <pthaugen@us.ibm.com>
14734 Peter Bergner <bergner@vnet.ibm.com>
14736 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14737 Document new power8 builtins.
14739 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14740 condition code register, to allow 128-bit logical operations to be
14741 done in the VSX or GPR registers.
14742 (nor<mode>3): Use the canonical form for nor.
14743 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14744 vclz*, and vpopcnt* vector instructions.
14745 (nand<mode>3): Likewise.
14746 (orc<mode>3): Likewise.
14747 (clz<mode>2): LIkewise.
14748 (popcount<mode>2): Likewise.
14750 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14751 that only the GPRs are recognized.
14753 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14754 support for new power8 builtins.
14756 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14758 (xscvdpspn): Likewise.
14764 (vpopcnt): Likewise.
14765 (vpopcntb): Likewise.
14766 (vpopcnth): Likewise.
14767 (vpopcntw): Likewise.
14768 (vpopcntd): Likewise.
14770 (vmrgew): Likewise.
14771 (vmrgow): Likewise.
14773 (eqv_v16qi3): Likewise.
14774 (eqv_v8hi3): Likewise.
14775 (eqv_v4si3): Likewise.
14776 (eqv_v2di3): Likewise.
14777 (eqv_v4sf3): Likewise.
14778 (eqv_v2df3): Likewise.
14780 (nand_v16qi3): Likewise.
14781 (nand_v8hi3): Likewise.
14782 (nand_v4si3): Likewise.
14783 (nand_v2di3): Likewise.
14784 (nand_v4sf3): Likewise.
14785 (nand_v2df3): Likewise.
14787 (orc_v16qi3): Likewise.
14788 (orc_v8hi3): Likewise.
14789 (orc_v4si3): Likewise.
14790 (orc_v2di3): Likewise.
14791 (orc_v4sf3): Likewise.
14792 (orc_v2df3): Likewise.
14794 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14795 allow power8 quad mode in 64-bit.
14796 (rs6000_builtin_vectorized_function): Add support to vectorize
14797 ISA 2.07 count leading zeros, population count builtins.
14798 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14799 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14800 (builtin_function_type): Add vgbbd builtin function which takes an
14802 (altivec_expand_vec_perm_const): Add support for new power8 merge
14805 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14806 that does not include TImdoe for use with 32-bit.
14807 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14809 (UNSPEC_VSX_CVDPSPN): Likewise.
14810 (vsx_xscvdpspn): Likewise.
14811 (vsx_xscvspdpn): Likewise.
14812 (vsx_xscvdpspn_scalar): Likewise.
14813 (vsx_xscvspdpn_directmove): Likewise.
14814 (vsx_and<mode>3): Split logical operations into 32-bit and
14815 64-bit. Add support to do logical operations on TImode as well as
14816 VSX vector types. Allow logical operations to be done in either
14817 VSX registers or in general purpose registers in 64-bit mode. Add
14818 splitters if GPRs were used. For AND, add clobber of CCmode to
14819 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14821 (vsx_and<mode>3_32bit): Likewise.
14822 (vsx_and<mode>3_64bit): Likewise.
14823 (vsx_ior<mode>3): Likewise.
14824 (vsx_ior<mode>3_32bit): Likewise.
14825 (vsx_ior<mode>3_64bit): Likewise.
14826 (vsx_xor<mode>3): Likewise.
14827 (vsx_xor<mode>3_32bit): Likewise.
14828 (vsx_xor<mode>3_64bit): Likewise.
14829 (vsx_one_cmpl<mode>2): Likewise.
14830 (vsx_one_cmpl<mode>2_32bit): Likewise.
14831 (vsx_one_cmpl<mode>2_64bit): Likewise.
14832 (vsx_nor<mode>3): Likewise.
14833 (vsx_nor<mode>3_32bit): Likewise.
14834 (vsx_nor<mode>3_64bit): Likewise.
14835 (vsx_andc<mode>3): Likewise.
14836 (vsx_andc<mode>3_32bit): Likewise.
14837 (vsx_andc<mode>3_64bit): Likewise.
14838 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14839 and xxlorc instructions.
14840 (vsx_eqv<mode>3_64bit): Likewise.
14841 (vsx_nand<mode>3_32bit): Likewise.
14842 (vsx_nand<mode>3_64bit): Likewise.
14843 (vsx_orc<mode>3_32bit): Likewise.
14844 (vsx_orc<mode>3_64bit): Likewise.
14846 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14848 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14850 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14851 (p8_vmrgow): Likewise.
14852 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14853 GPRs to be split under VSX.
14854 (p8v_clz<mode>2): Add power8 count leading zero support.
14855 (p8v_popcount<mode>2): Add power8 population count support.
14856 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14859 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14862 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14864 (vec_nand): Likewise.
14865 (vec_vclz): Likewise.
14866 (vec_vclzb): Likewise.
14867 (vec_vclzd): Likewise.
14868 (vec_vclzh): Likewise.
14869 (vec_vclzw): Likewise.
14870 (vec_vgbbd): Likewise.
14871 (vec_vmrgew): Likewise.
14872 (vec_vmrgow): Likewise.
14873 (vec_vpopcnt): Likewise.
14874 (vec_vpopcntb): Likewise.
14875 (vec_vpopcntd): Likewise.
14876 (vec_vpopcnth): Likewise.
14877 (vec_vpopcntw): Likewise.
14879 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14881 PR rtl-optimization/57468
14882 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14885 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14887 PR rtl-optimization/57459
14888 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14889 type when setting live regs.
14891 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14893 * config/s390/s390.opt (mlra): New option.
14894 * config/s390/s390.c (s390_decompose_address): Check displacement
14895 for all registers for LRA.
14896 (s390_secondary_reload): Don't used secondary reloads for LRA.
14897 (s390_lra_p): New function.
14898 (TARGET_LRA_P): Define.
14899 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14900 of attribute cpu_facility to zarch for the last alternative.
14901 (*cmpmem_short): Ditto.
14903 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14905 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14906 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14907 (arm_expand_prologue): Likewise.
14909 2013-06-06 Teresa Johnson <tejohnson@google.com>
14912 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14913 as this is now done by redirect_edge_and_branch_force.
14914 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14915 barriers, and fix interaction with splitting.
14916 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14917 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14918 reflect changes made in the routine.
14919 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14920 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14921 since this is called in cfglayout mode, and replace partition fixup
14922 with assert as that is now done by force_nonfallthru_and_redirect.
14923 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14924 already be marked with region crossing note.
14925 (insert_section_boundary_note): Make non-static, gate on flag
14926 has_bb_partition, rewrite to also check for multiple partitions.
14927 (rest_of_handle_reorder_blocks): Remove call to
14928 insert_section_boundary_note, now done later during free_cfg.
14929 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14930 * bb-reorder.h (insert_section_boundary_note): Declare.
14931 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14932 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14933 invoke insert_section_boundary_note.
14934 (try_redirect_by_replacing_jump): Remove unnecessary
14935 check for region crossing note.
14936 (fixup_partition_crossing): New function.
14937 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14938 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14939 in non-cfglayout mode.
14940 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14941 remove old code that tried to do this. Emit barrier correctly
14942 when we are in cfglayout mode.
14943 (last_bb_in_partition): New function.
14944 (rtl_split_edge): Correctly fixup partition boundaries.
14945 (commit_one_edge_insertion): Remove old code that tried to
14946 fixup region crossing edge since this is now handled in
14947 split_block, and set up insertion point correctly since
14948 block may now end in a jump.
14949 (verify_hot_cold_block_grouping): Guard against checking when not in
14950 linearized RTL mode.
14951 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14953 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14954 rtl_verify_flow_info, so not called in cfglayout mode.
14955 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14956 (fixup_reorder_chain): Remove old code that attempted to fixup region
14957 crossing note as this is now handled in force_nonfallthru_and_redirect.
14958 (duplicate_insn_chain): Don't duplicate switch section notes.
14959 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14961 * basic-block.h (emit_barrier_after_bb): Declare.
14963 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14965 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14966 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14967 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14969 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14971 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14973 * config/arm/ldmstm.md: Regenerate.
14975 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14977 * config/arm/sync.md (atomic_loaddi_1):
14978 Disable predication for arm_restrict_it.
14979 (arm_load_exclusive<mode>): Likewise.
14980 (arm_load_exclusivesi): Likewise.
14981 (arm_load_exclusivedi): Likewise.
14982 (arm_load_acquire_exclusive<mode>): Likewise.
14983 (arm_load_acquire_exclusivesi): Likewise.
14984 (arm_load_acquire_exclusivedi): Likewise.
14985 (arm_store_exclusive<mode>): Likewise.
14986 (arm_store_exclusive<mode>): Likewise.
14987 (arm_store_release_exclusivedi): Likewise.
14988 (arm_store_release_exclusive<mode>): Likewise.
14990 2013-06-06 Richard Biener <rguenther@suse.de>
14992 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14994 (lto_tag_is_tree_code_p): Adjust.
14995 (lto_tag_is_gimple_code_p): Likewise.
14996 (lto_gimple_code_to_tag): Likewise.
14997 (lto_tag_to_gimple_code): Likewise.
14998 (lto_tree_code_to_tag): Likewise.
14999 (lto_tag_to_tree_code): Likewise.
15000 * data-streamer.h (streamer_write_hwi_in_range): Use
15001 uhwi streaming to stream the normalized range.
15002 (streamer_read_hwi_in_range): Likewise.
15004 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15006 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15007 (predicable_short_it): Likewise.
15008 (predicated): Likewise.
15009 (enabled): Handle above.
15010 (define_cond_exec): Set predicated attribute to yes.
15012 2013-06-05 Mike Stump <mikestump@comcast.net>
15014 * gdbinit.in (__FUNCTION__): Add.
15016 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15018 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15019 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15021 2013-06-05 Jan Hubicka <jh@suse.cz>
15023 * varasm.c (mark_decl_referenced): Revert the removal until targets
15026 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15028 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15029 instead of mark_decl_referenced.
15031 2013-06-05 Jan Hubicka <jh@suse.cz>
15033 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15034 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15035 and symtab_used_from_object_file_p.
15036 (cgraph_make_node_local_1): Clear forced_by_abi.
15037 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15038 * cgraph.h (symtab_node_base): Add forced_by_abi.
15039 (decide_is_variable_needed): Remove.
15040 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15041 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15042 (decide_is_symbol_needed): ... this one; handle symbols in general;
15043 always analyze virtuals; honnor forced_by_abi.
15044 (cgraph_finalize_function): Update.
15045 (varpool_finalize_decl): Update.
15046 (symbol_defined_and_needed): Remove.
15047 (analyze_functions): Update.
15048 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15049 output_refs, input_overwrite_node): Handle forced_by_abi.
15050 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15051 (address_taken_from_non_vtable_p): ... this one.
15052 (comdat_can_be_unshared_p_1): New function.
15053 (cgraph_comdat_can_be_unshared_p): Rename to ...
15054 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15055 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15056 (function_and_variable_visibility): Clear forced_by_abi as needed.
15057 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15058 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15059 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15060 * varpool.c (decide_is_variable_needed): Remove.
15062 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15064 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15065 (arm_option_override): Override arm_restrict_it where appropriate.
15066 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15067 * config/arm/arm.opt (mrestrict-it): New command-line option.
15068 * doc/invoke.texi: Document -mrestrict-it.
15070 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15072 * tsan.c (tsan_atomic_table): Make const.
15074 2013-06-05 Richard Biener <rguenther@suse.de>
15076 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15077 index associated with the tree we are supposed to replace.
15078 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15079 TREE_ASM_WRITTEN as zero for everything but SSA names.
15081 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15083 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15085 2013-06-04 Jan Hubicka <jh@suse.cz>
15087 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15088 (input_node, input_varpool_node): Handle correctly external same
15090 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15091 nodes at ltrans stage.
15093 2013-06-04 Jan Hubicka <jh@suse.cz>
15095 * ipa-inline.c (update_caller_keys): Fix availability test.
15096 (update_callee_keys): Likewise.
15097 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15098 to follow ELF standard.
15100 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15102 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15103 (mips64r5900el-*-elf*): New configurations.
15104 * config/mips/mips-cpus.def (r5900): New processor.
15105 * config/mips/mips-tables.opt: Regenerate.
15106 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15107 (mips_issue_rate): Handle PROCESSOR_R5900.
15108 (mips_reorg_process_insns): Force reorder mode for the R5900.
15109 * config/mips/mips.h (TARGET_MIPS5900): Define.
15110 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15112 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15114 * config/mips/mips.md (processor): Add r5900.
15115 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15117 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15119 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15120 into function to generate MOVI instruction.
15121 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15122 (aarch64_preferred_simd_mode): Turn into wrapper.
15123 (aarch64_output_scalar_simd_mov_immediate): New function.
15124 * config/aarch64/aarch64-protos.h: Add prototype for above.
15126 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15128 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15129 element_char member.
15130 (sizetochar): Return signed char.
15131 (aarch64_simd_valid_immediate): Remove elchar and other
15132 unnecessary variables.
15133 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15134 Calculate element_char as required.
15135 * config/aarch64/aarch64-protos.h: Update and move prototype
15136 for aarch64_output_simd_mov_immediate.
15137 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15140 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15142 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15143 information completed by aarch64_simd_valid_immediate.
15144 (aarch64_legitimate_constant_p): Update arguments.
15145 (aarch64_simd_valid_immediate): Work with struct rather than many
15147 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15148 (aarch64_simd_make_constant): Update arguments.
15149 (aarch64_output_simd_mov_immediate): Work with struct rather than
15150 many pointers. Output immediate directly rather than as operand.
15151 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15153 * config/aarch64/constraints.md (Dn): Update arguments.
15155 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15157 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15159 (aarch64_simd_immediate_valid_for_move): Remove.
15160 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15161 (aarch64_simd_make_constant): Update call.
15162 (aarch64_output_simd_mov_immediate): Update call.
15163 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15165 * config/aarch64/constraints.md (Dn): Update call.
15167 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15169 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15170 return type to bool for prototype.
15171 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15172 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15173 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15175 2013-06-04 Catherine Moore <clm@codesourcery.com>
15177 * config/mips/mips.opt (meva): New.
15178 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15179 (ASM_SPEC): Handle -meva.
15180 * doc/invoke.texi (meva): Document.
15182 2013-06-04 Alan Modra <amodra@gmail.com>
15184 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15187 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15189 * rtl.def: Add extra fourth optional field to define_cond_exec.
15190 * gensupport.c (process_one_cond_exec): Process attributes from
15192 * doc/md.texi: Document fourth field in define_cond_exec.
15194 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15196 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15197 out the processing order as in store_bit_field_1.
15199 2013-06-04 Jan Hubicka <jh@suse.cz>
15201 PR middle-end/57500
15202 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15203 non-VAR_DECL node if it does not exist yet.
15205 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15207 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15208 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15209 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15210 target_cpu_default setting.
15212 2013-06-03 Teresa Johnson <tejohnson@google.com>
15214 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15215 default to -fopt-info=optimized instead of all.
15216 * doc/invoke.texi: Ditto.
15217 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15218 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15219 (execute_vect_slp): Emit BB vectorization success under
15220 MSG_OPTIMIZED_LOCATIONS.
15221 * tree-vect-slp.c (vect_slp_transform_bb): Change
15222 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15223 * tree-vect-loop.c (vect_transform_loop): Ditto.
15225 2013-06-03 Jason Merrill <jason@redhat.com>
15228 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15229 Use TARGET_EXPR for C++.
15231 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15233 PR rtl-optimization/57268
15234 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15235 if DEBUG_INSN_P (insn).
15238 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15240 PR rtl-optimization/57268
15241 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15242 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15244 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15246 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15247 (ix86_avoid_lea_for_addr): Likewise.
15248 (exact_dependency_1): Likewise.
15249 (ix86_adjust_cost): Likewise.
15250 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15252 (do_reorder_for_imul): Fix typo, formatting and
15253 !reload_completed check removed.
15254 (ix86_sched_reorder): Fix typo and formatting.
15255 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15258 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15260 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15262 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15264 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15265 <STRING_CST>: Likewise.
15266 <VECTOR_CST>: Likewise.
15268 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15269 Mikael Morin <mikael@gcc.gnu.org>
15271 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15272 * config.in: Regenerated.
15273 * configure: Regenerated.
15275 2013-06-01 Jan Hubicka <jh@suse.cz>
15277 PR middle-end/57366
15278 * cgraphunit.c (compile): When weakref is not supported,
15279 set up transparent aliases before final output pass.
15280 * varasm.c (assemble_alias): Do not try to do it here.
15282 2013-06-01 Jan Hubicka <jh@suse.cz>
15284 PR middle-end/57467
15285 * passes.c (for_per_function): Skip unanalyzed functions.
15287 2013-06-01 Jan Hubicka <jh@suse.cz>
15289 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15290 (lto_symtab_merge_symbols_1): ... this one.
15291 (lto_symtab_merge_cgraph_nodes): Rename to ...
15292 (lto_symtab_merge_symbols): ... this one; simplify.
15293 * cgraph.c (same_body_aliases_done): Rename to ...
15294 (cpp_implicit_aliases_done): ... this one.
15295 (cgraph_create_function_alias): Update.
15296 (cgraph_same_body_alias): Update.
15297 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15298 (verify_edge_corresponds_to_fndecl): Simplify.
15299 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15300 (cgraph_node): Remove same_body_alias.
15301 (varpool_node): Remove alias_of and extra_name_alias.
15302 (same_body_aliases_done): Rename to ..
15303 (cpp_implicit_aliases_done): ... this one.
15304 (symtab_alias_ultimate_target): Add default parameter.
15305 (symtab_resolve_alias): New function.
15306 (fixup_same_cpp_alias_visibility): Declare.
15307 (cgraph_function_node): Add default parameter.
15308 (cgraph_node_asm_name): Likewise.
15309 (cgraph_function_or_thunk_node): Add default parameter; do
15310 not ICE when it is NULL.
15311 (varpool_variable_node): Likewise.
15312 * tree-emutls.c (create_emultls_var): Update.
15313 (ipa_lower_emutls): Update.
15314 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15315 (cgraph_reset_node): Reset alias info.
15316 (cgraph_finalize_function): Update.
15317 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15318 (analyze_function): Simplify.
15319 (cgraph_process_same_body_aliases): Simplify.
15320 (analyze_functions): Fixup same body aliases.
15321 (handle_alias_pairs): Simplify.
15322 (assemble_thunk): Update.
15323 (assemble_thunks_and_aliases): Update.
15324 (output_weakrefs): Rewrite.
15325 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15326 (lto_output_varpool_node): Likewise.
15327 (compute_ltrans_boundary): Remve assert.
15328 (get_alias_symbol): New functoin.
15329 (input_node): Rewrite alias handling.
15330 (input_varpool_node): Likewise.
15331 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15332 * ipa.c (process_references): Handle weakrefs correctly.
15333 (symtab_remove_unreachable_nodes): Likewise.
15334 * trans-mem.c (get_cg_data): Update.
15335 (ipa_tm_create_version_alias): Update.
15336 (ipa_tm_execute): Update.
15337 * symtab.c (dump_symtab_base): Dump aliases.
15338 (verify_symtab_base): Verify aliases.
15339 (symtab_node_availability): New function.
15340 (symtab_alias_ultimate_target): Simplify.
15341 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15342 handle all the fixup cases.
15343 (symtab_resolve_alias): New function.
15344 * passes.c (ipa_write_summaries): Handle weakrefs.
15345 * varpool.c (varpool_analyze_node): Simplify.
15346 (assemble_aliases): Update.
15347 (varpool_create_variable_alias): Simplify.
15348 (varpool_extra_name_alias): Simplify.
15349 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15350 (lto_symtab_merge_symbols): ... this one.
15352 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15355 PR rtl-optimization/57268
15356 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15357 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15359 2013-06-01 Tobias Burnus <burnus@net-b.de>
15361 Partially reverted:
15362 2013-05-31 Tobias Burnus <burnus@net-b.de>
15364 PR middle-end/57073
15365 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15368 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15370 PR rtl-optimization/57268
15371 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15372 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15374 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15376 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15377 unordered comparison operators when -fno-trapping-math is in effect
15379 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15380 and implement unordered comparison operators properly on the e500.
15382 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15384 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15385 for constant scalar integers.
15386 (simplify_relational_operation_1): Likewise.
15388 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15390 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15391 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15394 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15395 Igor Zamyatin <igor.zamyatin@intel.com>
15397 Silvermont (SLM) architecture performance tuning.
15398 * config/i386/i386.h (enum ix86_tune_indices): Add
15399 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15400 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15402 * config/i386/i386.c (initial_ix86_tune_features)
15403 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15404 (ix86_lea_outperforms): Handle Silvermont tuning.
15405 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15407 (ix86_use_lea_for_mov): Likewise.
15408 (ix86_avoid_lea_for_addr): Likewise.
15409 (ix86_lea_for_add_ok): Likewise.
15410 (exact_dependency_1): New function.
15411 (exact_store_load_dependency): Likewise.
15412 (ix86_adjust_cost): Handle Silvermont tuning.
15413 (do_reoder_for_imul): Likewise.
15414 (swap_top_of_ready_list): New function.
15415 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15417 * config/i386/i386.md (peepholes that split memory operand in fp
15420 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15422 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15423 Remove un-necessary braces.
15425 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15427 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15428 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15430 2013-05-31 Tobias Burnus <burnus@net-b.de>
15432 PR middle-end/57073
15433 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15435 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15438 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15439 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15440 * config/arm/neon.md (iordi3_neon): Remove.
15441 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15442 * config/arm/predicates.md (imm_for_neon_logic_operand):
15443 Move to earlier in the file.
15444 (neon_logic_op2): Likewise.
15445 (arm_iordi_operand_neon): New predicate.
15447 2013-05-31 Richard Biener <rguenther@suse.de>
15449 PR tree-optimization/57478
15450 PR tree-optimization/57453
15451 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15454 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15456 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15457 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15459 2013-05-30 Tobias Burnus <burnus@net-b.de>
15460 Thomas Koenig <tkoenig@gcc.gnu.org>
15462 PR middle-end/57073
15463 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15464 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15466 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15468 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15470 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15472 * target.def (register_usage_leveling_p): New hook.
15473 * targhooks.c (default_register_usage_leveling_p): New.
15474 * targhooks.h (default_register_usage_leveling_p): New prototype.
15475 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15476 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15477 * doc/tm.texi: Update.
15478 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15480 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15482 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15483 (*insv_reg<mode>): New define_insn.
15485 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15487 PR rtl-optimization/57439
15488 * postreload.c (move2add_valid_value_p): Check that we have
15489 a zero subreg_regno_offset when accessing the register in
15490 the requested mode.
15492 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15493 Igor Zamyatin <igor.zamyatin@intel.com>
15495 Silvermont (SLM) architecture pipeline model, tuning and
15497 * config.gcc: Add slm config options and target.
15499 * config/i386/slm.md: New.
15501 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15503 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15505 (ix86_target_macros_internal): Likewise.
15507 * config/i386/i386.c (slm_cost): New cost.
15508 (m_SLM): New macro flag.
15509 (initial_ix86_tune_features): Set m_SLM.
15510 (x86_accumulate_outgoing_args): Likewise.
15511 (x86_arch_always_fancy_math_387): Likewise.
15512 (processor_target_table): Add slm cost.
15513 (cpu_names): Add slm cpu name.
15514 (x86_option_override_internal): Set SLM ISA.
15515 (ix86_issue_rate): New case PROCESSOR_SLM.
15516 (ia32_multipass_dfa_lookahead): Likewise.
15517 (fold_builtin_cpu): Add slm.
15519 * config/i386/i386.h (TARGET_SLM): New target macro.
15520 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15521 (processor_type): Add PROCESSOR_SLM.
15523 * config/i386/i386.md (cpu): Add new value "slm".
15524 (slm.md): Include slm.md.
15526 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15527 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15529 * config/arm/arm-protos.h: Add and update function protos.
15530 * config/arm/arm.c (use_simple_return_p): New added.
15531 (thumb2_expand_return): Check simple_return flag.
15532 * config/arm/arm.md: Add simple_return and conditional simple_return.
15533 * config/arm/iterators.md: Add iterator for return and simple_return.
15535 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15537 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15538 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15539 (arm_emit_vfp_multi_reg_pop): Likewise.
15540 (thumb2_emit_ldrd_pop): Likewise.
15541 (arm_expand_epilogue): Add misc REG_CFA notes.
15542 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15544 2013-05-29 Lawrence Crowl <crowl@google.com>
15546 * config/arm/t-arm: Update for below.
15548 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15549 Change type to hash_table. Update dependent calls and types.
15551 * config/i386/t-cygming: Update for below.
15553 * config/i386/t-interix: Update for below.
15555 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15556 Change type to hash_table. Update dependent calls and types.
15557 (i386_find_on_wrapper_list::wrappers): Likewise.
15559 * config/ia64/t-ia64: Update for below.
15561 * config/ia64/ia64.c (bundle_state_table):
15562 Change type to hash_table. Update dependent calls and types.
15564 * config/mips/mips.c (mips_reorg_process_insns::htab):
15565 Change type to hash_table. Update dependent calls and types.
15567 * config/sol2.c (solaris_comdat_htab):
15568 Change type to hash_table. Update dependent calls and types.
15570 * config/t-sol2: Update for above.
15572 2013-05-29 Teresa Johnson <tejohnson@google.com>
15574 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15575 functions are not yet marked as defined.
15577 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15578 Pat Haugen <pthaugen@us.ibm.com>
15579 Peter Bergner <bergner@vnet.ibm.com>
15581 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15585 (vrotl<mode>3): Likewise.
15586 (vashl<mode>3): Likewise.
15587 (vlshr<mode>3): Likewise.
15588 (vashr<mode>3): Likewise.
15590 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15591 support for power8 V2DI builtins.
15593 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15594 power8 V2DI builtins.
15595 (vupkhsw): Likewise.
15596 (vupklsw): Likewise.
15597 (vaddudm): Likewise.
15598 (vminsd): Likewise.
15599 (vmaxsd): Likewise.
15600 (vminud): Likewise.
15601 (vmaxud): Likewise.
15602 (vpkudum): Likewise.
15603 (vpksdss): Likewise.
15604 (vpkudus): Likewise.
15605 (vpksdus): Likewise.
15610 (vsubudm): Likewise.
15611 (vcmpequd): Likewise.
15612 (vcmpgtsd): Likewise.
15613 (vcmpgtud): Likewise.
15614 (vcmpequd_p): Likewise.
15615 (vcmpgtsd_p): Likewise.
15616 (vcmpgtud_p): Likewise.
15617 (vupkhsw): Likewise.
15618 (vupklsw): Likewise.
15619 (vaddudm): Likewise.
15620 (vmaxsd): Likewise.
15621 (vmaxud): Likewise.
15622 (vminsd): Likewise.
15623 (vminud): Likewise.
15624 (vpksdss): Likewise.
15625 (vpksdus): Likewise.
15626 (vpkudum): Likewise.
15627 (vpkudus): Likewise.
15632 (vsubudm): Likewise.
15634 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15635 support for power8 V2DI instructions.
15637 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15638 power8 V2DI instructions. Combine pack and unpack insns to use an
15639 iterator for each mode. Check whether a particular mode supports
15640 Altivec instructions instead of just checking TARGET_ALTIVEC.
15641 (UNSPEC_VPKUWUM): Likewise.
15642 (UNSPEC_VPKSHSS): Likewise.
15643 (UNSPEC_VPKSWSS): Likewise.
15644 (UNSPEC_VPKUHUS): Likewise.
15645 (UNSPEC_VPKSHUS): Likewise.
15646 (UNSPEC_VPKUWUS): Likewise.
15647 (UNSPEC_VPKSWUS): Likewise.
15648 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15649 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15650 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15651 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15652 (UNSPEC_VUPKHSB): Likewise.
15653 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15654 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15655 (UNSPEC_VUPKHSH): Likewise.
15656 (UNSPEC_VUPKLSB): Likewise.
15657 (UNSPEC_VUPKLSH): Likewise.
15659 (VI_char): Likewise.
15660 (VI_scalar): Likewise.
15661 (VI_unit): Likewise.
15663 (VP_small): Likewise.
15664 (VP_small_lc): Likewise.
15665 (VU_char): Likewise.
15666 (add<mode>3): Likewise.
15667 (altivec_vaddcuw): Likewise.
15668 (altivec_vaddu<VI_char>s): Likewise.
15669 (altivec_vadds<VI_char>s): Likewise.
15670 (sub<mode>3): Likewise.
15671 (altivec_vsubcuw): Likewise.
15672 (altivec_vsubu<VI_char>s): Likewise.
15673 (altivec_vsubs<VI_char>s): Likewise.
15674 (altivec_vavgs<VI_char>): Likewise.
15675 (altivec_vcmpbfp): Likewise.
15676 (altivec_eq<mode>): Likewise.
15677 (altivec_gt<mode>): Likewise.
15678 (altivec_gtu<mode>): Likewise.
15679 (umax<mode>3): Likewise.
15680 (smax<mode>3): Likewise.
15681 (umin<mode>3): Likewise.
15682 (smin<mode>3): Likewise.
15683 (altivec_vpkuhum): Likewise.
15684 (altivec_vpkuwum): Likewise.
15685 (altivec_vpkshss): Likewise.
15686 (altivec_vpkswss): Likewise.
15687 (altivec_vpkuhus): Likewise.
15688 (altivec_vpkshus): Likewise.
15689 (altivec_vpkuwus): Likewise.
15690 (altivec_vpkswus): Likewise.
15691 (altivec_vpks<VI_char>ss): Likewise.
15692 (altivec_vpks<VI_char>us): Likewise.
15693 (altivec_vpku<VI_char>us): Likewise.
15694 (altivec_vpku<VI_char>um): Likewise.
15695 (altivec_vrl<VI_char>): Likewise.
15696 (altivec_vsl<VI_char>): Likewise.
15697 (altivec_vsr<VI_char>): Likewise.
15698 (altivec_vsra<VI_char>): Likewise.
15699 (altivec_vsldoi_<mode>): Likewise.
15700 (altivec_vupkhsb): Likewise.
15701 (altivec_vupkhs<VU_char>): Likewise.
15702 (altivec_vupkls<VU_char>): Likewise.
15703 (altivec_vupkhsh): Likewise.
15704 (altivec_vupklsb): Likewise.
15705 (altivec_vupklsh): Likewise.
15706 (altivec_vcmpequ<VI_char>_p): Likewise.
15707 (altivec_vcmpgts<VI_char>_p): Likewise.
15708 (altivec_vcmpgtu<VI_char>_p): Likewise.
15709 (abs<mode>2): Likewise.
15710 (vec_unpacks_hi_v16qi): Likewise.
15711 (vec_unpacks_hi_v8hi): Likewise.
15712 (vec_unpacks_lo_v16qi): Likewise.
15713 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15714 (vec_unpacks_lo_v8hi): Likewise.
15715 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15716 (vec_pack_trunc_v8h): Likewise.
15717 (vec_pack_trunc_v4si): Likewise.
15718 (vec_pack_trunc_<mode>): Likewise.
15720 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15722 (vec_vmaxsd): Likewise.
15723 (vec_vmaxud): Likewise.
15724 (vec_vminsd): Likewise.
15725 (vec_vminud): Likewise.
15726 (vec_vpksdss): Likewise.
15727 (vec_vpksdus): Likewise.
15728 (vec_vpkudum): Likewise.
15729 (vec_vpkudus): Likewise.
15730 (vec_vrld): Likewise.
15731 (vec_vsld): Likewise.
15732 (vec_vsrad): Likewise.
15733 (vec_vsrd): Likewise.
15734 (vec_vsubudm): Likewise.
15735 (vec_vupkhsw): Likewise.
15736 (vec_vupklsw): Likewise.
15738 2013-05-29 Jan Hubicka <jh@suse.cz>
15740 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15741 flags; reorder rest of fields in more consistent way.
15742 (varpool_node): Remove analyzed, finalized and alias.
15743 (cgraph_ndoe): Likewise.
15744 (symtab_alias_ultimate_target): New function.
15745 (cgraph_function_node): Move offline.
15746 (cgraph_reset_node): Declare.
15747 (cgraph_comdat_can_be_unshared_p): Remove.
15748 (varpool_remove_initializer): Declare.
15749 (varpool_first_defined_variable, varpool_next_defined_variable
15750 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15751 (cgraph_function_with_gimple_body_p): Update.
15752 (varpool_all_refs_explicit_p): Update.
15753 (symtab_alias_target): New function.
15754 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15755 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15756 (cgraph_function_or_thunk_node): Simplify using
15757 symtab_alias_ultimate_target.
15758 (varpool_variable_node): Likewise.
15759 * cgraph.c (cgraph_create_function_alias): Update.
15760 (cgraph_add_thunk): Update.
15761 (cgraph_remove_node): Update.
15762 (dump_cgraph_node): Do not dump removed flags.
15763 (cgraph_function_body_availability): Update.
15764 (cgraph_propagate_frequency): Update.
15765 (verify_cgraph_node): Check sanity of local flag.
15766 (cgraph_function_node): Move here from cgraph.h; revamp for
15767 cgraph_function_or_thunk_node.
15768 * lto-symtab.c (lto_varpool_replace_node): Update.
15769 (lto_symtab_resolve_can_prevail_p): Update.
15770 (lto_symtab_merge_cgraph_nodes): Update.
15771 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15772 propagate_constants_accross_call, devirtualization_time_bonus,
15773 ipcp_propagate_stage): Update.
15774 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15775 * ipa-inline-transform.c (clone_inlined_nodes,
15776 preserve_function_body_p): Update.
15777 * ipa-reference.c (propagate): Update.
15778 (write_node_summary_p): Update.
15779 * toplev.c (wrapup_global_declaration_2): Update.
15780 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15781 (analyze_function) ... this one.
15782 (cgraph_process_new_functions): Update.
15783 (cgraph_reset_node): Export.
15784 (cgraph_finalize_function): Update.
15785 (cgraph_add_new_function): Update.
15786 (process_function_and_variable_attributes): Update.
15787 (varpool_finalize_decl): Update.
15788 (symbol_finalized): Remove.
15789 (symbol_finalized_and_needed): Rename to ...
15790 (symbol_defined_and_needed): ... update.
15791 (cgraph_analyze_functions): Update.
15792 (handle_alias_pairs): Update.
15793 (mark_functions_to_output): Update.
15794 (assemble_thunk): Update.
15795 (output_in_order): Update.
15796 (output_weakrefs): Update.
15797 (finalize_compilation_unit): Update.
15798 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15799 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15800 input_node, input_varpool_node): Update.
15801 * dbxout.c (dbxout_expand_expr): Update.
15802 * cgraphclones.c (cgraph_clone_node): Update.
15803 (cgraph_copy_node_for_versioning): Update.
15804 (cgraph_materialize_clone): Update.
15805 (cgraph_materialize_all_clones): Update.
15806 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15807 propagate_pure_const, propagate_nothrow): Update.
15808 * lto-streamer-out.c (lto_output, write_symbol): Update.
15809 * ipa-utils.c (ipa_reverse_postorder): Update.
15810 * ipa-inline.c (can_inline_edge_p): Update.
15811 (update_caller_keys, ipa_inline): Update.
15812 * dwarf2out.c (reference_to_unused,
15813 premark_types_used_by_global_vars_helper): Update.
15814 * tree-eh.c (tree_could_trap_p): Update.
15815 * ipa-split.c (consider_split, execute_split_functions): Update.
15816 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15817 has_addr_references_p): Update; move ahead in file for better
15819 (process_references): Simplify.
15820 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15821 bodies are removed.
15822 (cgraph_comdat_can_be_unshared_p): Make static.
15823 (cgraph_externally_visible_p): Update.
15824 (varpool_externally_visible_p): Update.
15825 (function_and_variable_visibility): Update.
15826 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15827 ipa_tm_mark_force_output_node): Update.
15828 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15829 estimate_edge_devirt_benefit, inline_generate_summary,
15830 inline_write_summary): Update.
15831 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15832 * ipa-prop.c (ipa_compute_jump_functions): Update.
15833 (ipa_print_node_params, ipa_prop_read_section,
15834 ipa_update_after_lto_read, read_replacements_section): Update.
15835 * varasm.c (mark_decl_referenced): Update.
15836 (assemble_alias, dump_tm_clone_pairs): Update.
15837 * tree-inline.c (copy_bb): Update.
15838 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15840 * symtab.c (dump_symtab_base): Print new flags.
15841 (verify_symtab_base): Verify new flags.
15842 (symtab_alias_ultimate_target): New function.
15843 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15844 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15846 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15848 * i386.c (ix86_get_function_versions_dispatcher,
15849 ix86_generate_version_dispatcher_body): Update.
15850 (fold_builtin_cpu): Use varpool_add_new_variable.
15851 * varpool.c (varpool_remove_initializer): Break out from ...
15852 (varpool_remove_node): ... this one.
15853 (dump_varpool_node, varpool_node_for_asm,
15854 cgraph_variable_initializer_availability, varpool_analyze_node,
15855 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15856 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15859 2013-05-29 Jan Hubicka <jh@suse.cz>
15861 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15863 2013-05-29 Easwaran Raman <eraman@google.com>
15865 PR tree-optimization/57442
15866 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15867 when control exits the main loop.
15869 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15871 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15873 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15874 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15875 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15878 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15880 PR tree-optimization/57441
15881 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15882 Don't limit size of incr_vec to number of candidates.
15884 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15886 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15887 and mips16 directories.
15888 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15889 (MULTILIB_DIRNAMES): Ditto.
15890 (MULTILIB_EXCEPTIONS): Add new exceptions.
15891 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15892 (MULTILIB_DIRNAMES): Ditto.
15893 (MULTILIB_EXCEPTIONS): Add new exceptions.
15895 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15896 Marcus Shawcroft <marcus.shawcroft@arm.com>
15898 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15899 SYMBOL_TINY_ABSOLUTE.
15900 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15901 SYMBOL_TINY_ABSOLUTE.
15902 (aarch64_expand_mov_immediate): Likewise.
15903 (aarch64_classify_symbol): Likewise.
15904 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15905 Permit SYMBOL_TINY_ABSOLUTE.
15906 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15908 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15909 Marcus Shawcroft <marcus.shawcroft@arm.com>
15911 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15912 Refactor if/switch. Replace gcc_assert with if.
15914 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15916 * config/i386/i386.c (initial_ix86_tune_features): Enable
15917 FP Reassociation for AMD bdver1 and bdver2.
15919 2013-05-29 Martin Jambor <mjambor@suse.cz>
15921 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15922 and IMAGPART_EXPR do not occur within other handled_components.
15924 2013-05-29 Richard Biener <rguenther@suse.de>
15926 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15927 access on whether the use is in the BB we currently try to
15929 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15930 vectorize to vect_bb_slp_scalar_cost.
15932 2013-05-29 Richard Biener <rguenther@suse.de>
15934 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15935 computing scalar cost offsetted by stmts that are kept live
15937 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15938 for computation of scalar cost.
15940 2013-05-28 Steve Ellcey <sellcey@mips.com>
15942 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15944 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15946 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15947 array notation built-in reduction functions.
15948 * doc/passes.texi (Passes): Added documentation about changes done
15950 * doc/invoke.texi (C Dialect Options): Added documentation about
15951 the -fcilkplus flag.
15952 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15953 (BUILTINS_DEF): Depend on cilkplus.def.
15954 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15955 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15956 * cilkplus.def: New file.
15958 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15960 PR rtl-optimization/57439
15961 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15963 2013-05-28 Easwaran Raman <eraman@google.com>
15965 PR tree-optimization/57337
15966 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15967 (find_insert_point): Correctly identify the insertion point
15968 when two statements with the same UID is compared.
15970 2013-05-28 Richard Biener <rguenther@suse.de>
15972 PR tree-optimization/56787
15973 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15974 from the list of data references.
15975 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15977 (vect_analyze_loop_operations): Likewise.
15978 (vect_transform_loop): Remove clobbers.
15980 2013-05-28 Martin Jambor <mjambor@suse.cz>
15982 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15983 and REALPART_EXPRs have scalar type.
15985 2013-05-28 Richard Biener <rguenther@suse.de>
15987 PR tree-optimization/57411
15988 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15990 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15991 virtual operand propagation.
15993 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15995 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15996 destination register for bmasksi_vis.
15997 (vector_init_bshuffle): Likewise.
15998 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16000 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16002 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16003 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16004 mode if the instruction isn't available in the original mode.
16005 * config/sparc/sparc.opt (mfix-ut699): New option.
16006 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16007 (divdf3): Turn into expander.
16008 (divdf3_nofix): New insn.
16009 (divdf3_fix): Likewise.
16010 (divsf3): Disable if -mfix-ut699.
16011 (sqrtdf2): Turn into expander.
16012 (sqrtdf2_nofix): New insn.
16013 (sqrtdf2_fix): Likewise.
16014 (sqrtsf2): Disable if -mfix-ut699.
16016 2013-05-27 Richard Biener <rguenther@suse.de>
16018 PR middle-end/57412
16019 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16020 block for the new loop.
16022 2013-05-27 Richard Biener <rguenther@suse.de>
16024 PR tree-optimization/57343
16025 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16026 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16027 (number_of_iterations_cond): Do not build the folded tree.
16029 2013-05-27 Richard Biener <rguenther@suse.de>
16032 PR middle-end/57381
16033 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16034 OEP_CONSTANT_ADDRESS_OF retained.
16036 PR tree-optimization/57417
16037 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16038 for unchanged base.
16039 (set_ssa_val_to): Compare addresses using
16040 get_addr_base_and_unit_offset.
16042 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16044 PR rtl-optimization/56833
16045 * postreload.c (move2add_record_mode): New function.
16046 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16047 (move2add_use_add2_insn): Use move2add_record_sym_value.
16048 (move2add_use_add3_insn): Likewise.
16049 (reload_cse_move2add): Use move2add_valid_value_p and
16050 move2add_record_mode. Invalidate call-clobbered and REG_INC
16051 affected regs by setting reg_mode to VOIDmode.
16052 (move2add_note_store): Don't pretend the inside of a SUBREG is
16053 the actual destination. Invalidate single/leading registers by
16054 setting reg_mode to VOIDmode.
16055 Use move2add_record_sym_value, move2add_valid_value_p and
16056 move2add_record_mode.
16058 2013-05-27 Richard Biener <rguenther@suse.de>
16060 PR tree-optimization/57396
16061 * tree-affine.c (double_int_constant_multiple_p): Properly
16062 return false for val == 0 and div != 0.
16064 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16066 * config/mips/mips.h: Use #elif in preprocessor conditions.
16068 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16071 * config/mips/constraints.md (kl): New constraint.
16072 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16073 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16074 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16075 from LO for MIPS16.
16076 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16078 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16081 * config/mips/mips.c (mips_can_inline_p): New function.
16082 (TARGET_CAN_INLINE_P): Define.
16084 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16086 * sched-int.h (ds_t, dw_t): Make unsigned int.
16087 Fix documentation that describes how all the ds_t bits are used.
16088 Reserve the last bit for delayed-branch scheduling.
16089 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16090 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16091 (gen_dep_weak_1): Remove prototype.
16092 * sched-deps.c (get_dep_weak_1): Make static.
16093 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16094 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16095 * doc/tm.texi: Regenerate.
16096 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16098 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16101 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16103 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16104 Sandra Loosemore <sandra@codesourcery.com>
16106 * config.gcc (powerpc-*): Allow native for with-cpu.
16108 2013-05-24 Jeff Law <law@redhat.com>
16110 PR tree-optimization/57124
16111 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16112 conversion feeding a condition if the range has an overflow
16113 if -fstrict-overflow. Add warnings for when we do make the
16116 2013-05-24 Dehao Chen <dehao@google.com>
16118 * tree-cfg.c (locus_discrim_map): Fix the typo.
16119 (locus_discrim_hasher): Likewise.
16120 (locus_discrim_hasher::hash): Likewise.
16121 (locus_discrim_hasher::equal): Likewise.
16123 2013-05-24 Martin Jambor <mjambor@suse.cz>
16125 PR tree-optimization/57294
16126 * cgraph.h (ipa_record_stmt_references): Declare.
16127 * cgraphbuild.c (ipa_record_stmt_references): New function.
16128 (build_cgraph_edges): Use ipa_record_stmt_references.
16129 (rebuild_cgraph_edges): Likewise.
16130 (cgraph_rebuild_references): Likewise.
16131 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16132 associated with the old statement and build references from the
16133 newly built statements.
16134 * ipa-ref.c (ipa_remove_stmt_references): New function.
16135 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16137 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16139 * lra-constraints.c (emit_spill_move): Use smaller mode for
16141 (check_and_process_move): Consider mem-reg moves for secondary
16143 (curr_insn_transform): Don't lose insns emitted before for
16144 secondary memory moves.
16145 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16146 reg set up in the current insn.
16148 2013-05-24 Dehao Chen <dehao@google.com>
16150 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16152 (locus_descrim_hasher::equal): Likewise.
16153 (build_gimple_cfg): New discriminator assignment algorithm.
16154 (make_edges): Likewise.
16155 (next_discriminator_for_locus): Likewise.
16156 (same_line_p): Likewise.
16157 (assign_discriminators): Likewise.
16158 (make_cond_expr_edges): Likewise.
16159 (make_gimple_switch_edges): Likewise.
16160 (make_goto_expr_edges): Likewise.
16161 (make_gimple_asm_edges): Likewise.
16163 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16165 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16166 X format specifier to only display bottom 16 bits.
16167 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16168 immediate to match for operand 2, since it will be masked.
16170 2013-05-24 Richard Biener <rguenther@suse.de>
16172 PR tree-optimization/57287
16173 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16174 all SSA names that occur in abnormal PHIs.
16176 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16179 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16180 that index is not negative.
16182 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16184 PR rtl-optimization/55177
16185 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16186 (simplify_byte_swapping_operation): New.
16187 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16188 (simplify_relational_operation_1): Deal with BSWAP.
16190 2013-05-23 Richard Henderson <rth@redhat.com>
16193 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16194 (ix86_reorg): Call it.
16196 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16199 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16200 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16201 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16203 2013-05-23 Christian Bruel <christian.bruel@st.com>
16206 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16208 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16209 Marcus Shawcroft <marcus.shawcroft@arm.com>
16211 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16212 * config/aarch64/constraints.md (Usa): Remove.
16213 * doc/md.texi (AArch64 Usa): Remove.
16215 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16216 Marcus Shawcroft <marcus.shawcroft@arm.com>
16218 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16219 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16220 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16221 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16223 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16225 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16226 instruction (AdvSIMD).
16227 * config/aarch64/aarch64-builtins.c
16228 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16229 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16231 2013-05-23 Martin Jambor <mjambor@suse.cz>
16233 PR middle-end/57347
16234 * tree.h (contains_bitfld_component_ref_p): Declare.
16235 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16236 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16238 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16239 not access a bit-field. Assert all final offsets are byte-aligned.
16241 2013-05-23 Richard Biener <rguenther@suse.de>
16243 PR tree-optimization/57380
16244 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16245 least one invariant or re-used load.
16246 * passes.c (init_optimization_passes): Move pass_phiprop before
16249 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16251 * config/aarch64/aarch64-simd.md
16252 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16254 2013-05-23 Richard Biener <rguenther@suse.de>
16256 PR middle-end/57381
16257 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16258 OEP_CONSTANT_ADDRESS_OF retained.
16260 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16262 PR middle-end/57344
16263 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16264 don't lower unit. Handle unit not being always BITS_PER_WORD.
16266 2013-05-23 Richard Biener <rguenther@suse.de>
16268 PR rtl-optimization/57341
16269 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16270 instead of true_dependence.
16272 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16274 * bb-reorder.c (branch_threshold): Make const.
16275 (exec_threshold): Ditto.
16277 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16278 Pat Haugen <pthaugen@us.ibm.com>
16279 Peter Bergner <bergner@vnet.ibm.com>
16281 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16282 documentation for the power8 crypto builtins.
16284 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16286 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16287 macros for defining power8 builtin functions.
16288 (BU_P8V_AV_2): Likewise.
16289 (BU_P8V_AV_P): Likewise.
16290 (BU_P8V_VSX_1): Likewise.
16291 (BU_P8V_OVERLOAD_1): Likewise.
16292 (BU_P8V_OVERLOAD_2): Likewise.
16293 (BU_CRYPTO_1): Likewise.
16294 (BU_CRYPTO_2): Likewise.
16295 (BU_CRYPTO_3): Likewise.
16296 (BU_CRYPTO_OVERLOAD_1): Likewise.
16297 (BU_CRYPTO_OVERLOAD_2): Likewise.
16298 (XSCVSPDP): Fix typo, point to the correct instruction.
16299 (VCIPHER): Add power8 crypto builtins.
16300 (VCIPHERLAST): Likewise.
16301 (VNCIPHER): Likewise.
16302 (VNCIPHERLAST): Likewise.
16303 (VPMSUMB): Likewise.
16304 (VPMSUMH): Likewise.
16305 (VPMSUMW): Likewise.
16306 (VPERMXOR_V2DI): Likewise.
16307 (VPERMXOR_V4SI: Likewise.
16308 (VPERMXOR_V8HI: Likewise.
16309 (VPERMXOR_V16QI: Likewise.
16310 (VSHASIGMAW): Likewise.
16311 (VSHASIGMAD): Likewise.
16312 (VPMSUM): Likewise.
16313 (VPERMXOR): Likewise.
16314 (VSHASIGMA): Likewise.
16316 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16317 __CRYPTO__ if the crypto instructions are available.
16318 (altivec_overloaded_builtins): Add support for overloaded power8
16321 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16322 support for power8 crypto builtins.
16323 (builtin_function_type): Likewise.
16324 (altivec_init_builtins): Add support for builtins that take vector
16325 long long (V2DI) arguments.
16327 * config/rs6000/crypto.md: New file, define power8 crypto
16330 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16331 Pat Haugen <pthaugen@us.ibm.com>
16332 Peter Bergner <bergner@vnet.ibm.com>
16334 * doc/invoke.texi (Option Summary): Add power8 options.
16335 (RS/6000 and PowerPC Options): Likewise.
16337 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16338 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16339 wm, wn, wr documentation.
16341 * config/rs6000/constraints.md (wm): New constraint for VSX
16342 registers if direct move instructions are enabled.
16343 (wn): New constraint for no registers.
16344 (wq): New constraint for quad word even GPR registers.
16345 (wr): New constraint if 64-bit instructions are enabled.
16346 (wv): New constraint if power8 vector instructions are enabled.
16347 (wQ): New constraint for quad word memory locations.
16349 * config/rs6000/predicates.md (const_0_to_15_operand): New
16350 constraint for 0..15 for crypto instructions.
16351 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16352 as GPR and floating point registers.
16353 (int_reg_operand): New predicate to match only GPR registers.
16354 (base_reg_operand): New predicate to match base registers.
16355 (quad_int_reg_operand): New predicate to match even GPR registers
16356 for quad memory operations.
16357 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16358 operations in both GPR and VSX registers.
16359 (quad_memory_operand): New predicate for quad memory operations.
16360 (reg_or_indexed_operand): New predicate for direct move support.
16362 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16363 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16364 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16365 (POWERPC_MASKS): Add power8 options.
16366 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16369 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16370 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16372 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16373 (-mpower8-fusion): New power8 options.
16374 (-mpower8-fusion-sign): Likewise.
16375 (-mpower8-vector): Likewise.
16376 (-mcrypto): Likewise.
16377 (-mdirect-move): Likewise.
16378 (-mquad-memory): Likewise.
16380 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16381 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16383 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16384 (rs6000_debug_vector_unit): Add p8_vector.
16385 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16386 definitions. Also print fusion state.
16387 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16388 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16389 (rs6000_option_override_internal): Add support for power8.
16390 (rs6000_common_init_builtins): Add debugging for skipped builtins
16391 if -mdebug=builtin.
16392 (rs6000_adjust_cost): Add power8 support.
16393 (rs6000_issue_rate): Likewise.
16394 (insn_must_be_first_in_group): Likewise.
16395 (insn_must_be_last_in_group): Likewise.
16396 (force_new_group): Likewise.
16397 (rs6000_register_move_cost): Likewise.
16398 (rs6000_opt_masks): Likewise.
16400 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16401 power8 capable assembler, default to power7 options.
16402 (TARGET_DIRECT_MOVE): Likewise.
16403 (TARGET_CRYPTO): Likewise.
16404 (TARGET_P8_VECTOR): Likewise.
16405 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16406 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16407 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16408 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16409 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16410 (TARGET_XSCVDPSPN): Likewise.
16411 (TARGET_XSCVSPDPN): Likewsie.
16412 (TARGET_SYNC_HI_QI): Likewise.
16413 (TARGET_SYNC_TI): Likewise.
16414 (MASK_CRYPTO): Likewise.
16415 (MASK_DIRECT_MOVE): Likewise.
16416 (MASK_P8_FUSION): Likewise.
16417 (MASK_P8_VECTOR): Likewise.
16418 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16419 temporary used by some of the direct move instructions to get two FP
16420 temporary registers does not force creation of a stack frame.
16421 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16422 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16423 that any VSX registers are tieable, even if they are also an
16424 Altivec vector mode.
16425 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16426 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16427 (RS6000_BTM_CRYPTO): Likewise.
16428 (RS6000_BTM_COMMON): Likewise.
16430 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16431 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16432 (enum rs6000_vector): Add power8 vector support.
16434 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16438 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16439 (any_sibcall_could_use_r3): this and handle indirect calls.
16440 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16442 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16444 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16446 2013-05-22 Richard Biener <rguenther@suse.de>
16448 PR middle-end/57349
16449 * profile.c (branch_prob): Do not split blocks that are
16450 abnormally receiving from ECF_RETURNS_TWICE functions.
16452 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16454 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16455 address mode. Move pointer mode initialization to the same place.
16457 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16459 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16460 while it has any effect.
16462 2013-05-21 Easwaran Raman <eraman@google.com>
16464 PR tree-optimization/57322
16465 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16466 UID of the statement added to the BB to be 1.
16468 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16470 PR tree-optimization/57331
16471 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16472 of conversion from pointer type to integral type with integer.
16474 2013-05-21 Martin Jambor <mjambor@suse.cz>
16477 * ipa-prop.c (ipa_read_node_info): Process param_used and
16478 controlled_uses in the same order as when writing.
16480 2013-05-21 Magnus Granberg <baldrick@free.fr>
16483 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16485 2013-05-21 Richard Biener <rguenther@suse.de>
16487 PR tree-optimization/57318
16488 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16489 estimate stmts with side-effects as likely eliminated.
16491 2013-05-21 Richard Biener <rguenther@suse.de>
16493 PR tree-optimization/57330
16494 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16495 preserve the call stmts fntype.
16497 2013-05-21 Richard Biener <rguenther@suse.de>
16499 PR tree-optimization/57303
16500 * tree-ssa-sink.c (statement_sink_location): Improve killing
16501 stmt detection and properly handle self-assignments.
16503 2013-05-21 Christian Bruel <christian.bruel@st.com>
16505 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16506 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16507 registers. Set register size out of the PARALLEL loop.
16509 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16512 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16513 (*fmasf4, *fmasf4_media): New insns.
16515 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16517 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16518 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16519 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16520 (mips_idiv_insns): Update the comments to say that the returned
16521 instruction counts are in units of BASE_INSN_LENGTH.
16522 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16523 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16524 using 2 rather than 4 as the length of indirect MIPS16 and
16525 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16526 length of a NOP. Don't divide MIPS16 lengths by 2.
16527 (mips16_split_long_branches): Assume a branch is long if the
16528 length is greater than 4 rather than 8.
16529 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16530 rather than multiplying them by 2. Multiply instruction counts
16531 by BASE_INSN_LENGTH rather than 4.
16532 (*jump_mips16, tls_get_tp_mips16_<mode>)
16533 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16535 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16537 * config/mips/mips.md (extended_mips16): Remove branch case.
16538 (length): Remove duplicated extended_mips16 test.
16540 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16542 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16544 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16546 * recog.h (Recog_data): Rename to...
16547 (recog_data_d): ...this.
16548 (recog_data): Update accordingly.
16549 * recog.c (recog_data): Likewise.
16550 * reload.c (save_recog_data): Likewise.
16551 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16552 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16554 2013-05-17 Julian Brown <julian@codesourcery.com>
16556 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16557 found in a REG_EQUAL note, invalidate it.
16559 2013-05-17 Easwaran Raman <eraman@google.com>
16561 * tree-ssa-reassoc.c (find_insert_point): New function.
16562 (insert_stmt_after): Likewise.
16563 (get_def_stmt): Likewise.
16564 (ensure_ops_are_available): Likewise.
16565 (not_dominated_by): Likewise.
16566 (rewrite_expr_tree): Do not move statements beyond what is
16567 necessary. Remove call to swap_ops_for_binary_stmt...
16568 (reassociate_bb): ... and move it here.
16569 (build_and_add_sum): Assign UIDs for new statements.
16570 (linearize_expr): Likewise.
16571 (do_reassoc): Renumber gimple statement UIDs.
16573 2013-05-17 Jan Hubicka <jh@suse.cz>
16575 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16577 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16578 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16580 (output_weakrefs): Update.
16582 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16583 Martin Jambor <mjambor@suse.cz>
16585 PR middle-end/57276
16586 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16587 value that corresponds to the given aggval is found in values vector.
16589 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16591 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16592 sse, sse2, sse3, ssse3 and sse4a flags to options.
16594 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16596 * gengtype-state.c: (s_expr_writer): New class, to handle
16597 prettifying of output layout of s-expressions.
16598 (state_writer): New class, to write out gtype.state.
16599 (state_written_type_count): Move this variable into member data of
16601 (s_expr_writer::s_expr_writer): New code: constructor for new class
16602 (state_writer::state_writer(): ditto
16603 (s_expr_writer::write_new_line): New function
16604 (s_expr_writer::write_any_indent): ditto
16605 (s_expr_writer::begin_s_expr): ditto
16606 (s_expr_writer::end_s_expr): ditto
16607 (write_state_fileloc): convert to method of state_writer...
16608 (state_writer:: write_state_fileloc): ...and use methods of
16609 s_expr_writer to write indentation into the gtype.state output file
16610 to visually represent the hierarchical structure of the list
16612 (write_state_fields): ditto, renaming to...
16613 (state_writer::write_state_fields)
16614 (write_state_a_string): ditto, renaming to...
16615 (state_writer::write_state_a_string)
16616 (write_state_string_option): ditto, renaming to...
16617 (state_writer::write_state_string_option)
16618 (write_state_type_option): ditto, renaming to...
16619 (state_writer::write_state_type_option)
16620 (write_state_nested_option): ditto, renaming to...
16621 (state_writer::write_state_nested_option)
16622 (write_state_option): ditto, renaming to...
16623 (state_writer::write_state_option)
16624 (write_state_options): ditto, renaming to...
16625 (state_writer::write_state_options)
16626 (write_state_lang_bitmap): ditto, renaming to...
16627 (state_writer::write_state_lang_bitmap)
16628 (write_state_version): ditto, renaming to...
16629 (state_writer::write_state_version)
16630 (write_state_scalar_type): ditto, renaming to...
16631 (state_writer::write_state_scalar_type)
16632 (write_state_string_type): ditto, renaming to...
16633 (state_writer::write_state_string_type)
16634 (write_state_undefined_type): ditto, renaming to...
16635 (state_writer::write_state_undefined_type)
16636 (write_state_struct_union_type): ditto, renaming to...
16637 (state_writer::write_state_struct_union_type)
16638 (write_state_struct_type): ditto, renaming to...
16639 (state_writer::write_state_struct_type)
16640 (write_state_user_struct_type): ditto, renaming to...
16641 (state_writer::write_state_user_struct_type)
16642 (write_state_lang_struct_type): ditto, renaming to...
16643 (state_writer::write_state_lang_struct_type)
16644 (write_state_param_struct_type): ditto, renaming to...
16645 (state_writer::write_state_param_struct_type)
16646 (write_state_pointer_type): ditto, renaming to...
16647 (state_writer::write_state_pointer_type)
16648 (write_state_array_type): ditto, renaming to...
16649 (state_writer::write_state_array_type)
16650 (write_state_gc_used): ditto, renaming to...
16651 (state_writer::write_state_gc_used)
16652 (write_state_common_type_content): ditto, renaming to...
16653 (state_writer::write_state_common_type_content)
16654 (write_state_type): ditto, renaming to...
16655 (state_writer::write_state_type)
16656 (write_state_pair_list): ditto, renaming to...
16657 (state_writer::write_state_pair_list)
16658 (write_state_pair): ditto, renaming to...
16659 (state_writer::write_state_pair)
16660 (write_state_typedefs): ditto, renaming to...
16661 (state_writer::write_state_typedefs)
16662 (write_state_structures): ditto, renaming to...
16663 (state_writer::write_state_structures)
16664 (write_state_param_structs): ditto, renaming to...
16665 (state_writer::write_state_param_structs)
16666 (write_state_variables): ditto, renaming to...
16667 (state_writer::write_state_variables)
16668 (write_state_srcdir): ditto, renaming to...
16669 (state_writer::write_state_srcdir)
16670 (write_state_files_list): ditto, renaming to...
16671 (state_writer::write_state_files_list)
16672 (write_state_languages): ditto, renaming to...
16673 (state_writer::write_state_languages)
16674 (write_state): create a state_writer instance and use it when
16675 writing out the state file
16677 2013-05-17 Mike Stump <mikestump@comcast.net>
16679 PR rtl-optimization/57304
16680 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16681 accessing DF_REF_REAL_LOC.
16683 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16685 PR rtl-optimization/57281
16686 PR rtl-optimization/57300
16687 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16688 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16689 what the other splitter did if the registers are dead.
16691 2013-05-17 Richard Biener <rguenther@suse.de>
16693 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16696 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16698 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16701 2013-05-17 Marek Polacek <polacek@redhat.com>
16703 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16704 length when doing non-zero store of storing '\0' to '\0'.
16706 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16708 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16709 vect_external_def oprnd1 with loop_vinfo, try to emit
16710 optional cast, negation and and stmts on the loop preheader
16711 edge instead of into the pattern def seq.
16713 PR tree-optimization/57051
16714 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16715 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16717 2013-05-16 Nick Clifton <nickc@redhat.com>
16719 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16720 (rl78_is_naked_func): New function.
16721 (rl78_expand_prologue): Skip prologue generation for naked functions.
16722 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16723 * doc/extend.texi (naked): Add RL78 to the list of processors
16724 that supports this attribute.
16726 2013-05-16 Jeff Law <law@redhat.com>
16728 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16730 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16732 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16733 cache parameters using detect_caches_amd also for CYRIX,
16734 NSC and TM2 signatures.
16736 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16737 Dzianis Kahanovich <mahatma@eu.by>
16741 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16742 VIA/Centaur processors and determine their cache parameters
16743 using detect_caches_amd.
16745 2013-05-16 Teresa Johnson <tejohnson@google.com>
16747 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16748 (rtl_verify_edges): New function.
16749 (rtl_verify_bb_insns): Ditto.
16750 (rtl_verify_bb_pointers): Ditto.
16751 (rtl_verify_bb_insn_chain): Ditto.
16752 (rtl_verify_fallthru): Ditto.
16753 (rtl_verify_bb_layout): Ditto.
16754 (rtl_verify_flow_info_1): Outline checks into new functions.
16755 (rtl_verify_flow_info): Ditto.
16757 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16759 * cfghooks.c (copy_bbs): Add update_dominance argument.
16760 * cfghooks.h (copy_bbs): Update prototype.
16761 * tree-cfg.c (gimple_duplicate_sese_region):
16762 Add update_dominance argument.
16763 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16764 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16765 gimple_duplicate_sese_region call.
16766 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16767 Update copy_bbs call.
16768 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16769 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16771 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16773 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16774 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16775 vect_recog_rotate_pattern.
16776 (vect_recog_rotate_pattern): New function.
16778 2013-05-16 Jason Merrill <jason@redhat.com>
16780 * Makefile.in (LLINKER): New variable.
16781 (mostlyclean): Remove link mutex.
16782 * configure.ac: Handle --enable-link-mutex.
16783 * lock-and-run.sh: New script.
16785 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16788 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16791 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16793 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16795 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16797 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16798 * config/arm/arm.c (next_consecutive_mem): New function.
16799 (gen_movmem_ldrd_strd): Likewise.
16800 * config/arm/arm.md (movmemqi): Update condition and code.
16801 (unaligned_loaddi, unaligned_storedi): New patterns.
16803 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16805 * config.gcc: Obsolete *-*-solaris2.9*.
16806 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16808 2013-05-16 Richard Biener <rguenther@suse.de>
16810 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16811 earlier, after GRAPHITE transforms and IV canonicalization.
16813 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16815 * omp-low.c (extract_omp_for_data): For collapsed loops,
16816 if at least one of the loops is known at compile time to
16817 iterate zero times, set count to 0.
16818 (expand_omp_regimplify_p): New function.
16819 (expand_omp_for_generic): For collapsed loops, if at least
16820 one of the loops isn't known to iterate at least once,
16821 add runtime check with setting count to 0.
16822 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16823 For unsigned types if it isn't known at compile time that
16824 the loop will iterate at least once, add runtime check to bypass
16825 the whole loop if initial condition isn't true.
16827 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16829 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16831 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16833 PR middle-end/57286
16834 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16835 transformations to avoid an infinite loop.
16837 2013-05-16 Marek Polacek <polacek@redhat.com>
16839 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16841 2013-05-15 Leif Ekblad <leif@rdos.net>
16843 * config/i386/i386.c (ix86_decompose_address): Use
16844 DEFAULT_TLS_SEG_REG to access TLS segment register.
16845 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16846 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16847 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16849 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16852 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16853 sibling calls to functions that would normally be lazily bound,
16854 unless $gp is call-clobbered.
16856 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16858 * config/i386/i386.c (ix86_option_override_internal): Update
16859 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16860 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16861 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16862 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16864 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16866 2013-05-15 Andreas Schwab <schwab@suse.de>
16868 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16869 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16872 2013-05-15 Teresa Johnson <tejohnson@google.com>
16874 * loop-unroll.c (report_unroll_peel): Check decision before
16875 emitting unroll/peel message.
16877 2013-05-15 Teresa Johnson <tejohnson@google.com>
16879 * function.h (has_bb_partition): New rtl_data flag.
16880 (bb_reorder_complete): Ditto.
16881 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16882 instead of flag_reorder_blocks_and_partition.
16883 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16884 with some enhancements.
16885 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16886 * bb-reorder.c (connect_traces): Check for has_bb_partition
16887 instead of flag_reorder_blocks_and_partition.
16888 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16889 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16890 verify_hot_cold_block_grouping.
16891 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16893 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16896 * config/arm/predicates.md (call_insn_operand): New predicate.
16897 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16898 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16899 if insn is not a tail call.
16900 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16902 * config/arm/arm.h (enum reg_class): New caller save register class.
16903 (REG_CLASS_NAMES): Likewise.
16904 (REG_CLASS_CONTENTS): Likewise.
16905 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16908 2013-05-15 Richard Biener <rguenther@suse.de>
16910 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16911 of MSG_OPTIMIZED_LOCATIONS.
16912 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16913 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16915 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16916 of MSG_OPTIMIZED_LOCATIONS.
16917 (execute_vect_slp): Likewise.
16918 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16919 (vect_create_cond_for_alias_checks): Likewise.
16920 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16921 (vect_recog_widen_mult_pattern): Likewise.
16922 (vect_recog_widen_sum_pattern): Likewise.
16923 (vect_recog_over_widening_pattern): Likewise.
16924 (vect_recog_widen_shift_pattern): Likewise.
16925 (vect_recog_vector_vector_shift_pattern): Likewise.
16926 (vect_recog_divmod_pattern): Likewise.
16927 (vect_recog_mixed_size_cond_pattern): Likewise.
16928 (vect_recog_bool_pattern): Likewise.
16929 (vect_pattern_recog_1): Likewise.
16931 2013-05-15 Martin Jambor <mjambor@suse.cz>
16933 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16934 non-functions to builtin_unreachable.
16935 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16936 correct when new direct edges were discovered.
16938 2013-05-15 Martin Jambor <mjambor@suse.cz>
16940 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16941 header, print symbol order instead of node uid, print more information
16942 about indirect edge targets.
16943 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16945 (ipa_make_edge_direct_to_target): Likewise.
16946 (remove_described_reference): Likewise.
16947 (propagate_controlled_uses): Likewise.
16948 (ipa_print_node_params): Also print symbol order.
16949 (ipcp_transform_function): Print symbol order instead of node uids.
16950 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16951 (cgraph_get_create_real_symbol_node): Likewise.
16952 * ipa-cp.c (print_lattice): Likewise.
16953 (print_all_lattices): Likewise.
16954 (determine_versionability): Likewise.
16955 (initialize_node_lattices): Likewise.
16956 (estimate_local_effects): Likewise.
16957 (update_profiling_info): Likewise.
16958 (create_specialized_node): Likewise.
16959 (perhaps_add_new_callers): Likewise.
16960 (decide_about_value): Likewise.
16961 (decide_whether_version_node): Likewise.
16962 (identify_dead_nodes): Likewise.
16963 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16964 (dump_inline_summary): Likewise.
16965 (estimate_node_size_and_time): Likewise.
16966 (inline_analyze_function): Likewise.
16967 * ipa-inline.c (report_inline_failed_reason): Likewise.
16968 (want_early_inline_function_p): Likewise.
16969 (edge_badness): Likewise.
16970 (update_edge_key): Likewise.
16971 (inline_small_functions): Likewise. Add dumping of order to two other
16973 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16974 instead of node uids.
16975 (propagate_pure_const): Likewise.
16976 (propagate_pure_const): Likewise.
16977 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16978 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16979 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16981 * tree-pretty-print.c (dump_function_header): Likewise.
16982 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16983 Print symbol order instead of node uids.
16985 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16987 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16988 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16990 2013-05-15 Richard Biener <rguenther@suse.de>
16992 PR tree-optimization/57275
16993 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16994 return value for fail to do runtime alias checks for gather loads.
16996 2013-05-15 Jan Hubicka <jh@suse.cz>
17000 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17001 weakrefs are not external.
17002 (lto_symtab_merge_decls): Fix thinko when dealing with
17003 non-lto_symtab decls.
17004 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17005 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17006 * varpool.c (dump_varpool_node): Dump more flags.
17008 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17010 * config/i386/i386.c (processor_alias_table): Add instruction
17011 FSGSBASE for AMD bdver3 architecture.
17013 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17015 * tree.c (warn_deprecated_use): Print file:line using locus color.
17016 * diagnostic.c (diagnostic_report_current_module): Print file:line
17017 and file:line:column using locus color.
17019 2013-05-14 Mike Stump <mikestump@comcast.net>
17021 * gdbinit.in: Add __null.
17023 2013-05-14 Mike Stump <mikestump@comcast.net>
17025 * recog.h: Rename struct recog_data to Recog_data.
17026 * recog.c: Likewise.
17027 * reload.c (can_reload_into): Likewise.
17028 * config/picochip/picochip.c: Likewise.
17030 2013-05-14 Mike Stump <mikestump@comcast.net>
17032 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17034 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17036 * resource.h (struct resources): Remove unch_memory member.
17037 (CLEAR_RESOURCE): Don't clear unch_memory.
17038 * resource.c (mark_referenced_resources): Don't set it.
17039 (mark_set_resources): Likewise.
17040 (mark_target_live_regs): Don't clear it.
17041 (init_resource_info): Likewise.
17042 * reorg.c (resource_conflicts_p): Don't compare it.
17043 (redundant_insn): Don't set it.
17045 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17047 * emit-rtl.c (next_label): Remove unused function.
17048 (skip_consecutive_labels, link_cc0_insns): Move to ...
17049 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17050 only place where these functions are used, and make them static.
17052 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17054 * fold-const.c (fold_negate_expr): Handle vectors.
17055 (fold_truth_not_expr): Make it static.
17056 (fold_invert_truthvalue): New static function.
17057 (invert_truthvalue_loc): Handle vectors. Do not call
17058 fold_truth_not_expr directly.
17059 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17060 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17061 (fold_comparison): Handle vector constants.
17062 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17063 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17064 * tree.h (fold_truth_not_expr): Remove declaration.
17066 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17068 * config/aarch64/aarch64-simd.md
17069 (aarch64_vcond_internal<mode>): Rename to...
17070 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17071 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17072 float modes. Clarify all iterator modes.
17073 (vcond<mode><mode>): Use new name for vcond expanders.
17074 (vcond<v_cmp_result><mode>): Likewise.
17075 (vcondu<mode><mode>: Likewise.
17076 * config/aarch64/iterators.md (VDQF_COND): New.
17078 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17081 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17082 variable for the shift amount. Check that we shift by non-negative
17085 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17088 * config/arm/arm.h (EPILOGUE_USES): Only return true
17089 for LR_REGNUM after epilogue_completed.
17091 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17093 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17094 is error_mark_node.
17096 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17099 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17100 and Solaris 11+/x86 with gld.
17101 * configure: Regenerate.
17103 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17105 * expmed.c (expand_shift_1): Canonicalize rotates by
17106 constant bitsize / 2 to bitsize - 1.
17107 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17108 case ROTATERT>: Likewise.
17111 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17113 * config/i386/i386.md (rotateinv): New code attr.
17114 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17115 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17116 roll $31, %eax, etc.
17118 2013-05-14 Richard Biener <rguenther@suse.de>
17120 PR middle-end/57235
17121 * tree-eh.c (sink_clobbers): Give up for successors with
17122 multiple predecessors and no virtual uses.
17124 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17126 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17127 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17129 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17131 PR middle-end/57251
17132 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17133 the case when both op0 and op1 have VOIDmode.
17135 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17137 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17138 in multiply-accumulate mode.
17140 2013-05-13 Guozhi Wei <carrot@google.com>
17142 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17144 2013-05-13 Kai Tietz <ktietz@redhat.com>
17147 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17148 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17149 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17150 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17151 (ix86_expand_prologue): Likewise.
17152 (ix86_expand_split_stack_prologue): Likewise.
17153 (legitimate_pic_address_disp_p): Likewise.
17154 (legitimize_pic_address): Likewise.
17155 (legitimize_tls_address): Likewise.
17156 (legitimize_pe_coff_symbol): Likewise.
17157 (output_pic_addr_const): Likewise.
17158 (construct_plt_address): Likewise.
17159 (ix86_expand_call): Likewise.
17160 (x86_output_mi_thunk): Likewise.
17161 (x86_function_profiler): Likewise.
17163 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17165 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17166 similar switch cases.
17167 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17168 (aarch64_simd_mov_to_<mode>low): Delete.
17169 (aarch64_simd_mov_to_<mode>high): Delete.
17170 (move_lo_quad_<mode>): Add w<-r alternative.
17171 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17172 (aarch64_simd_mov_from_*): Update type attribute.
17173 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17176 2013-05-13 Jan Hubicka <jh@suse.cz>
17178 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17179 * config/i386/i386.c (ix86_compute_frame_layout,
17180 ix86_expand_epilogue, emit_i387_cw_initialization,
17181 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17182 ix86_local_alignment): Fix use of size/speed predicates.
17184 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17186 PR tree-optimization/45216
17187 PR tree-optimization/57157
17188 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17189 the (-Y) & (B - 1) variant if OP is |.
17190 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17191 return shifted. Use (-op1) & (prec - 1) as other_amount
17192 instead of prec - op1.
17194 2013-05-13 Martin Jambor <mjambor@suse.cz>
17196 PR middle-end/42371
17197 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17198 (ipa_constant_data): New type.
17199 (ipa_jump_func): Use ipa_constant_data to hold information about
17200 constant jump functions.
17201 (ipa_get_jf_constant): Adjust to jump function type changes.
17202 (ipa_get_jf_constant_rdesc): New function.
17203 (ipa_param_descriptor): New field controlled_uses.
17204 (ipa_get_controlled_uses): New function.
17205 (ipa_set_controlled_uses): Likewise.
17206 * ipa-ref.h (ipa_find_reference): Declare.
17207 * ipa-prop.c (ipa_cst_ref_desc): New type.
17208 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17210 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17211 New parameter cs. Adjust all callers.
17212 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17213 (remove_described_reference): New function.
17214 (jfunc_rdesc_usable): Likewise.
17215 (try_make_edge_direct_simple_call): Decrement controlled use count,
17216 attempt to remove reference if it hits zero.
17217 (combine_controlled_uses_counters): New function.
17218 (propagate_controlled_uses): Likewise.
17219 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17220 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17221 (ipa_print_node_params): Print described use counter.
17222 (ipa_write_jump_function): Adjust to jump function type changes.
17223 (ipa_read_jump_function): New parameter CS, pass it to
17224 ipa_set_jf_constant. Adjust caller.
17225 (ipa_write_node_info): Stream controlled use count
17226 (ipa_read_node_info): Likewise.
17227 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17229 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17230 count. Remove cloning-added reference if it reaches zero.
17231 * ipa-ref.c (ipa_find_reference): New function.
17233 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17235 * config/i386/i386.c (processor_target_table): Modified default
17236 alignment values for AMD BD and BT architectures.
17238 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17240 * tree-vect-generic.c (uniform_vector_p): Move ...
17241 * tree.c (uniform_vector_p): ... here.
17242 * tree.h (uniform_vector_p): Declare it.
17243 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17246 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17248 PR tree-optimization/57230
17249 * tree-ssa-strlen.c (handle_char_store): Record length for
17250 array store from STRING_CST.
17252 PR tree-optimization/57230
17253 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17256 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17258 * config/epiphany/epiphany.c (epiphany_init): Check size of
17259 NUM_MODES_FOR_MODE_SWITCHING.
17260 (epiphany_expand_prologue):
17261 Remove CONFIG_REGNUM initial value handling code.
17262 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17263 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17264 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17265 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17266 Don't return 1 for FP_MODE_NONE.
17267 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17268 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17269 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17270 * config/epiphany/epiphany.md (save_config): New pattern.
17272 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17274 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17276 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17278 * config/i386/i386.md (memory): Handle sseishft1.
17279 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17280 (*vec_extractv2di_1): Ditto.
17282 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17284 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17287 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17289 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17290 Add mthumb/march=armv7-a multilib.
17291 Add mthumb/march=armv7-r multilib.
17292 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17294 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17296 * config/v850/t-rtems: Add more multilibs.
17298 2013-05-10 Richard Biener <rguenther@suse.de>
17300 PR tree-optimization/57214
17301 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17302 not propagate from SSA names that occur in abnormal PHI nodes.
17304 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17306 * stor-layout.c (element_precision): New function.
17307 * machmode.h (element_precision): Declare it.
17308 * tree.c (build_minus_one_cst): New function.
17309 (element_precision): Likewise.
17310 * tree.h (build_minus_one_cst): Declare new function.
17311 (element_precision): Likewise.
17312 * fold-const.c (operand_equal_p): Use element_precision.
17313 (fold_binary_loc): Handle vector types.
17314 * convert.c (convert_to_integer): Use element_precision.
17315 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17318 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17320 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17321 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17322 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17323 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17324 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17325 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17326 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17327 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17328 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17329 (Uuw8): New constraints.
17330 (Usb4): Move into alphabetical order.
17331 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17332 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17333 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17334 previously unnamed patterns.
17335 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17336 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17337 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17338 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17339 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17340 of specific lengths.
17342 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17344 * config/i386/i386.md (rotateinv): New code attr.
17345 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17346 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17347 roll $31, %eax, etc.
17349 PR tree-optimization/45216
17350 PR tree-optimization/57157
17351 * tree-ssa-forwprop.c (simplify_rotate): New function.
17352 (ssa_forward_propagate_and_combine): Call it.
17354 2013-05-10 Richard Biener <rguenther@suse.de>
17356 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17357 disable peeling when we version for aliasing.
17358 (vector_alignment_reachable_p): Honor explicit user alignment.
17359 (vect_supportable_dr_alignment): Likewise.
17360 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17361 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17362 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17363 then peeling to arrange for the cost-model check to come first.
17365 2013-05-10 Alan Modra <amodra@gmail.com>
17367 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17368 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17369 * configure: Regenerate.
17371 2013-05-10 Alan Modra <amodra@gmail.com>
17374 * varasm.c (default_elf_select_section): Move !DECL_P check..
17375 (get_named_section): ..to here before calling get_section_name.
17377 (default_section_type_flags): Add DECL_P check.
17378 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17379 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17381 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17383 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17384 When using gen_stack_adjust_str with a register offset, add a
17385 REG_FRAME_RELATED_EXPR note.
17387 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17389 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17390 (*vec_extractv4si_zext_mem): Ditto.
17391 (*vec_extractv2di): Add 0->x and x->x alternatives.
17392 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17393 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17395 2013-05-09 Jason Merrill <jason@redhat.com>
17397 N3639 C++1y VLA support
17398 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17401 * tree.c (build_constructor_va): New.
17402 * tree.h: Declare it.
17404 2013-05-09 Martin Jambor <mjambor@suse.cz>
17407 * gimple-fold.c (canonicalize_constructor_val): Call
17408 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17410 2013-05-09 Jan Hubicka <jh@suse.cz>
17411 Richard Biener <rguenther@suse.de>
17414 * symtab.c (symtab_make_decl_local): Do not add private names.
17416 2013-05-09 Jan Hubicka <jh@suse.cz>
17419 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17420 (unlink_from_assembler_name_hash): Likewise.
17421 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17422 symtab_unregister_node, symtab_initialize_asm_name_hash,
17423 change_decl_assembler_name): Update.
17425 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17427 * config/aarch64/aarch64.md: New movtf split.
17428 (*movtf_aarch64): Update.
17429 (aarch64_movdi_tilow): Handle TF modes and rename to
17430 aarch64_movdi_<mode>low.
17431 (aarch64_movdi_tihigh): Handle TF modes and rename to
17432 aarch64_movdi_<mode>high
17433 (aarch64_movtihigh_di): Handle TF modes and rename to
17434 aarch64_mov<mode>high_di
17435 (aarch64_movtilow_di): Handle TF modes and rename to
17436 aarch64_mov<mode>low_di
17437 (aarch64_movtilow_tilow): Remove spurious whitespace.
17438 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17440 (aarch64_print_operand): Update.
17442 2013-05-09 Alan Modra <amodra@gmail.com>
17444 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17446 * configure: Regenerate.
17448 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17450 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17451 splitter preparation statements.
17452 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17453 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17456 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17458 * gimple-ssa-strength-reduction.c (count_candidates): Change
17459 return value to int.
17460 (analyze_candidates_and_replace): Change type of length to int.
17462 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17464 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17465 (*vec_extract<mode>): Use VI12_128 mode iterator.
17466 (*vec_extract<mode>_mem): Ditto.
17467 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17470 2013-05-08 Diego Novillo <dnovillo@google.com>
17475 2012-08-17 Diego Novillo <dnovillo@google.com>
17478 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17479 * config.in: Regenerate.
17480 * configure: Regenerate.
17481 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17483 2013-05-08 Jan Hubicka <jh@suse.cz>
17486 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17487 * cgraph.h (symtab_node_base): Add unique_name.
17488 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17489 input_overwrite_node, input_varpool_node): Stream unique_name.
17490 * cgraphclones.c (cgraph_create_virtual_clone,
17491 cgraph_function_versioning): Set unique_name.
17492 * ipa.c (function_and_variable_visibility): Set unique_name.
17494 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17496 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17497 (alloc_cand_and_find_basis): Restrict conditional candidate
17498 processing to CAND_MULTs.
17500 2013-05-08 Jan Hubicka <jh@suse.cz>
17503 lto-symtab.c (lto_symtab_symbol_p): New function.
17504 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17505 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17506 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17507 Skip static symbols.
17509 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17511 PR tree-optimization/57200
17512 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17513 Only call inform if the preceding warning_at returns true.
17515 2013-05-07 Han Shen <shenhan@google.com>
17517 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17518 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17519 * common.opt (fstack-protector-strong): New option.
17520 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17521 * doc/invoke.texi (Optimization Options): Document
17522 "-fstack-protector-strong".
17523 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17525 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17527 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17529 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17531 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17532 it is not lhs of the insn.
17533 (setup_reg_equiv): Remove insn having equiv memory which it is not
17535 * lra-constraints.c (process_address): Try to improve generation
17536 code for address base + disp.
17537 (lra_constraints): Make correct the code for checking insn setting
17538 up backward equivalence. Remove insn only if it is in the init
17540 * lra-eliminations.c (update_reg_eliminate): Change return value.
17541 (lra_eliminate): Use the result.
17543 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17545 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17546 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17547 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17548 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17549 register target operands.
17550 (*vec_extractv8hi_sse2): New pattern.
17551 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17552 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17553 (*vec_extract<mode>_mem): New insn and split pattern.
17555 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17557 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17558 (TARGET_ASAN_SHADOW_OFFSET): Define.
17559 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17560 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17562 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17564 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17565 (incr_vec_index): Return -1 if increment not found.
17566 (create_add_on_incoming_edge): Assert if increment not found.
17567 (record_increment): Limit number of increments recorded.
17568 (all_phi_incrs_profitable): Return false if an increment not found.
17569 (replace_profitable_candidates): Don't process increments that were
17571 (analyze_candidates_and_replace): Limit size of incr_vec.
17573 2013-05-07 Richard Biener <rguenther@suse.de>
17575 * calls.c (special_function_p): setjmp-like functions are leaf.
17576 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17577 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17579 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17581 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17582 (aarch64_simd_mov<mode>): New expander.
17583 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17584 (aarch64_simd_mov_to_<mode>high): Likewise.
17585 (aarch64_simd_mov_from_<mode>low): Likewise.
17586 (aarch64_simd_mov_from_<mode>high): Likewise.
17587 (aarch64_dup_lane<mode>): Update.
17588 (aarch64_dup_lanedi): New instruction pattern.
17589 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17590 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17592 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17594 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17595 (replace_mult_candidate): Remove unnecessary argument; remove
17596 unnecessary parameter from call to introduce_cast_before_cand.
17597 (replace_unconditional_candidate): Remove unnecessary parameter
17598 from call to replace_mult_candidate.
17599 (replace_conditional_candidate): Likewise.
17600 (insert_initializers): Use make_temp_ssa_name.
17601 (introduce_cast_before_cand): Remove unnecessary argument; use
17602 make_temp_ssa_name.
17603 (replace_one_candidate): Remove unnecessary argument; remove
17604 unnecessary parameter from calls to introduce_cast_before_cand.
17605 (replace_profitable_candidates): Remove unnecessary parameters
17606 from calls to replace_one_candidate.
17608 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17610 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17611 phi def as possibly hiding a basis for a CAND_ADD whose operands
17612 have been commuted in the analysis.
17613 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17615 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17617 * config/aarch64/aarch64.md
17618 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17619 shift value between 0-4.
17621 2013-05-07 Richard Biener <rguenther@suse.de>
17623 * double-int.h (rshift): New overload.
17624 * double-int.c (rshift): New function.
17625 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17626 (create_reference_ops_from_ref): Remove.
17627 (vn_reference_insert): Use shared ops for constructing the
17628 reference and copy it.
17630 2013-05-07 Richard Biener <rguenther@suse.de>
17632 PR middle-end/57190
17633 * tree-eh.c (sink_clobbers): Properly propagate
17634 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17636 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17638 PR tree-optimization/57149
17639 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17640 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17641 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17642 uninit_undefined_value_p instead of ssa_undefined_value_p.
17645 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17646 for modifier == EXPAND_INITIALIZER.
17648 2013-05-07 Anton Blanchard <anton@samba.org>
17650 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17651 for powerpc64 little endian.
17652 * configure: Regenerate.
17654 2013-05-06 Graham Stott <grahams@btinternet.com>
17656 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17657 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17658 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17659 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17661 2013-05-06 Graham Stott <grahams@btinternet.com>
17663 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17664 codes which allow non-lvalues.
17666 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17668 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17669 components are all 1s.
17670 (integer_minus_onep): New function.
17671 * tree.h (integer_minus_onep): Declare it.
17672 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17673 integer_minus_onep instead of integer_all_onesp.
17675 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17678 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17679 variations of these patterns.
17681 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17683 * config/i386/i386.md (isa): Add x64_sse4 member.
17684 (enabled): Handle x64_sse4.
17685 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17686 instruction for 64bit SSE4_1 targets. Update insn attributes.
17687 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17688 instruction for SSE4_1 targets. Update insn attributes.
17689 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17690 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17691 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17693 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17694 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17695 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17697 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17700 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17702 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17704 * final.c (do_assembler_dialects): Don't handle curly braces and
17705 vertical bar escaped by % as dialect delimiters.
17706 (output_asm_insn): Print curly braces and vertical bar if escaped
17707 by % and ASSEMBLER_DIALECT defined.
17708 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17709 * doc/tm.texi: Regenerated.
17711 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17713 * config/mips/mips.c: Include tree-pass.h.
17714 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17715 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17716 (pass_mips_machine_reorg2): New machine specific pass.
17717 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17718 (mips_option_override): Register the new pass.
17719 * rtl.h (cleanup_barriers): Remove prototype.
17720 (dbr_schedule): Likewise.
17721 * jump.c (cleanup_barriers): Make static.
17722 * reorg.c (dbr_schedule): Likewise.
17724 2013-05-06 Richard Biener <rguenther@suse.de>
17726 PR tree-optimization/57185
17727 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17728 SSA names properly.
17730 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17733 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17734 between "first_arg" and "insn", not "last" and "insn".
17736 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17738 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17739 (find_candidates_in_block): Re-enable slsr_process_phi.
17740 (create_phi_basis): Fix double counting of candidate adjustment.
17742 2013-05-06 Richard Biener <rguenther@suse.de>
17744 PR middle-end/57147
17745 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17746 the edge is also fallthru, preserve it and just clear the
17748 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17749 also complex, preserve that and just clear the fallthru flag.
17750 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17751 update virtual operands.
17753 2013-05-06 Alan Modra <amodra@gmail.com>
17755 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17756 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17757 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17758 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17759 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17760 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17762 2013-05-06 Alan Modra <amodra@gmail.com>
17764 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17765 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17766 (DEFAULT_ASM_ENDIAN): Define.
17767 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17768 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17769 Update -K PIC clause from sysv4.h.
17770 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17771 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17773 2013-05-06 Alan Modra <amodra@gmail.com>
17775 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17776 twice for little-endian.
17777 (ashrdi3_no_power, ashrdi3): Support little-endian.
17779 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17782 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17783 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17785 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17786 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17787 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17790 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17791 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17793 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17794 * config/pa/pa.opt: Make mbig-switch a no-op.
17795 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17796 (CASE_VECTOR_MODE): Always return SImode.
17797 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17798 for the !TARGET_BIG_SWITCH case.
17799 * config/pa/pa-linux.h: Likewise.
17800 * config/pa/pa-openbsd.h: Likewise.
17801 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17802 * config/pa/pa.md (short_jump): Remove define_insn.
17803 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17804 (casesi0): Remove define_insn.
17805 (type): Remove btable_branch.
17806 (pa_combine_type): Likewise.
17807 (in_nullified_branch_delay): Likewise.
17808 (in_call_delay): Likewise.
17809 (define_delay): Likewise.
17810 (define_insn_reservation "Z3"): Likewise.
17811 (define_insn_reservation "Z4"): Likewise.
17812 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17813 (pa_adjust_insn_length): Remove adjustment for btable branches.
17814 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17817 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17819 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17820 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17821 Add m->r,x alternatives.
17822 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17823 splitters using SWI48x mode iterator.
17824 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17825 TARGET_64BIT. Add m->x alternative.
17826 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17827 Add o->x alternative. Enable for TARGET_SSE.
17828 (sse_storeq): Remove expander.
17829 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17830 with memory input operand.
17831 (*vec_extractv2di_1 splitter): New.
17832 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17833 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17835 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17837 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17838 (INT_LOWPART): Delete.
17839 (extract_MB): Adjust.
17840 (extract_ME): Adjust.
17841 (print_operand): Adjust.
17843 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17845 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17846 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17847 (reg_or_logical_cint_operand, easy_fp_constant,
17848 logical_const_operand): Delete "CONST_DOUBLE" case.
17849 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17850 "HOST_BITS_PER_WIDE_INT == 64" test.
17851 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17852 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17853 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17854 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17856 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17857 CONST_DOUBLE DImode/VOIDmode case.
17858 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17859 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17860 CONST_DOUBLE VOIDmode case.
17861 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17862 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17863 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17864 Delete CONST_DOUBLE case.
17865 (splitters for mov FMOVE64 const_double): Delete
17866 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17867 "HOST_BITS_PER_WIDE_INT >= 64" test.
17868 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17870 (mov DI const_double): Delete.
17872 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17874 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17875 on op shows all bits zero in mode of a lowpart subreg, return zero.
17877 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17880 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17881 to save TFmode registers and DImode to save TImode registers for
17882 caller save operations.
17883 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17884 mark being partially clobbered since they only use the first
17887 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17888 and TDmode only use the upper 64-bits of each VSX register.
17890 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17892 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17893 (find_candidates_in_block): Disable slsr_process_phi.
17895 2013-05-03 Guozhi Wei <carrot@google.com>
17897 * coverage.c (coverage_obj_init): Move the construction of gcov
17899 (build_init_ctor): ... here.
17901 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17903 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17904 (slsr_cand_d): Redefine def_phi.
17905 (stride_status, phi_adjust_status, count_phis_status): New enums.
17906 (find_phi_def): New.
17907 (find_basis_for_base_expr): New.
17908 (find_basis_for_candidate): Handle hidden bases.
17909 (alloc_cand_and_find_basis): Handle phi candidates.
17910 (slsr_process_phi): New.
17911 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17912 (create_mul_imm_cand): Likewise.
17913 (create_add_ssa_cand): Exclude phi base candidates.
17914 (create_add_imm_cand): Likewise.
17915 (slsr_process_cast): Likewise.
17916 (slsr_process_copy): Likewise.
17917 (find_candidates_in_block): Handle phi candidates.
17918 (dump_candidate): Likewise.
17919 (unconditional_cands): Delete.
17920 (unconditional_cands_with_known_stride_p): Delete.
17921 (phi_dependent_cand_p): New.
17922 (cand_increment): Handle phi-dependent candidates.
17923 (replace_dependent): Delete.
17924 (replace_mult_candidate): New.
17925 (replace_unconditional_candidate): New.
17926 (incr_vec_index): Move to avoid forward reference.
17927 (create_add_on_incoming_edge): New.
17928 (create_phi_basis): New.
17929 (replace_dependents): Delete.
17930 (replace_conditional_candidate): New.
17931 (phi_add_costs): New.
17932 (replace_uncond_cands_and_profitable_phis): New.
17933 (record_increment): Handle phi adjustments.
17934 (record_phi_increments): New.
17935 (record_increments): Handle phi adjustments.
17936 (phi_incr_cost): New.
17937 (lowest_cost_path): Handle phis.
17938 (total_savings): Likewise.
17939 (analyze_increments): Likewise.
17940 (ncd_with_phi): New.
17941 (ncd_of_cand_and_phis): New.
17942 (nearest_common_dominator_for_cands): Handle phi increments.
17943 (all_phi_incrs_profitable): New.
17944 (replace_profitable_candidates): Handle phi-dependent candidates.
17945 (analyze_candidates_and_replace): Likewise.
17947 2013-05-03 Teresa Johnson <tejohnson@google.com>
17950 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17951 do not exceed REG_BR_PROB_BASE.
17953 2013-05-03 Jeff Law <law@redhat.com>
17955 PR tree-optimization/57144
17956 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17957 operand of the condition will bit into the new type when eliminating
17958 a cast feeding a condition.
17960 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17962 PR rtl-optimization/57130
17963 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17964 of COMPARE as in_code to the recursive call if needed.
17966 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17968 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17969 (enabled): Handle new members.
17970 * config/i386/sse.md (*vec_concatv2si): Merge from
17971 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17972 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17974 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17976 PR tree-optimization/57027
17977 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17978 for fnms opportunity, check we got the prerequisite kind
17979 of tree / gimple before using accessor functions.
17981 2013-05-03 Richard Biener <rguenther@suse.de>
17983 * double-int.h (lshift): New overload without precision
17984 and arith argument.
17985 (operator *=, operator +=, operator -=): Move ...
17986 * double-int.c (operator *=, operator +=, operator -=): ... here
17987 and implement more efficiently.
17988 (mul_double_with_sign): Remove.
17989 (lshift_double): Adjust to take unsinged shift argument, push
17990 dispatching code to callers.
17991 (mul_double_wide_with_sign): Add early out for callers that
17992 are not interested in high parts or overflow.
17993 (lshift): New function.
17994 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17995 dispatch code here.
17996 (lrotate, rrotate): Use logical shifts.
17997 * expr.c (get_inner_reference): Use lshift.
17998 * fixed-value.c (do_fixed_divide): Likewise.
17999 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18000 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18001 (indirect_refs_may_alias_p): Likewise.
18002 (stmt_kills_ref_p_1): Likewise.
18004 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18006 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18008 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18010 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18011 scalar form of FABD instruction.
18013 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18015 * lra-constraints.c (process_alt_operands): Add checking alt
18016 number to choose the best alternative.
18018 2013-05-02 Richard Biener <rguenther@suse.de>
18020 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18021 bitmap and its handling.
18022 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18024 2013-05-02 Richard Biener <rguenther@suse.de>
18026 PR middle-end/57140
18027 * tree-inline.c (copy_loops): Properly handle removed loops.
18028 (copy_cfg_body): Mark destination loops for fixup if source
18029 loops needed fixup.
18031 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18034 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18035 generating simple_return for naked functions.
18037 2013-05-02 Martin Jambor <mjambor@suse.cz>
18039 PR middle-end/56988
18040 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18041 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18043 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18044 ipa_agg_replacement_value structures.
18045 (known_aggs_to_agg_replacement_list): Likewise.
18046 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18047 (read_agg_replacement_chain): Likewise.
18048 (ipcp_transform_function): Also check that by_ref flags match.
18050 2013-05-02 Richard Biener <rguenther@suse.de>
18052 * graphds.h (struct graph): Add obstack member.
18053 * graphds.c (new_graph): Initialize obstack and allocate
18055 (add_edge): Allocate edge from the obstack.
18056 (free_graph): Free the obstack instead of all edges and vertices.
18058 2013-05-02 Teresa Johnson <tejohnson@google.com>
18060 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18062 * cfg.c (update_bb_profile_for_threading): Ditto.
18063 * tree-inline.c (copy_bb): Ditto.
18064 (copy_edges_for_bb): Ditto.
18065 (initialize_cfun): Ditto.
18066 (copy_cfg_body): Ditto.
18067 (expand_call_inline): Ditto.
18068 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18069 (estimate_node_size_and_time): Ditto.
18070 (inline_merge_summary): Ditto.
18071 * cgraphclones.c (cgraph_clone_edge): Ditto.
18072 (cgraph_clone_node): Ditto.
18073 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18074 (compute_trg_info): Ditto.
18076 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18078 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18079 S reg when fp attribute set.
18080 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18082 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18084 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18086 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18087 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18088 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18090 2013-05-02 Richard Biener <rguenther@suse.de>
18092 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18093 (struct instantiate_cache_entry): New type.
18094 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18095 (struct instantiate_cache_type): New type.
18096 (set_instantiated_value, get_instantiated_value): Remove.
18097 (get_instantiated_value_entry): New function.
18098 (instantiate_scev_name): Use the new cache and adjust.
18099 (instantiate_scev_poly): Adjust.
18100 (instantiate_scev_binary): Likewise.
18101 (instantiate_array_ref): Likewise.
18102 (instantiate_scev_convert): Likewise.
18103 (instantiate_scev_not): Likewise.
18104 (instantiate_scev_3): Likewise.
18105 (instantiate_scev_2): Likewise.
18106 (instantiate_scev_r): Likewise.
18107 (instantiate_scev): Likewise.
18108 (resolve_mixers): Likewise.
18110 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18113 * lra-constraints.c (best_small_class_operands_num): Remove.
18114 (process_alt_operands): Remove small_class_operands_num. Take
18115 small classes operands into losers and only if the operand is not
18116 matched. Modify debugging output.
18117 (curr_insn_transform): Remove best_small_class_operands_num.
18120 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18122 * config/aarch64/aarch64-builtins.c
18123 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18124 * config/aarch64/aarch64-simd-builtins.def
18125 (reduc_splus_): Add new modes.
18126 (reduc_uplus_): New.
18127 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18128 (reduc_uplus_v4sf): Likewise.
18129 (reduc_splus_v4sf): Likewise.
18130 (aarch64_addv<mode>): Likewise.
18131 (reduc_uplus_<mode>): Likewise.
18132 (reduc_splus_<mode>): Likewise.
18133 (aarch64_addvv2di): Likewise.
18134 (reduc_uplus_v2di): Likewise.
18135 (reduc_splus_v2di): Likewise.
18136 (aarch64_addvv2si): Likewise.
18137 (reduc_uplus_v2si): Likewise.
18138 (reduc_splus_v2si): Likewise.
18139 (reduc_<sur>plus_<mode>): New.
18140 (reduc_<sur>plus_v2di): Likewise.
18141 (reduc_<sur>plus_v2si): Likewise.
18142 (reduc_<sur>plus_v4sf): Likewise.
18143 (aarch64_addpv4sf): Likewise.
18144 * config/aarch64/arm_neon.h
18145 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18146 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18147 add UNSPEC_SADDV, UNSPEC_UADDV.
18149 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18151 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18153 * config/aarch64/arm_neon.h
18154 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18156 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18158 * config/aarch64/aarch64-builtins
18159 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18161 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18163 * config/aarch64/aarch64-simd-builtins.def
18164 (reduc_smax_): New.
18165 (reduc_smin_): Likewise.
18166 (reduc_umax_): Likewise.
18167 (reduc_umin_): Likewise.
18168 (reduc_smax_nan_): Likewise.
18169 (reduc_smin_nan_): Likewise.
18172 (smax): Update for V2SF, V4SF and V2DF modes.
18175 (smin_nan): Likewise.
18176 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18177 (<su><maxmin><mode>3): ...This, refactor.
18178 (s<maxmin><mode>3): New.
18179 (<maxmin_uns><mode>3): Likewise.
18180 (reduc_<maxmin_uns>_<mode>): Refactor.
18181 (reduc_<maxmin_uns>_v4sf): Likewise.
18182 (reduc_<maxmin_uns>_v2si): Likewise.
18183 (aarch64_<fmaxmin><mode>: Remove.
18184 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18186 (vmin<q>_f<32,64>): Likewise.
18187 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18189 (su): Add mappings for smax, smin, umax, umin.
18191 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18192 (FMAXMIN): Rename as...
18193 (FMAXMIN_UNS): ...This.
18195 (fmaxminv): Likewise.
18196 (fmaxmin): Likewise.
18198 (maxmin_uns_op): Likewise.
18200 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18202 * config/aarch64/arm_neon.h
18203 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18204 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18205 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18207 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18209 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18210 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18212 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18214 * config/aarch64/aarch64-simd.md
18215 (vcond<mode>_internal): Handle special cases for constant masks.
18216 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18217 (vcondu<mode><mode>): Likewise.
18218 (vcond<v_cmp_result><mode>): New.
18220 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18222 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18223 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18224 * config/aarch64/aarch64-simd-builtins.def
18225 (cmeq): Update to BUILTIN_VALLDI.
18230 * config/aarch64/arm_neon.h
18231 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18232 to builtins or C as appropriate.
18234 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18236 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18238 (cmhi): Rename to...
18240 * config/aarch64/aarch64-simd.md
18241 (simd_mode): Add SF.
18242 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18243 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18244 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18245 (cstore<mode>_neg): ...This.
18246 * config/aarch64/iterators.md
18248 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18249 (COMPARISONS): New.
18250 (UCOMPARISONS): Likewise.
18251 (optab): Add missing comparisons.
18257 (VCMP_S): Likewise.
18258 (VCMP_U): Likewise.
18259 (V_cmp_result): Add DF, SF modes.
18260 (v_cmp_result): Likewise.
18262 (vmtype): Likewise.
18263 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18265 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18267 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18268 define_insn to define_insn_and_split.
18269 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18270 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18271 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18272 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18273 (thumb2_negscc): Likewise.
18275 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18277 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18279 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18281 * config/arm/thumb2.md: Remove trailing whitespaces.
18283 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18285 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18286 Use gen_int_mode rather than GEN_INT.
18288 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18290 * value-prof.c (stream_in_histogram_value): Remove the strayed
18293 2013-04-30 Richard Biener <rguenther@suse.de>
18295 PR middle-end/57122
18296 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18298 2013-04-30 Richard Biener <rguenther@suse.de>
18300 PR middle-end/57107
18301 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18303 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18305 PR rtl-optimization/56957
18306 PR rtl-optimization/57105
18307 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18308 variable. Use just INSN_UID for determining whether an insn
18309 should be only disconnected from the insn stream.
18310 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18312 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18314 PR tree-optimization/57104
18315 * tsan.c (instrument_expr): Don't instrument accesses to
18316 DECL_HARD_REGISTER VAR_DECLs.
18318 2013-04-30 Richard Biener <rguenther@suse.de>
18320 * function.h (loops_for_fn): New inline function.
18321 (set_loops_for_fn): Likewise.
18322 * cfgloop.h (place_new_loop): Add struct function parameter.
18323 (get_loop): Likewise.
18324 (get_loops): Likewise.
18325 (number_of_loops): Likewise.
18326 (fel_next): Adjust.
18327 (fel_init): Likewise.
18328 * cfg.c (get_loop_copy): Adjust.
18329 * cfgloop.c (flow_loops_dump): Likewise.
18330 (record_loop_exits): Likewise.
18331 (verify_loop_structure): Likewise.
18332 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18333 (estimate_reg_pressure_cost): Likewise.
18334 (mark_loop_exit_edges): Likewise.
18335 * cfgloopmanip.c (place_new_loop): Likewise.
18336 (add_loop): Likewise.
18337 (duplicate_loop): Likewise.
18338 * graph.c (draw_cfg_nodes): Likewise.
18339 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18340 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18341 (extract_affine_chrec): Likewise.
18342 (build_scop_iteration_domain): Likewise.
18343 * graphite.c (graphite_initialize): Likewise.
18344 * ira-build.c (create_loop_tree_nodes): Likewise.
18345 (more_one_region_p): Likewise.
18346 (rebuild_regno_allocno_maps): Likewise.
18347 (mark_loops_for_removal): Likewise.
18348 (mark_all_loops_for_removal): Likewise.
18349 (remove_unnecessary_regions): Likewise.
18350 (ira_build): Likewise.
18351 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18352 * loop-init.c (fix_loop_structure): Likewise.
18353 (gate_rtl_move_loop_invariants): Likewise.
18354 (gate_rtl_unswitch): Likewise.
18355 (gate_rtl_unroll_and_peel_loops): Likewise.
18356 (rtl_doloop): Likewise.
18357 * lto-streamer-in.c (input_cfg): Likewise.
18358 * lto-streamer-out.c (output_cfg): Likewise.
18359 * modulo-sched.c (sms_schedule): Likewise.
18360 * predict.c (tree_estimate_probability): Likewise.
18361 (tree_estimate_probability_driver): Likewise.
18362 (estimate_loops): Likewise.
18363 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18364 (move_sese_region_to_fn): Likewise.
18365 (debug_loop_num): Likewise.
18366 * tree-chrec.c (chrec_evaluate): Likewise.
18367 (hide_evolution_in_other_loops_than_loop): Likewise.
18368 (chrec_component_in_loop_num): Likewise.
18369 (reset_evolution_in_loop): Likewise.
18370 (evolution_function_is_invariant_rec_p): Likewise.
18371 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18372 * tree-inline.c (copy_loops): Likewise.
18373 (copy_cfg_body): Likewise.
18374 (tree_function_versioning): Likewise.
18375 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18376 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18378 (add_to_evolution_1): Likewise.
18379 (scev_const_prop): Likewise.
18380 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18381 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18382 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18383 (tree_ssa_lim_initialize): Likewise.
18384 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18385 (verify_loop_closed_ssa): Likewise.
18386 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18387 (tree_ssa_loop_im): Likewise.
18388 (tree_ssa_loop_unswitch): Likewise.
18389 (tree_vectorize): Likewise.
18390 (check_data_deps): Likewise.
18391 (tree_ssa_loop_ivcanon): Likewise.
18392 (tree_ssa_loop_bounds): Likewise.
18393 (tree_complete_unroll): Likewise.
18394 (tree_complete_unroll_inner): Likewise.
18395 (tree_parallelize_loops): Likewise.
18396 (tree_ssa_loop_prefetch): Likewise.
18397 (tree_ssa_loop_ivopts): Likewise.
18398 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18399 * tree-vectorizer.c (vectorize_loops): Likewise.
18401 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18403 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18404 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18405 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18406 with EABI_LINK_SPEC.
18408 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18411 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18414 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18417 * lra-constraints.c (process_alt_operands): Discourage a bit more
18418 using memory for pseudos. Print cost dump for alternatives.
18419 Modify cost values for conflicts with early clobbers.
18420 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18422 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18425 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18427 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18429 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18430 from/to S register.
18431 (movdi_aarch64): Support LDR/STR from/to D register.
18433 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18435 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18436 or higher by default.
18438 2013-04-29 Richard Biener <rguenther@suse.de>
18440 PR middle-end/57075
18441 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18442 even if not adding abnormal edges for calls that can make
18445 2013-04-29 Richard Biener <rguenther@suse.de>
18447 PR middle-end/57103
18448 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18450 (move_stmt_r): Remove redundant checking.
18452 2013-04-29 Teresa Johnson <tejohnson@google.com>
18455 * basic-block.h (apply_scale): New function.
18456 (apply_probability): Use apply_scale.
18457 * gimple-streamer-in.c (input_bb): Ditto.
18458 * lto-streamer-in.c (input_cfg): Ditto.
18459 * lto-cgraph.c (merge_profile_summaries): Ditto.
18460 * tree-optimize.c (execute_fixup_cfg): Ditto.
18461 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18462 (copy_edges_for_bb): Ditto.
18463 (copy_cfg_body): Ditto.
18465 2013-04-29 Tom de Vries <tom@codesourcery.com>
18467 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18468 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18469 (tail_merge_optimize): Handle current_loops == NULL.
18471 2013-04-26 Jeff Law <law@redhat.com>
18473 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18474 (simplify_cond_using_ranges): Generalize code to simplify
18475 COND_EXPRs where one argument is a constant and the other
18476 is an SSA_NAME created by an integral type conversion.
18478 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18480 * config/arm/arm.md (store_minmaxsi): Use only when
18481 optimize_insn_for_size_p.
18483 2013-04-29 Christian Bruel <christian.bruel@st.com>
18486 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18488 2013-04-29 Richard Biener <rguenther@suse.de>
18490 PR middle-end/57089
18491 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18492 loop tree make sure to schedule a fixup for the child as well.
18493 (expand_omp_for_generic): Properly add loops.
18494 (expand_omp_for_static_nochunk): Likewise.
18495 (expand_omp_for_static_chunk): Likewise.
18496 (expand_omp_for): For the degenerate case fixup loops.
18497 (expand_omp_sections): Fix default bb placement in loops.
18498 (expand_omp_atomic_pipeline): Properly add loops.
18500 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18502 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18504 2013-04-29 Tom de Vries <tom@codesourcery.com>
18506 * tree-ssa-tail-merge.c: Update header comment.
18508 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18510 * config/aarch64/arm_neon.h
18511 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18512 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18513 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18514 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18515 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18516 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18517 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18518 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18520 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18522 * config/aarch64/aarch64-simd.md
18523 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18524 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18525 fix_trunc, fixuns_trunc.
18526 (ftrunc<VDQF:mode>2): New.
18527 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18528 (fix_trunc_optab): New.
18530 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18532 * config/aarch64/aarch64-builtins.c
18533 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18534 iceilf, lround, iroundf.
18536 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18539 * config/i386/i386.h (enum ix86_tune_indices)
18540 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18541 New, split from X86_TUNE_INTER_UNIT_MOVES.
18542 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18543 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18544 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18545 (TARGET_INTER_UNIT_MOVES): Remove.
18546 * config/i386/i386.c (initial_ix86_tune_features): Update.
18547 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18548 (ix86_expand_convert_uns_didf_sse): Use
18549 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18550 (ix86_expand_vector_init_one_nonzero): Ditto.
18551 (ix86_expand_vector_init_interleave): Ditto.
18552 (inline_secondary_memory_needed): Return true for moves from SSE class
18553 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18554 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18555 * config/i386/constraints.md (Yi, Ym): Depend on
18556 TARGET_INTER_UNIT_MOVES_TO_VEC.
18557 (Yj, Yn): New constraints.
18558 * config/i386/i386.md (*movdi_internal): Change constraints of
18559 operand 1 from Yi to Yj and from Ym to Yn.
18560 (*movsi_internal): Ditto.
18561 (*movdf_internal): Ditto.
18562 (*movsf_internal): Ditto.
18563 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18564 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18565 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18566 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18567 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18568 * config/i386/sse.md (movdi_to_sse): Ditto.
18569 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18570 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18571 TARGET_INTER_UNIT_MOVES.
18572 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18573 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18574 instead of TARGET_INTER_UNIT_MOVES.
18575 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18576 operand 1 from Yi to Yj and from Ym to Yn.
18578 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18580 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18581 (float_truncate_hi_): Likewise.
18582 (float_extend_lo_): Likewise.
18583 (float_truncate_lo_): Likewise.
18584 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18585 (aarch64_float_extend_lo_v2df): Likewise.
18586 (vec_unpacks_hi_v4sf): Likewise.
18587 (aarch64_float_truncate_lo_v2sf): Likewise.
18588 (aarch64_float_truncate_hi_v4sf): Likewise.
18589 (vec_pack_trunc_v2df): Likewise.
18590 (vec_pack_trunc_df): Likewise.
18592 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18594 * config/aarch64/aarch64-builtins.c
18595 (aarch64_fold_builtin): Fold float conversions.
18596 * config/aarch64/aarch64-simd-builtins.def
18597 (floatv2si, floatv4si, floatv2di): New.
18598 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18599 * config/aarch64/aarch64-simd.md
18600 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18601 * config/aarch64/iterators.md (FLOATUORS): New.
18602 (optab): Add float, floatuns.
18603 (su_optab): Likewise.
18605 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18607 * config/aarch64/aarch64-builtins.c
18608 (aarch64_builtin_vectorized_function): Use new names for
18610 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18611 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18612 (fcvtzu): Split as...
18613 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18614 (fcvtas): Split as...
18615 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18616 (fcvtau): Split as...
18617 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18618 (fcvtps): Split as...
18619 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18620 (fcvtpu): Split as...
18621 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18622 (fcvtms): Split as...
18623 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18624 (fcvtmu): Split as...
18625 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18626 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18627 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18628 (lfrintnusf, lfrintnudf): Likewise.
18629 * config/aarch64/aarch64-simd.md
18630 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18632 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18633 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18634 (fcvt_pattern): Likewise.
18636 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18638 * config/aarch64/aarch64-simd.md
18639 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18640 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18642 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18644 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18645 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18646 (vrnd<a,m,n,p>_f32): Implement using builtins.
18647 (vrnd<i,x><q>_f<32, 64>): New.
18649 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18651 * config/aarch64/aarch64-builtins.c
18652 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18653 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18654 (frintz): Rename to...
18656 (frintp): Rename to...
18658 (frintm): Rename to...
18660 (frinti): Rename to...
18661 (nearbyint): ...this.
18662 (frintx): Rename to...
18664 (frinta): Rename to...
18666 * config/aarch64/aarch64-simd.md
18667 (aarch64_frint<frint_suffix><mode>): Delete.
18668 (<frint_pattern><mode>2): Convert to insn.
18669 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18670 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18671 (frint_pattern): Likewise.
18672 (frint_suffix): Likewise.
18674 2013-04-29 Richard Biener <rguenther@suse.de>
18676 PR tree-optimization/57081
18677 * loop-init.c: Include tree-flow.h.
18678 (loop_optimizer_finalize): Free number of iteration estimates.
18679 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18681 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18683 PR tree-optimization/57083
18684 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18685 non-singleton shift count range, zero extend low_bound for uns case.
18687 * config/i386/predicates.md (general_vector_operand): New predicate.
18688 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18689 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18690 if they aren't nonimmediate operands. If their original values
18691 satisfy const_vector_equal_evenodd_p, don't shift them.
18692 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18693 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18694 if not nonimmediate_operand.
18695 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18696 instead of register_operand.
18697 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18699 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18701 * stor-layout.c (finalize_size_functions): Allocate a structure and
18702 reset cfun before dumping the functions.
18704 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18706 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18709 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18710 use xop_pmacsdqh if uns_p.
18711 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18712 the immediate rotate count.
18714 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18716 * rtl.h (struct rtx_def): Add comment for field jump.
18717 (LRA_SUBREG_P): New macro.
18718 * recog.c (register_operand): Check LRA_SUBREG_P.
18719 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18721 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18722 (lra_final_code_change): Skip subreg change for operators.
18723 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18724 if there are no operand changes.
18725 * lra-constraints.c (curr_insn_set): New.
18726 (match_reload): Set LRA_SUBREG_P.
18727 (emit_spill_move): Ditto.
18728 (check_and_process_move): Use curr_insn_set. Process only single
18729 set insns. Don't initialize sec_mem_p and change_p.
18730 (simplify_operand_subreg): Use LRA_SUBREG_P.
18731 (reg_in_class_p): New function.
18732 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18733 of #ifdef. Add code to remove cycling.
18734 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18735 non-null disp. Reload inner instead of disp when base and index
18736 are null. Try to put lo_sum into register.
18737 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18738 (check_and_process_move): Move code for move cost check to
18739 simple_move_p. Remove equiv_substitution.
18740 (simple_move_p): New function.
18741 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18742 curr_insn_set. Call check_and_process_move only for single set
18743 insns. Use the new function. Move call of check_and_process_move
18744 after operand equiv substitution and address process.
18746 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18749 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18750 with nonlocal goto receivers or returns twice calls, ignore
18751 unininitialized values from abnormal edges to nl goto receiver
18752 or returns twice call.
18754 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18756 PR tree-optimization/57051
18757 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18758 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18761 2013-04-26 Richard Biener <rguenther@suse.de>
18763 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18764 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18765 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18766 (fixup_loop_arrays_after_move): New function.
18767 (move_sese_region_to_fn): Properly outline the loop tree parts
18768 of the SESE region.
18770 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18772 * config/i386/i386.md (type, unit): Fix long lines.
18774 2013-04-26 Richard Biener <rguenther@suse.de>
18776 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18777 (lto-streamer-out.o): Likewise.
18778 * cfgloop.c (init_loops_structure): Export, add struct function
18779 argument and adjust.
18780 (flow_loops_find): Adjust.
18781 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18782 (init_loops_structure): Declare.
18783 * lto-streamer-in.c: Include cfgloop.h.
18784 (input_cfg): Input the loop tree.
18785 * lto-streamer-out.c: Include cfgloop.h.
18786 (output_cfg): Output the loop tree.
18787 (output_struct_function_base): Do not drop PROP_loops.
18789 2013-03-26 Richard Biener <rguenther@suse.de>
18791 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18792 (pass_build_cfg): Provide PROP_loops.
18793 (move_sese_region_to_fn): Remove loops that are outlined into fn
18795 * tree-inline.c: Include cfgloop.h.
18796 (initialize_cfun): Do not drop PROP_loops.
18797 (copy_loops): New function.
18798 (copy_cfg_body): Copy loop structure.
18799 (tree_function_versioning): Initialize destination loop tree.
18800 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18801 (pass_parallelize_loops): Do IL verification.
18802 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18803 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18804 the CFG make sure we fixup loops as well.
18805 * tree-ssa-tail-merge.c: Include cfgloop.h.
18806 (replace_block_by): When merging loop latches mark loops for fixup.
18807 * lto-streamer-out.c (output_struct_function_base): Drop
18808 PROP_loops for now.
18809 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18810 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18811 * ipa-split.c: Include cfgloop.h.
18812 (split_function): Add the new return block to the loop tree root.
18813 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18814 whether we have removed the forwarder block.
18815 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18816 * cfgloop.h (place_new_loop): Declare.
18817 * cfgloopmanip.c (place_new_loop): Export.
18818 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18819 (tree-switch-conversion.o): Likewise.
18820 (tree-complex.o): Likewise.
18821 (tree-inline.o): Likewise.
18822 (tree-ssa-tailmerge.o): Likewise.
18823 (ipa-split.o): Likewise.
18824 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18825 (tree-ssa-copy.o): Likewise.
18826 * tree-switch-conversion.c: Include cfgloop.h
18827 (process_switch): If we emit a bit-test cascade, schedule loops
18829 * tree-complex.c: Include cfgloop.h.
18830 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18831 * asan.c: Include cfgloop.h.
18832 (create_cond_insert_point): Properly add new basic-blocks to
18833 loops, schedule loop fixup.
18834 * cfgloop.c (verify_loop_structure): Check that looks are not
18836 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18838 (expand_omp_for_generic): Likewise.
18839 (expand_omp_sections): Likewise.
18840 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18841 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18842 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18843 is initialized, not when loops are present.
18844 * tree-parloops.c (parallelize_loops): Remove checking here.
18845 * passes.c (init_optimization_passes): Schedule a copy-propagation
18846 pass before complete unrolling of inner loops.
18848 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18850 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18851 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18852 (colorize_init): Add argument to _WIN32 version.
18853 * toplev.c: Include diagnostic-color.h.
18854 (process_options): Default to -fdiagnostics-color=auto if
18855 GCC_COLORS env var is in the environment.
18856 * common.opt (fdiagnostics-color=): Add Var and Init.
18857 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18858 env var is in the environment, the default is auto rather than never.
18860 * diagnostic.h (file_name_as_prefix): Add context argument.
18861 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18862 the string as locus.
18863 * langhooks.c (lhd_print_error_function): Adjust caller.
18865 2013-04-25 Lawrence Crowl <crowl@google.com>
18867 * var-tracking.c (shared_hash_def::htab):
18868 Change type to hash_table. Update dependent calls and types.
18870 2013-04-25 Lawrence Crowl <crowl@google.com>
18872 * Makefile.in: Update as needed below.
18874 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18875 Move declaration to after the type's method definitons.
18877 * attribs.c (htab_t scoped_attributes::attribute_hash):
18878 Change type to hash_table. Update dependent calls and types.
18880 * bitmap.c (htab_t bitmap_desc_hash):
18881 Change type to hash_table. Update dependent calls and types.
18883 * cselib.c (htab_t cselib_hash_table):
18884 Change type to hash_table. Update dependent calls and types.
18886 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18887 (hash_string_slot_node): Move implementation into lto-streamer.h
18888 struct string_slot_hasher.
18889 (eq_string_slot_node): Likewise.
18891 * data-streamer-out.c: Update output_block::string_hash_table
18892 dependent calls and types.
18894 * dwarf2cfi.c (htab_t trace_index):
18895 Change type to hash_table. Update dependent calls and types.
18897 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18898 Change type to hash_table. Update dependent calls and types.
18899 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18900 (htab_t optimize_external_refs::map): Likewise.
18901 (htab_t output_comp_unit::extern_map): Likewise.
18902 (htab_t output_comdat_type_unit::extern_map): Likewise.
18903 (htab_t output_macinfo::macinfo_htab): Likewise.
18904 (htab_t optimize_location_lists::htab): Likewise.
18905 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18907 * except.c (htab_t ehspec_hash_type):
18908 Change type to hash_table. Update dependent calls and types.
18909 (assign_filter_values::ttypes): Likewise.
18910 (assign_filter_values::ehspec): Likewise.
18911 (sjlj_assign_call_site_values::ar_hash): Likewise.
18912 (convert_to_eh_region_ranges::ar_hash): Likewise.
18914 * gcse.c (htab_t pre_ldst_table):
18915 Change type to hash_table. Update dependent calls and types.
18917 * ggc-common.c (htab_t saving_htab):
18918 Change type to hash_table. Update dependent calls and types.
18919 (htab_t loc_hash): Likewise.
18920 (htab_t ptr_hash): Likewise.
18921 (call_count): Rename ggc_call_count.
18922 (call_alloc): Rename ggc_call_alloc.
18923 (loc_descriptor): Rename make_loc_descriptor.
18924 (add_statistics): Rename ggc_add_statistics.
18926 * ggc-common.c (saving_htab):
18927 Change type to hash_table. Update dependent calls and types.
18929 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18930 (push_gimplify_context): Likewise.
18931 (pop_gimplify_context): Likewise.
18932 (struct gimple_temp_hash_elt): Added.
18933 (struct gimplify_hasher): Likewise.
18934 (struct gimplify_ctx.temp_htab):
18935 Change type to hash_table. Update dependent calls and types.
18937 * gimple-fold.c: Include gimplify-ctx.h.
18939 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18940 Change type to hash_table. Update dependent calls and types.
18941 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18942 avoid potential global name collision.
18944 * gimplify.c: Include gimplify-ctx.h.
18945 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18946 (htab_t gimplify_ctx::temp_htab):
18947 Update dependent calls and types for new type hash_table.
18948 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18949 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18951 * gimplify-ctx.h: New.
18952 (struct gimple_temp_hash_elt): Move from gimplify.c.
18953 (class gimplify_hasher): New.
18954 (struct gimplify_ctx): Move from gimple.h.
18955 (htab_t gimplify_ctx::temp_htab):
18956 Change type to hash_table. Update dependent calls and types.
18958 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18959 (htab_t ivs_params::newivs_index):
18960 Change type to hash_table. Update dependent calls and types.
18961 (htab_t ivs_params::params_index): Likewise.
18962 (htab_t print_generated_program::params_index): Likewise.
18963 (htab_t gloog::newivs_index): Likewise.
18964 (htab_t gloog::params_index): Likewise.
18966 * graphite.c: Include graphite-htab.h.
18967 4htab_t graphite_transform_loops::bb_pbb_mapping):
18968 Change type to hash_table. Update dependent calls and types.
18970 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18971 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18972 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18974 * graphite-dependences.c: Include graphite-htab.h.
18975 (loop_is_parallel_p): Change hash table type of parameter.
18977 * graphite-htab.h: New.
18978 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18979 (extern find_pbb_via_hash): Move from graphite-poly.h.
18980 (extern loop_is_parallel_p): Move from graphite-poly.h.
18981 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18983 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18984 (extern loop_is_parallel_p): Move to graphite-htab.h.
18985 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18987 * haifa-sched.c (htab_t delay_htab):
18988 Change type to hash_table. Update dependent calls and types.
18989 (htab_t delay_htab_i2): Likewise.
18991 * ira-color.c (htab_t allocno_hard_regs_htab):
18992 Change type to hash_table. Update dependent calls and types.
18994 * ira-costs.c (htab_t cost_classes_htab):
18995 Change type to hash_table. Update dependent calls and types.
18997 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18998 Change type to hash_table. Update dependent calls and types.
19000 * loop-iv.c (htab_t bivs):
19001 Change type to hash_table. Update dependent calls and types.
19003 * loop-unroll.c (htab_t opt_info::insns_to_split):
19004 Change type to hash_table. Update dependent calls and types.
19005 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19007 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19008 (struct string_slot_hasher): New.
19009 (htab_t output_block::string_hash_table):
19010 Change type to hash_table. Update dependent calls and types.
19012 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19013 (htab_t file_name_hash_table):
19014 Change type to hash_table. Update dependent calls and types.
19016 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19019 * lto-streamer.c (htab_t tree_htab):
19020 Change type to hash_table. Update dependent calls and types.
19022 * omp-low.c: Include gimplify-ctx.h.
19024 * passes.c (htab_t name_to_pass_map):
19025 Change type to hash_table. Update dependent calls and types.
19026 (pass_traverse): Rename to passes_pass_traverse.
19028 * plugin.c (htab_t event_tab):
19029 Change type to hash_table. Update dependent calls and types.
19031 * postreload-gcse.c (htab_t expr_table):
19032 Change type to hash_table. Update dependent calls and types.
19033 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19035 * sese.c (debug_rename_map_1): Make extern.
19036 (htab_t copy_bb_and_scalar_dependences::rename_map):
19037 Change type to hash_table. Update dependent calls and types.
19039 * sese.h (extern debug_rename_map): Move to .c file.
19041 * store-motion.c (htab_t store_motion_mems_table):
19042 Change type to hash_table. Update dependent calls and types.
19044 * trans-mem.c (htab_t tm_new_mem_hash):
19045 Change type to hash_table. Update dependent calls and types.
19047 * tree-browser.c (htab_t TB_up_ht):
19048 Change type to hash_table. Update dependent calls and types.
19050 * tree-cfg.c (htab_t discriminator_per_locus):
19051 Change type to hash_table. Update dependent calls and types.
19053 * tree-complex.c: Include tree-hasher.h
19054 (htab_t complex_variable_components):
19055 Change type to hash_table. Update dependent calls and types.
19057 * tree-eh.c (htab_t finally_tree):
19058 Change type to hash_table. Update dependent calls and types.
19060 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19061 struct int_tree_hasher.
19062 (extern int_tree_map_eq): Likewise.
19063 (uid_decl_map_hash): Removed.
19064 (extern decl_tree_map_eq): Likewise.
19066 * tree-hasher.h: New.
19067 (struct int_tree_hasher): New.
19068 (typedef int_tree_htab_type): New.
19070 * tree-inline.c: Include gimplify-ctx.h.
19072 * tree-mudflap.c: Include gimplify-ctx.h.
19074 * tree-parloops.c: Include tree-hasher.h.
19075 (htab_t eliminate_local_variables_stmt::decl_address):
19076 Change type to hash_table. Update dependent calls and types.
19077 (htab_t separate_decls_in_region::decl_copies): Likewise.
19079 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19080 Change type to hash_table. Update dependent calls and types.
19082 * tree-sra.c (candidates):
19083 Change type to hash_table. Update dependent calls and types.
19085 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19087 (int_tree_map_hash): Likewise.
19089 * tree-ssa-dom.c (htab_t avail_exprs):
19090 Change type to hash_table. Update dependent calls and types.
19092 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19093 Change type to hash_table. Update dependent calls and types.
19095 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19096 Change type to hash_table. Update dependent calls and types.
19098 * tree-ssa-phiopt.c (seen_ssa_names):
19099 Change type to hash_table. Update dependent calls and types.
19101 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19102 Change type to hash_table. Update dependent calls and types.
19104 * tree-ssa-uncprop.c (equiv):
19105 Change type to hash_table. Update dependent calls and types.
19107 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19109 PR rtl-optimization/57003
19110 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19111 call note_stores with kill_clobbered_value callback again after
19112 killing regs_invalidated_by_call.
19114 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19116 * config/aarch64/aarch64-simd.md
19117 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19118 (aarch64_simd_bsl<mode>): Likewise.
19119 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19121 2013-04-25 Marek Polacek <polacek@redhat.com>
19123 PR tree-optimization/57066
19124 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19126 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19128 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19130 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19132 * config/aarch64/aarch64-builtins.c
19133 (aarch64_fold_builtin): New.
19134 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19135 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19136 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19137 * config/aarch64/arm_neon.h
19138 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19140 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19141 Tejas Belagod <tejas.belagod@arm.com>
19143 * config/aarch64/aarch64-builtins.c
19144 (aarch64_gimple_fold_builtin): New.
19145 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19146 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19147 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19148 (addvv4sf): Update.
19149 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19151 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19153 * config/aarch64/aarch64.md
19154 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19156 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19158 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19159 (*ngcsi_uxtw): New pattern.
19161 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19162 Julian Brown <julian@codesourcery.com>
19164 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19165 (TB_DREG): Add T_V4HF.
19166 (v4hf_UP): New macro.
19167 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19168 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19169 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19171 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19172 (arm_vector_mode_supported_p): Handle V4HF.
19173 (arm_mangle_map): Handle V4HFmode.
19174 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19175 * config/arm/arm_neon_builtins.def: Add entries for
19176 vcvtv4hfv4sf, vcvtv4sfv4hf.
19177 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19178 (neon_vcvtv4hfv4sf): Likewise.
19179 * config/arm/neon-gen.ml: Handle half-precision floating point
19181 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19182 * config/arm/arm_neon.h: Regenerate.
19183 * config/arm/neon.ml (type elts): Add F16.
19184 (type vectype): Add T_float16x4, T_floatHF.
19185 (type vecmode): Add V4HF.
19186 (type features): Add Requires_FP_bit feature.
19187 (elt_width): Handle F16.
19188 (elt_class): Likewise.
19189 (elt_of_class_width): Likewise.
19190 (mode_of_elt): Refactor.
19191 (type_for_elt): Handle F16, fix error messages.
19192 (vectype_size): Handle T_float16x4.
19193 (vcvt_sh): New function.
19194 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19195 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19196 (string_of_mode): Handle V4HF.
19197 * doc/arm-neon-intrinsics.texi: Regenerate.
19199 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19201 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19202 format specifier in 'X' case.
19204 2013-04-25 Alan Modra <amodra@gmail.com>
19207 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19208 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19209 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19210 Repeat for many other rotate/shift and mask patterns using subregs.
19211 Name lshiftrt insns.
19212 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19213 on WORDS_BIG_ENDIAN.
19215 2013-04-25 Alan Modra <amodra@gmail.com>
19217 * config.gcc: Support little-endian powerpc-linux targets.
19218 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19219 (LINK_OS_LINUX_SPEC): Define.
19220 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19221 Preserve MASK_LITTLE_ENDIAN.
19222 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19223 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19224 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19225 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19226 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19227 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19228 Correct fp word order for little-endian. Don't shift toc entries
19229 smaller than a word for little-endian.
19230 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19231 (bswapdi2 splits): Correct low-part subreg for little-endian.
19232 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19233 low/high where such is correct only for be.
19234 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19235 little-endian for -mcall-aixdesc.
19237 2013-04-25 Alan Modra <amodra@gmail.com>
19239 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19240 replace_equiv_address_nv.
19242 2013-04-25 Alan Modra <amodra@gmail.com>
19244 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19246 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19249 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19250 * rtl.h (struct rtx_def): ...
19252 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19254 PR rtl-optimizations/57046
19255 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19256 for multi-reg splits.
19258 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19260 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19262 2013-04-24 Sterling Augustine <saugustine@google.com>
19264 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19265 (comp_dir_string, debug_str_dwo_section): New.
19266 (DEBUG_STR_DWO_SECTION): Rename to ...
19267 (DEBUG_DWO_STR_SECTION): ... this.
19268 (DEBUG_NORM_STR_SECTION): Delete.
19269 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19270 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19271 (find_AT_string): Move most logic to ...
19272 (find_AT_string_in_table): ... here. New.
19273 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19274 add_skeleton_AT_string. Delete logic.
19275 (output_skeleton_debug_sections): Remove call to
19276 add_top_level_skeleton_die_attrs.
19277 (add_comp_dir_attribute): Move logic to comp_dir_string.
19278 (dwarf2out_init): Initialize debug_str_dwo_section.
19279 (output_indirect_string): Call find_string_form.
19280 (output_indirect_strings): Rewrite.
19281 (prune_unused_types): Empty skeleton_debug_str_hash.
19282 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19283 (dwarf2out_finish): Call output_indirect_strings.
19285 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19287 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19289 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19291 * rtl.h (struct rtx_def): Add comment for field jump.
19292 (LRA_SUBREG_P): New macro.
19293 * recog.c (register_operand): Check LRA_SUBREG_P.
19294 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19296 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19297 (lra_final_code_change): Skip subreg change for operators.
19298 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19299 if there are no operand changes.
19300 * lra-constraints.c (curr_insn_set): New.
19301 (match_reload): Set LRA_SUBREG_P.
19302 (emit_spill_move): Ditto.
19303 (check_and_process_move): Use curr_insn_set. Process only single
19304 set insns. Don't initialize sec_mem_p and change_p.
19305 (simplify_operand_subreg): Use LRA_SUBREG_P.
19306 (reg_in_class_p): New function.
19307 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19308 of #ifdef. Add code to remove cycling.
19309 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19310 non-null disp. Reload inner instead of disp when base and index
19311 are null. Try to put lo_sum into register.
19312 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19313 (check_and_process_move): Move code for move cost check to
19314 simple_move_p. Remove equiv_substitution.
19315 (simple_move_p): New function.
19316 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19317 curr_insn_set. Call check_and_process_move only for single set
19318 insns. Use the new function. Move call of check_and_process_move
19319 after operand equiv substitution and address process.
19321 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19323 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19324 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19325 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19327 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19329 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19331 2013-04-24 Marek Polacek <polacek@redhat.com>
19333 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19334 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19335 (select_loops_exit_conditions): Likewise.
19336 (number_of_iterations_for_all_loops): Likewise.
19337 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19338 (scev_analysis): Likewise.
19340 2013-04-02 Catherine Moore <clm@codesourcery.com>
19341 Chao-ying Fu <fu@mips.com>
19343 * config/mips/micromips.md (jraddiusp): New pattern.
19344 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19345 instruction if possible.
19347 2013-04-24 Alan Modra <amodra@gmail.com>
19349 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19351 2013-04-24 Julian Brown <julian@codesourcery.com>
19352 Chung-Lin Tang <cltang@codesourcery.com>
19354 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19355 dependency behavior in enumeration type DIE generation. Add TODO note
19356 to comments about future DW_FORM_sdata/udata re-work of related code.
19358 2013-04-23 Lawrence Crowl <crowl@google.com>
19360 * Makefile.in: Update as needed below.
19362 * hash-table.h (class hash_table):
19363 Correct many methods with parameter types compare_type to the correct
19364 value_type. (Correct code was unlikely to notice the change.)
19365 (hash_table::elements_with_deleted) New.
19366 (class hashtable::iterator): New.
19367 (hashtable::begin()): New.
19368 (hashtable::end()): New.
19369 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19371 * statistics.c (statistics_hashes):
19372 Change type to hash_table. Update dependent calls and types.
19374 * tree-into-ssa.c (var_infos):
19375 Change type to hash_table. Update dependent calls and types.
19377 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19378 Change type to hash_table. Update dependent calls and types.
19380 * tree-ssa-loop-im.c (struct mem_ref.refs):
19381 Change type to hash_table. Update dependent calls and types.
19383 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19384 Change type to hash_table. Update dependent calls and types.
19386 * tree-ssa-sccvn.c (vn_tables_s::nary):
19387 Change type to hash_table. Update dependent calls and types.
19388 (vn_tables_s::phis): Likewise.
19389 (vn_tables_s::references): Likewise.
19391 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19392 (vn_reference_eq): Update parameter and return types.
19394 * tree-ssa-structalias.c (pointer_equiv_class_table):
19395 Change type to hash_table. Update dependent calls and types.
19396 (location_equiv_class_table): Likewise.
19398 * tree-vect-data-refs.c: Consequential changes for making
19399 peeling a hash_table.
19401 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19402 (destroy_loop_vec_info): Dependent hash_table update.
19404 * tree-vectorizer.h (peeling_htab):
19405 Change type to hash_table. Update dependent calls and types.
19407 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19409 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19410 to check the register content is equal or not.
19411 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19412 to assign register content record.
19413 * lra-eliminations.c (update_reg_eliminate): Use
19414 lra_update_reg_val_offset to update register content offset.
19415 * lra-int.h (struct lra_reg): Add offset member.
19416 (lra_reg_val_equal_p): New static inline function.
19417 (lra_update_reg_val_offset): New static inline function.
19418 (lra_assign_reg_val): New static inline function.
19419 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19420 to assign register content record.
19421 (initialize_lra_reg_info_element): Initial offset to zero.
19423 2013-04-23 Catherine Moore <clm@codesourcery.com>
19425 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19426 operands. Record compression.
19428 2013-04-23 Xinliang David Li <davidxl@google.com>
19430 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19432 2013-04-23 Richard Biener <rguenther@suse.de>
19434 PR middle-end/57036
19435 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19436 parameter, only add abnormal goto edges from the copied body
19437 if the call could perform abnormal gotos.
19438 (copy_cfg_body): Adjust.
19440 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19442 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19444 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19446 * coretypes.h (gimple_stmt_iterator): Add struct to make
19449 2013-04-23 Richard Biener <rguenther@suse.de>
19451 PR tree-optimization/57026
19452 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19453 from SSA names occuring in abnormal PHI nodes.
19455 2013-04-22 Andi Kleen <ak@linux.intel.com>
19457 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19459 2013-04-22 Andi Kleen <ak@linux.intel.com>
19461 * lto/lto.c (print_lto_report_1): Declare early.
19462 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19464 2013-04-22 Andi Kleen <ak@linux.intel.com>
19466 * common.opt (-flto-report-wpa): Add.
19467 * doc/invoke.texi (-flto-report-wpa): Add.
19468 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19471 2013-04-22 Xinliang David Li <davidxl@google.com>
19473 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19474 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19475 * Makefile.in: New dependency
19477 David Daney <ddaney.cavm@gmail.com>
19479 * configure.ac (gcc_cv_as_micromips_support): Use the
19480 --fatal-warnings option.
19481 * configure: Regenerate.
19483 2013-04-22 Marek Polacek <polacek@redhat.com>
19486 * tsan.c (instrument_expr): Don't instrument expression
19487 in case its size is zero.
19489 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19493 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19495 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19497 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19499 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19500 (gimple_stmt_iterator): New typedef.
19501 * gimple.h (gimple_stmt_iterator): Rename to...
19502 (gimple_stmt_iterator_d): ... This.
19503 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19504 trees be valid for GIMPLE and GENERIC.
19505 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19506 * gimple-fold.c (gimple_fold_call): Call target hook
19507 gimple_fold_builtin.
19508 * hooks.c (hook_bool_gsiptr_false): New.
19509 * hooks.h (hook_bool_gsiptr_false): New.
19510 * target.def (fold_stmt): New.
19511 * doc/tm.texi: Regenerate.
19513 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19516 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19517 a set sp if no stack realignment.
19519 2013-04-22 Nick Clifton <nickc@redhat.com>
19521 * config.gcc (tilegx-linux): Extend extra_objs rather than
19523 (tilepro-linux): Likewise.
19525 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19527 * config/aarch64/aarch64-builtins.c
19529 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19530 (VAR<1-12>): Add MAP parameter.
19531 (BUILTIN_*): Likewise.
19532 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19533 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19534 (aarch64_ushl_n<mode>): Likewise.
19535 (aarch64_sshr_n<mode>): Likewise.
19536 (aarch64_ushr_n<mode>): Likewise.
19537 (aarch64_<maxmin><mode>): Likewise.
19538 (aarch64_sqrt<mode>): Likewise.
19539 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19540 (vshr<q>_n_*): Likewise.
19542 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19544 * config/aarch64/aarch64-builtins.c
19545 (aarch64_simd_builtin_type_mode): Handle SF types.
19547 (BUILTIN_GPF): Define.
19548 (aarch64_init_simd_builtins): Handle SF types.
19549 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19550 (frecps): Likewise.
19551 (frecpx): Likewise.
19552 * config/aarch64/aarch64-simd.md
19553 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19554 (aarch64_frecpe<mode>): New.
19555 (aarch64_frecps<mode>): Likewise.
19556 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19557 (v8type): Add frecp<esx>.
19558 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19559 (aarch64_frecps<mode>): Likewise.
19560 * config/aarch64/iterators.md (FRECP): New.
19561 (frecp_suffix): Likewise.
19562 * config/aarch64/arm_neon.h
19563 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19565 2013-04-22 Christian Bruel <christian.bruel@st.com>
19568 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19569 (REG_CLASS_NAMES): Idem.
19570 (REG_CLASS_CONTENTS): Idem.
19571 (REGCLASS_HAS_FP_REG): Idem.
19572 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19573 (sh_conditional_register_usage): Idem.
19575 2013-04-21 Jeff Law <law@redhat.com>
19577 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19578 (ssa_forward_propagate_and_combine): Use it.
19580 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19582 * lra.c: Update the flow chart diagram.
19584 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19586 PR rtl-optimization/56847
19587 * lra-constraints.c (process_alt_operands): Discourage alternative
19588 with non-matche doffsettable memory constraint fro memory with
19591 2013-04-19 Richard Biener <rguenther@suse.de>
19593 PR tree-optimization/56982
19594 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19596 * gimplify.c (gimplify_call_expr): Notice special calls.
19597 (gimplify_modify_expr): Likewise.
19598 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19599 abnormal control flow receivers.
19600 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19601 in the same way as cfun->has_nonlocal_labels.
19602 (gimple_purge_dead_abnormal_call_edges): Likewise.
19603 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19604 receivers start a basic-block.
19606 2013-04-19 Richard Biener <rguenther@suse.de>
19608 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19610 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19611 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19612 (SLP_TREE_LOAD_PERMUTATION): Add.
19613 (vect_transform_slp_perm_load): Adjust prototype.
19614 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19615 (vect_free_slp_instance): Likewise.
19616 (vect_create_new_slp_node): Likewise.
19617 (vect_supported_slp_permutation_p): Remove.
19618 (vect_slp_rearrange_stmts): Adjust.
19619 (vect_supported_load_permutation_p): Likewise. Inline
19620 vect_supported_slp_permutation_p here.
19621 (vect_analyze_slp_instance): Compute load permutations per
19622 slp node instead of per instance.
19623 (vect_get_slp_defs): Adjust.
19624 (vect_transform_slp_perm_load): Likewise.
19625 (vect_schedule_slp_instance): Remove redundant code.
19626 (vect_schedule_slp): Remove hack for PR56270, add it ...
19627 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19628 CSE loads for SLP. Adjust.
19630 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19632 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19633 spelling in two comments.
19635 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19638 * config/arm/arm.c (load_multiple_sequence): Require SP
19639 as base register for loads if SP is in the register list.
19641 2013-04-19 Martin Jambor <mjambor@suse.cz>
19643 PR tree-optimization/56718
19644 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19645 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19646 and made public. Adjusted all callers.
19647 (ipa_intraprocedural_devirtualization): New function.
19648 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19649 (ipa_intraprocedural_devirtualization): Likewise.
19650 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19652 2013-04-19 Richard Biener <rguenther@suse.de>
19654 PR tree-optimization/57000
19655 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19657 2013-04-19 Terry Guo <terry.guo@arm.com>
19659 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19661 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19662 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19663 (cortex_m4_fmacs): Use new reservations.
19664 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19666 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19668 PR rtl-optimization/56999
19669 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19671 (lra_coalesce): Remove split_origin_bitmap and related code.
19672 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19673 ranges if necessary.
19675 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19677 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19679 (ix86_expand_call): Remove clobbered_registers array and use
19680 x86_64_ms_sysv_extra_clobbered_registers instead.
19681 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19683 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19685 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19686 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19687 (*call_value_rex64_ms_sysv): Ditto.
19689 2013-04-18 Cary Coutant <ccoutant@google.com>
19691 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19692 parent instead of die_mark.
19694 2013-04-18 Diego Novillo <dnovillo@google.com>
19696 * gimple.c (create_gimple_tmp): New.
19697 (get_expr_type): New.
19698 (build_assign): New.
19699 (build_type_cast): New.
19700 * gimple.h (enum ssa_mode): Define.
19701 (gimple_seq_set_location): New.
19702 * asan.c (build_check_stmt): Change some gimple_build_* calls
19703 to use build_assign and build_type_cast.
19705 2013-04-18 Richard Biener <rguenther@suse.de>
19707 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19708 handle negative step. Remove redundant checks.
19709 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19710 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19711 for negative step and grouped loads fail to vectorize.
19713 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19715 * emit-rtl.c (reset_insn_used_flags): New function.
19716 (reset_all_used_flags): Use it.
19717 (verify_insn_sharing): New function.
19718 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19720 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19722 PR tree-optimization/56984
19723 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19724 and (x >> M) >= N don't register any assertion if N << M is the
19727 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19729 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19730 do not delete_insn insns that have not yet been emitted, only
19731 unlink them with remove_insn.
19732 * df-scan.c (df_insn_delete): Revert r197492.
19734 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19736 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19737 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19739 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19741 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19742 define_insn_and_split.
19743 (and_scc,ior_scc,negscc): Likewise.
19744 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19746 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19748 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19749 can benefit from using a sequence of LDRD instructions in epilogue
19750 instead of a single LDM instruction.
19752 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19755 * doc/extend.texi: Fix typo.
19757 2013-04-17 Richard Biener <rguenther@suse.de>
19759 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19760 (vect_build_slp_tree): ... here.
19761 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19762 match. Remove special-casing of mismatched complex loads.
19763 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19764 re-try the match with swapped commutative operands.
19765 (vect_supported_load_permutation_p): Remove special-casing of
19766 mismatched complex loads.
19767 (vect_analyze_slp_instance): Adjust.
19769 2013-04-17 Richard Biener <rguenther@suse.de>
19771 PR rtl-optimization/56921
19772 * cfgloop.h (struct loop): Add simple_loop_desc member.
19773 (struct niter_desc): Mark with GTY(()).
19774 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19775 * loop-iv.c (get_simple_loop_desc): Likewise.
19776 (free_simple_loop_desc): Likewise.
19779 2013-04-16 Richard Biener <rguenther@suse.de>
19781 PR rtl-optimization/56921
19782 * loop-init.c (pass_rtl_move_loop_invariants): Add
19783 TODO_do_not_ggc_collect to todo_flags_finish.
19784 (pass_rtl_unswitch): Same.
19785 (pass_rtl_unroll_and_peel_loops): Same.
19786 (pass_rtl_doloop): Same.
19788 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19790 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19791 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19792 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19794 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19795 * tree-streamer.c (record_common_node): Adjust reference in comment.
19797 2013-04-17 Terry Guo <terry.guo@arm.com>
19799 * config/arm/cortex-m4.md: Add a new bypass.
19801 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19803 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19805 (*subs_<optab><mode>_multp2): New pattern.
19806 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19807 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19809 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19811 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19812 (*subs_mul_imm_<mode>): New pattern.
19814 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19817 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19818 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19819 (vsx_movti_32bit): Change j->wa to O->wa.
19821 2013-04-16 Richard Biener <rguenther@suse.de>
19823 PR rtl-optimization/56921
19824 * loop-init.c (pass_rtl_move_loop_invariants): Add
19825 TODO_do_not_ggc_collect to todo_flags_finish.
19826 (pass_rtl_unswitch): Same.
19827 (pass_rtl_unroll_and_peel_loops): Same.
19828 (pass_rtl_doloop): Same.
19830 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19832 * config/arm/arm.c (emit_multi_reg_push): New declaration
19833 for an existing function.
19834 (arm_emit_strd_push): New function.
19835 (arm_expand_prologue): Used here.
19836 (arm_emit_ldrd_pop): New function.
19837 (arm_expand_epilogue): Used here.
19838 (arm_get_frame_offsets): Update condition.
19839 (arm_emit_multi_reg_pop): Add a special case for load of a single
19840 register with writeback.
19842 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19844 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19847 2013-04-16 Richard Biener <rguenther@suse.de>
19849 PR tree-optimization/56756
19850 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19851 (first_mem_ref_loc): New.
19852 (execute_sm): Place the load temporarily before a previous
19853 access instead of in the latch edge to ensure its SSA dependencies
19854 are defined at points dominating the load.
19856 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19858 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19859 correct fix by moving header and footer insn to the footer of
19860 the merged basic block. Clear BB_END of the merged-away block.
19862 PR middle-end/43631
19863 * emit-rtl.c (make_note_raw): New function.
19864 (link_insn_into_chain): New static inline function.
19865 (add_insn): Use it.
19866 (add_insn_before, add_insn_after): Factor insn chain linking code...
19867 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19868 using link_insn_into_chain.
19869 (note_outside_basic_block_p): New helper function for emit_note_after
19870 and emit_note_before.
19871 (emit_note_after): Use nobb variant of add_insn_after if the note
19872 should not be contained in a basic block.
19873 (emit_note_before): Use nobb variant of add_insn_before if the note
19874 should not be contained in a basic block.
19875 (emit_note_copy): Use make_note_raw.
19876 (emit_note): Likewise.
19877 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19878 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19879 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19880 the moved barrier the tail of the basic block it follows.
19881 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19883 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19885 PR tree-optimization/56962
19886 * gimple-ssa-strength-reduction.c (record_increment): Only set
19887 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19888 either rhs1 or rhs2 is equal to c->base_expr.
19890 2013-04-15 Richard Biener <rguenther@suse.de>
19892 PR tree-optimization/56933
19893 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19895 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19896 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19897 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19898 dependence check ...
19899 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19901 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19902 GROUP_READ_WRITE_DEPENDENCE.
19904 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19906 * emit-rtl.c (reset_all_used_flags): New function.
19907 (verify_rtl_sharing): Call reset_all_used_flags before and after
19908 performing the checks.
19910 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19912 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19913 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19914 * config/arm/constraints.md (De): New constraint.
19915 * config/arm/neon.md (anddi3_neon): Delete.
19916 (neon_vand<mode>): Expand to standard anddi3 pattern.
19917 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19918 Move earlier in the file.
19919 (neon_inv_logic_op2): Likewise.
19920 (arm_anddi_operand_neon): New predicate.
19922 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19924 * configure.ac (gcc_cv_ld_as_needed): Set
19925 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19926 Use -z ignore, -z record on *-*-solaris2*.
19927 (HAVE_LD_AS_NEEDED): Update comment.
19928 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19929 * configure: Regenerate.
19930 * config.in: Regenerate.
19931 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19932 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19933 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19934 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19935 equivalents. Fix markup.
19936 * doc/tm.texi: Regenerate.
19938 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19940 * config/i386/i386.opt: New option mstack-protector-guard=.
19941 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19942 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19943 TARGET_SSP_TLS_GUARD.
19944 * config/i386/i386.c (ix86_option_override_internal): Set
19945 ix86_stack_protector_guard.
19946 * config/i386/i386.md (stack_protect_set): Enable for
19947 TARGET_SSP_TLS_GUARD only.
19948 (stack_protect_set_<mode>): Ditto.
19949 (stack_protect_test): Ditto.
19950 (stack_protect_test_<mode>): Ditto.
19951 * doc/invoke.texi (i386 Option): Document.
19953 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19956 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19957 (S_MODES): Set H_MODE bit.
19958 (SF_MODES): Set only S_MODE and SF_MODE bits.
19959 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19960 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19961 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19962 <MODE_FLOAT>: Likewise.
19964 2013-04-15 Joey Ye <joey.ye@arm.com>
19966 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19968 2013-04-15 Joey Ye <joey.ye@arm.com>
19970 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19972 (thumb_far_jump_used_p): Count instruction size and set
19975 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19977 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19978 * resource.c (mark_target_live_regs): Reformat conditional block.
19980 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19982 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19983 notes, they are emitted only just before final.
19984 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19986 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19988 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19989 * cfgrtl.c (delete_insn): Call it here instead.
19990 * lra-spills.c (lra_final_code_change): Use delete_insn.
19991 * haifa-sched.c (sched_remove_insn): Likewise.
19992 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19993 returning to the nop pool.
19994 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19995 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19997 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19999 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20000 * doc/tm.texi: Regenerated.
20002 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20004 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20007 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20009 * df-core.c (df_find_def): Compare register numbers.
20010 (df_find_use): Likewise.
20012 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20015 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20016 lra_in_progress for return.
20018 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20020 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20021 define_insn into define_insn_and_split and emit movsicc patterns.
20023 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20025 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20027 2013-04-12 Richard Biener <rguenther@suse.de>
20029 * tree-pass.h (TODO_do_not_ggc_collect): New.
20030 * passes.c (execute_one_ipa_transform_pass): Honor
20031 TODO_do_not_ggc_collect.
20032 (execute_one_pass): Likewise.
20035 2013-04-10 Richard Biener <rguenther@suse.de>
20037 * passes.c (init_optimization_passes): Remove reload pass.
20038 * ira.c (do_reload): Merge into ...
20040 (rest_of_handle_reload): Remove.
20041 (pass_reload): Likewise.
20042 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20043 of reload for vzeroupper pass placement.
20045 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20047 PR tree-optimization/56918
20048 PR tree-optimization/56920
20049 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20050 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20051 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20052 use wide_mul_with_sign method.
20054 2013-04-12 Richard Biener <rguenther@suse.de>
20056 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20057 not be considered a gimple constant.
20059 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20061 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20062 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20065 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20066 Jakub Jelinek <jakub@redhat.com>
20068 * opts.c: Include diagnostic-color.h.
20069 (common_handle_option): Handle OPT_fdiagnostics_color_.
20070 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20071 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20072 (diagnostic-color.o): New.
20073 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20074 (diagnostic_color_rule): New enum.
20075 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20076 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20077 the location string.
20078 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20079 either NULL, or color kind.
20080 * diagnostic-color.c: New file.
20081 * diagnostic-color.h: New file.
20082 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20084 * doc/invoke.texi (-fdiagnostics-color): Document.
20085 * pretty-print.h (pp_show_color): Define.
20086 (struct pretty_print_info): Add show_color field.
20087 * diagnostic.c: Include diagnostic-color.h.
20088 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20089 macros. Colorize error:, warning: etc. strings and also the location
20091 (diagnostic_show_locus): Colorize the caret line.
20092 * pretty-print.c: Include diagnostic-color.h.
20093 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20094 inside of %< %> quotes or quoted through q format modifier.
20096 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20098 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20100 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20102 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20103 code in CC_NZ mode.
20104 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20107 2013-04-11 Marek Polacek <polacek@redhat.com>
20109 PR tree-optimization/48184
20110 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20112 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20114 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20115 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20116 (skip_simple_arithmetic): Tidy up.
20117 * tree.h (skip_simple_constant_arithmetic): Declare.
20119 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20121 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20123 2013-04-11 Richard Biener <rguenther@suse.de>
20125 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20126 generate vector constants.
20128 2013-04-11 Richard Biener <rguenther@suse.de>
20130 PR tree-optimization/56878
20131 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20132 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20134 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20135 Prefer to align the DR with the most invariant base address.
20137 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20139 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20142 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20144 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20145 floating-point vector comparisons against 0.
20147 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20149 PR tree-optimization/56899
20150 * fold-const.c (extract_muldiv_1): Apply distributive law
20151 only if TYPE_OVERFLOW_WRAPS (ctype).
20153 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20156 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20157 loading has side effect.
20159 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20161 * config/sparc/sparc.c: Include tree-pass.h.
20162 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20163 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20164 head of file. Change return type. Split off gate function.
20165 (sparc_gate_work_around_errata): New function.
20166 (pass_work_around_errata): New pass definition.
20167 (insert_pass_work_around_errata) New pass insert definition to
20168 insert pass_work_around_errata just after delayed-branch scheduling.
20169 (sparc_option_override): Insert the pass.
20170 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20172 2013-04-10 David S. Miller <davem@davemloft.net>
20174 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20175 or -mcpu=hypersparc.
20177 * target.def (cstore_mode): New hook.
20178 * target.h: Include insn-codes.h
20179 * targhooks.c: Likewise.
20180 (default_cstore_mode): New function.
20181 * targhooks.h: Declare it.
20182 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20183 * doc/tm.texi: Rebuild.
20184 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20185 target hook, rather than inspecting the insn_data.
20186 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20187 (TARGET_CSTORE_MODE): Redefine.
20188 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20190 * config/sparc/predicates.md (cstore_result_operand): New special
20192 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20193 Use it for operand 0.
20194 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20195 (*snesi_special): Likewise.
20196 (*snesi_zero): Likewise.
20197 (*seqsi_zero): Likewise.
20198 (*sltu_insn): Likewise.
20199 (*sgeu_insn): Likewise.
20200 (*seqdi_special): Make operand 0 and comparison operation be of
20202 (*snedi_special): Likewise.
20203 (*snedi_special_vis3): Likewise.
20204 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20205 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20206 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20207 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20208 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20209 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20210 (*sltu_extend_sp64): Likewise.
20211 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20212 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20213 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20214 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20215 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20217 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20219 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20220 (aarch64_start_file): Use the new function.
20222 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20223 Jason Merrill <jason@redhat.com>
20225 * common.opt: Add -gdwarf.
20226 * opts.c (common_handle_option): Handle it.
20227 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20229 2013-04-10 Richard Biener <rguenther@suse.de>
20231 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20232 (execute_one_ipa_transform_pass): But unconditionally here.
20233 (execute_one_pass): And here.
20234 (init_optimization_passes): Remove reload pass.
20235 * tree-pass.h (TODO_ggc_collect): Remove.
20236 (pass_reload): Likewise.
20237 * ira.c (do_reload): Merge into ...
20239 (rest_of_handle_reload): Remove.
20240 (pass_reload): Likewise.
20241 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20242 of reload for vzeroupper pass placement.
20243 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20244 and todo_flags_finish of all passes.
20246 2013-04-10 Richard Biener <rguenther@suse.de>
20248 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20249 first_const_oprnd field, rename first_def_type to first_op_type.
20250 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20251 (vect_get_and_check_slp_defs): Always use the type of the
20252 operand. Allow mixed vect_external_def, vect_constant_def types.
20253 (vect_get_constant_vectors): Handle mixed vect_external_def,
20254 vect_constant_def types.
20256 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20258 PR tree-optimization/55524
20259 * tree-ssa-math-opts.c
20260 (convert_mult_to_fma): Don't use an fms construct
20261 when we don't have an fms operation, but fnma, and it looks
20262 likely that we'll be able to use the latter.
20264 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20266 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20268 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20269 inline fail caused by overwritable functions.
20271 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20273 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20274 unnecessary bits in the constant power of two case.
20276 2013-04-10 Richard Biener <rguenther@suse.de>
20278 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20279 broken code swapping operands.
20280 (vect_build_slp_tree): Do not compute load permutations here.
20281 (vect_analyze_slp_instance): Compute load permutations here,
20282 after building the SLP tree.
20284 2013-04-09 Christian Bruel <christian.bruel@st.com>
20286 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20287 of next/prev_real_insn.
20289 2013-04-09 Jan Hubicka <jh@suse.cz>
20291 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20292 Drop aliased parameter.
20293 (function_and_variable_visibility): Do not handle alias pairs.
20294 * cgraph.c (varpool_externally_visible_p): Update prototype.
20295 * varpool.c (varpool_add_new_variable): Update.
20297 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20299 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20301 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20303 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20305 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20307 2013-04-09 Marek Polacek <polacek@redhat.com>
20309 PR tree-optimization/48762
20310 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20312 2013-04-09 Richard Biener <rguenther@suse.de>
20314 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20316 (vect_build_slp_tree): Likewise.
20317 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20318 calculating the cost of a SLP instance.
20319 (vect_analyze_slp_instance): Use it from here, after building
20322 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20324 PR middle-end/56883
20325 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20326 expand_omp_for_static_chunk): Use simple_p = true in
20327 force_gimple_operand_gsi calls when assigning to addressable decls.
20329 2013-04-09 Jeff Law <law@redhat.com>
20331 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20332 when the boolean was created by converting a wider object which
20333 had a boolean range.
20335 2013-04-09 Richard Biener <rguenther@suse.de>
20337 * tree-vectorizer.h (slp_void_p): Remove.
20338 (slp_tree): Typedef before _slp_tree declaration.
20339 (struct _slp_tree): Use a vector of slp_tree as children.
20340 (vect_get_place_in_interleaving_chain): Remove.
20341 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20343 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20345 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20346 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20347 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20348 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20349 Use slp_node instead of slp_void_p and adjust.
20351 2013-04-09 Richard Biener <rguenther@suse.de>
20353 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20354 work that is not necessary.
20356 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20358 PR tree-optimization/56854
20359 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20360 forward into clobber stmts if it would change MEM_REF lhs into
20363 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20365 * tree.c (type_hash_lookup, type_hash_add): Make static.
20366 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20368 2013-04-09 Richard Biener <rguenther@suse.de>
20370 * tree.h (unsave_expr_now): Remove.
20371 * tree-inline.c (mark_local_for_remap_r): Remove.
20372 (unsave_expr_1): Likewise.
20373 (unsave_r): Likewise.
20374 (unsave_expr_now): Likewise.
20375 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20376 (propagate_tree_value): Likewise.
20378 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20380 * doc/rtl.texi (sequence): Rewrite documentation to match the
20381 current use of SEQUENCE rtl objects.
20382 * rtl.def (SEQUENCE): Likewise.
20384 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20385 Update documentation.
20386 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20387 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20389 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20391 2013-04-08 Teresa Johnson <tejohnson@google.com>
20393 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20394 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20396 (estimate_edge_size_and_time): Add comment to suggest using rounding
20398 (estimate_node_size_and_time): Ditto.
20399 (remap_edge_change_prob): Use helper rounding divide methods.
20400 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20401 (gimple_mod_pow2_value_transform): Ditto.
20402 (gimple_mod_subtract_transform): Ditto.
20403 (gimple_ic_transform): Ditto.
20404 (gimple_stringops_transform): Ditto.
20405 * stmt.c (conditional_probability): Ditto.
20406 (emit_case_dispatch_table): Ditto.
20407 * lto-cgraph.c (merge_profile_summaries): Ditto.
20408 * tree-optimize.c (execute_fixup_cfg): Ditto.
20409 * cfgcleanup.c (try_forward_edges): Ditto.
20410 * cfgloopmanip.c (scale_loop_profile): Ditto.
20412 (duplicate_loop_to_header_edge): Ditto.
20413 (lv_adjust_loop_entry_edge): Ditto.
20414 * tree-vect-loop.c (vect_transform_loop): Ditto.
20415 * profile.c (compute_branch_probabilities): Ditto.
20416 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20417 * lto-streamer-in.c (input_cfg): Ditto.
20418 * gimple-streamer-in.c (input_bb): Ditto.
20419 * ipa-cp.c (update_profiling_info): Ditto.
20420 (update_specialized_profile): Ditto.
20421 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20422 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20424 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20425 (compute_trg_info): Ditto.
20426 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20427 (purge_dead_edges): Ditto.
20428 * loop-unswitch.c (unswitch_loop): Ditto.
20429 * cgraphclones.c (cgraph_clone_edge): Ditto.
20430 (cgraph_clone_node): Ditto.
20431 * tree-inline.c (copy_bb): Ditto.
20432 (copy_edges_for_bb): Ditto.
20433 (initialize_cfun): Ditto.
20434 (copy_cfg_body): Ditto.
20435 (expand_call_inline): Ditto.
20437 2013-04-08 Kai Tietz <ktietz@redhat.com>
20439 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20440 TARGET_CYGWIN64 by TARGET_64BIT.
20442 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20444 * config/epiphany/epiphany.md (GPR_1): New constant.
20445 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20446 * config/epiphany/epiphany.c (gen_compare_reg):
20447 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20448 is already in place.
20449 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20450 Don't require being called during rtl expansion; If y operlaps r0,
20452 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20453 (epiphany_expand_epilogue): Likewise.
20455 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20456 Don't use CC_FPmode for ORDERED / UNORDERED.
20457 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20459 * config/epiphany/constraints.md (CnL): New constraint.
20460 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20461 * config/epiphany/predicates.md (add_operand): Allow 1024.
20463 * config/epiphany/epiphany.md (logical_op): New code iterator.
20464 (op_mnc): New code attribute.
20465 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20466 (mov_f+1, mov_f+2): New peephole2 patterns.
20468 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20469 (cstoresi4): Also allow re-use of zero result when doing a NE
20470 comparison to a non-zero operand.
20471 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20473 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20474 Use gen_addsi3_i / gen_subsi3_i.
20476 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20480 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20481 contain anything but clobbers, at most one __builtin_stack_restore,
20482 optionally debug stmts and final resx, and if it has at least one
20483 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20484 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20485 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20486 which isn't defaut definition, remove them.
20487 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20488 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20489 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20490 with MEM_REF LHS with SSA_NAME address.
20492 2013-04-08 Jeff Law <law@redhat.com>
20494 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20496 2013-04-08 Richard Biener <rguenther@suse.de>
20498 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20500 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20501 determined vector type.
20502 (vect_analyze_data_refs): Likewise.
20503 (vect_get_new_vect_var): Adjust.
20504 (vect_create_destination_var): Preserve SSA name versions.
20505 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20506 not dump anything here.
20508 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20510 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20511 Add member lr_slot_known.
20512 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20514 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20515 Remove code that sets lr_slot_offset according to what a previous
20516 version of epiphany_emit_save_restore used to do.
20517 (epiphany_emit_save_restore): When doing an lr save or restore,
20518 set/verify lr_slot_known and lr_slot_offset.
20520 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20523 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20526 2013-04-08 Richard Biener <rguenther@suse.de>
20528 * alias.c (find_base_term): Fix thinko in previous change.
20530 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20532 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20533 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20534 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20535 if possible to compute val.
20536 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20537 For QImode integers don't require anything about precision. Use
20538 const_with_all_bytes_same to find out if the constant doesn't have
20539 repeated bytes in it.
20541 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20543 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20546 2013-04-08 Marek Polacek <polacek@redhat.com>
20548 PR rtl-optimization/48182
20549 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20552 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20555 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20556 nuses, make sure we have a label.
20558 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20561 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20562 (rs6000_emit_swdiv_low_precision): Remove.
20563 (rs6000_emit_swdiv): Rewrite to handle between one and four
20564 iterations of Newton-Raphson generally; modify required number of
20565 iterations for some cases.
20566 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20568 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20570 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20571 set-but-unused variable.
20573 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20574 basic blocks of released function bodies garbage-collectable.
20576 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20577 (struct rtl_opt_pass): Add TODO_df_finish.
20579 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20581 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20583 * config/arm/constraints.md (q): New constraint.
20584 * config/arm/ldrdstrd.md: New file.
20585 * config/arm/arm.md (ldrdstrd.md) New include.
20586 (arm_movdi): Use "q" instead of "r" constraint
20587 for double-word memory access.
20588 (movdf_soft_insn): Likewise.
20589 * config/arm/vfp.md (movdi_vfp): Likewise.
20590 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20591 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20592 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20593 (mem_ok_for_ldrd_strd): Likewise.
20594 (output_move_double): Update assertion.
20596 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20598 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20600 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20602 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20603 define_insn_and_split.
20604 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20606 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20608 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20609 define_insn_and_split.
20610 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20611 (shiftsi3_compare): New pattern.
20612 (rrx): New pattern.
20613 * config/arm/unspecs.md (UNSPEC_RRX): New.
20615 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20617 * config/arm/arm.md (negdi_extendsidi): New pattern.
20618 (negdi_zero_extendsidi): Likewise.
20620 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20622 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20623 define_insn_and_split.
20624 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20625 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20627 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20629 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20630 define_insn_and_split.
20631 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20632 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20634 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20636 * config/arm/arm.md (subsi3_carryin): New pattern.
20637 (subsi3_carryin_const): Likewise.
20638 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20639 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20641 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20643 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20645 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20647 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20648 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20650 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20652 * config/arm/arm.c (arm_expand_builtin): Change fcode
20653 type to unsigned int.
20655 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20657 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20659 2013-04-04 Ian Lance Taylor <iant@google.com>
20661 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20664 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20666 PR middle-end/56729
20667 * df-scan.c (df_insn_delete): Disable failing assert.
20669 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20671 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20672 New function prototype.
20673 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20674 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20675 (arm_builtin_vectorized_function): New function.
20677 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20679 * config/arm/arm_neon_builtins.def: New file.
20680 * config/arm/arm.c (neon_builtin_data): Move contents to
20681 arm_neon_builtins.def.
20682 (enum arm_builtins): Include neon builtin definitions.
20683 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20684 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20686 2013-04-04 Marek Polacek <polacek@redhat.com>
20688 PR tree-optimization/48186
20689 * predict.c (maybe_hot_frequency_p): Return false if
20690 HOT_BB_FREQUENCY_FRACTION is 0.
20691 (cgraph_maybe_hot_edge_p): Likewise.
20693 2013-04-04 Richard Biener <rguenther@suse.de>
20695 PR tree-optimization/56826
20696 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20699 2013-04-04 Richard Biener <rguenther@suse.de>
20701 PR tree-optimization/56213
20702 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20703 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20705 2013-04-04 Richard Biener <rguenther@suse.de>
20707 PR tree-optimization/56837
20708 * tree-loop-distribution.c (classify_partition): For non-zero
20709 values require that the value has the same precision as its
20710 mode to be useful as memset value.
20712 2013-04-03 Nick Clifton <nickc@redhat.com>
20714 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20715 (fmssf4): Use fmsf.s on E3V5 architectures.
20716 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20717 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20719 2013-04-03 Jeff Law <law@redhat.com>
20721 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20722 (lra-eliminations.o): Likewise.
20724 2013-04-03 Teresa Johnson <tejohnson@google.com>
20726 * gcov-io.c (compute_working_sets): Moved most of body of old
20727 compute_working_sets here from profile.c.
20728 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20729 (gcov_working_set_t): Moved typedef here from basic-block.h
20730 (compute_working_set): Declare.
20731 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20732 (get_working_sets): Renamed from compute_working_set,
20733 replace most of body with call to new compute_working_sets.
20734 (get_exec_counts): Replace call to compute_working_sets
20735 to get_working_sets.
20736 * profile.h (get_working_sets): Renamed from compute_working_set.
20737 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20738 to get_working_sets.
20739 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20740 * gcov-dump.c (dump_working_sets): New function.
20742 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20744 * hwint.c (sext_hwi, zext_hwi): New functions.
20745 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20746 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20747 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20748 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20749 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20750 (sext_hwi, zext_hwi): New functions.
20752 2013-04-03 Jeff Law <law@redhat.com>
20754 PR tree-optimization/56799
20755 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20756 back test for widening conversion erroneously dropped in prior change.
20758 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20761 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20762 instead of next_real_insn.
20764 2013-04-03 Marek Polacek <polacek@redhat.com>
20767 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20769 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20772 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20774 (thumb1_output_casesi): Likewise.
20775 (thumb2_output_casesi): Likewise.
20777 2013-04-03 Richard Biener <rguenther@suse.de>
20779 PR tree-optimization/56817
20780 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20782 (tree_unroll_loops_completely_1): ... new function to manually
20783 walk the loop tree, properly defering outer loops of unrolled
20784 loops to later iterations.
20786 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20788 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20789 (vectorizable_load): Likewise.
20790 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20791 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20793 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20795 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20798 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20800 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20802 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20804 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20806 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20808 PR tree-optimization/56790
20809 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20812 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20814 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20816 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20817 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20820 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20823 * tree.h (force_folding_builtin_constant_p): New decl.
20824 * builtins.c (force_folding_builtin_constant_p): New variable.
20825 (fold_builtin_constant_p): Fold immediately also if
20826 force_folding_builtin_constant_p.
20828 2013-04-03 Richard Biener <rguenther@suse.de>
20830 PR tree-optimization/56812
20831 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20832 DRs of the same interleaving chain are independent.
20834 2013-04-02 Jason Merrill <jason@redhat.com>
20836 * gdbinit.in (pbb): Use debug fn.
20838 2013-04-02 Lawrence Crowl <crowl@google.com>
20840 * sese.h (struct ivtype_map_elt_s): Remove unused.
20841 (extern debug_ivtype_map): Remove unused.
20842 (extern eq_ivtype_map_elts): Remove unused.
20843 * sese.c (debug_ivtype_map): Removed unused.
20844 (debug_ivtype_map_1): Removed unused.
20845 (debug_ivtype_elt): Remove unused.
20846 (eq_ivtype_map_elts): Remove unused.
20849 2013-04-02 Kai Tietz <ktietz@redhat.com>
20852 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20853 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20854 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20856 (legitimize_pe_coff_symbol): Likewise.
20857 (is_imported_p): New helper-function.
20858 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20859 for Windows x64 targets.
20860 (ix86_expand_prologue): Optimize for pe-coff targets.
20861 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20862 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20863 medium/large code-model.
20864 (legitimize_pic_address): Likewise.
20865 (legitimize_tls_address): Likewise.
20866 (ix86_expand_call): Likewise.
20867 (x86_output_mi_thunk): Likewise.
20868 (get_dllimport_decl): Add new beimport argument.
20869 (construct_plt_address): Don't assert for x64 pe-coff targets.
20870 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20872 (SYMBOL_FLAG_STUBVAR): New macro.
20873 (SYMBOL_REF_STUBVAR_P): Likewise.
20874 * config/i386/winnt.c (stub_list): New structure.
20875 (stub_head): New local variable.
20876 (i386_pe_record_stub): New function.
20877 (i386_pe_file_end): Emit refptr-stubs.
20879 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20881 PR rtl-optimization/56745
20882 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20883 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20886 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20887 and both of them are MEM_REFs, just compare first argument for
20888 equality and attempt to deal even with differing offsets.
20891 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20892 of gimple_clobber_p to be MEM_REF.
20893 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20894 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20895 after gimplification.
20896 * asan.c (get_mem_ref_of_assignment): Don't instrument
20897 gimple_clobber_p stmts.
20898 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20899 gimple_clobber_p stmt if they have MEM_REF lhs and
20900 are dead because of another gimple_clobber_p stmt.
20901 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20902 gimple_clobber_p stmts like debug stmts.
20903 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20904 that refer to unused VAR_DECLs or uninitialized values.
20905 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20906 gimple_clobber_p stmts if they refer to removed parameters.
20907 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20910 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20912 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20913 using SWI48 mode attribute.
20915 2013-04-02 Wei Mi <wmi@google.com>
20917 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20918 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20919 *<rotate_insn><mode>3_mask in i386.md.
20921 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20923 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20925 2013-04-02 Richard Biener <rguenther@suse.de>
20927 PR tree-optimization/56778
20928 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20929 Runtime alias tests are not supported for gather loads.
20930 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20931 stmts referenced from SSA operands before updating SSA form.
20933 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20934 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20936 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20937 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20938 * config/arm/cortex-a53.md: New file.
20939 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20940 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20941 * config/arm/arm.c (arm_issue_rate): Likewise.
20942 * config/arm/arm-tune.md: Regenerate
20943 * config/arm/arm-tables.opt: Regenerate.
20944 * config/arm/arm-cores.def: Add cortex-a53.
20946 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20948 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20951 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20953 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20954 scalar load/store operations using B/H registers.
20955 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20957 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20959 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20961 * config/aarch64/aarch64.c
20962 (aarch64_simd_scalar_immediate_valid_for_move): New.
20963 * config/aarch64/aarch64-protos.h
20964 (aarch64_simd_scalar_immediate_valid_for_move): New.
20965 * config/aarch64/constraints.md (Dh, Dq): New.
20966 * config/aarch64/iterators.md (hq): New.
20968 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20970 * reorg.c (get_branch_condition): Deal with conditional returns.
20971 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20973 2013-04-01 Wei Mi <wmi@google.com>
20975 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20976 Truncate operand 2 using %b asm operand modifier.
20977 (*<shift_insn><mode>3_mask): Ditto.
20978 (*<rotate_insn><mode>3_mask): Ditto.
20980 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20982 PR middle-end/56798
20983 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20985 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20987 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20989 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20991 2013-03-30 Lawrence Crowl <crowl@google.com>
20993 * dse.c (clear_alias_sets): Remove never set.
20994 (disqualified_clear_alias_sets): Remove never set.
20995 (clear_alias_mode_pool): Remove never set.
20996 (dse_step0): Remove condition that is never true.
20997 (canon_address): Remove condition that is never true.
20998 (dse_step7): Remove condition that is never true.
20999 (rest_of_handle_dse): Remove condition that is never true.
21000 (rest_of_handle_dse::did_global): Remove never read from above.
21001 (dse_step2_spill): Remove never called from above.
21002 (dse_step5_spill): Remove never called from above.
21004 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21006 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21007 JUMP_TABLE_DATA changes.
21008 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21009 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21010 (Insns) <jump_table_data>: New entry.
21011 * doc/tm.texi: Regenerate.
21013 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21015 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21016 for table jump at the end of a basic block using tablejump_p.
21017 * targhooks.c (default_invalid_within_doloop): Likewise.
21018 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21019 target hook implementation that is identical to the default hook.
21020 (rs6000_invalid_within_doloop): Remove.
21022 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21023 unused variable from tablejump_p call.
21025 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21026 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21027 (INSN_DELETED_P): Likewise.
21028 (emit_jump_table_data): New prototype.
21029 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21030 after 4th as unused.
21031 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21032 * sched-vis.c (print_insn): Likewise.
21033 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21034 insn for compatibility with back ends that use next_active_insn to
21035 identify jump table data.
21036 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21037 (remove_insn): Likewise.
21038 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21040 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21041 (emit_jump_table_data): New function.
21043 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21044 basic block, a JUMP_TABLE_DATA never is.
21045 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21046 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21047 off from code handling real insns.
21048 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21049 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21050 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21051 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21052 is not a NONDEBUG_INSN_P.
21053 * ira-costs.c (scan_one_insn): Likewise.
21054 * jump.c (mark_all_labels): Likewise.
21055 (mark_jump_label_1): Likewise.
21056 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21057 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21059 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21060 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21061 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21062 JUMP_TABLE_DATA_P insns.
21063 (calculate_elim_costs_all_insns): Likewise.
21064 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21065 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21066 (delete_output_reload): Code style fixups.
21067 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21068 insn flags on this non-insn.
21069 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21070 as scheduling barriers, for pre-change compatibility.
21071 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21072 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21074 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21075 redundant JUMP_TABLE_DATA_P test.
21076 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21077 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21078 (frv_for_each_packet): Likewise.
21079 * config/i386/i386.c (min_insn_size): Likewise.
21080 (ix86_avoid_jump_mispredicts): Likewise.
21081 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21082 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21083 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21084 (mips16_insn_length): Robustify.
21085 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21086 (mips16_split_long_branches): Likewise.
21087 * config/pa/pa.c (pa_combine_instructions): Likewise.
21088 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21089 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21090 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21091 as contributing to pool range lengths.
21092 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21093 Remove redundant JUMP_TABLE_DATA_P test.
21094 (sh_loop_align): Likewise.
21095 (split_branches): Likewise.
21096 (sh_insn_length_adjustment): Likewise.
21097 * config/spu/spu.c (get_branch_target): Likewise.
21099 2013-03-29 Jan Hubicka <jh@suse.cz>
21101 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21102 gcov streaming; stream hot bb threshold to ltrans.
21103 * predict.c (get_hot_bb_threshold): Break out from ....
21104 (maybe_hot_count_p): ... here.
21105 (set_hot_bb_threshold): New function.
21106 * lto-section-in.c (lto_section_name): Add profile.
21107 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21108 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21109 and data-streamer.h
21110 (histogram_entry): New structure.
21111 (histogram, histogram_pool): New global vars.
21112 (histogram_hash): New structure.
21113 (histogram_hash::hash): New method.
21114 (histogram_hash::equal): Likewise.
21115 (account_time_size): New function.
21116 (cmp_counts): New function.
21117 (dump_histogram): New function.
21118 (ipa_profile_generate_summary): New function.
21119 (ipa_profile_write_summary): New function.
21120 (ipa_profile_read_summary): New function.
21121 (ipa_profile): Decide on threshold.
21122 (pass_ipa_profile): Add ipa_profile_write_summary and
21123 ipa_profile_read_summary.
21124 * Makefile.in (ipa.o): Update dependencies.
21125 * lto-streamer.h (LTO_section_ipa_profile): New section.
21127 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21129 * tree.h (VAR_P): New.
21131 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21134 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21136 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21138 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21139 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21140 (control_flow_insn_p): Likewise.
21141 * cfgrtl.c (duplicate_insn_chain): Likewise.
21142 * final.c (get_attr_length_1): Likewise.
21143 (shorten_branches): Likewise.
21144 (final_scan_insn): Likewise.
21145 * function.c (instantiate_virtual_regs): Likewise.
21146 * gcse.c (insert_insn_end_basic_block): Likewise.
21147 * ira-costs.c (scan_one_insn): Likewise.
21148 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21149 * lra.c (check_rtl): Likewise.
21150 * reload1.c (elimination_costs_in_insn): Likewise.
21151 * reorg.c (follow_jumps): Likewise.
21153 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21154 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21155 (thumb_far_jump_used_p): Likewise.
21156 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21157 (workaround_speculation): Likewise.
21158 (add_sched_insns_for_speculation): Likewise.
21159 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21160 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21161 (frv_for_each_packet): Likewise.
21162 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21163 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21164 (final_emit_insn_group_barriers): Likewise.
21165 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21166 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21167 (mips16_insn_length): Likewise.
21168 * config/pa/pa.c (pa_reorg): Likewise.
21169 (pa_combine_instructions): Likewise.
21170 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21171 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21172 (sh_reorg): Likewise.
21173 (split_branches): Likewise.
21174 * config/spu/spu.c (get_branch_target): Likewise.
21176 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21179 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21181 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21182 Fix declaration name.
21184 2013-03-28 Lawrence Crowl <crowl@google.com>
21186 * graphds.h (struct graph.indicies): Remove unused.
21187 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21188 (SCOP_ORIGINAL_PDDRS): Remove unused.
21189 * sese.h (extern insert_loop_close_phis): Removed unused.
21190 (extern insert_guard_phis): Removed unused.
21191 (extern ivtype_map_elt_info): Removed unused.
21192 (new_ivtype_map_elt): Removed unused.
21193 * sese.c (ivtype_map_elt_info): Removed unused.
21195 2013-03-28 Lawrence Crowl <crowl@google.com>
21197 * Makefile.in: Add several missing include dependences.
21199 (test-dump.o): New. This object is not added to any executable,
21200 but is present for ad-hoc testing.
21202 (debug (const bitmap_head_def &)): New.
21203 (debug (const bitmap_head_def *)): New.
21205 (extern debug (const bitmap_head_def &)): New.
21206 (extern debug (const bitmap_head_def *)): New.
21208 (debug (edge_def &)): New.
21209 (debug (edge_def *)): New.
21211 (debug (basic_block_def &)): New.
21212 (debug (basic_block_def *)): New.
21214 (dump_node (const_tree, int, FILE *)): Correct source file.
21216 (debug (die_struct &)): New.
21217 (debug (die_struct *)): New.
21219 (extern debug (die_struct &)): New.
21220 (extern debug (die_struct *)): New.
21221 * gimple-pretty-print.c
21222 (debug (gimple_statement_d &)): New.
21223 (debug (gimple_statement_d *)): New.
21224 * gimple-pretty-print.h
21225 (extern debug (gimple_statement_d &)): New.
21226 (extern debug (gimple_statement_d *)): New.
21228 (debug (ira_allocno_copy &)): New.
21229 (debug (ira_allocno_copy *)): New.
21230 (debug (ira_allocno &)): New.
21231 (debug (ira_allocno *)): New.
21233 (extern debug (ira_allocno_copy &)): New.
21234 (extern debug (ira_allocno_copy *)): New.
21235 (extern debug (ira_allocno &)): New.
21236 (extern debug (ira_allocno *)): New.
21238 (debug (live_range &)): New.
21239 (debug (live_range *)): New.
21241 (debug (lra_live_range &)): New.
21242 (debug (lra_live_range *)): New.
21244 (debug (lra_live_range &)): New.
21245 (debug (lra_live_range *)): New.
21247 (debug (omega_pb_d &)): New.
21248 (debug (omega_pb_d *)): New.
21250 (extern debug (omega_pb_d &)): New.
21251 (extern debug (omega_pb_d *)): New.
21253 (debug (const rtx_def &)): New.
21254 (debug (const rtx_def *)): New.
21256 (debug_tree (tree): Move within file.
21257 (debug_raw (const tree_node &)): New.
21258 (debug_raw (const tree_node *)): New.
21259 (dump_tree_via_hooks (const tree_node *, int)): New.
21260 (debug (const tree_node &)): New.
21261 (debug (const tree_node *)): New.
21262 (debug_verbose (const tree_node &)): New.
21263 (debug_verbose (const tree_node *)): New.
21264 (debug_head (const tree_node &)): New.
21265 (debug_head (const tree_node *)): New.
21266 (debug_body (const tree_node &)): New.
21267 (debug_body (const tree_node *)): New.
21268 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21269 (debug (vec<tree, va_gc> &)): New.
21270 (debug (vec<tree, va_gc> *)): New.
21272 (extern debug (const rtx_def &)): New.
21273 (extern debug (const rtx_def *)): New.
21275 (debug_raw (simple_bitmap_def &)): New.
21276 (debug_raw (simple_bitmap_def *)): New.
21277 (debug (simple_bitmap_def &)): New.
21278 (debug (simple_bitmap_def *)): New.
21280 (extern debug (simple_bitmap_def &)): New.
21281 (extern debug (simple_bitmap_def *)): New.
21282 (extern debug_raw (simple_bitmap_def &)): New.
21283 (extern debug_raw (simple_bitmap_def *)): New.
21285 (debug (vinsn_def &)): New.
21286 (debug (vinsn_def *)): New.
21287 (debug_verbose (vinsn_def &)): New.
21288 (debug_verbose (vinsn_def *)): New.
21289 (debug (expr_def &)): New.
21290 (debug (expr_def *)): New.
21291 (debug_verbose (expr_def &)): New.
21292 (debug_verbose (expr_def *)): New.
21293 (debug (vec<rtx> &)): New.
21294 (debug (vec<rtx> *)): New.
21296 (extern debug (vinsn_def &)): New.
21297 (extern debug (vinsn_def *)): New.
21298 (extern debug_verbose (vinsn_def &)): New.
21299 (extern debug_verbose (vinsn_def *)): New.
21300 (extern debug (expr_def &)): New.
21301 (extern debug (expr_def *)): New.
21302 (extern debug_verbose (expr_def &)): New.
21303 (extern debug_verbose (expr_def *)): New.
21304 (extern debug (vec<rtx> &)): New.
21305 (extern debug (vec<rtx> *)): New.
21307 (_list_iter_cond_expr): Make inline instead of static.
21309 (debug (sreal &)): New.
21310 (debug (sreal *)): New.
21312 (extern debug (sreal &)): New.
21313 (extern debug (sreal *)): New.
21315 (extern debug_raw (const tree_node &)): New.
21316 (extern debug_raw (const tree_node *)): New.
21317 (extern debug (const tree_node &)): New.
21318 (extern debug (const tree_node *)): New.
21319 (extern debug_verbose (const tree_node &)): New.
21320 (extern debug_verbose (const tree_node *)): New.
21321 (extern debug_head (const tree_node &)): New.
21322 (extern debug_head (const tree_node *)): New.
21323 (extern debug_body (const tree_node &)): New.
21324 (extern debug_body (const tree_node *)): New.
21325 (extern debug (vec<tree, va_gc> &)): New.
21326 (extern debug (vec<tree, va_gc> *)): New.
21328 (debug (struct loop &)): New.
21329 (debug (struct loop *)): New.
21330 (debug_verbose (struct loop &)): New.
21331 (debug_verbose (struct loop *)): New.
21332 * tree-dump.c: Add header dependence.
21334 (extern debug (struct loop &)): New.
21335 (extern debug (struct loop *)): New.
21336 (extern debug_verbose (struct loop &)): New.
21337 (extern debug_verbose (struct loop *)): New.
21339 (debug (data_reference &)): New.
21340 (debug (data_reference *)): New.
21341 (debug (vec<data_reference_p> &)): New.
21342 (debug (vec<data_reference_p> *)): New.
21343 (debug (vec<ddr_p> &)): New.
21344 (debug (vec<ddr_p> *)): New.
21346 (extern debug (data_reference &)): New.
21347 (extern debug (data_reference *)): New.
21348 (extern debug (vec<data_reference_p> &)): New.
21349 (extern debug (vec<data_reference_p> *)): New.
21350 (extern debug (vec<ddr_p> &)): New.
21351 (extern debug (vec<ddr_p> *)): New.
21353 (debug (pt_solution &)): New.
21354 (debug (pt_solution *)): New.
21356 (extern debug (pt_solution &)): New.
21357 (extern debug (pt_solution *)): New.
21359 (debug (_var_map &)): New.
21360 (debug (_var_map *)): New.
21361 (debug (tree_live_info_d &)): New.
21362 (debug (tree_live_info_d *)): New.
21364 (extern debug (_var_map &)): New.
21365 (extern debug (_var_map *)): New.
21366 (extern debug (tree_live_info_d &)): New.
21367 (extern debug (tree_live_info_d *)): New.
21369 2013-03-28 Jan Hubicka <jh@suse.cz>
21371 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21373 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21375 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21376 record only when desired or required.
21378 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21380 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21381 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21383 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21385 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21386 (*andsi3_compare0_uxtw): New pattern.
21387 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21388 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21390 2013-03-28 Jan Hubicka <jh@suse.cz>
21392 * data-streamer-in.c (streamer_read_gcov_count): New function.
21393 * gimple-streamer-out.c: Include value-prof.h.
21394 (output_gimple_stmt): Output histogram.
21395 (output_bb): Use streamer_write_gcov_count.
21396 * value-prof.c: Include data-streamer.h
21397 (dump_histogram_value): Add HIST_TYPE_MAX.
21398 (stream_out_histogram_value): New function.
21399 (stream_in_histogram_value): New function.
21400 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21401 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21402 * data-streamer-out.c (streamer_write_gcov_count): New function.
21403 (streamer_write_gcov_count_stream): New function.
21404 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21405 (lto_output_node): Likewise.
21406 (input_node, input_edge): Likewise.
21407 * lto-streamer-out.c (output_cfg): Update streaming.
21408 * lto-streamer-in.c (input_cfg): Likewise.
21409 * data-streamer.h (streamer_write_gcov_count,
21410 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21411 * gimple-streamer-in.c: Include value-prof.h
21412 (input_gimple_stmt): Input histograms.
21413 (input_bb): Update profile streaming.
21415 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21417 * genmodes.c (emit_max_int): New function.
21418 (emit_insn_modes_h): Added call to emit_max_function.
21419 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21421 * machmode.def: Fixed comment.
21423 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21425 * combine.c (try_combine): Removed useless assert.
21426 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21428 2013-03-28 Marek Polacek <polacek@redhat.com>
21429 Richard Biener <rguenther@suse.de>
21431 PR tree-optimization/56695
21432 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21433 build signed result of a vector comparison.
21434 * tree-cfg.c (verify_gimple_comparison): Check that a result
21435 of a vector comparison has signed type.
21437 2013-03-28 Richard Biener <rguenther@suse.de>
21439 PR tree-optimization/37021
21440 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21441 do not restrict gaps between groups.
21442 * tree-vect-stmts.c (vectorizable_load): Properly account for
21443 a gap between groups.
21445 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21447 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21448 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21451 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21453 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21454 * doc/extend.texi (Named Address Spaces): Ditto.
21455 (Variable Attributes): Ditto.
21457 2013-03-27 Kai Tietz <ktietz@redhat.com>
21459 * config.build: Add support for cygwin x64 target.
21460 * config.gcc: Likewise.
21461 * config.host: Likewise.
21462 * configure.ac: Likewise
21463 * configure: Regenerated.
21465 2013-03-27 Kai Tietz <ktietz@redhat.com>
21467 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21468 * config/i386/t-cygwin-w64: New file.
21469 * config/i386/cygwin-w64.h: New file.
21470 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21471 and add support for x64-cygwin target.
21472 (CPP_SPEC): Likewise.
21473 (CXX_WRAP_SPEC_LIST): Undefine before define.
21474 (LIBGCJ_SONAME): Use 15 as version.
21476 2013-03-27 Richard Biener <rguenther@suse.de>
21478 PR tree-optimization/56716
21479 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21480 dumping for ref nodes.
21482 2013-03-27 Martin Jambor <mjambor@suse.cz>
21484 PR tree-optimization/55334
21485 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21486 restricted pointers to arrays.
21488 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21490 * Makefile.in (.SUFFIXES): Add .cc.
21491 (.c.o): Apply same recipe for implicit rule .cc.o.
21493 2013-03-27 Richard Biener <rguenther@suse.de>
21495 PR tree-optimization/37021
21496 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21497 REALPART/IMAGPART_EXPRs around the supported refs.
21498 * tree-ssa-structalias.c (find_func_aliases): Assume that
21499 floating-point values are not used to transfer pointers.
21501 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21503 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21504 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21505 * doc/tm.texi: Regenerate.
21506 * targhooks.h (default_has_ifunc_p): New.
21507 * targhooks.c (default_has_ifunc_p): Ditto.
21508 * config/linux-protos.h: New file.
21509 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21510 hook for linux which disables support of indirect functions in android.
21511 * config/linux-android.c: New file.
21512 * config/t-linux-android.c: Ditto.
21513 * config.gcc: Added new object file linux-android.o.
21514 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21515 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21516 * varasm.c (do_assemble_alias): Likewise.
21517 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21518 doesn't support indirect functions.
21519 * configure: Regenerate.
21521 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21524 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21525 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21528 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21530 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21532 2013-03-27 Terry Guo <terry.guo@arm.com>
21534 * config/arm/arm-cores.def: Added core cortex-r7.
21535 * config/arm/arm-tune.md: Regenerated.
21536 * config/arm/arm-tables.opt: Regenerated.
21537 * doc/invoke.texi: Added entry for core cortex-r7.
21539 2013-03-27 Walter Lee <walt@tilera.com>
21541 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21542 double-decrement of next_scratch_regno.
21544 2013-03-27 Walter Lee <walt@tilera.com>
21546 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21548 (insn_v1mulus): Ditto.
21549 (insn_v2muls): Ditto.
21551 2013-03-27 Walter Lee <walt@tilera.com>
21553 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21554 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21556 2013-03-27 Walter Lee <walt@tilera.com>
21558 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21559 (*sibcall_value): Ditto.
21561 2013-03-27 Walter Lee <walt@tilera.com>
21563 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21564 (insn_mnz_v8qi): ... this ...
21565 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21567 (insn_v<n>mnz): Replaced by ...
21568 (insn_v1mnz): ... this ...
21569 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21571 (insn_mz_<mode>): Replaced by ...
21572 (insn_mz_v8qi): ... this ...
21573 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21575 (insn_v<n>mz): Replaced by ...
21576 (insn_v1mz): ... this ...
21577 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21580 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21582 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21584 2013-03-26 Roland McGrath <mcgrathr@google.com>
21586 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21587 than fprintf with a non-constant, non-format string.
21589 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21591 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21592 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21593 operand 0 predicate.
21594 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21595 attribute. Use general_x64nomem_operand as operand 1 predicate.
21596 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21597 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21598 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21599 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21600 (mov<mode>_insv_1): Remove expander. Merge insn with
21601 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21602 Use general_x64nomem_operand as operand 1 predicate.
21603 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21604 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21605 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21606 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21607 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21608 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21609 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21610 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21611 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21612 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21613 (general_x64nomem_operand): Ditto.
21615 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21617 * config/rtems.opt: Add -pthread option.
21619 2013-03-26 Richard Biener <rguenther@suse.de>
21621 * alias.c (find_base_term): Avoid redundant and not used recursion.
21622 (base_alias_check): Get the initial base term from the caller.
21623 (true_dependence_1): Compute and pass base terms to base_alias_check.
21624 (write_dependence_p): Likewise.
21625 (may_alias_p): Likewise.
21627 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21629 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21630 PC-relative load in SI modes and above only.
21632 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21634 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21635 * config/arm/iwmmxt.md (WCGR0): Update.
21636 (WCGR1, WCGR2, WCGR3): Likewise.
21638 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21640 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21641 Use x64 and nox64 isa attributes.
21643 2013-03-26 Richard Biener <rguenther@suse.de>
21645 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21646 alignment computations and rely on get_object_alignment_1
21647 for the !TYPE_P case.
21648 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21650 2013-03-26 Walter Lee <walt@tilera.com>
21652 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21653 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21655 2013-03-25 Jeff Law <law@redhat.com>
21657 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21658 check for INTEGRAL_TYPE_P that was missing due to checking in
21659 wrong version of prior patch.
21661 2013-03-25 Walter Lee <walt@tilera.com>
21663 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21664 TILEGX_INSN_SHUFFLEBYTES1.
21665 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21667 (tilegx_builtins): Ditto.
21668 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21670 2013-03-25 Walter Lee <walt@tilera.com>
21672 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21673 (floatunssisf2): New pattern.
21674 (floatsidf2): New pattern.
21675 (floatunssidf2): New pattern.
21677 2013-03-25 Walter Lee <walt@tilera.com>
21679 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21680 tests for constraint J, K, N, P.
21682 2013-03-25 Walter Lee <walt@tilera.com>
21684 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21685 Use indirect/pcrel encoding.
21686 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21689 2013-03-25 Steve Ellcey <sellcey@mips.com>
21691 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21692 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21693 * config/mips/mips.c (mips_option_override): Set IMADD default.
21694 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21695 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21696 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21697 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21698 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21700 2013-03-25 Jeff Law <law@redhat.com>
21702 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21703 slightly to avoid creating and folding useless trees. Simplify
21704 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21706 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21708 * config/i386/i386.md (*zero_extendsidi2): Merge with
21709 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21710 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21711 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21712 targets. Clarify comment.
21714 2013-03-25 Martin Jambor <mjambor@suse.cz>
21716 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21717 pass-through jump functions differently.
21718 (ipa_read_jump_function): Likewise. Also use setter functions to set
21721 2013-03-25 Martin Jambor <mjambor@suse.cz>
21723 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21724 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21726 (ipa_get_indirect_edge_target): New function.
21727 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21728 ipa_get_indirect_edge_target. Update all callers.
21729 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21730 ipa_get_indirect_edge_target_1 instead of calling
21731 ipa_get_indirect_edge_target.
21732 (create_specialized_node): Pass aggvlas to
21733 ipcp_discover_new_direct_edges.
21735 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21737 * config/arm/arm.md (f_sels, f_seld): New types.
21738 (*cmov<mode>): New pattern.
21739 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21742 2013-03-25 Kai Tietz <ktietz@redhat.com>
21744 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21745 POSIX-printf for mingw-hosted builds.
21747 2013-03-25 Richard Biener <rguenther@suse.de>
21749 PR middle-end/56694
21750 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21751 must-not-throw stmt location.
21753 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21755 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21756 Emit load-acquire versions when acq is true.
21757 (arm_emit_store_exclusive): Add rel parameter.
21758 Emit store-release versions when rel is true.
21759 (arm_split_compare_and_swap): Use acquire-release instructions
21761 of barriers when appropriate.
21762 (arm_split_atomic_op): Likewise.
21763 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21764 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21765 (VUNSPEC_SLX): Likewise.
21766 (VUNSPEC_LDA): Likewise.
21767 (VUNSPEC_STL): Likewise.
21768 * config/arm/sync.md (atomic_load<mode>): New pattern.
21769 (atomic_store<mode>): Likewise.
21770 (arm_load_acquire_exclusive<mode>): Likewise.
21771 (arm_load_acquire_exclusivesi): Likewise.
21772 (arm_load_acquire_exclusivedi): Likewise.
21773 (arm_store_release_exclusive<mode>): Likewise.
21775 2013-03-25 Catherine Moore <clm@codesourcery.com>
21777 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21778 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21779 * config/mip/predicates.md (lwsp_swsp_operand,
21780 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21781 sb16_operand, db4_operand, db7_operand, ib3_operand,
21782 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21783 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21784 andi16_operand): New predicates.
21785 * config/mips/mips.md (compression): New attribute.
21786 (enabled): New attribute.
21787 (length): Consider compression in computing length.
21788 (shift_compression): New code attribute.
21789 (*add<mode>3): New operands. Record compression.
21790 (sub<mode>3): Likewise.
21791 (one_cmpl<mode>2): Likewise.
21792 (*and<mode>3): Likewise.
21793 (*ior<mode>3): Likewise.
21794 (unnamed pattern for xor): Likewise.
21795 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21796 (*<optab><mode>3): Likewise.
21797 (*mov<mode>_internal: Likewise.
21798 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21799 (mips_unsigned_immediate_p): New.
21800 (umips_lwsp_swsp_address_p): New.
21801 (m16_based_address_p): New.
21802 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21803 (mips_unsigned_immediate_p): New prototype.
21804 (lwsp_swsp_address_p): New prototype.
21805 (m16_based_address_p): New prototype.
21806 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21807 (mips_signed_immediate_p): New function.
21808 (m16_based_address_p): New function.
21809 (lwsp_swsp_address_p): New function.
21810 (mips_print_operand_punctuation): Recognize short delay slot insns
21811 for microMIPS.add<mode>3"
21813 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21816 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21817 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21819 2013-03-25 Richard Biener <rguenther@suse.de>
21821 PR tree-optimization/56689
21822 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21825 2013-03-25 Richard Biener <rguenther@suse.de>
21827 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21829 (memory_references): Likewise.
21830 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21831 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21832 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21833 (gather_mem_refs_in_loops): Fold into ...
21834 (analyze_memory_references): ... this. Move initialization
21835 to tree_ssa_lim_initialize.
21836 (fill_always_executed_in): Rename to ...
21837 (fill_always_executed_in_1): ... this.
21838 (fill_always_executed_in): Move contains_call computation to
21839 this new function from ...
21840 (tree_ssa_lim_initialize): ... here.
21841 (tree_ssa_lim): Call fill_always_executed_in.
21843 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21845 * postreload.c (reload_combine): Fix code detecting returns.
21847 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21849 * function.c (emit_use_return_register_into_block): On cc0 targets,
21850 do not emit the sequence between cc0 setter and user.
21852 2013-03-25 Kai Tietz <ktietz@redhat.com>
21854 * config/i386/predicates.md (local_symbolic_operand): Interpret
21855 dll-imported symbols as none-local.
21857 2013-03-25 Richard Biener <rguenther@suse.de>
21859 * tree-ssa-loop-im.c (struct depend): Remove.
21860 (struct lim_aux_data): Make depends a vec of gimples.
21861 (free_lim_aux_data): Adjust.
21862 (add_dependency): Likewise.
21863 (set_level): Likewise.
21865 2013-03-25 Richard Biener <rguenther@suse.de>
21867 PR middle-end/56434
21868 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21869 the pointer returned by calls with ECF_MALLOC set.
21871 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21873 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21875 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21877 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21878 using MMXMODE mode iterator.
21879 (*move<mode>_internal): Merge with *movv2sf_internal and
21880 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21882 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21884 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21885 (record_last_mem_set_info): Likewise.
21887 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21888 of XNEWVEC followed by memset.
21889 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21891 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21893 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21894 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21895 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21896 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21897 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21898 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21899 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21900 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21901 BARRIER_P instead of GET_CODE.
21903 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21905 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21906 inaccuracy in the probing code.
21908 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21909 (ctrapdi4): Likewise.
21911 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21913 * calls.c (expand_call): Add missing guard to code handling return
21914 of non-BLKmode structures in MSB.
21915 * function.c (expand_function_end): Likewise.
21917 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21919 * combine.c (try_combine): Adjust comment. Do not add the set of
21920 insn #0 if the destination indirectly is set or dies in insn #2.
21921 Tidy up code to distribute a new note.
21923 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21925 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21926 also for alternatives 16 and 17.
21928 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21930 * config/i386/sse.md (*mov<mode>_internal): Merge with
21931 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21932 Emit insn template depending on type attribute. Use
21933 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21934 movd instead of movq mnemonic for interunit moves. Rewrite mode
21935 attribute calculation. Remove unit attribute calculation.
21936 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21937 Set prefix_data16 attribute for DImode ssemov types.
21938 Use Ym instead of y for SSE-MMX conversion alternatives.
21939 Reorder operand constraints.
21941 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21943 * df.h (df_insn_delete): Adjust prototype.
21944 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21945 and let it decide whether mark the basic block dirty.
21946 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21947 * df-scan.c (df_insn_info_delete): New helper function, split
21948 off from df_insn_delete.
21949 (df_scan_free_bb_info): Use it.
21950 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21952 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21953 that the insn is actually an insn and it has a non-NULL basic block.
21954 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21956 2013-03-22 Richard Biener <rguenther@suse.de>
21958 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21960 (mem_ref_alloc): Do not allocate them.
21961 (refs_independent_p): Do not query or maintain a cache.
21963 2013-03-22 Richard Biener <rguenther@suse.de>
21965 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21966 (gather_mem_refs_in_loops): Do not compute it.
21967 (analyze_memory_references): Do not allocate it.
21968 (tree_ssa_lim_finalize): Do not free it.
21969 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21971 2013-03-22 Richard Biener <rguenther@suse.de>
21973 * is-a.h (as_a): Use gcc_checking_assert.
21975 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21977 * config/aarch64/aarch64.c (aarch64_print_operand): New
21978 format specifier for printing a constant in hex.
21979 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21980 format specifier for printing second operand.
21982 2013-03-22 Richard Biener <rguenther@suse.de>
21984 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21986 (gather_mem_refs_in_loops): Perform store accumulation here.
21987 (create_vop_ref_mapping_loop): Remove.
21988 (create_vop_ref_mapping): Likewise.
21989 (analyze_memory_references): Initialize refs_stored_in_loop.
21990 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21991 (record_indep_loop): Remove.
21992 (record_dep_loop): New function.
21993 (ref_indep_loop_p_1): Adjust to only walk over references
21994 in the loop, not its subloops.
21995 (ref_indep_loop_p): Rename to ...
21996 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21997 maintaining a more fine-grained cache.
21998 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21999 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22001 2013-03-22 Richard Biener <rguenther@suse.de>
22003 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22004 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22005 aggregate mem_ref_loc.
22006 (free_mem_ref_locs): Inline into ...
22007 (memref_free): ... this and adjust.
22008 (mem_ref_alloc): Adjust.
22009 (mem_ref_locs_alloc): Remove.
22010 (record_mem_ref_loc): Adjust.
22011 (get_all_locs_in_loop): Rewrite into ...
22012 (for_all_locs_in_loop): ... this iterator.
22013 (rewrite_mem_ref_loc): New functor.
22014 (rewrite_mem_refs): Use for_all_locs_in_loop.
22015 (sm_set_flag_if_changed): New functor.
22016 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22017 (ref_always_accessed): New functor.
22018 (ref_always_accessed_p): Use for_all_locs_in_loop.
22020 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22022 * tree-pass.h (PROP_gimple_lvec): New.
22023 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22024 (init_optimization_passes): Move pass_lower_vector.
22025 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22027 (pass_lower_vector): Provide PROP_gimple_lvec.
22028 (pass_lower_vector_ssa): Likewise.
22029 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22031 2013-03-21 Mark Wielaard <mjw@redhat.com>
22033 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22035 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22037 * config/i386/i386.md (*movdi_internal): Disparage slightly
22038 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22039 conversion alternatives.
22041 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22043 PR middle-end/48087
22044 * diagnostic.def (DK_WERROR): New kind.
22045 * diagnostic.h (werrorcount): Define.
22046 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22047 promoted to DK_ERROR, increment DK_WERROR counter instead of
22049 * toplev.c (toplev_main): Call print_ignored_options even if
22050 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22051 even if just werrorcount is non-zero.
22054 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22056 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22057 (string_cst_pool_decl): New function.
22058 (optimize_one_addr_into_implicit_ptr): New function.
22059 (resolve_addr_in_expr): Optimize DWARF location expression
22060 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22061 which doesn't live in memory, but has DW_AT_location or
22062 DW_AT_const_value, or refers to a string literal, into
22063 DW_OP_GNU_implicit_pointer.
22064 (optimize_location_into_implicit_ptr): New function.
22065 (resolve_addr): If removing DW_AT_location of a variable because
22066 it was DW_OP_addr of address of the variable, but the variable doesn't
22067 live in memory, try to emit const value attribute for the initializer.
22069 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22071 * tree.h (VECTOR_TYPE_P): New macro.
22072 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22073 TYPE_MODE): Use it.
22074 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22075 VEC_COND_EXPR cannot be lvalues.
22076 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22078 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22080 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22081 Restrict the transformation to equal modes.
22083 2013-03-21 Richard Biener <rguenther@suse.de>
22085 PR tree-optimization/39326
22086 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22087 (MEM_ANALYZABLE): Adjust.
22088 (record_mem_ref_loc): Move bitmap ops ...
22089 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22090 unanalyzable refs, do not record locations for it.
22091 (analyze_memory_references): Allocate ref zero as shared
22093 (refs_independent_p): Do not test for unanalyzed mems here.
22094 (ref_indep_loop_p_1): Special-case disambiguation against
22095 the unanalyzed ref.
22096 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22098 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22100 * config/arm/arm-protos.h (tune_params): Add
22101 prefer_neon_for_64bits field.
22102 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22103 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22104 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22105 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22106 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22107 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22108 (arm_option_override): Handle -mneon-for-64bits new option.
22109 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22110 (prefer_neon_for_64bits): Declare new variable.
22111 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22112 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22113 (arch_enabled): Handle new arch types. Remove support for onlya8
22115 (one_cmpldi2): Use new arch names.
22116 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22117 * config/arm/arm.opt (mneon-for-64bits): Add option.
22118 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22119 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22120 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22122 * doc/invoke.texi (-mneon-for-64bits): Document.
22124 2013-03-21 Richard Biener <rguenther@suse.de>
22126 PR tree-optimization/39326
22127 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22128 (sort_bbs_in_loop_postorder_cmp): New function.
22129 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22131 2013-03-21 Richard Biener <rguenther@suse.de>
22133 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22134 (vect_insert_into_interleaving_chain): Likewise.
22135 (vect_drs_dependent_in_basic_block): Inline ...
22136 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22138 (vect_analyze_data_ref_dependence): ... here. Simplify.
22139 (vect_check_interleaving): Simplify.
22140 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22141 (vect_slp_analyze_data_ref_dependences): ... this new function.
22142 (dr_group_sort_cmp): New function.
22143 (vect_analyze_data_ref_accesses): Compute data-reference groups
22144 here instead of in vect_analyze_data_ref_dependence. Use
22145 a more efficient algorithm.
22146 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22147 vect_slp_analyze_data_ref_dependences. Call
22148 vect_analyze_data_ref_accesses earlier.
22149 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22150 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22151 (vect_slp_analyze_data_ref_dependences): New prototype.
22153 2013-03-21 Richard Biener <rguenther@suse.de>
22155 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22156 ref is stored in the loop.
22157 (find_refs_for_sm): Walk only over all stores.
22158 (store_motion_loop): Allocate from lim_bitmap_obstack.
22159 (store_motion): Likewise.
22161 2013-03-21 Richard Biener <rguenther@suse.de>
22163 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22164 Update virtual SSA form.
22166 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22168 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22169 * configure: Regenerate.
22170 * config.in: Regenerate.
22171 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22172 if !HAVE_LD_EH_FRAME_CIEV3.
22174 2013-03-21 Richard Biener <rguenther@suse.de>
22176 * tree-cfg.c (verify_expr_no_block): New function.
22177 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22178 nor DECL_VALUE_EXPR have locations with associated blocks.
22179 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22180 (clear_unused_block_pointer): Remove code dealing with
22181 blocks in DECL_DEBUG_EXPR locations.
22183 2013-03-21 Richard Biener <rguenther@suse.de>
22185 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22186 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22187 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22188 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22189 instead of DECL_DEBUG_EXPR_IS_FROM.
22190 * gimplify.c (gimplify_modify_expr): Likewise.
22191 * tree-cfg.c (verify_expr_location_1): Likewise.
22192 * tree-complex.c (create_one_component_var): Likewise.
22193 * tree-sra.c (create_access_replacement): Likewise.
22194 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22195 (clear_unused_block_pointer): Likewise.
22196 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22197 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22198 * var-tracking.c (var_debug_decl): Likewise.
22199 (track_expr_p): Likewise.
22200 * tree-inline.c (add_local_variables): Likewise. Set
22201 DECL_HAS_DEBUG_EXPR_P after copying it.
22202 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22203 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22205 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22208 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22209 * configure: Regenerate.
22210 * config.in: Regenerate.
22211 * config/i386/i386.md (*movdf_internal): Use
22212 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22213 movd instead of movq mnemonic for interunit moves.
22214 (*movdi_internal): Ditto.
22216 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22218 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22219 (abd<mode>_3): New pattern.
22220 (aba<mode>_3): New pattern.
22221 (fabd<mode>_3): New pattern.
22223 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22225 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22226 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22227 occurrence of REGISTER_PREFIX as its empty string.
22229 2013-03-20 Jeff Law <law@redhat.com>
22231 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22232 addititional equivalences for equality comparisons between an SSA_NAME
22233 and a constant where the SSA_NAME was set from a widening conversion.
22235 2013-03-20 Walter Lee <walt@tilera.com>
22237 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22239 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22241 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22242 depending on type attribute.
22243 (*movti_internal): Ditto.
22244 (*movtf_internal): Ditto.
22245 (*movxf_internal): Ditto.
22246 (*movdf_internal): Ditto.
22247 (*movsf_internal): Ditto.
22249 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22251 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22252 maybe_vex for sselog1 and ssemov types.
22253 (*movdi_internal): Reorder operand constraints.
22254 (*movsi_internal): Ditto. Set prefix attribute to
22255 maybe_vex for sselog1 and ssemov types.
22256 (*movtf_internal): Set prefix attribute to maybe_vex
22257 for sselog1 and ssemov types.
22258 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22259 DImode ssemov types. Reorder operand constraints.
22260 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22261 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22262 attribute for SImode ssemov types. Reorder operand constraints.
22264 2013-03-20 Martin Jambor <mjambor@suse.cz>
22266 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22267 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22269 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22271 * config/rs6000/predicates.md (indexed_address, update_address_mem
22272 update_indexed_address_mem): New predicates.
22273 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22274 attribute for load/store instructions.
22275 * config/rs6000/dfp.md (movsd_store): Likewise.
22276 (movsd_load): Likewise.
22277 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22278 (unnamed HI->DI extend define_insn): Likewise.
22279 (unnamed SI->DI extend define_insn): Likewise.
22280 (unnamed QI->SI extend define_insn): Likewise.
22281 (unnamed QI->HI extend define_insn): Likewise.
22282 (unnamed HI->SI extend define_insn): Likewise.
22283 (unnamed HI->SI extend define_insn): Likewise.
22284 (extendsfdf2_fpr): Likewise.
22285 (movsi_internal1): Likewise.
22286 (movsi_internal1_single): Likewise.
22287 (movhi_internal): Likewise.
22288 (movqi_internal): Likewise.
22289 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22290 attribute for load/store instructions.
22291 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22293 (mov<mode>_softfloat): Likewise.
22294 (mov<mode>_hardfloat32): Likewise.
22295 (mov<mode>_hardfloat64): Likewise.
22296 (mov<mode>_softfloat64): Likewise.
22297 (movdi_internal32): Likewise.
22298 (movdi_internal64): Likewise.
22299 (probe_stack_<mode>): Likewise.
22301 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22303 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22304 floating point, and decimal floating point to reload iterator.
22306 * config/rs6000/constraints.md (wl constraint): New constraints to
22307 return FLOAT_REGS if certain options are used to reduce the number
22308 of separate patterns that exist in the file.
22309 (wx constraint): Likewise.
22310 (wz constraint): Likewise.
22312 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22313 -mdebug=reg, print wg, wl, wx, and wz constraints.
22314 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22315 Initialize the reload functions for 64-bit binary/decimal floating
22317 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22318 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22319 create the buffer on the stack to overcome not having a 32-bit
22321 (rs6000_emit_move): Likewise.
22322 (rs6000_secondary_memory_needed_rtx): Likewise.
22323 (rs6000_alloc_sdmode_stack_slot): Likewise.
22324 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22325 via xxlxor, just like DFmode 0.0.
22327 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22328 define as 1 if we are running on a power7 or newer.
22329 (enum r6000_reg_class_enum): Add new constraints.
22331 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22332 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22333 with other moves by using conditional constraits (wg). Use LFIWZX
22334 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22335 (movsd splitter): Likewise.
22336 (movsd_hardfloat): Likewise.
22337 (movsd_softfloat): Likewise.
22339 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22340 binary and decimal floating point moves.
22341 (fmove_ok): New attributes to combine binary and decimal floating
22342 point moves, and to combine power6x (mfpgpr) moves along normal
22344 (real_value_to_target): Likewise.
22345 (f32_lr): Likewise.
22346 (f32_lm): Likewise.
22347 (f32_li): Likewise.
22348 (f32_sr): Likewise.
22349 (f32_sm): Likewise.
22350 (f32_si): Likewise.
22351 (movsf): Combine binary and decimal floating point moves. Combine
22352 power6x (mfpgpr) moves with other moves by using conditional
22353 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22354 (mov<mode> for SFmode/SDmode); Likewise.
22355 (SFmode/SDmode splitters): Likewise.
22356 (movsf_hardfloat): Likewise.
22357 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22358 (movsf_softfloat): Likewise.
22359 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22361 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22362 wx and wz constraints.
22364 * config/rs6000/constraints.md (wg constraint): New constraint to
22365 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22367 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22370 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22371 -mdebug=reg, print wg, wl, wx, and wz constraints.
22372 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22373 Initialize the reload functions for 64-bit binary/decimal floating
22375 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22376 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22377 create the buffer on the stack to overcome not having a 32-bit
22379 (rs6000_emit_move): Likewise.
22380 (rs6000_secondary_memory_needed_rtx): Likewise.
22381 (rs6000_alloc_sdmode_stack_slot): Likewise.
22382 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22383 via xxlxor, just like DFmode 0.0.
22385 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22386 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22387 with other moves by using conditional constraits (wg). Use LFIWZX
22388 and STFIWX for loading SDmode on power7.
22389 (movdd splitters): Likewise.
22390 (movdd_hardfloat32): Likewise.
22391 (movdd_softfloat32): Likewise.
22392 (movdd_hardfloat64_mfpgpr): Likewise.
22393 (movdd_hardfloat64): Likewise.
22394 (movdd_softfloat64): Likewise.
22396 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22397 64-bit binary and decimal floating point moves.
22398 (FMOVE64X): Likewise.
22399 (movdf): Combine 64-bit binary and decimal floating point moves.
22400 Combine power6x (mfpgpr) moves with other moves by using
22401 conditional constraits (wg).
22402 (mov<mode> for DFmode/DDmode): Likewise.
22403 (DFmode/DDmode splitters): Likewise.
22404 (movdf_hardfloat32): Likewise.
22405 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22406 (movdf_softfloat32): Likewise.
22407 (movdf_hardfloat64_mfpgpr): Likewise.
22408 (movdf_hardfloat64): Likewise.
22409 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22410 (movdf_softfloat64): Likewise.
22411 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22412 (reload_<mode>_load): Move to later in the file so they aren't in
22413 the middle of the floating point move insns.
22414 (reload_<mode>_store): Likewise.
22416 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22419 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22420 constraint if -mdebug=reg.
22421 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22422 Enable using dd reload support if needed.
22424 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22425 binary and decimal floating point moves in rs6000.md.
22426 (movtd_internal): Likewise.
22428 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22429 decimal floating point moves.
22431 (movtf_internal): Likewise.
22432 (mov<mode>_internal, TDmode/TFmode): Likewise.
22433 (movtf_softfloat): Likewise.
22434 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22436 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22437 movdi_internal64, using wg constraint for move direct operations.
22438 (movdi_internal64): Likewise.
22440 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22441 MODES_TIEABLE_P for selected modes. Print the numerical value of
22442 the various virtual registers. Use GPR/FPR first/last values,
22443 instead of hard coding the register numbers. Print which modes
22444 have reload functions registered.
22445 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22446 settings before/after setting cpu, target and subtarget settings.
22447 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22448 and for secondary reload failures in rs6000_secondary_reload_inner.
22449 (rs6000_secondary_reload_fail): Likewise.
22450 (rs6000_secondary_reload_inner): Likewise.
22452 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22453 macros for first/last GPR and FPR registers.
22454 (LAST_GPR_REGNO): Likewise.
22455 (FIRST_FPR_REGNO): Likewise.
22456 (LAST_FPR_REGNO): Likewise.
22458 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22459 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22460 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22461 (vcond<mode><mode>): Likewise.
22462 (vcondu<mode><mode>): Likewise.
22463 (vector_gtu<mode>): Likewise.
22464 (vector_gte<mode>): Likewise.
22465 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22466 to prevent the compiler from converting DImode operations to TImode.
22467 (ior<mode>3): Likewise.
22468 (and<mode>3): Likewise.
22469 (one_cmpl<mode>2): Likewise.
22470 (nor<mode>3): Likewise.
22471 (andc<mode>3): Likewise.
22473 * config/rs6000/constraints.md (wt constraint): New constraint
22474 that returns VSX_REGS if TImode is allowed in VSX registers.
22476 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22477 constant under VSX.
22479 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22480 similar to TImode, but it is restricted to being in the GPRs.
22482 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22483 TImode to occupy a single VSX register.
22485 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22486 -mvsx-timode for power7/power8.
22487 (power7 cpu): Likewise.
22488 (power8 cpu): Likewise.
22490 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22491 sure that TFmode/TDmode take up two registers if they are ever
22492 allowed in the upper VSX registers.
22493 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22495 (rs6000_init_hard_regno_mode_ok): Likewise.
22496 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22497 constraint. Print if LRA is turned on.
22498 (rs6000_option_override_internal): Give an error if -mvsx-timode
22499 and VSX is not enabled.
22500 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22501 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22502 to reg+offset addressing. Use PTImode when checking offset
22503 addresses for validity.
22504 (reg_offset_addressing_ok_p): Likewise.
22505 (rs6000_legitimate_offset_address_p): Likewise.
22506 (rs6000_legitimize_address): Likewise.
22507 (rs6000_legitimize_reload_address): Likewise.
22508 (rs6000_legitimate_address_p): Likewise.
22509 (rs6000_eliminate_indexed_memrefs): Likewise.
22510 (rs6000_emit_move): Likewise.
22511 (rs6000_secondary_reload): Likewise.
22512 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22513 reloads to fpr registers to continue to use reg+offset addressing,
22514 but 64-bit reloads to altivec registers need reg+reg addressing.
22515 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22516 it. Treat LO_SUM like a PLUS operation.
22517 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22518 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22519 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22520 registers to share a register with a smaller sized type, since VSX
22521 puts scalars in the upper 64-bits.
22522 (print_operand): Add support for PTImode.
22523 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22524 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22525 registers, but don't have arithmetic support.
22526 (rs6000_memory_move_cost): Add test for VSX.
22527 (rs6000_opt_masks): Add -mvsx-timode.
22529 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22532 (VSr): Use wt constraint for TImode.
22533 (VSv): Drop TImode support.
22534 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22535 (vsx_movti_64bit): Likewise.
22536 (vsx_movti_32bit): Likewise.
22537 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22538 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22539 one '?' on the appropriate output constraint. Do not allow TImode
22540 logical operations on 32-bit systems.
22541 (vsx_ior<mode>3): Likewise.
22542 (vsx_xor<mode>3): Likewise.
22543 (vsx_one_cmpl<mode>2): Likewise.
22544 (vsx_nor<mode>3): Likewise.
22545 (vsx_andc<mode>3): Likewise.
22546 (vsx_concat_<mode>): Likewise.
22547 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22549 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22550 OPTION_MASK_VSX_TIMODE.
22551 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22552 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22554 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22555 (TI2 iterator): New iterator for TImode, PTImode.
22556 (wd mode attribute): Add values for vector types.
22557 (movti_string): Replace TI move operations with operations for TImode
22558 and PTImode. Add support for TImode being allowed in VSX registers.
22559 (mov<mode>_string, TImode/PTImode): Likewise.
22560 (movti_ppc64): Likewise.
22561 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22562 (TI mode splitters): Likewise.
22564 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22567 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22569 PR tree-optimization/56355
22570 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22571 Also handle integers with undefined overflow.
22573 2013-03-20 Catherine Moore <clm@codesourcery.com>
22574 Maciej W. Rozycki <macro@codesourcery.com>
22575 Tom de Vries <tom@codesourcery.com>
22576 Nathan Sidwell <nathan@codesourcery.com>
22577 Iain Sandoe <iain@codesourcery.com>
22578 Nathan Froyd <froydnj@codesourcery.com>
22579 Chao-ying Fu <fu@mips.com>
22581 * doc/extend.texi: (micromips, nomicromips, nocompression):
22582 Document new function attributes.
22583 * doc/invoke.texi (minterlink-compressed, mmicromips,
22584 m14k, m14ke, m14kec): Document new options.
22585 (minterlink-mips16): Update documentation.
22586 * doc/md.texi (ZC, ZD): Document new constraints.
22587 * configure.ac (gcc_cv_as_micromips): Check if linker
22588 supports the .set micromips directive.
22589 * configure: Regenerate.
22590 * config.in: Regenerate.
22591 * config/mips/mips-tables.opt: Regenerate.
22592 * config/mips/micromips.md: New file.
22593 * constraints.md (ZC, ZD): New constraints.
22594 * config/mips/predicates.md (movep_src_register): New predicate.
22595 (movep_src_operand): New predicate.
22596 (non_volatile_mem_operand): New predicate.
22597 * config/mips/mips.md (multimem): New type.
22598 (length): Differentiate between 17-bit and 18-bit branch offsets.
22599 (MOVEP1, MOVEP2): New mode iterator.
22600 (mov_<load>l): Use ZC constraint.
22601 (mov_<load>r): Likewise.
22602 (mov_<store>l): Likewise.
22603 (mov_<store>r): Likewise.
22604 (*branch_equality<mode>_inverted): Add microMIPS support.
22605 (*branch_equality<mode>): Likewise.
22606 (*jump_absolute): Likewise.
22607 (indirect_jump_<mode>): Likewise.
22608 (tablejump_<mode>): Likewise.
22609 (<optab>_internal): Likewise.
22610 (sibcall_internal): Likewise.
22611 (sibcall_value_internal): Likewise.
22612 (prefetch): Use constraint ZD.
22613 * config/mips/mips.opt (minterlink-compressed): New option.
22614 (minterlink-mips16): Now an alias for minterlink-compressed.
22615 (mmicromips): New option.
22616 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22617 (compare_and_swap_12): Likewise.
22618 (sync_add<mode>): Likewise.
22619 (sync_<optab>_12): Likewise.
22620 (sync_old_<optab>_12): Likewise.
22621 (sync_new_<optab>_12): Likewise.
22622 (sync_nand_12): Likewise.
22623 (sync_old_nand_12): Likewise.
22624 (sync_new_nand_12): Likewise.
22625 (sync_sub<mode>): Likewise.
22626 (sync_old_add<mode>): Likewise.
22627 (sync_old_sub<mode>): Likewise.
22628 (sync_new_add<mode>): Likewise.
22629 (sync_new_sub<mode>): Likewise.
22630 (sync_<optab><mode>): Likewise.
22631 (sync_old_<optab><mode>): Likewise.
22632 (sync_new_<optab><mode>): Likewise.
22633 (sync_nand<mode>): Likewise.
22634 (sync_old_nand<mode>): Likewise.
22635 (sync_new_nand<mode>): Likewise.
22636 (sync_lock_test_and_set<mode>): Likewise.
22637 (test_and_set_12): Likewise.
22638 (atomic_compare_and_swap<mode>): Likewise.
22639 (atomic_exchange<mode>_llsc): Likewise.
22640 (atomic_fetch_add<mode>_llsc): Likewise.
22641 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22642 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22643 (umips_save_restore_pattern_p): Likewise.
22644 (umips_load_store_pair_p): Likewise.
22645 (umips_output_load_store_pair): Likewise.
22646 (umips_movep_target_p): Likewise.
22647 (umips_12bit_offset_address_p): Likewise.
22648 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22649 (mips_base_mips16): Rename this...
22650 (mips_base_compression_flags): ...to this. Update all uses.
22651 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22652 (mips_mips16_decl_p): Delete.
22653 (mips_nomips16_decl_p): Delete.
22654 (mips_get_compress_on_flags): New function.
22655 (mips_get_compress_off_flags): New function.
22656 (mips_get_compress_mode): New function.
22657 (mips_get_compress_on_name): New function.
22658 (mips_get_compress_off_name): New function.
22659 (mips_insert_attributes): Support multiple compression types.
22660 (mips_merge_decl_attributes): Likewise.
22661 (umips_12bit_offset_address_p): New function.
22662 (mips_start_function_definition): Emit .set micromips directive.
22663 (mips_call_may_need_jalx_p): New function.
22664 (mips_function_ok_for_sibcall): Add microMIPS support.
22665 (mips_print_operand_punctuation): Support short delay slots and
22667 (umips_swm_mask, umips_swm_encoding): New.
22668 (umips_build_save_restore): New function.
22669 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22670 (was_mips16_p): Remove.
22671 (old_compression_mode): New.
22672 (mips_set_compression_mode): New function.
22673 (mips_set_current_function): Add microMIPS support.
22674 (mips_option_override): Likewise.
22675 (umips_save_restore_pattern_p): New function.
22676 (umips_output_save_restore): New function.
22677 (umips_load_store_pair_p_1): New function.
22678 (umips_load_store_pair_p): New function.
22679 (umips_output_load_store_pair_1): New function.
22680 (umips_output_load_store_pair): New function.
22681 (umips_movep_target_p) New function.
22682 (mips_prepare_pch_save): Add microMIPS support.
22683 * config/mips/mips.h (TARGET_COMPRESSION): New.
22684 (TARGET_CPU_CPP_BUILTINS): Update macro
22685 to use new compression flags and to support microMIPS.
22686 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22687 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22688 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22689 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22690 (ASM_SPEC): Support mmicromips and mno-micromips.
22691 (M16STORE_REG_P): New macro.
22692 (MIPS_CALL): Support TARGET_MICROMIPS.
22693 (MICROMIPS_J): New macro.
22694 (mips_base_mips16): Rename this...
22695 (mips_base_compression_flags): ...to this.
22696 (UMIPS_12BIT_OFFSET_P): New macro.
22697 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22698 (MULTILIB_DIRNAMES): Likewise.
22699 2013-03-20 Richard Biener <rguenther@suse.de>
22701 PR tree-optimization/56661
22702 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22703 the result does not have to be distinct.
22705 2013-03-20 Richard Biener <rguenther@suse.de>
22707 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22710 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22711 Steven Bosscher <steven@gcc.gnu.org>
22713 PR rtl-optimization/56605
22714 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22716 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22719 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22720 that require movd instead of movq.
22722 2013-03-20 Richard Biener <rguenther@suse.de>
22724 * tree-ssa-structalias.c (struct variable_info): Add pointer
22725 to the first field of an aggregate with sub-vars. Make
22726 this and the pointer to the next subfield its ID.
22727 (vi_next): New function.
22728 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22729 storedanything_id, integer_id): Increment by one.
22730 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22731 get_call_clobber_vi): Adjust.
22732 (solution_set_expand): Simplify and speedup.
22733 (solution_set_add): Inline into ...
22734 (set_union_with_increment): ... this. Adjust accordingly.
22735 (do_sd_constraint): Likewise.
22736 (do_ds_constraint): Likewise.
22737 (do_complex_constraint): Simplify.
22738 (build_pred_graph): Adjust.
22739 (solve_graph): Likewise. Simplify and speedup.
22740 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22741 get_constraint_for_component_ref, get_constraint_for_1,
22742 first_vi_for_offset, first_or_preceding_vi_for_offset,
22743 create_function_info_for, create_variable_info_for_1,
22744 create_variable_info_for, intra_create_variable_infos): Adjust.
22745 (init_base_vars): Push NULL for ID zero.
22746 (compute_points_to_sets): Adjust.
22748 2013-03-20 Richard Biener <rguenther@suse.de>
22750 * cfgloop.c (verify_loop_structure): Streamline and avoid
22751 ICEing on corrupt loop tree.
22752 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22755 2013-03-20 Richard Biener <rguenther@suse.de>
22757 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22758 check whether an SSA update is needed.
22760 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22762 * config/mips/constraints.md (T): Rename to...
22766 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22767 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22769 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22771 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22772 (*subsi3_carryin_uxtw): Likewise.
22774 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22776 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22777 (*rorsi3_insn_uxtw): Likewise.
22779 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22781 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22782 (*extrsi5_insn_uxtw): Likewise.
22784 2013-03-19 Richard Biener <rguenther@suse.de>
22786 PR tree-optimization/56273
22787 * passes.c (init_optimization_passes): Move second VRP after DOM.
22789 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22791 * config/i386/i386.md (*movti_internal): Merge from
22792 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22793 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22794 nox64 isa attributes.
22796 2013-03-18 Richard Biener <rguenther@suse.de>
22798 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22800 (merge_node_constraints): Likewise.
22801 (build_succ_graph): Likewise.
22802 (valid_graph_edge): Inline into single caller.
22803 (unify_nodes): Likewise. Use bitmap_set_bit return value
22805 (scc_visit): Fix formatting and variable use.
22806 (do_sd_constraint): Use gcc_checking_assert.
22807 (do_ds_constraint): Likewise.
22808 (do_complex_constraint): Likewise.
22809 (condense_visit): Likewise. Cleanup.
22810 (dump_pred_graph): New function.
22811 (perform_var_substitution): Dump the pred-graph before
22812 variable substitution.
22813 (find_equivalent_node): Use gcc_checking_assert.
22814 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22816 2013-03-18 Richard Biener <rguenther@suse.de>
22818 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22819 Remove cond_expr_stmt_list argument and do not gimplify the
22821 (vect_loop_versioning): Adjust.
22822 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22823 Cleanup to use less temporaries.
22824 (vect_create_data_ref_ptr): Cleanup.
22826 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22828 PR tree-optimization/56635
22829 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22830 require types_compatible_p types.
22832 2013-03-18 Nick Clifton <nickc@redhat.com>
22834 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22835 spurious backslash.
22837 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22838 Add missing line to comment describing function.
22840 2013-03-18 Richard Biener <rguenther@suse.de>
22842 PR tree-optimization/56210
22843 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22844 Handle string / character search functions.
22845 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22847 2013-03-18 Richard Biener <rguenther@suse.de>
22849 PR middle-end/56483
22850 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22851 and implement properly.
22852 * gimple.h (gimple_cond_single_var_p): Remove.
22854 2013-03-18 Richard Biener <rguenther@suse.de>
22856 * tree-data-ref.h (find_data_references_in_loop): Declare.
22857 * tree-data-ref.c (get_references_in_stmt): Use a stack
22858 vector pre-allocated in the callers.
22859 (find_data_references_in_stmt): Adjust.
22860 (graphite_find_data_references_in_stmt): Likewise.
22861 (create_rdg_vertices): Likewise.
22862 (find_data_references_in_loop): Export.
22863 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22864 Compute dependences here...
22865 (vect_analyze_data_refs): ...not here. When we encounter
22866 a non-vectorizable data reference in basic-block vectorization
22867 truncate the data reference vector. Do not bother to
22868 fixup data-dependence information for gather loads.
22869 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22870 of data references, as reported.
22872 2013-03-18 Richard Biener <rguenther@suse.de>
22874 PR tree-optimization/3713
22875 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22876 has_constants and expr.
22877 (stmt_has_constants): Properly valueize SSA names when deciding
22878 whether the stmt has constants.
22880 2013-03-18 Richard Biener <rguenther@suse.de>
22882 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22883 whole function when there is nothing to do.
22884 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22885 * tree-vectorizer.c (vectorize_loops): Update virtual and
22886 loop-closed SSA once.
22887 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22889 2013-03-18 Richard Biener <rguenther@suse.de>
22891 PR middle-end/56113
22892 * domwalk.c (bb_postorder): New global static.
22893 (cmp_bb_postorder): New function.
22894 (walk_dominator_tree): Replace scheme imposing an order for
22895 visiting dominator sons by one sorting them at the time they
22896 are pushed on the stack.
22898 2013-03-18 Richard Biener <rguenther@suse.de>
22900 PR tree-optimization/39326
22901 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22902 (struct mem_ref): Replace mem member with ao_ref typed member.
22903 (MEM_ANALYZABLE): Adjust.
22904 (memref_eq): Likewise.
22905 (mem_ref_alloc): Likewise.
22906 (gather_mem_refs_stmt): Likewise.
22907 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22908 (execute_sm_if_changed_flag_set): Adjust.
22909 (execute_sm): Likewise.
22910 (ref_always_accessed_p): Likewise.
22911 (refs_independent_p): Likewise.
22912 (can_sm_ref_p): Likewise.
22914 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22917 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22918 return 1 even for !unsignedp.
22920 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22922 * config/i386/i386.md (isa): Add x64 and nox64.
22923 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22924 (*pushtf): Enable *roF alternative for x64 isa only.
22925 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22926 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22927 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22928 (*movtf_internal): Merge from *movtf_internal_rex64 and
22929 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22930 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22931 nox64 isa attributes.
22932 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22933 nox64 isa attributes.
22934 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22936 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22938 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22940 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22943 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22944 class names. Remove trailing comma after "ALL_REGS".
22946 2013-03-16 Jan Hubicka <jh@suse.cz>
22948 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22949 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22950 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22951 of cgraph_get_create_node.
22952 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22954 2013-03-16 Jason Merrill <jason@redhat.com>
22957 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22958 with DW_AT_default_value.
22960 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22962 * BASE-VER: Set to 4.9.0.
22964 2013-03-14 Andi Kleen <ak@linux.intel.com>
22967 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22968 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22969 Document _x* TSX intrinsics.
22971 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22972 David Holsgrove <david.holsgrove@xilinx.com>
22974 * configure.ac: Add MicroBlaze TLS support detection.
22975 * configure: Regenerate.
22976 * config/microblaze/microblaze-protos.h
22977 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22978 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22979 * config/microblaze/microblaze.c (microblaze_address_type): Add
22980 ADDRESS_TLS and tls_reloc address types.
22981 (microblaze_address_info): Add tls_reloc.
22982 (TARGET_HAVE_TLS): Define.
22983 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22984 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22985 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22986 load_tls_operand, microblaze_call_tls_get_addr,
22987 microblaze_legitimize_tls_address): New functions.
22988 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22989 (get_base_reg): Use microblaze_tls_symbol_p.
22990 (microblaze_classify_address): Handle TLS.
22991 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22992 label_mentioned_p and microblaze_tls_referenced_p.
22993 (microblaze_legitimize_address): Handle TLS.
22994 (microblaze_address_insns): Handle ADDRESS_TLS.
22995 (pic_address_needs_scratch): Handle TLS.
22996 (print_operand_address): Handle TLS.
22997 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22998 (microblaze_expand_move): Handle TLS.
22999 (microblaze_legitimate_constant_p): Check
23000 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23001 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23002 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23003 (PIC_OFFSET_TABLE_REGNUM): Set.
23004 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23005 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23006 (addsi3, movsi_internal2, movdf_internal): Update constraints
23007 * config/microblaze/predicates.md (arith_plus_operand): Define
23008 (move_operand): Redefine as move_src_operand,
23009 check microblaze_tls_referenced_p.
23011 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23013 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23014 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23016 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23018 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23021 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23023 PR tree-optimization/53265
23024 * common.opt (Waggressive-loop-optimizations): New option.
23025 * tree-ssa-loop-niter.c: Include tree-pass.h.
23026 (do_warn_aggressive_loop_optimizations): New function.
23027 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23028 if number_of_latch_executions returned constant.
23029 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23030 early. If number_of_latch_executions returned constant, set
23031 nb_iterations_upper_bound back to it.
23032 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23034 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23035 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23037 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23038 (MULTILIB_OSDIRNAMES): Set.
23039 * genmultilib: If defaultosdirname doesn't start with :: , set
23040 defaultosdirname2 instead, clear it and emit two . multilib_raw
23041 entries instead of just one.
23043 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23045 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23046 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23047 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23048 (SUBTARGET_OVERRIDE_OPTIONS): New.
23050 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23053 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23054 (musermode): Convert to Var(TARGET_USERMODE).
23055 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23056 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23057 * config/sh/sh.c (sh_option_override): Use
23058 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23059 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23061 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23063 (udivsi3_i4_single, divsi3_i4_single): Use
23064 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23066 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23068 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23071 2013-03-13 Richard Biener <rguenther@suse.de>
23073 PR tree-optimization/56608
23074 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23075 calls when vectorizing basic-blocks.
23077 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23080 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23083 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23085 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23087 2013-03-11 Jan Hubicka <jh@suse.cz>
23090 * lto-streamer-out.c (output_symbol_p): Skip references from
23091 constructors of external variables.
23093 2013-03-11 Jan Hubicka <jh@suse.cz>
23095 PR middle-end/56571
23096 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23098 * emit-rtl.c (verify_rtx_sharing): Likewise.
23099 (copy_insn_1): Likewise.
23100 * rtl.c (copy_rtx): Likewise.
23102 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23105 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23106 output_operand_lossage message.
23108 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23111 * arm.c (shift_op): Validate RTL pattern on the fly.
23112 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23115 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23118 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23119 new call patterns that clobber %fr12.
23120 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23121 split and postreload patterns.
23122 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23123 registers %fr12 and %fr12R as call used.
23125 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23127 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23128 (canon_address, record_store, replace_read, check_mem_read_rtx,
23129 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23130 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23131 rest_of_handle_dse): Likewise.
23133 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23135 PR middle-end/56524
23136 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23138 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23139 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23140 (save_optabs_if_changed): Replace with...
23141 (init_tree_optimization_optabs): ...this.
23142 * optabs.c (save_optabs_if_changed): Rename to...
23143 (init_tree_optimization_optabs): ...this. Take the optimization node
23144 as argument. Do nothing if the base optabs are already correct.
23145 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23146 to recompute optabs.
23147 * function.h (function): Remove optabs field.
23148 * function.c (invoke_set_current_function_hook): Call
23149 init_tree_optimization_optabs. Use the result to initialize
23152 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23154 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23155 if GTMA_HAS_NO_INSTRUMENTATION.
23156 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23157 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23158 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23159 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23160 GTMA_HAS_NO_INSTRUMENTATION.
23162 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23164 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23167 2013-03-08 Marek Polacek <polacek@redhat.com>
23168 Jakub Jelinek <jakub@redhat.com>
23170 PR tree-optimization/56478
23171 * predict.c (is_comparison_with_loop_invariant_p): Change the
23172 type of loop_step to tree.
23173 (predict_loops): Adjust.
23174 (predict_iv_comparison): Perform the computations on double_ints.
23176 2013-03-08 Richard Biener <rguenther@suse.de>
23178 PR tree-optimization/56570
23179 * tree-cfg.c (verify_expr_location_1): Verify locations for
23181 * tree-sra.c (create_access_replacement): Strip locations
23182 from DECL_DEBUG_EXPRs.
23184 2013-03-08 Richard Biener <rguenther@suse.de>
23186 * tree-inline.c (expand_call_inline): Do not associate
23187 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23188 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23190 2013-03-08 Richard Biener <rguenther@suse.de>
23192 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23193 or block changes with -Og. Fix for location / block encoding
23194 changes and PHI arguments with locations.
23196 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23198 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23200 (struct output_info): Likewise.
23201 (register_overhead): Remove bad gcc_assert.
23202 (bitmap_find_bit): If there is only a single bitmap element, do not
23203 count a miss as a search.
23204 (print_statistics): Update for counter type changes.
23205 (dump_bitmap_statistics): Likewise. Print headers such that they
23206 are properly lined up with the printed counters.
23208 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23210 PR tree-optimization/56559
23211 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23212 check that it has only a single use.
23214 2013-03-07 Richard Biener <rguenther@suse.de>
23216 * doc/invoke.texi (fwhole-program): Discourage use in combination
23219 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23221 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23223 PR tree-optimization/56539
23224 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23225 instead of GSI_CONTINUE_LINKING as last argument to
23226 force_gimple_operand_gsi. Adjust function comment.
23228 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23231 PR middle-end/56548
23232 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23233 promoted mode, convert the result back to the original mode.
23235 2013-03-06 Richard Biener <rguenther@suse.de>
23237 PR middle-end/56294
23238 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23239 (insert_updated_phi_nodes_compare_uids): New function.
23240 (update_ssa): Sort symbols_to_rename after UID before
23241 traversing it to insert PHI nodes.
23243 2013-03-06 Richard Biener <rguenther@suse.de>
23245 PR middle-end/50494
23246 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23247 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23250 2013-02-13 Richard Biener <rguenther@suse.de>
23253 * varasm.c (output_constant_def_1): Get the decl representing
23254 the constant as argument.
23255 (output_constant_def): Wrap output_constant_def_1.
23256 (make_decl_rtl): Use output_constant_def_1 with the decl
23257 representing the constant.
23258 (build_constant_desc): Optionally re-use a decl already
23259 representing the constant.
23260 (tree_output_constant_def): Adjust.
23262 2013-03-06 Joey Ye <joey.ye@arm.com>
23265 * gcc.c (convert_white_space): New function.
23266 (main): Handles white space in function name.
23268 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23271 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23272 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23273 to SH_DIV_CALL_TABLE for TARGET_SH2.
23274 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23276 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23277 call-table options.
23279 2013-03-05 Sterling Augustine <saugustine@google.com>
23280 Cary Coutant <ccoutant@google.com>
23283 * dwarf2out.c (resolve_addr): Don't call
23284 remove_loc_list_addr_table_entries a second time for the same
23287 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23290 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23291 (avoid_complex_debug_insns): New function.
23292 (expand_debug_locations): Call it.
23294 PR rtl-optimization/56484
23295 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23296 lifetimes of hard registers on small register class machines.
23298 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23300 * config/microblaze/microblaze-protos.h: Rename
23301 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23302 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23304 (microblaze_fast_interrupt_function_p): New function.
23305 (microblaze_is_interrupt_handler): Rename to
23306 microblaze_is_interrupt_variant and add fast_interrupt check.
23307 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23308 (save_restore_insns): Likewise.
23309 (compute_frame_size): Likewise.
23310 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23311 (microblaze_globalize_label): Likewise.
23312 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23313 * config/microblaze/microblaze.md: Use wrapper
23314 microblaze_is_interrupt_variant.
23316 2013-03-05 Kai Tietz <ktietz@redhat.com>
23318 * sdbout.c (sdbout_one_type): Switch to current function's section
23319 supporting cold/hot.
23321 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23323 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23326 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23328 PR middle-end/56461
23329 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23330 if VALGRIND_GET_VBITS is defined, temporarily make object
23331 memory all defined, and restore previous valgrind addressability
23332 and definability afterwards. Free this_object at the end.
23334 PR middle-end/56461
23335 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23336 right before calling lra_create_live_ranges, also call it
23337 when clearing live_p. Only call lra_clear_live_ranges
23338 at the end if live_p.
23340 PR middle-end/56461
23341 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23343 2013-03-05 Richard Biener <rguenther@suse.de>
23345 PR tree-optimization/56521
23346 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23349 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23352 * except.h (remove_unreachable_eh_regions): New prototype.
23353 * except.c (remove_eh_handler_splicer): New function, split out
23354 of remove_eh_handler.
23355 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23356 warning about running it on many EH regions one at a time.
23357 (remove_unreachable_eh_regions_worker): New function, walk the
23358 EH tree in depth-first order and remove non-marked regions.
23359 (remove_unreachable_eh_regions): New function.
23360 * tree-eh.c (mark_reachable_handlers): New function, split out
23361 from remove_unreachable_handlers.
23362 (remove_unreachable_handlers): Use mark_reachable_handlers and
23363 remove_unreachable_eh_regions.
23364 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23365 and remove_unreachable_eh_regions.
23367 2013-03-05 Richard Biener <rguenther@suse.de>
23369 PR middle-end/56525
23370 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23371 not freeing them until the end.
23373 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23375 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23377 2013-03-05 Richard Biener <rguenther@suse.de>
23379 PR tree-optimization/56270
23380 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23381 of loads after scheduling an SLP instance.
23383 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23385 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23387 (check_gcc_parallelize): Run guality.exp as a separate job from
23388 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23389 struct-layout-1.exp with stackalign.exp.
23391 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23393 PR middle-end/56461
23394 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23395 load_index sbitmap even if some bit in it isn't set.
23397 PR middle-end/56461
23398 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23399 (discover_iteration_bound_by_body_walk): Change queues to
23400 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23401 spelling in comment. Call safe_push on queues[bound_index] directly.
23402 Release queues[queue_index] in every iteration unconditionally.
23403 Release bounds vector.
23405 PR middle-end/56461
23406 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23407 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23408 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23411 2013-03-05 Richard Biener <rguenther@suse.de>
23414 * tree-inline.c (remap_blocks_to_null): New function.
23415 (expand_call_inline): When expanding a call stmt without
23416 an associated block inline remap all callee blocks to NULL.
23418 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23420 PR rtl-optimization/56494
23421 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23422 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23423 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23425 PR middle-end/56461
23426 * sel-sched-ir.c (free_sched_pools): Release
23427 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23428 if succs_info_pool.max_top isn't -1.
23431 * opts.c (opts_obstack, opts_concat): Moved to...
23432 * opts-common.c (opts_obstack, opts_concat): ... here.
23434 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23436 PR middle-end/56461
23437 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23439 2013-03-04 Martin Jambor <mjambor@suse.cz>
23441 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23442 all appropriate places.
23444 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23446 PR tree-optimization/56424
23447 * ipa-split.c (split_function): Do not set the RSO flag if result is
23448 not by reference and its type is a register type.
23450 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23452 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23453 (microblaze_legitimate_pic_operand): Likewise
23454 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23455 new function microblaze_legitimate_pic_operand
23456 * config/microblaze/microblaze-protos.h
23457 (microblaze_legitimate_pic_operand): Declare.
23459 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23461 * config/microblaze/predicates.md (call_insn_simple_operand):
23462 New predicate for supported rtx code types.
23463 * config/microblaze/microblaze.md (call_internal1): Use
23464 call_insn_simple_operand predicate.
23466 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23468 PR middle-end/56461
23469 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23470 partitions.ordered_remove.
23472 PR middle-end/56461
23473 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23474 vec_oprnds0.create (1) for modifier == NONE.
23476 PR middle-end/56461
23477 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23478 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23479 vec_oprnds1 right before pushing anything to it for
23482 PR middle-end/56461
23483 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23484 set nbbs to 0 instead of having separate code path.
23485 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23486 instead of false as last argument if returning NULL.
23488 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23490 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23491 the attribute is now called "target" instead of "option".
23492 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23493 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23494 attribute/pragma name for TARGET_OPTION_VALID_P and
23495 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23496 * doc/tm.texi: Regenerated.
23498 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23500 * config/microblaze/microblaze.c:
23501 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23502 * config/microblaze/microblaze.h: Add -mxl-reorder to
23504 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23505 instructions emitted if TARGET_REORDER.
23506 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23507 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23510 2013-03-01 Xinliang David Li <davidxl@google.com>
23512 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23515 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23517 PR middle-end/56461
23518 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23519 vector even when returning true. Fix up function comment formatting.
23521 PR middle-end/56461
23522 * ira-build.c (ira_loop_nodes_count): New variable.
23523 (create_loop_tree_nodes): Initialize it.
23524 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23526 PR middle-end/56461
23527 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23528 method on dr_chain and result_chain.
23529 * tree-vect-stmts.c (vectorizable_store): Only call
23530 result_chain.create if j == 0.
23532 PR middle-end/56461
23533 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23534 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23535 before overwriting it.
23537 2013-03-01 Tobias Burnus <burnus@net-b.de>
23539 * doc/extended.texi (C Extensions): Change order in @menu
23541 (Other MIPS Built-in Functions): Move last MIPS entry before
23542 "picoChip Built-in Functions".
23543 (SH Built-in Functions): Move after RX Built-in Functions.
23544 * doc/gcc.texi (Introduction): Change order in @menu
23546 * doc/md.texi (Constraints): Ditto.
23547 * gty.texi (Type Information): Ditto.
23548 (User-provided marking routines for template types): Make
23550 * doc/invoke.texi (AArch64 Options): Move before
23551 "Adapteva Epiphany Options".
23553 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23554 Jakub Jelinek <jakub@redhat.com>
23557 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23558 no_address_safety_analysis attribute.
23559 * doc/extend.texi (no_address_safety_attribute): Rename to
23560 no_sanitize_address attribute, mention no_address_safety_analysis
23561 attribute as deprecated alias.
23563 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23565 PR middle-end/56461
23566 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23567 type to vec<vec<tree> > *.
23568 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23569 to be vec<tree> instead of vec<tree> *, set vec_defs
23570 to vNULL and call vec_defs.create (number_of_vects), adjust other
23572 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23573 vectorizable_condition): Adjust vect_get_slp_defs callers.
23575 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23577 * config/aarch64/aarch64.c
23578 (aarch64_float_const_representable): Remove unused variable.
23580 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23582 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23584 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23586 * config/aarch64/aarch64-builtins.c
23587 (aarch64_init_simd_builtins): Make static.
23589 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23591 * config/aarch64/aarch64.c
23592 (aarch64_simd_make_constant): Make static.
23594 2013-02-28 Martin Jambor <mjambor@suse.cz>
23596 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23597 with no initialization to the RHS of debug statements.
23599 2013-02-28 Martin Jambor <mjambor@suse.cz>
23601 PR tree-optimization/56294
23602 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23604 (get_access_replacement): Do not call create_access_replacement.
23605 Assert a replacement exists.
23606 (get_repl_default_def_ssa_name): Create the replacement declaration
23609 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23611 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23612 final_end_function.
23614 2013-02-28 Marek Polacek <polacek@redhat.com>
23616 PR rtl-optimization/56466
23617 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23618 if we're changing a loop.
23619 (peel_loops_completely): Likewise.
23621 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23624 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23626 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23629 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23630 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23631 INTX_FTYPE_FX, FX_FTYPE_INTX.
23632 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23634 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23636 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23637 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23638 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23639 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23640 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23641 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23642 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23643 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23644 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23645 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23646 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23647 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23648 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23649 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23650 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23651 (avrxmega6): Increase max flash segments from 5 to 6.
23652 * config/avr/t-multilib: Regenerate.
23653 * config/avr/avr-tables.opt: Regenerate.
23654 * doc/avr-mmcu.texi: Regenerate.
23656 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23658 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23659 (avr_device_to_arch): Rename to avr_device_to_ld.
23660 (avr_device_to_as): New prototype.
23661 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23662 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23663 * config/avr/driver-avr.c (avr_device_to_as): New.
23664 (avr_device_to_arch): Rename to avr_device_to_ld.
23666 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23668 PR middle-end/56461
23669 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23670 method on dr_chain and result_chain.
23672 PR middle-end/56461
23673 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23674 pointer_set_destroy on not_executed_last_iteration.
23676 PR middle-end/56461
23677 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23679 PR middle-end/56461
23680 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23681 FOR_EACH_DEFINED_FUNCTION when freeing state.
23683 PR middle-end/56461
23684 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23686 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23689 PR middle-end/56461
23690 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23691 known_aggs[i].items and release known_aggs vector.
23693 PR middle-end/56461
23694 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23696 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23698 * config/microblaze/microblaze.c (microblaze_emit_compare):
23699 Use xor for EQ/NE comparisions.
23700 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23701 (cbranchsf4): Adjust operator to comparison_operator.
23703 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23705 PR middle-end/56461
23706 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23708 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23709 vec_safe_push, always update *slot.
23710 (redirect_edge_var_map_clear): Use vec_free.
23711 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23712 (free_var_map_entry): Use vec_free.
23713 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23714 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23716 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23718 PR middle-end/45472
23719 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23720 when the may_trap_p bit of the exprs being merged differs.
23721 Reorder tests for speculativeness in the logical and operator.
23723 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23725 * incpath.c (add_standard_paths): Use reconcat instead of concat
23726 where appropriate and avoid leaking memory.
23728 * opts.h: Include obstack.h.
23729 (opts_concat): New prototype.
23730 (opts_obstack): New declaration.
23731 * opts.c (opts_concat): New function.
23732 (opts_obstack): New variable.
23733 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23734 (finish_options): Use opts_concat instead of concat
23735 and XOBNEWVEC instead of XNEWVEC.
23736 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23737 generate_option): Likewise.
23738 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23739 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23742 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23743 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23745 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23747 PR middle-end/56461
23748 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23750 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23752 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23753 (arm_block_move_unaligned_straight): Likewise.
23754 (arm_adjust_block_mem): Likewise.
23756 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23759 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23760 temp, cond and label.
23761 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23764 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23765 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23768 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23769 of prologue/epilogue functions.
23772 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23773 Remove unused variable cfa_offset.
23774 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23777 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23778 type promotion to unsigned.
23781 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23782 for HOST_WIDE_INT of 32 bit / same size as int.
23783 (arm_block_move_unaligned_straight): Likewise.
23784 (arm_adjust_block_mem): Likewise.
23787 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23790 2013-02-26 Marek Polacek <polacek@redhat.com>
23792 PR tree-optimization/56426
23793 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23795 2013-02-26 Richard Biener <rguenther@suse.de>
23798 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23799 unused variable loops.
23801 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23803 PR tree-optimization/56448
23804 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23805 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23806 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23807 later operands of the references, or even first operand for
23808 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23810 PR tree-optimization/56443
23811 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23812 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23813 to type_for_mode langhook.
23815 2013-02-25 Matt Turner <mattst88@gmail.com>
23817 * doc/invoke.texi: Document r4700.
23819 2013-02-25 Richard Biener <rguenther@suse.de>
23821 PR tree-optimization/56175
23822 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23824 (simplify_bitwise_binary): ... here. Also guard the conversion
23825 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23827 2013-02-25 Catherine Moore <clm@codesourcery.com>
23830 2013-02-24 Catherine Moore <clm@codesourcery.com>
23831 Maciej W. Rozycki <macro@codesourcery.com>
23832 Tom de Vries <tom@codesourcery.com>
23833 Nathan Sidwell <nathan@codesourcery.com>
23834 Iain Sandoe <iain@codesourcery.com>
23835 Nathan Froyd <froydnj@codesourcery.com>
23836 Chao-ying Fu <fu@mips.com>
23838 * doc/extend.texi: (micromips, nomicromips, nocompression):
23839 Document new function attributes.
23840 * doc/invoke.texi (minterlink-compressed, mmicromips,
23841 m14k, m14ke, m14kec): Document new options.
23842 (minterlink-mips16): Update documentation.
23843 * doc/md.texi (ZC, ZD): Document new constraints.
23844 * configure.ac (gcc_cv_as_micromips): Check if linker
23845 supports the .set micromips directive.
23846 * configure: Regenerate.
23847 * config.in: Regenerate.
23848 * config/mips/mips-tables.opt: Regenerate.
23849 * config/mips/micromips.md: New file.
23850 * constraints.md (ZC, AD): New constraints.
23851 * config/mips/predicates.md (movep_src_register): New predicate.
23852 (movep_src_operand): New predicate.
23853 (non_volatile_mem_operand): New predicate.
23854 * config/mips/mips.md (multimem): New type.
23855 (length): Differentiate between 17-bit and 18-bit branch offsets.
23856 (MOVEP1, MOVEP2): New mode iterator.
23857 (mov_<load>l): Use ZC constraint.
23858 (mov_<load>r): Likewise.
23859 (mov_<store>l): Likewise.
23860 (mov_<store>r): Likewise.
23861 (*branch_equality<mode>_inverted): Add microMIPS support.
23862 (*branch_equality<mode>): Likewise.
23863 (*jump_absolute): Likewise.
23864 (indirect_jump_<mode>): Likewise.
23865 (tablejump_<mode>): Likewise.
23866 (<optab>_internal): Likewise.
23867 (sibcall_internal): Likewise.
23868 (sibcall_value_internal): Likewise.
23869 (prefetch): Use constraint ZD.
23870 * config/mips/mips.opt (minterlink-compressed): New option.
23871 (minterlink-mips16): Now an alias for minterlink-compressed.
23872 (mmicromips): New option.
23873 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23874 (compare_and_swap_12): Likewise.
23875 (sync_add<mode>): Likewise.
23876 (sync_<optab>_12): Likewise.
23877 (sync_old_<optab>_12): Likewise.
23878 (sync_new_<optab>_12): Likewise.
23879 (sync_nand_12): Likewise.
23880 (sync_old_nand_12): Likewise.
23881 (sync_new_nand_12): Likewise.
23882 (sync_sub<mode>): Likewise.
23883 (sync_old_add<mode>): Likewise.
23884 (sync_old_sub<mode>): Likewise.
23885 (sync_new_add<mode>): Likewise.
23886 (sync_new_sub<mode>): Likewise.
23887 (sync_<optab><mode>): Likewise.
23888 (sync_old_<optab><mode>): Likewise.
23889 (sync_new_<optab><mode>): Likewise.
23890 (sync_nand<mode>): Likewise.
23891 (sync_old_nand<mode>): Likewise.
23892 (sync_new_nand<mode>): Likewise.
23893 (sync_lock_test_and_set<mode>): Likewise.
23894 (test_and_set_12): Likewise.
23895 (atomic_compare_and_swap<mode>): Likewise.
23896 (atomic_exchange<mode>_llsc): Likewise.
23897 (atomic_fetch_add<mode>_llsc): Likewise.
23898 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23899 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23900 (umips_save_restore_pattern_p): Likewise.
23901 (umips_load_store_pair_p): Likewise.
23902 (umips_output_load_store_pair): Likewise.
23903 (umips_movep_target_p): Likewise.
23904 (umips_12bit_offset_address_p): Likewise.
23905 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23906 (mips_base_mips16): Rename this...
23907 (mips_base_compression_flags): ...to this. Update all uses.
23908 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23909 (mips_mips16_decl_p): Delete.
23910 (mips_nomips16_decl_p): Delete.
23911 (mips_get_compress_on_flags): New function.
23912 (mips_get_compress_off_flags): New function.
23913 (mips_get_compress_mode): New function.
23914 (mips_get_compress_on_name): New function.
23915 (mips_get_compress_off_name): New function.
23916 (mips_insert_attributes): Support multiple compression types.
23917 (mips_merge_decl_attributes): Likewise.
23918 (umips_12bit_offset_address_p): New function.
23919 (mips_start_function_definition): Emit .set micromips directive.
23920 (mips_call_may_need_jalx_p): New function.
23921 (mips_function_ok_for_sibcall): Add microMIPS support.
23922 (mips_print_operand_punctuation): Support short delay slots and
23924 (umips_swm_mask, umips_swm_encoding): New.
23925 (umips_build_save_restore): New function.
23926 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23927 (was_mips16_p): Remove.
23928 (old_compression_mode): New.
23929 (mips_set_compression_mode): New function.
23930 (mips_set_current_function): Add microMIPS support.
23931 (mips_option_override): Likewise.
23932 (umips_save_restore_pattern_p): New function.
23933 (umips_output_save_restore): New function.
23934 (umips_load_store_pair_p_1): New function.
23935 (umips_load_store_pair_p): New function.
23936 (umips_output_load_store_pair_1): New function.
23937 (umips_output_load_store_pair): New function.
23938 (umips_movep_target_p) New function.
23939 (mips_prepare_pch_save): Add microMIPS support.
23940 * config/mips/mips.h (TARGET_COMPRESSION): New.
23941 (TARGET_CPU_CPP_BUILTINS): Update macro
23942 to use new compression flags and to support microMIPS.
23943 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23944 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23945 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23946 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23947 (ASM_SPEC): Support mmicromips and mno-micromips.
23948 (M16STORE_REG_P): New macro.
23949 (MIPS_CALL): Support TARGET_MICROMIPS.
23950 (MICROMIPS_J): New macro.
23951 (mips_base_mips16): Rename this...
23952 (mips_base_compression_flags): ...to this.
23953 (UMIPS_12BIT_OFFSET_P): New macro.
23954 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23955 (MULTILIB_DIRNAMES): Likewise.
23957 2013-02-25 Tom de Vries <tom@codesourcery.com>
23959 PR rtl-optimization/56131
23960 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23961 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23962 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23964 2013-02-25 Tobias Burnus <burnus@net-b.de>
23966 * doc/invoke.texi (-fsanitize=): Move from optimization
23967 to debugging options.
23969 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23971 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23973 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23974 Alexander Monakov <amonakov@ispras.ru>
23976 PR middle-end/56077
23977 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23978 flush pending lists also on non-jumps. Adjust comment.
23980 2013-02-24 Catherine Moore <clm@codesourcery.com>
23981 Maciej W. Rozycki <macro@codesourcery.com>
23982 Tom de Vries <tom@codesourcery.com>
23983 Nathan Sidwell <nathan@codesourcery.com>
23984 Iain Sandoe <iain@codesourcery.com>
23985 Nathan Froyd <froydnj@codesourcery.com>
23986 Chao-ying Fu <fu@mips.com>
23988 * doc/extend.texi: (micromips, nomicromips, nocompression):
23989 Document new function attributes.
23990 * doc/invoke.texi (minterlink-compressed, mmicromips,
23991 m14k, m14ke, m14kec): Document new options.
23992 (minterlink-mips16): Update documentation.
23993 * doc/md.texi (ZC, ZD): Document new constraints.
23994 * configure.ac (gcc_cv_as_micromips): Check if linker
23995 supports the .set micromips directive.
23996 * configure: Regenerate.
23997 * config.in: Regenerate.
23998 * config/mips/mips-tables.opt: Regenerate.
23999 * config/mips/micromips.md: New file.
24000 * constraints.md (ZC, AD): New constraints.
24001 * config/mips/predicates.md (movep_src_register): New predicate.
24002 (movep_src_operand): New predicate.
24003 (non_volatile_mem_operand): New predicate.
24004 * config/mips/mips.md (multimem): New type.
24005 (length): Differentiate between 17-bit and 18-bit branch offsets.
24006 (MOVEP1, MOVEP2): New mode iterator.
24007 (mov_<load>l): Use ZC constraint.
24008 (mov_<load>r): Likewise.
24009 (mov_<store>l): Likewise.
24010 (mov_<store>r): Likewise.
24011 (*branch_equality<mode>_inverted): Add microMIPS support.
24012 (*branch_equality<mode>): Likewise.
24013 (*jump_absolute): Likewise.
24014 (indirect_jump_<mode>): Likewise.
24015 (tablejump_<mode>): Likewise.
24016 (<optab>_internal): Likewise.
24017 (sibcall_internal): Likewise.
24018 (sibcall_value_internal): Likewise.
24019 (prefetch): Use constraint ZD.
24020 * config/mips/mips.opt (minterlink-compressed): New option.
24021 (minterlink-mips16): Now an alias for minterlink-compressed.
24022 (mmicromips): New option.
24023 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24024 (compare_and_swap_12): Likewise.
24025 (sync_add<mode>): Likewise.
24026 (sync_<optab>_12): Likewise.
24027 (sync_old_<optab>_12): Likewise.
24028 (sync_new_<optab>_12): Likewise.
24029 (sync_nand_12): Likewise.
24030 (sync_old_nand_12): Likewise.
24031 (sync_new_nand_12): Likewise.
24032 (sync_sub<mode>): Likewise.
24033 (sync_old_add<mode>): Likewise.
24034 (sync_old_sub<mode>): Likewise.
24035 (sync_new_add<mode>): Likewise.
24036 (sync_new_sub<mode>): Likewise.
24037 (sync_<optab><mode>): Likewise.
24038 (sync_old_<optab><mode>): Likewise.
24039 (sync_new_<optab><mode>): Likewise.
24040 (sync_nand<mode>): Likewise.
24041 (sync_old_nand<mode>): Likewise.
24042 (sync_new_nand<mode>): Likewise.
24043 (sync_lock_test_and_set<mode>): Likewise.
24044 (test_and_set_12): Likewise.
24045 (atomic_compare_and_swap<mode>): Likewise.
24046 (atomic_exchange<mode>_llsc): Likewise.
24047 (atomic_fetch_add<mode>_llsc): Likewise.
24048 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24049 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24050 (umips_save_restore_pattern_p): Likewise.
24051 (umips_load_store_pair_p): Likewise.
24052 (umips_output_load_store_pair): Likewise.
24053 (umips_movep_target_p): Likewise.
24054 (umips_12bit_offset_address_p): Likewise.
24055 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24056 (mips_base_mips16): Rename this...
24057 (mips_base_compression_flags): ...to this. Update all uses.
24058 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24059 (mips_mips16_decl_p): Delete.
24060 (mips_nomips16_decl_p): Delete.
24061 (mips_get_compress_on_flags): New function.
24062 (mips_get_compress_off_flags): New function.
24063 (mips_get_compress_mode): New function.
24064 (mips_get_compress_on_name): New function.
24065 (mips_get_compress_off_name): New function.
24066 (mips_insert_attributes): Support multiple compression types.
24067 (mips_merge_decl_attributes): Likewise.
24068 (umips_12bit_offset_address_p): New function.
24069 (mips_start_function_definition): Emit .set micromips directive.
24070 (mips_call_may_need_jalx_p): New function.
24071 (mips_function_ok_for_sibcall): Add microMIPS support.
24072 (mips_print_operand_punctuation): Support short delay slots and
24074 (umips_swm_mask, umips_swm_encoding): New.
24075 (umips_build_save_restore): New function.
24076 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24077 (was_mips16_p): Remove.
24078 (old_compression_mode): New.
24079 (mips_set_compression_mode): New function.
24080 (mips_set_current_function): Add microMIPS support.
24081 (mips_option_override): Likewise.
24082 (umips_save_restore_pattern_p): New function.
24083 (umips_output_save_restore): New function.
24084 (umips_load_store_pair_p_1): New function.
24085 (umips_load_store_pair_p): New function.
24086 (umips_output_load_store_pair_1): New function.
24087 (umips_output_load_store_pair): New function.
24088 (umips_movep_target_p) New function.
24089 (mips_prepare_pch_save): Add microMIPS support.
24090 * config/mips/mips.h (TARGET_COMPRESSION): New.
24091 (TARGET_CPU_CPP_BUILTINS): Update macro
24092 to use new compression flags and to support microMIPS.
24093 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24094 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24095 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24096 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24097 (ASM_SPEC): Support mmicromips and mno-micromips.
24098 (M16STORE_REG_P): New macro.
24099 (MIPS_CALL): Support TARGET_MICROMIPS.
24100 (MICROMIPS_J): New macro.
24101 (mips_base_mips16): Rename this...
24102 (mips_base_compression_flags): ...to this.
24103 (UMIPS_12BIT_OFFSET_P): New macro.
24104 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24105 (MULTILIB_DIRNAMES): Likewise.
24107 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24110 * target-globals.c (save_target_globals): For init_reg_sets and
24111 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24113 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24115 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24116 * config/aarch64/t-aarch64
24117 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24119 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24121 PR inline-asm/56148
24122 * lra-constraints.c (process_alt_operands): Reload operand
24123 conflicting with earlier clobber only if no more other conflicting
24126 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24129 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24130 if not linking a shared library.
24132 2013-02-22 Seth LaForge <sethml@google.com>
24134 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24136 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24138 * config/arm/arm.md (split for extendsidi): Update condition.
24139 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24140 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24141 (qhs_zextenddi_cstr): Likewise.
24143 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24145 PR middle-end/56420
24146 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24147 avoid signed wrapping.
24148 (expand_mult): Handle properly multiplication by
24149 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24150 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24151 in the compiler if coeff is HOST_WIDE_INT_MIN.
24152 (expand_divmod): Don't make ext_op1 static, change it's type to
24153 uhwi. Avoid undefined behavior in -INTVAL (op1).
24155 PR rtl-optimization/50339
24156 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24158 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24159 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24160 into splitting_ashiftrt field.
24161 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24163 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24166 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24168 PR middle-end/56108
24169 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24170 are sure to go irrevocable.
24172 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24174 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24175 scalars are valid operands.
24177 2013-02-21 Martin Jambor <mjambor@suse.cz>
24179 PR tree-optimization/56310
24180 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24181 only matching indices and non-negative final offsets.
24182 (intersect_aggregates_with_edge): Pass src_idx to
24183 agg_replacements_to_vector. Pass src_idx insstead of index to
24184 intersect_with_agg_replacements.
24186 2013-02-21 Martin Jambor <mjambor@suse.cz>
24188 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24189 instead of hard-wired defaults.
24191 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24193 * doc/invoke.texi (MIPS Options): Update documentation of the
24194 floating-point multiply-accumulate instruction restrictions.
24196 2013-02-21 Kostya Serebryany <kcc@google.com>
24198 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24199 asan_shadow_offset on x86_64 linux.
24201 2013-02-21 Richard Biener <rguenther@suse.de>
24203 PR tree-optimization/56415
24205 2013-02-11 Richard Biener <rguenther@suse.de>
24207 PR tree-optimization/56273
24208 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24211 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24214 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24217 PR inline-asm/56405
24218 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24219 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24221 2013-02-20 Jan Hubicka <jh@suse.cz>
24223 PR tree-optimization/56265
24224 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24225 when target is referenced for first time.
24227 2013-02-20 Richard Biener <rguenther@suse.de>
24229 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24230 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24231 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24232 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24233 not return anything.
24234 (rename_ssa_copies): Do not remove unused locals.
24235 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24236 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24237 * passes.c (execute_function_todo): Do not schedule unused locals
24238 removal if cleanup_tree_cfg did something.
24239 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24240 about the number of removed locals.
24242 2013-02-20 Richard Biener <rguenther@suse.de>
24244 PR tree-optimization/56398
24245 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24247 2013-02-20 Martin Jambor <mjambor@suse.cz>
24249 PR tree-optimization/55334
24250 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24251 restricted pointers to arrays.
24253 2013-02-20 Richard Biener <rguenther@suse.de>
24254 Jakub Jelinek <jakub@redhat.com>
24256 PR tree-optimization/56396
24257 * tree-ssa-ccp.c (n_const_val): New static variable.
24258 (get_value): Return NULL for SSA names we don't have a lattice
24260 (ccp_initialize): Initialize n_const_val.
24261 * tree-ssa-copy.c (n_copy_of): New static variable.
24262 (init_copy_prop): Initialize n_copy_of.
24263 (get_value): Return NULL_TREE for SSA names we don't have a
24266 2013-02-20 Martin Jambor <mjambor@suse.cz>
24268 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24270 2013-02-20 Richard Biener <rguenther@suse.de>
24272 * genpreds.c (write_lookup_constraint): Do not compare first
24273 letter of the constraint again.
24275 2013-02-20 Richard Biener <rguenther@suse.de>
24277 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24279 (get_use_iv_cost): Terminate hashtable walk when coming across
24282 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24284 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24285 reassociation for avx2 targets.
24287 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24289 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24290 Add version check for v8.10.a to enable microblaze_has_clz
24291 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24292 version and TARGET_PATTERN_COMPARE check
24293 * config/microblaze/microblaze.md: New clzsi2 instruction
24295 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24297 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24298 function before branching.
24300 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24302 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24304 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24306 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24308 PR middle-end/55889
24309 * sel-sched.c: Include ira.h.
24310 (implicit_clobber_conflict_p): New function.
24311 (moveup_expr): Use it.
24312 * Makefile.in (sel-sched.o): Depend on ira.h.
24314 2013-02-19 Richard Biener <rguenther@suse.de>
24316 PR tree-optimization/56384
24317 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24318 (vn_hash_type): Split out from ...
24319 (vn_hash_constant_with_type): ... here.
24320 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24321 (vn_phi_eq): Compare types from vn_phi_s structure.
24322 (vn_phi_lookup): Populate vn_phi_s type.
24323 (vn_phi_insert): Likewise.
24325 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24327 PR tree-optimization/56350
24328 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24329 if haven't found reduction or nested cycle operand, rather than
24330 asserting we must find it.
24332 PR tree-optimization/56381
24333 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24336 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24337 Jakub Jelinek <jakub@redhat.com>
24340 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24341 (swap_optab_enable): Same.
24342 (init_all_optabs): Use argument instead of global.
24343 * tree.h (struct tree_optimization_option): New field target_optabs.
24344 * expr.h (init_all_optabs): Add argument to prototype.
24345 (TREE_OPTIMIZATION_OPTABS): New.
24346 (save_optabs_if_changed): Protoize.
24347 * optabs.h: Declare this_fn_optabs.
24348 * optabs.c (save_optabs_if_changed): New.
24349 Declare this_fn_optabs.
24350 (init_optabs): Add argument to init_all_optabs() call.
24351 * function.c (invoke_set_current_function_hook): Handle per
24353 * function.h (struct function): New field optabs.
24354 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24355 optimization_current_node has changed.
24356 * target-globals.h (save_target_globals_default_opts): Protoize.
24357 * target-globals.c (save_target_globals_default_opts): New.
24359 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24362 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24363 registers %fr12 and %fr12R as call used.
24366 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24367 and HImode, require all displacements to be an integer multiple of
24369 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24370 only allow QImode and HImode when reload is in progress and strict is
24371 true. Likewise for symbolic addresses. Use base14_operand to check
24372 displacements in REG+BASE addresses.
24374 2013-02-18 Richard Biener <rguenther@suse.de>
24376 PR tree-optimization/56366
24377 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24378 handle sign-conversion of outer-loop initial induction value.
24380 2013-02-18 Richard Biener <rguenther@suse.de>
24382 PR middle-end/56349
24383 * cfghooks.c (merge_blocks): If we merge a latch into another
24384 block adjust references to it.
24385 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24386 (verify_loop_structure): Verify that a recorded latch is in fact
24389 2013-02-18 Richard Biener <rguenther@suse.de>
24391 PR tree-optimization/56321
24392 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24393 order SSA name release and virtual operand unlinking.
24395 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24397 * config/microblaze/microblaze.md (save_stack_block): Define.
24398 (restore_stack_block): Likewise.
24400 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24402 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24403 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24404 * config/microblaze/microblaze.c (microblaze_option_override):
24405 Bail out early for PIC modes when target does not support PIC.
24407 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24409 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24410 Replace with a microblaze version.
24411 (microblaze_trampoline_init): Adapt for microblaze.
24412 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24415 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24416 Dodji Seketeli <dodji@redhat.com>
24419 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24420 (instrument_mem_region_access): Do not forget to always put
24421 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24422 0) statement, even for cases where either 'base' or 'base + len'
24423 are not instrumented -- because they have been previously
24424 instrumented. Simplify the logic by putting all the statements
24425 instrument 'base + len' inside a sequence, and then insert that
24426 sequence right before the current insertion point. Then, to
24427 instrument 'base + len', just get an iterator on that statement.
24428 And do not forget to update the pointer to iterator the function
24429 received as argument.
24431 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24433 PR rtl-optimization/56348
24434 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24436 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24438 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24439 (clean_graph_dump_file): Pass base to start_graph_dump.
24441 2013-02-14 Richard Henderson <rth@redhat.com>
24444 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24446 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24448 * collect2-aix.h: Define F_LOADONLY.
24450 2013-02-14 Richard Biener <rguenther@suse.de>
24453 * varasm.c (output_constant_def_1): Get the decl representing
24454 the constant as argument.
24455 (output_constant_def): Wrap output_constant_def_1.
24456 (make_decl_rtl): Use output_constant_def_1 with the decl
24457 representing the constant.
24458 (build_constant_desc): Optionally re-use a decl already
24459 representing the constant.
24460 (tree_output_constant_def): Adjust.
24462 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24465 * asan.c (instrument_builtin_call): Really put the length of the
24466 second source argument into src1_len.
24468 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24470 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24471 argument. If it is false, don't create edge from then_bb to
24473 (insert_if_then_before_iter): Pass true to it.
24474 (build_check_stmt): Pass false to it.
24475 (transform_statements): Flush hash table only on extended basic
24476 block boundaries, rather than at the beginning of every bb.
24477 Don't flush hash table on nonfreeing_call_p calls.
24478 * tree-flow.h (nonfreeing_call_p): New prototype.
24479 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24481 2013-02-13 David S. Miller <davem@davemloft.net>
24483 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24485 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24488 * ira.c (max_regno_before_ira): Move from ...
24490 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24493 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24495 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24497 2013-02-13 Richard Biener <rguenther@suse.de>
24500 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24501 globals in MEM_REFs.
24503 2013-02-13 Richard Biener <rguenther@suse.de>
24505 * loop-init.c (loop_optimizer_init): Clear loop state when
24506 re-initializing preserved loops.
24507 * loop-unswitch.c (unswitch_single_loop): Return whether
24508 we unswitched the loop. Do not verify loop state here.
24509 (unswitch_loops): When we unswitched a loop discover new loops.
24511 2013-02-13 Kostya Serebryany <kcc@google.com>
24513 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24515 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24517 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24519 Avoid instrumenting duplicated memory access in the same basic block
24520 * Makefile.in (asan.o): Add new dependency on hash-table.h
24521 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24522 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24523 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24524 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24525 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24526 (get_mem_ref_of_assignment): New functions.
24527 (get_mem_refs_of_builtin_call): Extract from
24528 instrument_builtin_call and tweak a little bit to make it fit with
24530 (instrument_builtin_call): Use the new
24531 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24532 of is_gimple_builtin_call.
24533 (instrument_derefs, instrument_mem_region_access): Insert the
24534 instrumented memory reference into the hash table.
24535 (maybe_instrument_assignment): Renamed instrument_assignment into
24536 this, and change it to advance the iterator when instrumentation
24537 actually happened and return true in that case. This makes it
24538 homogeneous with maybe_instrument_assignment, and thus give a
24539 chance to callers to be more 'regular'.
24540 (transform_statements): Clear the memory reference hash table
24541 whenever we enter a new BB, when we cross a function call, or when
24542 we are done transforming statements. Use
24543 maybe_instrument_assignment instead of instrumentation. No more
24544 need to special case maybe_instrument_assignment and advance the
24545 iterator after calling it; it's now handled just like
24546 maybe_instrument_call. Update comment.
24548 2013-02-13 Richard Biener <rguenther@suse.de>
24550 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24551 Fix loop discovery code.
24553 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24555 PR inline-asm/56148
24556 * lra-constraints.c (process_alt_operands): Match early clobber
24557 operand with itself. Check conflicts with earlyclobber only if
24558 the operand is not reloaded. Prefer to reload conflicting operand
24559 if earlyclobber and matching operands are the same.
24561 2013-02-12 Richard Biener <rguenther@suse.de>
24564 * lto-streamer-out.c (write_symbol): Do not output symbols
24565 for hard register variables.
24567 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24570 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24571 (umulsidi3_insn, mulsidi3_insn): New insns.
24573 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24575 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24576 (struct tune_params): Add vec_costs field.
24577 * config/arm/arm.c (arm_builtin_vectorization_cost)
24578 (arm_add_stmt_cost): New functions.
24579 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24580 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24581 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24582 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24583 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24584 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24585 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24587 2013-02-12 Richard Biener <rguenther@suse.de>
24590 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24591 decls again if possible.
24593 2013-02-12 Richard Biener <rguenther@suse.de>
24595 PR middle-end/56288
24596 * tree-ssa.c (verify_ssa_name): Fix check, move
24597 SSA_NAME_IN_FREE_LIST check up.
24599 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24600 Steven Bosscher <steven@gcc.gnu.org>
24602 PR rtl-optimization/56151
24603 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24604 equal to op0 or op1, and last_insn pattern is CODE operation
24605 with MEM dest and one of the operands matches that MEM.
24607 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24609 * doc/extend.texi: Document Function Multiversioning and "default"
24610 parameter string to target attribute.
24611 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24612 target attribute parameter is "default".
24613 (ix86_compare_version_priority): Remove checks for target attribute.
24614 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24615 Remove check for target attribute equal to NULL. Add assert.
24616 (ix86_generate_version_dispatcher_body): Change error to sorry.
24618 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24619 Jack Howarth <howarth@bromo.med.uc.edu>
24620 Patrick Marlier <patrick.marlier@gmail.com>
24623 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24624 define ENDFILE_SPEC as TM_DESTRUCTOR.
24625 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24627 2013-02-11 Alexander Potapenko <glider@google.com>
24628 Jack Howarth <howarth@bromo.med.uc.edu>
24629 Jakub Jelinek <jakub@redhat.com>
24632 * config/darwin.c (cdtor_record): Rename ctor_record.
24633 (sort_cdtor_records): Rename sort_ctor_records.
24634 (finalize_dtors): New routine to sort destructors by
24635 priority before use in assemble_integer.
24636 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24638 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24640 PR rtl-optimization/56275
24641 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24642 offset is non-negative and less than cmode size before
24643 calling simplify_subreg.
24645 2013-02-11 Richard Biener <rguenther@suse.de>
24647 PR tree-optimization/56264
24648 * cfgloop.h (fix_loop_structure): Adjust prototype.
24649 * loop-init.c (fix_loop_structure): Return the number of
24650 newly discovered loops.
24651 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24652 are discovered, do a full loop-closed SSA rewrite.
24654 2013-02-11 Richard Biener <rguenther@suse.de>
24656 PR tree-optimization/56273
24657 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24659 (check_array_ref): Fix missing newline in dumps.
24660 (search_for_addr_array): Likewise.
24662 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24664 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24666 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24669 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24671 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24673 PR rtl-optimization/56246
24674 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24676 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24677 constraints are satisfied.
24679 2013-02-08 Jeff Law <law@redhat.com>
24682 * emit-rtl.c (reg_is_parm_p): New function.
24683 * regs.h (reg_is_parm_p): New prototype.
24684 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24685 callee-clobbered registers.
24687 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24690 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24691 If there is no implicit builtin declaration, just return NULL.
24693 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24695 * config/i386/sse.md (FMAMODEM): New mode iterator.
24696 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24697 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24699 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24701 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24702 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24703 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24705 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24707 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24708 (microblaze*-*-elf): Likewise.
24709 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24711 * config/microblaze/microblaze-c.c: Add builtin defines for
24712 _LITTLE_ENDIAN and _BIG_ENDIAN.
24713 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24714 add to TARGET_DEFAULT flags.
24715 Expand ASM_SPEC and LINK_SPEC.
24716 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24717 * config/microblaze/microblaze.md: Update extendsidi2 and
24718 movdi_internal instructions to use low-order / high-order reg
24720 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24721 options and inversemask / mask of LITTLE_ENDIAN.
24722 * config/microblaze/t-microblaze: Expand multilib options to
24723 include mlittle-endian (le) and update exceptions patterns.
24725 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24727 PR rtl-optimization/56195
24728 * lra-constraints.c (get_reload_reg): Don't reuse regs
24729 if they have smaller mode than requested, if they have
24730 wider mode than requested, try to return a SUBREG.
24732 PR tree-optimization/56250
24733 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24734 if type is unsigned and code isn't MULT_EXPR.
24736 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24738 PR tree-optimization/56064
24739 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24740 bits according to mode.
24741 * fixed-value.h (fixed_from_double_int)
24742 (const_fixed_from_double_int): Adjust comments.
24744 2013-02-08 Richard Biener <rguenther@suse.de>
24747 * lto-streamer.h (struct data_in): Remove current_file, current_line
24748 and current_col members.
24749 * lto-streamer-out.c (lto_output_location): Stream changed bits
24750 en-block for efficiency.
24751 * lto-streamer-in.c (clear_line_info): Remove.
24752 (lto_input_location): Cache current file, line and column
24753 globally via local statics. Read changed bits en-block.
24754 (input_function): Do not call clear_line_info.
24755 (lto_read_body): Likewise.
24756 (lto_input_toplevel_asms): Likewise.
24758 2013-02-08 Michael Matz <matz@suse.de>
24760 PR tree-optimization/52448
24761 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24762 (nt_call_phase): New static.
24763 (add_or_mark_expr): Only mark accesses with newer phase than any
24765 (nonfreeing_call_p): New.
24766 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24767 (nt_fini_block): Keep blocks marked as visited.
24768 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24770 2013-02-08 Richard Biener <rguenther@suse.de>
24772 * ira.c (ira): Free broken dominator information.
24774 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24776 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24778 2013-02-08 Marek Polacek <polacek@redhat.com>
24780 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24782 2013-02-08 Richard Biener <rguenther@suse.de>
24784 PR middle-end/56181
24785 * cfgloop.h (flow_loops_find): Adjust.
24786 (bb_loop_header_p): Declare.
24787 * cfgloop.c (bb_loop_header_p): New function split out from ...
24788 (flow_loops_find): ... here. Adjust function signature,
24789 support incremental loop structure update.
24790 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24791 * cfgloopmanip.c (fix_loop_structure): Move ...
24792 * loop-init.c (fix_loop_structure): ... here.
24793 (apply_loop_flags): Split out from ...
24794 (loop_optimizer_init): ... here.
24795 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24796 in incremental mode, only remove dead loops here.
24798 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24801 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24802 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24803 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24804 (*round<mode>3.libgcc): New insns for fixed-modes.
24805 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24806 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24807 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24808 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24809 implementations. Define to __builtin_avr_absFX,
24810 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24811 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24812 __builtin_avr_countlsFX, respectively.
24813 * config/avr/avr-c.c (target.h): Include it.
24814 (enum avr_builtin_id): New enum.
24815 (avr_resolve_overloaded_builtin): New static function.
24816 (avr_register_target_pragmas): Use it to set
24817 targetm.resolve_overloaded_builtin.
24818 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24819 tree nodes used by DEF_BUILTIN.
24820 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24821 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24822 <AVR_BUILTIN_xxBITS>: Same.
24824 2013-02-08 Richard Biener <rguenther@suse.de>
24826 * cfgloop.c (verify_loop_structure): Properly handle
24827 a loop exiting to another loop header.
24828 * ira-int.h (ira_loops): Remove.
24829 * ira.c (ira_loops): Remove.
24830 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24831 (do_reload): Use loop_optimizer_finalize.
24832 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24833 number_of_loops to access the loop tree.
24834 (more_one_region_p): Likewise.
24835 (finish_loop_tree_nodes): Likewise.
24836 (rebuild_regno_allocno_maps): Likewise.
24837 (mark_loops_for_removal): Likewise.
24838 (mark_all_loops_for_removal): Likewise.
24839 (remove_unnecessary_regions): Likewise.
24840 (ira_build): Likewise.
24841 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24843 2013-02-08 Richard Biener <rguenther@suse.de>
24845 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24846 * ipa-pure-const.c (analyze_function): Avoid calling
24847 mark_irreducible_loops twice.
24848 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24850 2013-02-07 David S. Miller <davem@davemloft.net>
24852 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24854 * var-tracking.c (vt_add_function_parameter): Test the presence of
24855 HAVE_window_save properly and do not remap argument registers when
24856 we have a leaf function.
24858 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24861 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24863 * config/i386/i386.c (ix86_print_operand): Ditto.
24865 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24867 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24869 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24871 PR rtl-optimization/56225
24872 * lra-constraints.c (process_alt_operands): Check that reload hard
24873 reg can hold value for strict_low_part.
24875 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24878 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24879 dwarf2out_end_function.
24880 (in_first_function_p, maybe_at_text_label_p,
24881 first_loclabel_num_not_at_text_label): New variables.
24882 (dwarf2out_var_location): In the first function find out
24883 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24884 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24887 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24889 PR rtl-optimization/56178
24890 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24891 SUBREG of a register. Tidy up related block of code.
24892 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24893 note if the source is a register or a SUBREG of a register.
24895 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24898 * config/rs6000/rs6000.md (ptrm): New mode attr.
24899 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24900 call_value_indirect_aix<pttrsize>,
24901 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24904 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24906 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24907 if -bnortl. Convert to strcmp and strncmp.
24909 2013-02-07 Alan Modra <amodra@gmail.com>
24912 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24913 addresses won't wrap when offsetting.
24914 (rs6000_secondary_reload): Provide secondary reloads needed for
24915 wrapping LO_SUM addresses.
24917 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24919 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24920 MACH, just __MACH__.
24922 2013-02-06 Richard Biener <rguenther@suse.de>
24924 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24925 instead of calling fix_loop_structure.
24927 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24929 PR middle-end/56217
24930 * omp-low.c (use_pointer_for_field): Return false if
24931 lower_send_shared_vars doesn't generate any copy-out code.
24933 2013-02-06 Tom de Vries <tom@codesourcery.com>
24935 PR rtl-optimization/56131
24936 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24937 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24938 of the label is NULL. Add comment.
24940 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24942 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24945 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24946 (STATIC_LIBTSAN_LIBS): Likewise.
24947 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24948 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24949 is defined, don't add anything else beyond that.
24950 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24951 (LINK_COMMAND_SPEC): Use them.
24953 PR tree-optimization/56205
24954 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24955 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24956 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24958 2013-02-05 Richard Biener <rguenther@suse.de>
24960 PR tree-optimization/53342
24961 PR tree-optimization/53185
24962 * tree-vectorizer.h (vect_check_strided_load): Remove.
24963 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24964 not disallow peeling for vectorized strided loads.
24965 (vect_check_strided_load): Make static and simplify.
24966 (vect_analyze_data_refs): Adjust.
24967 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24968 correctly when vectorizing strided loads.
24970 2013-02-05 Richard Biener <rguenther@suse.de>
24972 * doc/install.texi: Refer to ISL, not PPL.
24974 2013-02-05 Jan Hubicka <jh@suse.cz>
24976 PR tree-optimization/55789
24977 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24979 2013-02-05 Jan Hubicka <jh@suse.cz>
24981 PR tree-optimization/55789
24982 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24983 the dead call anyway.
24985 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24988 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24990 2013-02-04 Alexander Potapenko <glider@google.com>
24991 Jack Howarth <howarth@bromo.med.uc.edu>
24992 Jakub Jelinek <jakub@redhat.com>
24995 * config/darwin.c (sort_ctor_records): Stabilized qsort
24996 on constructor priority by using original position.
24997 (finalize_ctors): New routine to sort constructors by
24998 priority before use in assemble_integer.
24999 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25001 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25004 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25005 about visibility on artificial decls.
25006 * config/sol2.c (solaris_assemble_visibility): Likewise.
25008 2013-02-04 Kai Tietz <ktietz@redhat.com>
25011 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25012 argument and improve checking of return-argument types for 16-byte
25014 (ix86_function_value_1): Add additional valtype argument on call
25015 of function_value_64.
25016 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25017 handling infunction_value_64 function.
25019 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25021 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25023 2013-02-04 Richard Biener <rguenther@suse.de>
25025 PR tree-optimization/56188
25026 * tree-ssa-structalias.c (label_visit): Consider case with
25027 initially non-empty points-to set.
25028 (perform_var_substitution): Dump node mapping and clean up.
25030 2013-02-04 Richard Guenther <rguenther@suse.de>
25033 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25034 node prevail as last resort.
25035 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25036 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25038 2013-02-04 Richard Biener <rguenther@suse.de>
25040 PR tree-optimization/56113
25041 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25043 (equiv_class_lookup_or_add): ... this.
25044 (label_visit): Adjust and fix error in previous patch.
25045 (perform_var_substitution): Adjust.
25047 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25049 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25050 * config/sh/sh4-300.md: Likewise.
25051 * config/sh/sh4a.md: Likewise.
25052 * config/sh/constraints.md: Likewise.
25053 * config/sh/sh.md: Likewise.
25054 * config/sh/netbsd-elf.h: Likewise.
25055 * config/sh/predicates.md: Likewise.
25056 * config/sh/sh-protos.h: Likewise.
25057 * config/sh/ushmedia.h: Likewise.
25058 * config/sh/linux.h: Likewise.
25059 * config/sh/sh.c: Likewise.
25060 * config/sh/superh.h: Likewise.
25061 * config/sh/elf.h: Likewise.
25062 * config/sh/sh4.md: Likewise.
25063 * config/sh/sh.h: Likewise.
25065 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25067 * config/pa/constraints.md: Adjust unused letters. Change "T"
25068 constraint to match_test floating_point_store_memory_operand().
25069 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25070 (base14_operand): New.
25071 (floating_point_store_memory_operand): New.
25072 (integer_store_memory_operand): Revise to use base14_operand and
25073 reg_plus_base_memory_operand.
25074 (move_dest_operand): Allow symbolic_memory_operands.
25075 (symbolic_memory_operand): Check for LO_SOM.
25076 (symbolic_operand): Change default case to break.
25077 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25078 CONST_DOUBLE values to be reloaded by putting them into memory when
25079 the destination is a floating point register.
25080 (movdf): Remove code to handle CONST_DOUBLE.
25082 (reload_indf_r1): New.
25083 (reload_insf_r1): New.
25084 Consistently use "Q" and "T" constraints with integer and floating
25085 point move instructions, respectively.
25086 (movdi): Remove FAIL.
25087 Change predicate for source operand unamed DImode move from
25088 general_operand to move_src_operand.
25089 (umulsidi3): Change predicate for destination operand to
25091 Likewise for similar unamed patterns.
25092 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25093 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25094 (hppa_legitimize_address): Simplify mask calculation.
25095 (pa_emit_move_sequence): Revised handling of secondary reloads from
25096 REG+D addresses for floating point loads and stores. Directly handle
25097 loading CONST0_RTX (mode) to a floating point register.
25098 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25099 to floating point registers. Don't restrict secondary reloads to
25100 floating point registers to integer modes. Revise some comments and
25102 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25103 (pa_legitimate_address_p): New.
25104 (pa_legitimize_reload_address): New.
25105 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25106 (STRICT_REG_OK_FOR_BASE_P): New.
25107 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25108 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25110 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25111 Andrew Dixie <andrewd@gentrack.com>
25113 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25116 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25118 * expmed.c (extract_bit_field_1): Pass the full width of the
25119 structure to get_best_reg_extraction_insn.
25121 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25124 * configure.ac (use_cxa_atexit): Add AIX.
25125 * configure: Regenerate.
25127 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25129 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25132 * final.c (need_profile_function): New variable.
25133 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25134 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25135 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25136 notes, targetm.asm_out.function_prologue doesn't emit anything,
25137 HAVE_prologue and profiler should be emitted before prologue,
25138 set need_profile_function instead of emitting it.
25139 (final_scan_insn): If need_profile_function, emit
25140 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25141 NOTE_INSN_FUNCTION_BEG note.
25143 2013-02-01 Richard Henderson <rth@redhat.com>
25145 * config/rs6000/rs6000.md (smulditi3): New.
25148 * config/alpha/alpha.md (umulditi3): New.
25150 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25152 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25153 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25155 2013-02-01 Richard Biener <rguenther@suse.de>
25157 PR tree-optimization/56113
25158 * tree-ssa-structalias.c (label_visit): Reduce work for
25159 single-predecessor nodes.
25161 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25163 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25164 range isn't testing for zero.
25166 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25168 PR middle-end/56113
25169 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25171 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25172 Nick Clifton <nickc@redhat.com>
25174 * config/v850/constraints.md (Q): Define as a memory constraint.
25175 * config/v850/predicates.md (label_ref_operand): New predicate.
25176 (e3v5_shift_operand): New predicate.
25177 (ior_operator): New predicate.
25178 * config/v850/t-v850: Add e3v5 multilib.
25179 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25180 (v850_gen_movdi): Prototype.
25181 * config/v850/v850.c: Add support for e3v5 architecture.
25182 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25184 (construct_save_jarl): Add e3v5 long JARL support.
25185 (v850_adjust_insn_length): New function. Adjust length of call
25186 insns when using e3v5 instructions.
25187 (v850_gen_movdi): New function: Generate instructions to move a
25189 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25190 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25191 (TARGET_USE_FPU): Enable for e3v5.
25192 (CONST_OK_FOR_W): New macro.
25193 (ADJUST_INSN_LENGTH): Define.
25194 * config/v850/v850.md (UNSPEC_LOOP): Define.
25195 (attr cpu): Add v850e3v5.
25196 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25197 (movdi): New pattern.
25198 (movdi_internal): New pattern.
25199 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25200 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25201 (cstoresf4): Likewise.
25202 (cstoredf4): Likewise.
25203 (insv): New pattern.
25204 (rotlso3_a): New pattern.
25205 (rotlsi3_b): New pattern
25206 (rotlsi3_v850e3v5): New pattern.
25207 (doloop_begin): New pattern.
25208 (fix_loop_counter): New pattern.
25209 (doloop_end): New pattern.
25210 (branch_normal): Add e3v5 long branch support.
25211 (branch_invert): Likewise.
25212 (branch_z_normal): Likewise.
25213 (branch_z_invert): Likewise.
25214 (branch_nz_normal): Likewise.
25215 (branch_nz_invert): Likewise.
25216 (call_internal_short): Add e3v5 register-indirect JARL support.
25217 (call_internal_long): Likewise.
25218 (call_value_internal_short): Likewise.
25219 (call_value_internal_long): Likewise.
25220 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25221 (mloop): New option.
25222 * config.gcc: Add support for configuring v840e3v5 target.
25223 * doc/invoke.texi: Document new v850 specific command line options.
25225 2013-01-31 Paul Koning <ni1d@arrl.net>
25229 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25230 children if parent is a class.
25231 (prune_unused_types_prune): Don't add DW_AT_declaration.
25233 2013-01-31 Richard Biener <rguenther@suse.de>
25235 PR tree-optimization/56157
25236 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25237 match up operand with SLP child.
25239 2013-01-31 Jason Merrill <jason@redhat.com>
25242 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25243 parameters the first time.
25244 (gen_scheduled_generic_parms_dies): Check completeness here.
25246 2013-01-31 Richard Biener <rguenther@suse.de>
25248 PR middle-end/53073
25249 * common.opt (faggressive-loop-optimizations): New flag,
25250 enabled by default.
25251 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25252 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25253 infer_loop_bounds_from_undefined by it.
25255 2013-01-31 Richard Biener <rguenther@suse.de>
25257 PR tree-optimization/56150
25258 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25259 visit virtual operands.
25260 (find_uses_to_rename_bb): Likewise.
25262 2013-01-31 Richard Biener <rguenther@suse.de>
25264 PR tree-optimization/56150
25265 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25266 mixed store non-store stmts.
25268 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25271 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25272 LIBASAN_EARLY_SPEC is defined.
25273 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25274 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25276 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25279 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25280 invalid args instead of ICEing on it.
25281 (ix86_valid_target_attribute_tree): Return error_mark_node if
25282 ix86_valid_target_attribute_inner_p failed.
25283 (ix86_valid_target_attribute_p): Return false only if
25284 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25285 target("default") attribute.
25286 (sorted_attr_string): Change argument from const char * to tree,
25287 merge in all target attribute arguments rather than just one.
25288 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25289 instead of free. Avoid using strcat.
25290 (ix86_mangle_function_version_assembler_name): Mangle
25291 target("default") as if no target attribute is present. Adjust
25292 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25293 instead of xmalloc and XDELETEVEC instead of free.
25294 (ix86_function_versions): Don't return true if one of the decls
25295 doesn't have target attribute. If they don't and one of the decls
25296 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25297 sorted_attr_string caller. Use XDELETEVEC instead of free.
25298 (ix86_supports_function_versions): Remove.
25299 (make_name): Fix up formatting.
25300 (make_dispatcher_decl): Remove resolver_name and its initialization.
25301 Avoid leaking memory.
25302 (is_function_default_version): Return true if there is
25303 target("default") attribute rather than no target attribute at all.
25304 (make_resolver_func): Avoid leaking memory.
25305 (ix86_generate_version_dispatcher_body): Likewise.
25306 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25307 * target.def (supports_function_versions): Remove.
25308 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25309 * doc/tm.texi: Regenerated.
25311 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25313 PR rtl-optimization/56144
25314 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25315 for values with side effects.
25317 2013-01-30 Richard Biener <rguenther@suse.de>
25319 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25320 (sparseset_pop): Likewise.
25321 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25322 to be able to use quick_push in the worker loop.
25324 2013-01-30 Marek Polacek <polacek@redhat.com>
25326 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25328 2013-01-30 Richard Biener <rguenther@suse.de>
25331 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25333 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25335 PR tree-optimization/56064
25336 * fixed-value.c (fixed_from_double_int): New function.
25337 * fixed-value.h (fixed_from_double_int): New prototype.
25338 (const_fixed_from_double_int): New static inline function.
25339 * fold-const.c (native_interpret_fixed): New static function.
25340 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25341 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25342 (native_encode_fixed): New static function.
25343 (native_encode_expr) <FIXED_CST>: Use it.
25344 (native_interpret_int): Move double_int worker code to...
25345 * double-int.c (double_int::from_buffer): ...this new static method.
25346 * double-int.h (double_int::from_buffer): Prototype it.
25348 2013-01-30 Richard Biener <rguenther@suse.de>
25350 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25351 New pointer-map and obstack.
25352 (init_alias_vars): Allocate pointer-map and obstack.
25353 (delete_points_to_sets): Free them.
25354 (find_what_var_points_to): Cache result.
25355 (find_what_p_points_to): Adjust for changed interface of
25356 find_what_var_points_to.
25357 (compute_points_to_sets): Likewise.
25358 (ipa_pta_execute): Likewise.
25360 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25362 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25363 * configure: Regenerate.
25364 * config.in: Regenerate.
25365 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25366 #nobits/#progbits if supported.
25368 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25371 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25372 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25373 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25375 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25377 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25378 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25380 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25382 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25383 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25385 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25387 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25389 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25390 * config/arm/cortex-a7.md: New bypasses using
25391 arm_mac_accumulator_is_result.
25393 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25395 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25396 (cortex_a7_neon_mla): Likewise.
25397 (cortex_a7_fpfmad): New reservation.
25398 (cortex_a7_fpmacs): Use ffmas and update required units.
25399 (cortex_a7_fpmuld): Update required units and latency.
25400 (cortex_a7_fpmacd): Likewise.
25401 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25402 (cortex_a7_neon). Likewise.
25403 (bypass) Update participating units.
25405 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25407 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25408 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25410 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25411 (vfp_fmul): Use ffmad.
25412 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25413 (cortex_r4_fmacd): Use ffmad.
25414 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25415 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25416 (cortex_a9_fmacd): Use ffmad.
25417 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25418 (cortex_a8_vfp_macd): Use ffmad.
25419 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25420 (cortex_a5_fpmacd): Use ffmad.
25421 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25422 (cortex_a15_vfp_macd): Use ffmad.
25423 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25425 2013-01-29 Jason Merrill <jason@redhat.com>
25428 * varasm.c (default_assemble_visibility): Don't warn about
25429 visibility on artificial decls.
25431 2013-01-29 Richard Biener <rguenther@suse.de>
25433 PR tree-optimization/56113
25434 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25436 (label_visit): Free duplicate bitmaps and record the leader instead.
25437 (perform_var_substitution): Adjust.
25439 2013-01-29 Richard Biener <rguenther@suse.de>
25441 PR tree-optimization/55270
25442 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25443 the CFG, schedule loops for fixup.
25445 2013-01-29 Nick Clifton <nickc@redhat.com>
25447 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25450 2013-01-28 Leif Ekblad <leif@rdos.net>
25452 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25453 * config/i386/i386.h (TARGET_RDOS): New macro.
25454 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25455 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25456 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25457 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25458 DEFAULT_LARGE_SECTION_THRESHOLD.
25459 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25460 * config/i386/rdos.h: New file.
25461 * config/i386/rdos64.h: New file.
25463 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25466 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25469 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25471 PR rtl-optimization/56117
25472 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25473 call cselib_lookup_from_insn on the MEM before calling
25474 add_insn_mem_dependence.
25476 2013-01-28 Richard Biener <rguenther@suse.de>
25478 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25479 to a stmt that didn't have one.
25480 (copy_phis_for_bb): Likewise for PHI arguments.
25481 (copy_debug_stmt): Likewise for debug stmts.
25483 2013-01-28 Richard Biener <rguenther@suse.de>
25485 PR tree-optimization/56034
25486 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25487 (partition_builtin_p): Adjust.
25488 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25489 it is the last partition.
25490 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25491 up the vertex for the definition.
25492 (classify_partition): Classify whether a partition is a
25493 PKIND_REDUCTION, thus has uses outside of the loop.
25494 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25495 Merge all PKIND_REDUCTION partitions into the last partition.
25496 (tree_loop_distribution): Seed partitions from reductions as well.
25498 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25500 PR tree-optimization/56125
25501 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25502 pow(x,c) into sqrt(x) * powi(x, n/2) or
25503 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25504 optimizing for size.
25505 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25506 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25509 PR tree-optimization/56094
25510 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25511 to UNKNOWN_LOCATION while gimplifying expr.
25513 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25516 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25517 operand 0 in movabs insn template for -masm=intel asm alternative.
25518 (*movabs<mode>_2): Ditto for operand 1.
25520 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25523 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25526 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25528 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25531 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25533 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25534 Undef to avoid warning.
25536 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25538 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25539 * configure: Regenerate.
25541 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25543 PR tree-optimization/56098
25544 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25545 for stmts with volatile ops.
25546 (cond_store_replacement): Don't optimize if assign has volatile ops.
25547 (cond_if_else_store_replacement_1): Don't optimize if either
25548 then_assign or else_assign have volatile ops.
25549 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25552 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25554 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25556 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25558 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25559 missing ':' in asm example.
25561 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25563 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25564 entries into lane and laneq entries.
25565 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25566 Remove AdvSIMD scalar modes.
25567 (aarch64_sq<r>dmulh_laneq<mode>): New.
25568 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25570 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25571 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25572 * config/aarch64/iterators.md (VCOND): New.
25575 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25578 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25579 Add NULL LIBNAME argument to existing definitions.
25580 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25581 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25582 * config/avr/avr.c (DEF_BUILTIN): Same.
25583 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25584 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25585 implementation is available (DECL_ASSEMBLER_NAME is set).
25586 (avr_fold_absfx): New static function.
25587 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25588 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25589 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25590 AVR_BUILTIN_ABSLLK.
25591 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25592 (abshk, absk, abslk, absllk): Provide as static inline functions.
25594 2013-01-25 Marek Polacek <polacek@redhat.com>
25596 PR tree-optimization/56035
25597 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25599 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25601 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25602 (*movtf_internal_rex64): Add (!o,C) alternative
25603 (*movxf_internal_rex64): Ditto.
25604 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25606 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25608 * doc/invoke.texi: fix typo.
25609 * doc/objc.texi: fix typo.
25611 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25613 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25614 for the first two alternatives.
25616 2013-01-24 Diego Novillo <dnovillo@google.com>
25618 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25619 (ggc-zone.o): Remove.
25620 * configure.ac: Remove option --with-gc.
25621 * configure: Re-generate.
25622 * doc/install.texi: Remove documentation for --with-gc.
25623 * gengtype.c (write_enum_defn): Remove. Update all users.
25624 (write_Types_process_field): Remove generation of gt_e_* argument.
25625 (output_type_enum): Remove. Update all users.
25626 (write_enum_defn): Remove. Update all users.
25627 (enum alloc_zone): Remove. Update all users.
25628 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25629 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25630 Update all callers.
25631 (struct ptr_data): Remove field TYPE. Update all users.
25632 (gt_pch_note_object): Remove argument TYPE. Update all users.
25633 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25635 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25636 (struct alloc_zone): Remove.
25637 (ggc_internal_alloc_zone_stat): Remove.
25638 (ggc_internal_cleared_alloc_zone_stat): Remove.
25639 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25640 (ggc_pch_count_object): Remove last argument. Update all users.
25641 (ggc_pch_alloc_object): Remove last argument. Update all users.
25642 (struct alloc_zone): Remove.
25643 * ggc-zone.c: Remove.
25644 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25645 (struct alloc_zone): Remove.
25646 (ggc_alloc_typed_stat): Remove.
25647 (ggc_alloc_typed): Remove.
25648 (ggc_splay_alloc): Remove first argument.
25649 (rtl_zone): Remove. Update all users.
25650 (tree_zone): Remove. Update all users.
25651 (tree_id_zone): Remove. Update all users.
25652 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25653 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25654 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25655 * tree-ssanames.c: Remove references to zone allocator in comments.
25657 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25659 * config/avr/avr.c (avr_out_fract): Make register numbers that
25660 might be outside of source operand signed.
25662 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25664 * config/i386/constraints.md (Yf): New constraint.
25665 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25666 of f constraint to conditionaly disable x87 register preferences.
25667 (*movdf_internal): Ditto.
25668 (*movsf_internal): Ditto.
25670 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25672 PR inline-asm/55934
25673 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25674 that have operands with impossible constraints.
25675 Add a FIXME for a speed-up opportunity.
25676 * lra-constraints.c (process_alt_operands): Verify that a class
25677 selected from constraints on asms is valid for the operand mode.
25678 (curr_insn_transform): Remove incorrect comment.
25680 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25682 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25683 TOC operand is a valid symbol ref in the constant pool.
25685 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25687 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25689 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25692 * config/avr/stdfix.h: New file.
25693 * t-avr (stdfix-gcc.h): New rule to build it.
25694 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25696 2013-01-23 Kostya Serebryany <kcc@google.com>
25698 * config/darwin.h: remove dependency on
25699 CoreFoundation (asan on Mac OS).
25701 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25704 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25705 instead of cmpdi_operand for first comparison operand.
25706 Don't assert that comparison operands aren't both constants.
25708 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25710 * doc/install.texi (Downloading the Source): Update references to
25711 downloading separate components.
25713 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25715 * doc/extend.texi (__int128): Improve grammar.
25717 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25720 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25721 alternative to (o,r).
25722 (*movdi_internal_rex64): Remove (!o,n) alternative.
25723 (DImode immediate->memory splitter): Remove.
25724 (DImode immediate->memory peephole2): Remove.
25725 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25726 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25727 alternative to (!o,*r).
25728 (*movtf_internal_sse): New pattern.
25729 (*movxf_internal_rex64): New pattern.
25730 (*movxf_internal): Disable for TARGET_64BIT.
25731 (*movdf_internal_rex64): Remove (!o,F) alternative.
25733 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25735 PR middle-end/56074
25736 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25737 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25738 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25739 stmt locations where LOCATION_LOCUS of the stmt location is
25740 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25743 * config/i386/i386.md (UNSPEC_STOS): New.
25744 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25745 *strsetqi_1): Add UNSPEC_STOS.
25747 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25750 * doc/invoke.texi: Remove left over -Wsynth example.
25752 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25754 PR tree-optimization/56051
25755 * fold-const.c (fold_binary_loc): Don't fold
25756 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25757 a narrowing conversion, or widening conversion from signed
25760 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25762 PR rtl-optimization/56023
25763 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25764 dependent on debug instruction.
25766 2013-01-21 Martin Jambor <mjambor@suse.cz>
25768 PR middle-end/56022
25769 * function.c (allocate_struct_function): Call
25770 invoke_set_current_function_hook earlier.
25772 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25774 * reload1.c (init_reload): Only initialize reload_obstack
25775 during the first call.
25777 2013-01-21 Marek Polacek <polacek@redhat.com>
25779 * cfgloop.c (verify_loop_structure): Fix up grammar.
25781 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25783 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25784 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25786 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25789 * config/arm/marvell-pj4.md: Update copyright year.
25790 Fix up use of alu to alu_reg and simple_alu_imm.
25792 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25794 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25796 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25799 * lra-constraints.c (curr_insn_transform): Don't reuse original
25800 insn for secondary memory move when memory mode should be different.
25802 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25804 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25805 atomic_storedi_1): New patterns.
25807 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25809 btver2 pipeline descriptions.
25810 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25812 * config/i386/i386.md (btver2_decode): New type attributes.
25813 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25815 * config/i386/btver2.md: New file describing btver2 pipelines.
25817 2013-01-19 Andrew Pinski <apinski@cavium.com>
25819 PR tree-optimization/52631
25820 * tree-ssa-sccvn (visit_use): Before looking up the original
25821 statement, try looking up the simplified expression.
25823 2013-01-19 Anthony Green <green@moxielogic.com>
25825 * config/moxie/moxie.c (moxie_expand_prologue): Set
25826 current_function_static_stack_size.
25828 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25830 PR tree-optimization/56029
25831 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25832 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25834 2013-01-18 Sharad Singhai <singhai@google.com>
25836 PR tree-optimization/55995
25837 * dumpfile.c (dump_loc): Print location only if available.
25838 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25840 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25843 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25844 secondary memory move.
25845 (inherit_reload_reg): Use rclass instead of cl for
25846 check_secondary_memory_needed_p.
25848 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25850 PR middle-end/56015
25851 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25852 the case where writing real complex part of target modifies op1.
25854 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25856 * config/aarch64/aarch64-simd.md
25857 (aarch64_vcond_internal<mode>): Handle unordered cases.
25858 * config/aarch64/iterators.md (v_cmp_result): New.
25860 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25861 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25863 * config/arm/marvell-pj4.md: New file.
25864 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25865 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25866 (generic_vfp): Likewise.
25867 * config/arm/arm-cores.def: Add marvell-pj4.
25868 * config/arm/arm-tune.md: Regenerate.
25869 * config/arm/arm-tables.opt: Regenerate.
25870 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25871 * doc/invoke.texi: Document marvell-pj4.
25873 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25875 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25877 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25882 * var-tracking.c (negative_power_of_two_p): New.
25883 (global_get_addr_cache, local_get_addr_cache): New.
25884 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25885 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25887 (vt_stack_offset_p): Remove.
25888 (vt_canon_true_dep): Always canonicalize loc's address.
25889 (clobber_overlapping_mems): Make sure we have a MEM.
25890 (local_get_addr_clear_given_value): New.
25891 (val_reset): Clear local cached entries.
25892 (compute_bb_dataflow): Create and release the local cache.
25893 Disable duplicate MEMs clobbering.
25894 (emit_notes_in_bb): Clobber MEMs likewise.
25895 (vt_emit_notes): Create and release the local cache.
25896 (vt_initialize, vt_finalize): Create and release the global
25897 cache, respectively.
25898 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25900 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25902 PR libmudflap/53359
25903 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25904 not found in the symtab.
25906 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25909 PR rtl-optimization/55547
25910 PR rtl-optimization/53827
25913 * alias.c (offset_overlap_p): New, factored out of...
25914 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25915 the conservative special case for symbolic constants. Don't
25916 adjust zero sizes on alignment.
25918 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25920 PR rtl-optimization/52573
25921 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25922 REG_UNUSED for the same register.
25924 2013-01-17 Richard Biener <rguenther@suse.de>
25925 Marek Polacek <polacek@redhat.com>
25927 PR rtl-optimization/55833
25928 * loop-unswitch.c (unswitch_loops): Move loop verification...
25929 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25930 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25931 Set it to true when we're removing a loop from hierarchy tree in
25932 an irreducible region.
25933 (fix_bb_placements): Adjust caller.
25934 (fix_loop_placements): Likewise.
25936 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25938 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25939 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25940 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25941 Remove ID. Adjust comments.
25942 * config/avr/avr-c.c (avr_builtin_name): Remove.
25943 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25944 * config/avr/avr.c (avr_tolower): New static function.
25945 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25946 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25947 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25950 2013-01-17 Jan Hubicka <jh@suse.cz>
25952 PR tree-optimization/55273
25953 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25955 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25958 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25959 store through atomic_store<mode>_1.
25960 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25962 2013-01-17 Martin Jambor <mjambor@suse.cz>
25964 PR tree-optimizations/55264
25965 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25966 for virtual methods.
25967 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25968 virtual methods before inlining is over.
25969 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25971 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25974 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25976 PR rtl-optimization/56005
25977 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25978 pending reads for prefetch.
25980 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25982 * config/aarch64/aarch64.md
25983 (*cstoresi_neg_uxtw): New pattern.
25984 (*cmovsi_insn_uxtw): New pattern.
25985 (*<optab>si3_uxtw): New pattern.
25986 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25987 (*<optab>si3_insn_uxtw): New pattern.
25988 (*bswapsi2_uxtw): New pattern.
25990 2013-01-16 Richard Biener <rguenther@suse.de>
25992 * tree-inline.c (tree_function_versioning): Remove set but
25993 never used variable.
25995 2013-01-16 Richard Biener <rguenther@suse.de>
25997 PR tree-optimization/55964
25998 * tree-flow.h (rename_variables_in_loop): Remove.
25999 (rename_variables_in_bb): Likewise.
26000 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26001 (copy_loop_before): Adjust and delete update-ssa status.
26002 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26003 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26004 (rename_variables_in_loop): Remove.
26005 (slpeel_update_phis_for_duplicate_loop): Likewise.
26006 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26007 use available cfg machinery instead of duplicating it.
26008 Update PHI nodes and perform poor-mans SSA update here.
26009 (slpeel_tree_peel_loop_to_edge): Adjust.
26011 2013-01-16 Richard Biener <rguenther@suse.de>
26013 PR tree-optimization/54767
26014 PR tree-optimization/53465
26015 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26016 (vrp_visit_phi_node): For PHI arguments coming via backedges
26017 drop all symbolical range information.
26018 (execute_vrp): Compute backedges.
26020 2013-01-16 Richard Biener <rguenther@suse.de>
26022 * doc/install.texi: Update CLooG and ISL requirements to
26025 2013-01-16 Christian Bruel <christian.bruel@st.com>
26028 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26029 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26030 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26032 2013-01-16 DJ Delorie <dj@redhat.com>
26034 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26035 (UNSPECV_SP_SWITCH_E): New.
26036 (sp_switch_1): Change to an unspec.
26037 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26040 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26042 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26043 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26044 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26045 (expand_mem_thread_fence): Ditto.
26046 (expand_mem_signal_fence): Ditto.
26047 (expand_atomic_load): Ditto.
26048 (expand_atomic_store): Ditto.
26050 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26052 PR rtl-optimization/55547
26053 PR rtl-optimization/53827
26056 * alias.c (memrefs_conflict_p): Set sizes to negative after
26059 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26062 * function.c (thread_prologue_and_epilogue_insns): Always
26063 add crtl->drap_reg to set_up_by_prologue.set, even if
26064 stack_realign_drap is false.
26066 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26068 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26069 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26070 *call): Fix indention.
26072 2013-01-15 Tom de Vries <tom@codesourcery.com>
26075 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26078 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26080 PR rtl-optimization/55153
26081 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26083 2013-01-15 Martin Jambor <mjambor@suse.cz>
26085 PR tree-optimization/55920
26086 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26087 accesses as grp_to_be_debug_replaced.
26089 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26091 PR tree-optimization/55920
26092 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26093 there is non-useless type conversion needed from debug rhs to lhs,
26094 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26096 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26097 Mikael Pettersson <mikpe@it.uu.se>
26100 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26102 (ASM_OUTPUT_CASE_LABEL): Remove.
26103 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26104 * final.c (shorten_branches): Update alignment of labels before
26105 jump tables if CASE_VECTOR_SHORTEN_MODE.
26107 2013-01-15 Richard Biener <rguenther@suse.de>
26110 * system.h: Do not include gmp.h for building host tools.
26112 2013-01-15 Richard Biener <rguenther@suse.de>
26114 PR middle-end/55882
26115 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26116 account for bitpos when computing alignment.
26118 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26120 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26121 (ix86_target_macros_internal): Likewise.
26123 * config/i386/i386.c (m_CORE2I7): Removed.
26124 (m_CORE_HASWELL): New macro.
26125 (m_CORE_ALL): Likewise.
26126 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26127 (initial_ix86_arch_features): Likewise.
26128 (processor_target_table): Initializations for Core avx2.
26129 (cpu_names): New names "core-avx2".
26130 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26131 PROCESSOR_CORE_HASWELL.
26132 (ix86_issue_rate): New case.
26133 (ia32_multipass_dfa_lookahead): Likewise.
26134 (ix86_sched_init_global): Likewise.
26136 * config/i386/i386.h (TARGET_HASWELL): New macro.
26137 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26138 (processor_type): New PROCESSOR_HASWELL.
26140 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26142 PR tree-optimization/55955
26143 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26144 *SHIFT_EXPR and *ROTATE_EXPR codes.
26146 PR tree-optimization/48766
26147 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26148 -ftrapv disable -fwrapv.
26150 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26153 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26154 etc. to 1 and not to __flash.
26155 Use LL suffix for __INT24_MAX__ with -mint8.
26156 Use ULL suffix for __UINT24_MAX__ with -mint8.
26158 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26160 * config/avr/avr-arch.h
26161 (struct base_arch_s): Use typedef avr_arch_t instead.
26162 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26163 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26164 * config/avr/avr.c: Same.
26165 * config/avr/avr-devices.c: Same.
26166 * config/avr/driver-avr.c: Same.
26167 * config/avr/gen-avr-mmcu-texi.c: Same.
26168 * config/avr/avr-mcus.def: Adjust comment.
26170 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26172 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26173 * config/aarch64/iterators.md (VALLDI): New.
26175 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26176 Andi Kleen <ak@linux.intel.com>
26179 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26180 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26183 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26185 * config/avr/avr-stdint.h: Remove trailing blanks.
26186 * config/avr/avr-log.h: Same.
26187 * config/avr/avr-arch.h: Same.
26188 * config/avr/avr-devices.c: Same.
26189 * config/avr/avr-dimode.md: Same.
26190 * config/avr/predicates.md: Same.
26191 * config/avr/avr-c.c: Same. And fix typo.
26193 * config/avr/avr-protos.h: Same. And:
26194 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26195 (init_cumulative_args): Rename to avr_init_cumulative_args.
26196 (expand_prologue): Rename to avr_expand_prologue.
26197 (expand_epilogue): Rename to avr_expand_epilogue.
26198 (adjust_insn_length): Rename to avr_adjust_insn_length.
26199 (notice_update_cc): Rename to avr_notice_update_cc.
26200 (final_prescan_insn): Rename to avr_final_prescan_insn.
26201 * config/avr/avr.c: Same.
26202 * config/avr/avr.h: Same.
26203 * config/avr/avr.md: Remove trailing blanks.
26204 (prologue): Use avr_expand_prologue.
26205 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26207 2013-01-14 Richard Biener <rguenther@suse.de>
26209 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26210 verify_location, collect_subblocks): New functions.
26211 (verify_gimple_in_cfg): Verify that locations only reference
26212 BLOCKs in the functions BLOCK tree.
26214 2013-01-14 Richard Biener <rguenther@suse.de>
26216 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26218 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26220 (insert_out_of_ssa_copy_on_edge): Likewise.
26221 (rewrite_close_phi_out_of_ssa): Likewise.
26222 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26224 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26225 propagated constants.
26226 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26229 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26231 * config/avr/avr-modes.def: Add GPL copyright notice.
26233 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26235 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26236 MEMMODEL_MASK to determine memory model.
26237 (atomic_store<mode>): Ditto from operands[2].
26238 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26240 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26243 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26244 (fold_gimple_assign): Don't call unshare_expr here.
26245 (fold_ctor_reference): Call unshare_expr.
26247 2013-01-13 Terry Guo <terry.guo@arm.com>
26249 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26250 * doc/fragments.texi: Document MULTILIB_REUSE.
26251 * gcc.c (multilib_reuse): New internal spec.
26252 (set_multilib_dir): Also search multilib from multilib_reuse.
26253 * genmultilib (tmpmultilib3): Refactor code.
26254 (tmpmultilib4): Ditto.
26255 (multilib_reuse): New multilib argument.
26257 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26259 * Makefile.in: Update copyright.
26261 2013-01-12 Tom de Vries <tom@codesourcery.com>
26263 PR middle-end/55890
26264 * calls.c (expand_call): Check if arg_nr is valid.
26266 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26268 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26269 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26270 documentation. Add missing '__' in front of
26271 __builtin_ia32_packssdw256.
26273 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26276 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26277 NO_REGS for larl operands.
26278 (s390_reload_larl_operand): Use s390_load_address instead of
26281 2013-01-11 Richard Biener <rguenther@suse.de>
26283 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26284 (verify_node_sharing): ... here.
26285 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26287 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26289 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26290 Substitute TREECHECKING.
26291 * configure: Regenerate.
26292 * Makefile.in (TREECHECKING): New.
26294 2013-01-11 Richard Guenther <rguenther@suse.de>
26296 PR tree-optimization/44061
26297 * tree-vrp.c (extract_range_basic): Compute zero as
26298 value-range for __builtin_constant_p of function parameters.
26300 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26302 Update copyright years.
26304 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26306 PR rtl-optimization/55672
26307 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26308 const to be eliminable.
26310 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26312 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26313 * configure: Regenerate.
26315 2013-01-10 Richard Biener <rguenther@suse.de>
26317 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26319 2013-01-10 Richard Biener <rguenther@suse.de>
26322 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26323 locations for virtual PHI arguments.
26324 (rewrite_update_phi_arguments): Likewise.
26326 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26328 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26331 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26333 PR tree-optimization/55921
26334 * tree-complex.c (expand_complex_asm): New function.
26335 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26337 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26340 * config/s390/s390.c (s390_symref_operand_p)
26341 (s390_loadrelative_operand_p): Merge the two functions.
26342 (s390_check_qrst_address, print_operand_address): Add parameters
26343 to s390_loadrelative_operand_p invokation.
26344 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26345 (s390_reload_larl_operand, s390_secondary_reload): Use
26346 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26347 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26349 2013-01-09 Mike Stump <mikestump@comcast.net>
26351 * dse.c (record_store): Remove unnecessary assert.
26353 2013-01-09 Jan Hubicka <jh@suse.cz>
26355 PR tree-optimization/55569
26356 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26357 * cfgloop.h (scale_loop_profile): Likewise.
26359 2013-01-09 Jan Hubicka <jh@suse.cz>
26362 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26364 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26366 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26368 PR middle-end/55114
26369 * expr.h (maybe_emit_group_store): Declare.
26370 * expr.c (maybe_emit_group_store): New function.
26371 * builtins.c (expand_builtin_int_roundingfn): Call it.
26372 (expand_builtin_int_roundingfn_2): Likewise.
26374 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26376 PR rtl-optimization/55829
26377 * lra-constraints.c (match_reload): Add code for absent output.
26378 (curr_insn_transform): Add code for reloads of matched inputs
26381 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26383 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26384 attribute of movddup insn to DF.
26385 (*vec_interleave_lowv2df): Ditto.
26386 (vec_dupv2df): Ditto.
26388 2013-01-09 Jan Hubicka <jh@suse.cz>
26390 PR tree-optimiation/55875
26391 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26392 EVERY_ITERATION parameter.
26393 (number_of_iterations_exit): Check if exit is executed every iteration.
26394 (idx_infer_loop_bounds): Similarly here.
26395 (n_of_executions_at_most): Simplify
26396 to only test for cases where statement is dominated by the
26397 particular bound; handle correctly the "postdominance" test.
26398 (scev_probably_wraps_p): Use max loop iterations info
26399 as a global bound first.
26401 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26402 Nick Clifton <nickc@redhat.com>
26404 * config/v850/v850.md (cbranchsf4): New pattern.
26405 (cstoresf4): New pattern.
26406 (cbranchdf4): New pattern.
26407 (cstoredf4): New pattern.
26408 (movsicc): Disallow floating point comparisons.
26409 (cmpsf_le_insn): Fix order of operators.
26410 (cmpsf_lt_insn): Likewise.
26411 (cmpsf_eq_insn): Likewise.
26412 (cmpdf_le_insn): Likewise.
26413 (cmpdf_lt_insn): Likewise.
26414 (cmpdf_eq_insn): Likewise.
26415 (cmpsf_ge_insn): Use LE comparison.
26416 (cmpdf_ge_insn): Likewise.
26417 (cmpsf_gt_insn): Use LT comparison.
26418 (cmpdf_gt_insn): Likewise.
26419 (cmpsf_ne_insn): Delete pattern.
26420 (cmpdf_ne_insn): Delete pattern.
26421 * config/v850/v850.c (v850_gen_float_compare): Use
26422 gen_cmpdf_eq_insn for NE comparison.
26423 (v850_float_z_comparison_operator)
26424 (v850_float_nz_comparison_operator): Move from here ...
26425 * config/v850/predicates.md: ... to here. Move GT and GE
26426 comparisons into v850_float_z_comparison_operator.
26427 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26429 (v850_float_nz_comparison_operator): Likewise.
26431 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26433 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26434 with calls to gen_insvsi/gen_insvdi.
26436 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26438 * config/i386/i386.c (initial_ix86_tune_features): Set up
26439 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26441 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26442 Jakub Jelinek <jakub@redhat.com>
26444 PR tree-optimization/48189
26445 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26446 If nitercst is 0, don't predict the exit edge.
26448 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26450 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26451 in asm_fprintf with reg_names.
26452 (aarch64_print_operand_address): Likewise.
26453 (aarch64_return_addr): Likewise.
26454 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26456 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26458 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26459 (INT_U6_BITS): Likewise.
26460 * config/pa/predicates.md (uint6_operand): New predicate.
26461 (shift5_operand, shift6_operand): Likewise.
26462 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26464 (lshrdi3): Use shift6_operand.
26465 (shrpsi4, shrpdi4): New insn patterns.
26466 (extzv): Delete expander.
26467 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26468 predicates in unamed zero extract patterns. Tighten common constraint.
26469 (extv): Delete expander.
26470 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26471 predicates in unamed sign extract patterns. Tighten common constraint.
26472 (insv): Delete expander.
26473 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26474 predicates in unamed insert patterns. Tighten common constraint.
26475 Change uint32_operand predicate to uint6_operand predicate in unamed
26476 DImode pattern to insert constant values of type 1...1xxxx.
26478 2013-01-04 Jan Hubicka <jh@suse.cz>
26480 PR tree-optimization/55823
26481 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26484 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26485 Uros Bizjak <ubizjak@gmail.com>
26487 PR rtl-optimization/55845
26488 * df-problems.c (can_move_insns_across): Stop scanning at
26489 volatile_insn_p source instruction or give up if
26490 across_from .. across_to range contains any volatile_insn_p
26493 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26495 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26496 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26498 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26499 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26501 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26504 * asan.c (asan_clear_shadow): New function.
26505 (asan_emit_stack_protection): Use it.
26507 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26509 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26510 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26511 with tab instead of space.
26513 2013-01-08 Nick Clifton <nickc@redhat.com>
26515 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26516 register bank 0 at the start of an interrupt handler.
26517 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26520 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26522 * config/aarch64/aarch64-simd.md
26523 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26524 (aarch64_simd_bsl): Likewise.
26525 (aarch64_vcond_internal<mode>): Likewise.
26526 (vcond<mode><mode>): Likewise.
26527 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26528 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26530 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26532 * config/aarch64/aarch64-builtins.c
26533 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26535 2013-01-08 Martin Jambor <mjambor@suse.cz>
26538 * tree-sra.c (analyze_access_subtree): Return true also after
26539 potentially creating a debug-only replacement.
26541 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26543 PR middle-end/55890
26544 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26546 PR tree-optimization/54120
26547 * tree-vrp.c (range_fits_type_p): Don't allow
26548 src_precision < precision from signed vr to unsigned_p
26549 if vr->min or vr->max is negative.
26550 (simplify_float_conversion_using_ranges): Test can_float_p
26551 against CODE_FOR_nothing.
26553 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26554 Richard Biener <rguenther@suse.de>
26556 PR middle-end/55851
26557 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26558 types instead of just INTEGER_TYPE types.
26560 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26562 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26565 2013-01-07 Steve Ellcey <sellcey@mips.com>
26568 * config/mips/mips.opt: Change mad to mmad to match documentation.
26570 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26573 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26574 .progmemx.data now.
26576 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26579 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26580 (avr_addrspace_t): Add .section_name field.
26581 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26583 (avr_addrspace): Same. Initialize .section_name. Remove last
26584 NULL entry. Put __memx into .progmemx.data.
26585 (progmem_section_prefix): Remove.
26586 (avr_asm_init_sections): No need to initialize progmem_section.
26587 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26588 section name prefix.
26589 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26590 retrieve the progmem section.
26591 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26592 boundary to run over avr_addrspace[].
26593 (avr_register_target_pragmas): Ditto.
26595 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26597 * varasm.c (output_constant_def_contents): For asan_protect_global
26598 protected strings, adjust DECL_ALIGN if needed, before testing for
26600 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26601 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26603 (output_object_block): For asan protected decls, emit asan padding
26604 after their contents.
26605 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26606 (asan_finish_file): Test it here instead.
26608 2013-01-07 Nick Clifton <nickc@redhat.com>
26609 Matthias Klose <doko@debian.org>
26610 Doug Kwan <dougkwan@google.com>
26611 H.J. Lu <hongjiu.lu@intel.com>
26614 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26616 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26618 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26620 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26622 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26624 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26627 * doc/install.texi (Cross-Compiler-Specific Options): Document
26630 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26632 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26633 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26634 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26635 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26636 vqmovun_high_s64): Fix source operand number and update copyright.
26638 2013-01-07 Richard Biener <rguenther@suse.de>
26640 PR middle-end/55890
26641 * gimple.h (gimple_call_builtin_p): New overload.
26642 * gimple.c (validate_call): New function.
26643 (gimple_call_builtin_p): Likewise.
26644 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26645 Use gimple_call_builtin_p.
26646 (find_func_clobbers): Likewise.
26647 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26648 (strlen_optimize_stmt): Likewise.
26650 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26652 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26653 (vld1q_dup_*): Likewise.
26654 (vld1_*): Likewise.
26655 (vld1q_*): Likewise.
26656 (vld1_lane_*): Likewise.
26657 (vld1q_lane_*): Likewise.
26659 2013-01-07 Richard Biener <rguenther@suse.de>
26661 * lto-streamer.h (LTO_minor_version): Bump to 2.
26663 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26665 * config/aarch64/aarch64-protos.h
26666 (aarch64_const_double_zero_rtx_p): Rename to...
26667 (aarch64_float_const_zero_rtx_p): ...this.
26668 (aarch64_float_const_representable_p): New.
26669 (aarch64_output_simd_mov_immediate): Likewise.
26670 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26671 move immediate case.
26672 * config/aarch64/aarch64.c
26673 (aarch64_const_double_zero_rtx_p): Rename to...
26674 (aarch64_float_const_zero_rtx_p): ...this.
26675 (aarch64_print_operand): Allow printing of new constants.
26676 (aarch64_valid_floating_const): New.
26677 (aarch64_legitimate_constant_p): Check for valid floating-point
26679 (aarch64_simd_valid_immediate): Likewise.
26680 (aarch64_vect_float_const_representable_p): New.
26681 (aarch64_float_const_representable_p): Likewise.
26682 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26683 (aarch64_output_simd_mov_immediate): New.
26684 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26685 (*movdf_aarch64): Likewise.
26686 * config/aarch64/constraints.md (Ufc): New.
26687 (Y): call aarch64_float_const_zero_rtx.
26688 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26690 2013-01-07 Richard Biener <rguenther@suse.de>
26692 PR tree-optimization/55888
26693 PR tree-optimization/55862
26694 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26695 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26696 not if it is contained therein.
26698 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26700 * config/avr/t-avr: Typo.
26702 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26705 * config/avr/t-avr: Don't automatically rebuild
26706 $(srcdir)/config/avr/t-multilib
26707 $(srcdir)/config/avr/avr-tables.opt
26708 $(srcdir)/doc/avr-mmcu.texi
26709 (avr-mcus): New phony target to build them on request.
26710 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26711 * avr/avr-mcus.def: Adjust comments.
26713 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26715 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26717 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26719 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26721 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26723 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26725 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26727 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26728 to generate profiling.
26729 * config/rs6000/aix64.h (LIB_SPEC): Same.
26731 2013-01-04 Andrew Pinski <apinski@cavium.com>
26733 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26735 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26737 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26739 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26741 (ix86_expand_move): Ditto.
26742 (ix86_zero_extend_to_Pmode): Ditto.
26743 (ix86_expand_call): Ditto.
26744 (ix86_expand_special_args_builtin): Ditto.
26745 (ix86_expand_builtin): Ditto.
26747 2013-01-04 Richard Biener <rguenther@suse.de>
26749 PR tree-optimization/55862
26750 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26751 translating them through PHI nodes.
26753 2013-01-04 Martin Jambor <mjambor@suse.cz>
26755 PR tree-optimization/55755
26756 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26757 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26759 2013-01-04 Marek Polacek <polacek@redhat.com>
26761 PR middle-end/55859
26762 * opts.c (default_options_optimization): Clarify error message.
26764 2013-01-04 Richard Biener <rguenther@suse.de>
26766 PR middle-end/55863
26767 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26770 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26773 * config/pa/pa.md (movsi): Revert previous change.
26774 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26777 2013-01-03 Richard Henderson <rth@redhat.com>
26779 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26780 after eliminating TLS symbols.
26782 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26785 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26786 * graphite-poly.c (debug_gmp_value): Likewise.
26788 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26791 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26792 selected code model, define __code_mode_small__, __code_model_medium__,
26793 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26794 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26795 xchg temporary register with %k. Declare temporary register as
26797 [__x86_64__]: For medium and large code models, preserve %rbx register.
26799 2013-01-03 Richard Biener <rguenther@suse.de>
26801 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26802 (dump_subscript): Adjust.
26803 (finalize_ddr_dependent): Do not dump redundant info.
26804 (analyze_siv_subscript): Adjust.
26805 (subscript_dependence_tester): Likewise.
26806 (compute_affine_dependence): Likewise.
26808 2013-01-03 Richard Biener <rguenther@suse.de>
26811 2013-01-03 Richard Biener <rguenther@suse.de>
26813 PR tree-optimization/55857
26814 * tree-vect-stmts.c (vectorizable_load): Do not setup
26815 re-alignment for invariant loads.
26817 2013-01-02 Richard Biener <rguenther@suse.de>
26819 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26820 invariant load do not generate a vector load from the scalar location.
26822 2013-01-03 Richard Biener <rguenther@suse.de>
26824 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26825 for not vectorizing.
26826 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26827 not build INDIRECT_REFs, call get_name once only.
26828 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26829 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26831 2013-01-03 Richard Biener <rguenther@suse.de>
26833 PR tree-optimization/55857
26834 * tree-vect-stmts.c (vectorizable_load): Do not setup
26835 re-alignment for invariant loads.
26837 2013-01-03 Richard Biener <rguenther@suse.de>
26840 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26841 prefer a built-in decl.
26843 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26845 * gcc.c (process_command): Update copyright notice dates.
26846 * gcov.c (print_version): Likewise.
26847 * gcov-dump.c (print_version): Likewise.
26849 PR rtl-optimization/55838
26850 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26851 iv0.step, iv1.step and step.
26853 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26854 Marc Glisse <marc.glisse@inria.fr>
26856 PR tree-optimization/55832
26857 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26858 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26859 integer_{one,zero}_node.
26861 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26864 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26865 * var-tracking.c (reverse_op): Don't add reverse ops to
26866 VALUEs that have already
26867 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26869 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26871 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26873 2013-01-02 Teresa Johnson <tejohnson@google.com>
26875 * dumpfile.c (dump_loc): Print filename with location.
26876 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26877 new location_t parameter to emit complete unroll message with
26878 new dump framework.
26879 (canonicalize_loop_induction_variables): Compute loops location
26880 and pass to try_unroll_loop_completely.
26881 * loop-unroll.c (report_unroll_peel): New function.
26882 (peel_loops_completely): Use new dump format with location
26883 for main dumpfile message, and invoke report_unroll_peel on success.
26884 (decide_unrolling_and_peeling): Ditto.
26885 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26886 by report_unroll_peel.
26887 (decide_peel_completely): Ditto.
26888 (decide_unroll_constant_iterations): Ditto.
26889 (decide_unroll_runtime_iterations): Ditto.
26890 (decide_peel_simple): Ditto.
26891 (decide_unroll_stupid): Ditto.
26892 * cfgloop.c (get_loop_location): New function.
26893 * cfgloop.h (get_loop_location): Declare.
26895 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26897 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26900 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26902 PR middle-end/55198
26903 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26904 BLKmode objects when EXPAND_MEMORY is specified.
26906 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26908 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26910 (fold_builtin_cpu): Do not share cpu model decls across statements.
26912 2013-01-02 Jason Merrill <jason@redhat.com>
26915 * tree.c (build_array_type_1): Revert earlier change.
26917 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26919 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26921 * config/aarch64/aarch64-tune.md: Re-generate.
26923 2013-01-02 Richard Biener <rguenther@suse.de>
26925 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26926 invariant load do not generate a vector load from the scalar location.
26928 2013-01-02 Richard Biener <rguenther@suse.de>
26931 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26932 * configure: Regenerate.
26934 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26936 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26937 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26938 (expand_builtin_int_roundingfn_2): Keep the original target around
26939 for the fallback case.
26941 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26943 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26944 to be clear for sign changes.
26946 2013-01-01 Jan Hubicka <jh@suse.cz>
26948 * ipa-inline-analysis.c: Fix formatting.
26950 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26952 PR tree-optimization/55831
26953 * tree-vect-loop.c (get_initial_def_for_induction): Use
26954 gsi_after_labels instead of gsi_start_bb.
26956 Copyright (C) 2013 Free Software Foundation, Inc.
26958 Copying and distribution of this file, with or without modification,
26959 are permitted in any medium without royalty provided the copyright
26960 notice and this notice are preserved.