1 2013-10-18 Richard Biener <rguenther@suse.de>
3 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
4 or TYPE_UNSIGNED of integral types.
5 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
6 NULL_TREE for zero-precision integral types.
8 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
10 * config/aarch64/arm_neon.h
11 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
13 2013-10-17 Sriraman Tallam <tmsriram@google.com>
15 * opth-gen.awk: Define target_flags_explicit.
17 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
19 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
20 fields to the reg_addr array that describes the valid addressing
21 mode for any register, general purpose registers, floating point
22 registers, and Altivec registers.
23 (FIRST_RELOAD_REG_CLASS): Likewise.
24 (LAST_RELOAD_REG_CLASS): Likewise.
25 (struct reload_reg_map_type): Likewise.
26 (reload_reg_map_type): Likewise.
27 (RELOAD_REG_VALID): Likewise.
28 (RELOAD_REG_MULTIPLE): Likewise.
29 (RELOAD_REG_INDEXED): Likewise.
30 (RELOAD_REG_OFFSET): Likewise.
31 (RELOAD_REG_PRE_INCDEC): Likewise.
32 (RELOAD_REG_PRE_MODIFY): Likewise.
34 (mode_supports_pre_incdec_p): New helper functions to say whether
35 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
36 (mode_supports_pre_modify_p): Likewise.
37 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
38 print the valid address mode bits for each mode.
39 (rs6000_debug_print_mode): Likewise.
40 (rs6000_debug_reg_global): Likewise.
41 (rs6000_setup_reg_addr_masks): New function to set up the address
42 mask bits for each type.
43 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
44 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
45 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
46 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
47 PRE_MODIFY are supported.
48 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
49 registers, instead of {src,dest}_av_p.
50 (rs6000_print_options_internal): Tweak the debug output slightly.
52 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
54 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
57 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
59 * tree-flow.h (struct omp_region): Move to omp-low.c.
60 Remove omp_ prototypes and variables.
61 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
62 (copy_var_decl): Relocate prototype from tree-flow.h.
63 * gimple.c (copy_var_decl): Relocate from omp-low.c.
64 * tree.h: Move prototype to omp-low.h.
65 * omp-low.h: New File. Relocate prototypes here.
66 * omp-low.c (struct omp_region): Make local here.
67 (root_omp_region): Make static.
68 (copy_var_decl) Move to gimple.c.
69 (new_omp_region): Make static.
70 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
71 * tree-cfg.c: Include omp-low.h.
72 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
73 * gimplify.c: Include omp-low.h.
74 * tree-parloops.c: Likewise.
76 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
78 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
79 operands are in memory, prefer to force non-matched operand 1 to
82 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
85 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
86 restrict TImode addresses to single indirect registers if both
87 -mquad-memory and -mvsx-timode are used.
88 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
89 we should emit load/store quad. Remove using %y for quad memory
92 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
93 constraints to allow load/store quad on machines where TImode is
94 not allowed in VSX registers. Use 'n' instead of 'F' constraint
95 for TImode to load integer constants.
97 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
99 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
101 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
103 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
104 handling of STACK_REG.
106 2013-10-17 Richard Biener <rguenther@suse.de>
108 PR tree-optimization/58143
109 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
111 (rewrite_to_defined_overflow): Likewise.
112 (move_computations_dom_walker::before_dom): Rewrite stmts
113 with undefined signed overflow that are not always executed
114 into unsigned arithmetic.
116 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
119 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
120 explicit isa flag to be an options variable, instead of using
121 global_options_set. Remove define from rs6000.h.
122 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
124 * config/rs6000/rs6000.c (rs6000_option_override_internal):
125 Initialize rs6000_isa_flags_explicit.
126 (rs6000_function_specific_save): Add gcc_options* parameter, so
127 that the powerpc builds after the 2013-10-15 changes.
128 (rs6000_function_specific_restore): Likewise.
130 2013-10-16 DJ Delorie <dj@redhat.com>
132 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
133 op is a REG before checking REGNO.
134 (rl78_alloc_physical_registers): Verify pattern is a SET before
137 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
139 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
141 (vec_unpacks_lo_v4sf): Likewise.
142 (vec_unpacks_float_hi_v4si): Likewise.
143 (vec_unpacks_float_lo_v4si): Likewise.
144 (vec_unpacku_float_hi_v4si): Likewise.
145 (vec_unpacku_float_lo_v4si): Likewise.
147 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
149 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
150 (vsx_concat_v2sf): Likewise.
152 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
154 * config/aarch64/aarch64.md
155 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
157 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
159 PR tree-optimization/58697
160 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
161 estimated_loop_iterations_int.
162 (max_stmt_executions_int): Call get_max_loop_iterations_int.
163 (get_max_loop_iterations_int): New. HWINT version of
164 get_max_loop_iterations.
165 * cfgloop.h: Add prototypes.
166 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
167 * loop-unroll.c (decide_peel_once_rolling): Call
168 get_estimated_loop_iterations_int.
169 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
170 * tree-ssa-loop-niter.h: Tweak prototypes.
172 2013-10-16 David Malcolm <dmalcolm@redhat.com>
174 * gengtype-parse.c (struct_field_seq): Ignore access-control
175 keywords ("public:" etc).
177 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
179 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
180 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
182 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
184 * config/arm/arm.opt (mlra): New option.
185 * config/arm/arm.c (arm_lra_p): New function.
186 (TARGET_LRA_P): Define.
188 2013-10-16 Paulo Matos <pmatos@broadcom.com>
190 * tree-core.h (tree_code_name): Remove.
191 * tree.h (get_tree_code_name): New prototype.
192 * tree.c (tree_code_name): Make static.
193 (get_tree_code_name): New function.
194 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
195 tree_class_check_failed, tree_range_check_failed,
196 tree_not_class_check_failed, omp_clause_check_failed,
197 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
198 wrapper get_tree_code_name instead of calling tree_code_name directly.
199 * tree-vrp.c (dump_asserts_for): Likewise.
200 * tree-dump.c (dequeue_and_dump): Likewise.
201 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
202 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
203 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
204 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
205 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
206 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
207 dump_gimple_omp_for): Likewise.
208 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
209 * tree-ssa-pre.c (print_pre_expr): Likewise.
210 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
211 * print-tree.c (print_node_brief, print_node): Likewise.
212 * gimple.c (gimple_check_failed): Likewise.
213 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
214 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
215 * config/mep/mep.c (mep_validate_vliw): Likewise.
216 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
217 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
219 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
221 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
224 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
226 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
227 (MULTILIB_MATCHES): Add multilib for -march=v8.
229 2013-10-15 Sriraman Tallam <tmsriram@google.com>
232 * optc-save-gen.awk: Add extra parameter to the save and restore
234 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
235 * tree.c (build_optimization_node): New parameter. Add extra parameter
236 to call to cl_optimization_save.
237 (build_target_option_node): New parameter. Add extra parameter
238 to call to cl_target_option_save.
239 * tree.h (build_optimization_node): New parameter.
240 (build_target_option_node): New parameter.
241 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
242 build_optimization_node and build_target_option_node.
243 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
244 (handle_pragma_push_options): Ditto.
245 * toplev.c (process_options): Ditto.
246 * opts.c (init_options_struct): Check for opts_set non-null.
247 * target.def (target_option.save): New parameter.
248 (target_option.restore): New parameter.
250 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
251 (ix86_pragma_target_parse): Ditto.
252 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
254 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
255 to build_optimization_node and build_target_option_node.
256 (rs6000_valid_attribute_p): Ditto.
257 (rs6000_pragma_target_parse): Ditto.
258 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
260 * config/i386/i386.h:
261 TARGET_64BIT_P: New Macro
262 TARGET_MMX_P: New Macro.
263 TARGET_3DNOW_P: New Macro.
264 TARGET_3DNOW_A_P: New Macro.
265 TARGET_SSE_P: New Macro.
266 TARGET_SSE2_P: New Macro.
267 TARGET_SSE3_P: New Macro.
268 TARGET_SSSE3_P: New Macro.
269 TARGET_SSE4_1_P: New Macro.
270 TARGET_SSE4_2_P: New Macro.
271 TARGET_AVX_P: New Macro.
272 TARGET_AVX2_P: New Macro.
273 TARGET_AVX512F_P: New Macro.
274 TARGET_AVX512PF_P: New Macro.
275 TARGET_AVX512ER_P: New Macro.
276 TARGET_AVX512CD_P: New Macro.
277 TARGET_FMA_P: New Macro.
278 TARGET_SSE4A_P: New Macro.
279 TARGET_FMA4_P: New Macro.
280 TARGET_XOP_P: New Macro.
281 TARGET_LWP_P: New Macro.
282 TARGET_ABM_P: New Macro.
283 TARGET_BMI_P: New Macro.
284 TARGET_BMI2_P: New Macro.
285 TARGET_LZCNT_P: New Macro.
286 TARGET_TBM_P: New Macro.
287 TARGET_POPCNT_P: New Macro.
288 TARGET_SAHF_P: New Macro.
289 TARGET_MOVBE_P: New Macro.
290 TARGET_CRC32_P: New Macro.
291 TARGET_AES_P: New Macro.
292 TARGET_PCLMUL_P: New Macro.
293 TARGET_CMPXCHG16B_P: New Macro.
294 TARGET_FSGSBASE_P: New Macro.
295 TARGET_RDRND_P: New Macro.
296 TARGET_F16C_P: New Macro.
297 TARGET_RTM_P: New Macro.
298 TARGET_HLE_P: New Macro.
299 TARGET_RDSEED_P: New Macro.
300 TARGET_PRFCHW_P: New Macro.
301 TARGET_ADX_P: New Macro.
302 TARGET_FXSR_P: New Macro.
303 TARGET_XSAVE_P: New Macro.
304 TARGET_XSAVEOPT_P: New Macro.
305 TARGET_LP64_P: New Macro.
306 TARGET_X32_P: New Macro.
307 TARGET_FPMATH_DEFAULT_P: New Macro.
308 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
309 * config/i386/i386.c (ix86_option_override_internal): New parameters.
311 Change ix86_tune_string to access opts->x_ix86_tune_string.
312 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
313 Change ix86_arch_string to access opts->x_ix86_arch_string.
314 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
315 Change ix86_pmode to access opts->x_ix86_pmode.
316 Change ix86_abi to access opts->x_ix86_abi.
317 Change ix86_cmodel to access opts->x_ix86_cmodel.
318 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
319 Change ix86_isa_flags_explicit to access
320 opts->x_ix86_isa_flags_explicit.
321 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
322 Change ix86_regparm to access opts->x_ix86_regparm.
323 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
324 Change ix86_preferred_stack_boundary_arg to access
325 opts->x_ix86_preferred_stack_boundary_arg.
326 Change ix86_force_align_arg_pointer to access
327 opts->x_ix86_force_align_arg_pointer.
328 Change ix86_incoming_stack_boundar_arg to access
329 opts->x_ix86_incoming_stack_boundar_arg.
330 Change ix86_fpmath to access opts->x_ix86_fpmath.
331 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
332 Change ix86_recip_name to access opts->x_ix86_recip_name.
333 Change ix86_stack_protector_guard to access
334 opts->x_ix86_stack_protector_guard.
335 Change ix86_tune_memcpy_strategy to access
336 opts->x_ix86_tune_memcpy_strategy.
337 Change ix86_tune_memset_strategy to access
338 opts->x_ix86_tune_memset_strategy.
339 Change global_options to access opts.
340 Change global_options_set to access opts_set.
341 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
342 Change TARGET_MMX to TARGET_MMX_P (opts->...).
343 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
344 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
345 Change TARGET_SSE to TARGET_SSE_P (opts->...).
346 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
347 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
348 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
349 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
350 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
351 Change TARGET_AVX to TARGET_AVX_P (opts->...).
352 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
353 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
354 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
355 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
356 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
357 Change TARGET_FMA to TARGET_FMA_P (opts->...).
358 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
359 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
360 Change TARGET_XOP to TARGET_XOP_P (opts->...).
361 Change TARGET_LWP to TARGET_LWP_P (opts->...).
362 Change TARGET_ABM to TARGET_ABM_P (opts->...).
363 Change TARGET_BMI to TARGET_BMI_P (opts->...).
364 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
365 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
366 Change TARGET_TBM to TARGET_TBM_P (opts->...).
367 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
368 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
369 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
370 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
371 Change TARGET_AES to TARGET_AES_P (opts->...).
372 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
373 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
374 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
375 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
376 Change TARGET_F16C to TARGET_F16C_P (opts->...).
377 Change TARGET_RTM to TARGET_RTM_P (opts->...).
378 Change TARGET_HLE to TARGET_HLE_P (opts->...).
379 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
380 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
381 Change TARGET_ADX to TARGET_ADX_P (opts->...).
382 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
383 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
384 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
385 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
386 Change TARGET_X32 to TARGET_X32_P (opts->...).
387 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
388 Change TARGET_FLOAT_RETURNS_IN_80387 to
389 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
390 (ix86_function_specific_save): New parameter. Use opts-> fields
391 to replace global fields.
392 (ix86_function_specific_restore): Ditto.
393 (ix86_valid_target_attribute_inner_p): New parameters.
395 Fix call to ix86_handle_option and set_option.
396 (ix86_valid_target_attribute_tree): New parameters.
397 Change global_options to access opts.
398 Change global_options_set to access opts_set.
399 Fix call to ix86_valid_target_attribute_inner_p.
400 Change ix86_tune_string to access opts->x_ix86_tune_string.
401 Change ix86_arch_string to access opts->x_ix86_arch_string.
402 Change ix86_fpmath to access opts->x_ix86_fpmath
403 Fix call to ix86_option_override_internal.
404 Fix call to ix86_add_new_builtins.
405 Fix calls to build_optimization_node and build_target_option_node.
406 (ix86_valid_target_attribute_p): Remove access to global_options.
407 Use new gcc_options structure func_options.
408 Fix call to ix86_valid_target_attribute_tree.
409 Fix call to build_optimization_node.
410 (get_builtin_code_for_version): Fix call to
411 ix86_valid_target_attribute_tree.
413 2013-10-15 David Malcolm <dmalcolm@redhat.com>
415 * Makefile.in (PICFLAG): New.
416 (enable_host_shared): New.
417 (INTERNAL_CFLAGS): Use PICFLAG.
418 (LIBIBERTY): Use pic build of libiberty.a if configured with
419 --enable-host-shared.
420 * configure.ac: Add --enable-host-shared, setting up new
422 * configure: Regenerate.
423 * doc/install.texi (--enable-shared): Add note contrasting it with ...
424 (--enable-host-shared): New option.
426 2013-10-15 Richard Biener <rguenther@suse.de>
428 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
429 for built-in functions.
431 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
433 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
434 (optimize_range_tests_1): New function,
435 extracted from optimize_range_tests.
436 (optimize_range_tests_xor): Similarly.
437 (optimize_range_tests_diff): New function.
438 (optimize_range_tests): Use optimize_range_tests_1.
440 2013-10-15 Cong Hou <congh@google.com>
442 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
443 requirement of the reduction pattern so that one operand of the
444 reduction operation can come from outside of the loop.
446 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
448 * config/arm/neon-schedgen.ml: Remove.
449 * config/arm/cortex-a9-neon.md: Remove comment regarding
452 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
454 * config/arm/types: Remove old neon types.
456 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
458 * config/arm/cortex-a7.md
459 (cortex_a7_neon_type): New.
460 (cortex_a7_neon_mul): Update for new types.
461 (cortex_a7_neon_mla): Likewise.
462 (cortex_a7_neon): Likewise.
464 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
466 * config/arm/cortex-a15-neon.md
467 (cortex_a15_neon_type): New,
469 (cortex_a15_neon_int_1): Remove.
470 (cortex_a15_neon_int_2): Likewise.
471 (cortex_a15_neon_int_3): Likewise.
472 (cortex_a15_neon_int_4): Likewise.
473 (cortex_a15_neon_int_5): Likewise.
474 (cortex_a15_neon_vqneg_vqabs): Likewise.
475 (cortex_a15_neon_vmov): Likewise.
476 (cortex_a15_neon_vaba): Likewise.
477 (cortex_a15_neon_vaba_qqq): Likewise.
478 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
479 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
480 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
482 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
483 (cortex_a15_neon_mla_qqq_8_16): Likewise.
484 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
485 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
486 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
487 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
488 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
489 (cortex_a15_neon_shift_1): Likewise.
490 (cortex_a15_neon_shift_2): Likewise.
491 (cortex_a15_neon_shift_3): Likewise.
492 (cortex_a15_neon_vshl_ddd): Likewise.
493 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
494 (cortex_a15_neon_vsra_vrsra): Likewise.
495 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
496 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
497 (cortex_a15_neon_bp_3cycle): Likewise.
498 (cortex_a15_neon_ldm_2): Likewise.
499 (cortex_a15_neon_stm_2): Likewise.
500 (cortex_a15_neon_mcr): Likewise.
501 (cortex_a15_neon_mrc): Likewise.
502 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
503 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
504 (cortex_a15_neon_fp_vmul_ddd): Likewise.
505 (cortex_a15_neon_fp_vmul_qqd): Likewise.
506 (cortex_a15_neon_fp_vmla_ddd): Likewise.
507 (cortex_a15_neon_fp_vmla_qqq): Likewise.
508 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
509 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
510 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
511 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
512 (cortex_a15_neon_bp_simple): Likewise.
513 (cortex_a15_neon_bp_2cycle): Likewise.
514 (cortex_a15_neon_bp_3cycle): Likewise.
515 (cortex_a15_neon_vld1_1_2_regs): Likewise.
516 (cortex_a15_neon_vld1_3_4_regs): Likewise.
517 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
518 (cortex_a15_neon_vld2_4_regs): Likewise.
519 (cortex_a15_neon_vld3_vld4): Likewise.
520 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
521 (cortex_a15_neon_vst1_3_4_regs): Likewise.
522 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
523 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
524 (cortex_a15_neon_vst3_vst4): Rename to...
525 (cortex_a15_neon_vst4): This, update for new attributes.
526 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
527 (cortex_a15_neon_vld3_vld4_lane): Likewise.
528 (cortex_a15_neon_vst1_vst2_lane): Likewise.
529 (cortex_a15_neon_vst3_vst4_lane): Likewise.
530 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
531 (cortex_a15_neon_ldm_2): Likewise.
532 (cortex_a15_neon_stm_2): Likewise.
533 (cortex_a15_neon_mcr): Likewise.
534 (cortex_a15_neon_mcr_2_mcrr): Likewise.
535 (cortex_a15_neon_mrc): Likewise.
536 (cortex_a15_neon_mrrc): Likewise.
538 (cortex_a15_neon_abd): New.
539 (cortex_a15_neon_abd_q): Likewise.
540 (cortex_a15_neon_aba): Likewise.
541 (cortex_a15_neon_aba_q): Likewise.
542 (cortex_a15_neon_acc): Likewise.
543 (cortex_a15_neon_acc_q): Likewise.
544 (cortex_a15_neon_arith_basic): Likewise.
545 (cortex_a15_neon_arith_complex): Likewise.
546 (cortex_a15_neon_multiply): Likewise.
547 (cortex_a15_neon_multiply_q): Likewise.
548 (cortex_a15_neon_mla): Likewise.
549 (cortex_a15_neon_mla_q): Likewise.
550 (cortex_a15_neon_sat_mla_long): Likewise.
551 (cortex_a15_neon_shift_acc): Likewise.
552 (cortex_a15_neon_shift_imm_basic): Likewise.
553 (cortex_a15_neon_shift_imm_complex): Likewise.
554 (cortex_a15_neon_shift_reg_basic): Likewise.
555 (cortex_a15_neon_shift_reg_basic_q): Likewise.
556 (cortex_a15_neon_shift_reg_complex): Likewise.
557 (cortex_a15_neon_shift_reg_complex_q): Likewise.
558 (cortex_a15_neon_fp_negabs): Likewise
559 (cortex_a15_neon_fp_arith): Likewise
560 (cortex_a15_neon_fp_arith_q): Likewise
561 (cortex_a15_neon_fp_cvt_int): Likewise
562 (cortex_a15_neon_fp_cvt_int_q): Likewise
563 (cortex_a15_neon_fp_cvt_16): Likewise
564 (cortex_a15_neon_fp_mul): Likewise
565 (cortex_a15_neon_fp_mul_q): Likewise
566 (cortex_a15_neon_fp_mla): Likewise
567 (cortex_a15_neon_fp_mla_q): Likewise
568 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
569 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
570 (cortex_a15_neon_bitops): Likewise.
571 (cortex_a15_neon_bitops_q): Likewise.
572 (cortex_a15_neon_from_gp): Likewise.
573 (cortex_a15_neon_from_gp_q): Likewise.
574 (cortex_a15_neon_tbl3_tbl4): Likewise.
575 (cortex_a15_neon_zip_q): Likewise.
576 (cortex_a15_neon_to_gp): Likewise.
577 (cortex_a15_neon_load_a): Likewise.
578 (cortex_a15_neon_load_b): Likewise.
579 (cortex_a15_neon_load_c): Likewise.
580 (cortex_a15_neon_load_d): Likewise.
581 (cortex_a15_neon_load_e): Likewise.
582 (cortex_a15_neon_load_f): Likewise.
583 (cortex_a15_neon_store_a): Likewise.
584 (cortex_a15_neon_store_b): Likewise.
585 (cortex_a15_neon_store_c): Likewise.
586 (cortex_a15_neon_store_d): Likewise.
587 (cortex_a15_neon_store_e): Likewise.
588 (cortex_a15_neon_store_f): Likewise.
589 (cortex_a15_neon_store_g): Likewise.
590 (cortex_a15_neon_store_h): Likewise.
591 (cortex_a15_vfp_to_from_gp): Likewise.
593 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
595 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
597 (cortex_a9_neon_vshl_ddd): Remove.
598 (cortex_a9_neon_vst3_vst4): Likewise.
599 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
601 (cortex_a9_neon_bit_ops_q): New.
603 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
604 (cortex_a9_neon_int_2): Likewise.
605 (cortex_a9_neon_int_3): Likewise.
606 (cortex_a9_neon_int_4): Likewise.
607 (cortex_a9_neon_int_5): Likewise.
608 (cortex_a9_neon_vqneg_vqabs): Likewise.
609 (cortex_a9_neon_vmov): Likewise.
610 (cortex_a9_neon_vaba): Likewise.
611 (cortex_a9_neon_vaba_qqq): Likewise.
612 (cortex_a9_neon_shift_1): Likewise.
613 (cortex_a9_neon_shift_2): Likewise.
614 (cortex_a9_neon_shift_3): Likewise.
615 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
616 (cortex_a9_neon_vsra_vrsra): Likewise.
617 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
618 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
619 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
621 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
622 (cortex_a9_neon_mla_qqq_8_16): Likewise.
623 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
625 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
626 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
627 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
628 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
629 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
630 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
631 (cortex_a9_neon_fp_vsum): Likewise.
632 (cortex_a9_neon_fp_vmul_ddd): Likewise.
633 (cortex_a9_neon_fp_vmul_qqd): Likewise.
634 (cortex_a9_neon_fp_vmla_ddd): Likewise.
635 (cortex_a9_neon_fp_vmla_qqq): Likewise.
636 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
637 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
638 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
639 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
640 (cortex_a9_neon_bp_simple): Likewise.
641 (cortex_a9_neon_bp_2cycle): Likewise.
642 (cortex_a9_neon_bp_3cycle): Likewise.
643 (cortex_a9_neon_ldr): Likewise.
644 (cortex_a9_neon_str): Likewise.
645 (cortex_a9_neon_vld1_1_2_regs): Likewise.
646 (cortex_a9_neon_vld1_3_4_regs): Likewise.
647 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
648 (cortex_a9_neon_vld2_4_regs): Likewise.
649 (cortex_a9_neon_vld3_vld4): Likewise.
650 (cortex_a9_neon_vld1_vld2_lane): Likewise.
651 (cortex_a9_neon_vld3_vld4_lane): Likewise.
652 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
653 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
654 (cortex_a9_neon_vst1_3_4_regs): Likewise.
655 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
656 (cortex_a9_neon_vst1_vst2_lane): Likewise.
657 (cortex_a9_neon_vst3_vst4_lane): Likewise.
658 (cortex_a9_neon_mcr): Likewise.
659 (cortex_a9_neon_mcr_2_mcrr): Likewise.
660 (cortex_a9_neon_mrc): Likewise.
661 (cortex_a9_neon_mrrc): Likewise.
663 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
665 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
667 (cortex_a8_neon_vshl_ddd): Remove.
668 (cortex_a8_neon_vst3_vst4): Likewise.
669 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
671 (cortex_a8_neon_bit_ops_q): New.
673 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
674 (cortex_a8_neon_int_2): Likewise..
675 (cortex_a8_neon_int_3): Likewise.
676 (cortex_a8_neon_int_5): Likewise.
677 (cortex_a8_neon_vqneg_vqabs): Likewise.
678 (cortex_a8_neon_int_4): Likewise.
679 (cortex_a8_neon_vaba): Likewise.
680 (cortex_a8_neon_vaba_qqq): Likewise.
681 (cortex_a8_neon_shift_1): Likewise.
682 (cortex_a8_neon_shift_2): Likewise.
683 (cortex_a8_neon_shift_3): Likewise.
684 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
685 (cortex_a8_neon_vsra_vrsra): Likewise.
686 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
687 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
688 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
690 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
691 (cortex_a8_neon_mla_qqq_8_16): Likewise.
692 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
694 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
695 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
696 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
697 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
698 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
699 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
700 (cortex_a8_neon_fp_vsum): Likewise.
701 (cortex_a8_neon_fp_vmul_ddd): Likewise.
702 (cortex_a8_neon_fp_vmul_qqd): Likewise.
703 (cortex_a8_neon_fp_vmla_ddd): Likewise.
704 (cortex_a8_neon_fp_vmla_qqq): Likewise.
705 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
706 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
707 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
708 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
709 (cortex_a8_neon_bp_simple): Likewise.
710 (cortex_a8_neon_bp_2cycle): Likewise.
711 (cortex_a8_neon_bp_3cycle): Likewise.
712 (cortex_a8_neon_ldr): Likewise.
713 (cortex_a8_neon_str): Likewise.
714 (cortex_a8_neon_vld1_1_2_regs): Likewise.
715 (cortex_a8_neon_vld1_3_4_regs): Likewise.
716 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
717 (cortex_a8_neon_vld2_4_regs): Likewise.
718 (cortex_a8_neon_vld3_vld4): Likewise.
719 (cortex_a8_neon_vld1_vld2_lane): Likewise.
720 (cortex_a8_neon_vld3_vld4_lane): Likewise.
721 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
722 (cortex_a8_neon_vst1_3_4_regs): Likewise.
723 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
724 (cortex_a8_neon_vst1_vst2_lane): Likewise.
725 (cortex_a8_neon_vst3_vst4_lane): Likewise.
726 (cortex_a8_neon_mcr): Likewise.
727 (cortex_a8_neon_mcr_2_mcrr): Likewise.
728 (cortex_a8_neon_mrc): Likewise.
729 (cortex_a8_neon_mrrc): Likewise.
731 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
733 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
735 * config/aarch64/aarch64-simd.md (neon_type): Remove.
736 (aarch64_simd_dup<mode>): Add "type" attribute.
737 (aarch64_dup_lane<mode>): Likewise.
738 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
739 (*aarch64_simd_mov<mode>): Likewise.
740 (aarch64_simd_mov_from_<mode>low): Likewise.
741 (aarch64_simd_mov_from_<mode>high): Likewise.
742 (orn<mode>3): Likewise.
743 (bic<mode>3): Likewise.
744 (add<mode>3): Likewise.
745 (sub<mode>3): Likewise.
746 (mul<mode>3): Likewise.
747 (*aarch64_mul3_elt<mode>): Likewise.
748 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
749 (*aarch64_mul3_elt_to_128df): Likewise.
750 (*aarch64_mul3_elt_to_64v2df): Likewise.
751 (neg<mode>2): Likewise.
752 (abs<mode>2): Likewise.
753 (abd<mode>_3): Likewise.
754 (aba<mode>_3): Likewise.
755 (fabd<mode>_3): Likewise.
756 (*fabd_scalar<mode>3): Likewise.
757 (and<mode>3): Likewise.
758 (ior<mode>3): Likewise.
759 (xor<mode>3): Likewise.
760 (one_cmpl<mode>2): Likewise.
761 (aarch64_simd_vec_set<mode>): Likewise.
762 (aarch64_simd_lshr<mode>): Likewise.
763 (aarch64_simd_ashr<mode>): Likewise.
764 (aarch64_simd_imm_shl<mode>): Likewise.
765 (aarch64_simd_reg_sshl<mode): Likewise.
766 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
767 (aarch64_simd_reg_shl<mode>_signed): Likewise.
768 (aarch64_simd_vec_setv2di): Likewise.
769 (aarch64_simd_vec_set<mode>): Likewise.
770 (aarch64_mla<mode>): Likewise.
771 (*aarch64_mla_elt<mode>): Likewise.
772 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
773 (aarch64_mls<mode>): Likewise.
774 (*aarch64_mls_elt<mode>): Likewise.
775 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
776 (<su><maxmin><mode>3): Likewise.
777 (move_lo_quad_<mode>): Likewise.
778 (aarch64_simd_move_hi_quad_<mode>): Likewise.
779 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
780 (vec_pack_trunc_<mode>): Likewise.
781 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
782 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
783 (*aarch64_<su>mlal_lo<mode>): Likewise.
784 (*aarch64_<su>mlal_hi<mode>): Likewise.
785 (*aarch64_<su>mlsl_lo<mode>): Likewise.
786 (*aarch64_<su>mlsl_hi<mode>): Likewise.
787 (*aarch64_<su>mlal<mode>): Likewise.
788 (*aarch64_<su>mlsl<mode>): Likewise.
789 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
790 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
791 (add<mode>3): Likewise.
792 (sub<mode>3): Likewise.
793 (mul<mode>3): Likewise.
794 (div<mode>3): Likewise.
795 (neg<mode>2): Likewise.
796 (abs<mode>2): Likewise.
797 (fma<mode>4): Likewise.
798 (*aarch64_fma4_elt<mode>): Likewise.
799 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
800 (*aarch64_fma4_elt_to_128df): Likewise.
801 (*aarch64_fma4_elt_to_64v2df): Likewise.
802 (fnma<mode>4): Likewise.
803 (*aarch64_fnma4_elt<mode>): Likewise.
804 (*aarch64_fnma4_elt_<vswap_width_name><mode>
805 (*aarch64_fnma4_elt_to_128df): Likewise.
806 (*aarch64_fnma4_elt_to_64v2df): Likewise.
807 (<frint_pattern><mode>2): Likewise.
808 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
809 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
810 (vec_unpacks_lo_v4sf): Likewise.
811 (aarch64_float_extend_lo_v2df): Likewise.
812 (vec_unpacks_hi_v4sf): Likewise.
813 (aarch64_float_truncate_lo_v2sf): Likewise.
814 (aarch64_float_truncate_hi_v4sf): Likewise.
815 (aarch64_vmls<mode>): Likewise.
816 (<su><maxmin><mode>3): Likewise.
817 (<maxmin_uns><mode>3): Likewise.
818 (reduc_<sur>plus_<mode>): Likewise.
819 (reduc_<sur>plus_v2di): Likewise.
820 (reduc_<sur>plus_v2si): Likewise.
821 (reduc_<sur>plus_<mode>): Likewise.
822 (aarch64_addpv4sf): Likewise.
823 (clz<mode>2): Likewise.
824 (reduc_<maxmin_uns>_<mode>): Likewise.
825 (reduc_<maxmin_uns>_v2di): Likewise.
826 (reduc_<maxmin_uns>_v2si): Likewise.
827 (reduc_<maxmin_uns>_<mode>): Likewise.
828 (reduc_<maxmin_uns>_v4sf): Likewise.
829 (aarch64_simd_bsl<mode>_internal): Likewise.
830 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
831 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
832 (aarch64_get_lane<mode>): Likewise.
833 (*aarch64_combinez<mode>): Likewise.
834 (aarch64_combine<mode>): Likewise.
835 (aarch64_simd_combine<mode>): Likewise.
836 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
837 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
838 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
839 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
840 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
841 (aarch64_<sur>h<addsub><mode>): Likewise.
842 (aarch64_<sur><addsub>hn<mode>): Likewise.
843 (aarch64_<sur><addsub>hn2<mode>): Likewise.
844 (aarch64_pmul<mode>): Likewise.
845 (aarch64_<su_optab><optab><mode>): Likewise.
846 (aarch64_<sur>qadd<mode>): Likewise.
847 (aarch64_sqmovun<mode>): Likewise.
848 (aarch64_<sur>qmovn<mode>): Likewise.
849 (aarch64_s<optab><mode>): Likewise.
850 (aarch64_sq<r>dmulh<mode>): Likewise.
851 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
852 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
853 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
854 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
855 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
856 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
857 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
858 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
859 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
860 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
861 (aarch64_sqdmull<mode>): Likewise.
862 (aarch64_sqdmull_lane<mode>_internal): Likewise.
863 (aarch64_sqdmull_n<mode>): Likewise.
864 (aarch64_sqdmull2<mode>_internal): Likewise.
865 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
866 (aarch64_sqdmull2_n<mode>_internal): Likewise.
867 (aarch64_<sur>shl<mode>): Likewise.
868 (aarch64_<sur>q<r>shl<mode>
869 (aarch64_<sur>shll_n<mode>): Likewise.
870 (aarch64_<sur>shll2_n<mode>): Likewise.
871 (aarch64_<sur>shr_n<mode>): Likewise.
872 (aarch64_<sur>sra_n<mode>): Likewise.
873 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
874 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
875 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
876 (aarch64_cm<optab><mode>): Likewise.
877 (aarch64_cm<optab>di): Likewise.
878 (aarch64_cm<optab><mode>): Likewise.
879 (aarch64_cm<optab>di): Likewise.
880 (aarch64_cmtst<mode>): Likewise.
881 (aarch64_cmtstdi): Likewise.
882 (aarch64_cm<optab><mode>): Likewise.
883 (*aarch64_fac<optab><mode>): Likewise.
884 (aarch64_addp<mode>): Likewise.
885 (aarch64_addpdi): Likewise.
886 (sqrt<mode>2): Likewise.
887 (vec_load_lanesoi<mode>): Likewise.
888 (vec_store_lanesoi<mode>): Likewise.
889 (vec_load_lanesci<mode>): Likewise.
890 (vec_store_lanesci<mode>): Likewise.
891 (vec_load_lanesxi<mode>): Likewise.
892 (vec_store_lanesxi<mode>): Likewise.
893 (*aarch64_mov<mode>): Likewise.
894 (aarch64_ld2<mode>_dreg): Likewise.
895 (aarch64_ld2<mode>_dreg): Likewise.
896 (aarch64_ld3<mode>_dreg): Likewise.
897 (aarch64_ld3<mode>_dreg): Likewise.
898 (aarch64_ld4<mode>_dreg): Likewise.
899 (aarch64_ld4<mode>_dreg): Likewise.
900 (aarch64_tbl1<mode>): Likewise.
901 (aarch64_tbl2v16qi): Likewise.
902 (aarch64_combinev16qi): Likewise.
903 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
904 (aarch64_st2<mode>_dreg): Likewise.
905 (aarch64_st2<mode>_dreg): Likewise.
906 (aarch64_st3<mode>_dreg): Likewise.
907 (aarch64_st3<mode>_dreg): Likewise.
908 (aarch64_st4<mode>_dreg): Likewise.
909 (aarch64_st4<mode>_dreg): Likewise.
910 (*aarch64_simd_ld1r<mode>): Likewise.
911 (aarch64_frecpe<mode>): Likewise.
912 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
913 (aarch64_frecps<mode>): Likewise.
915 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
917 * config/arm/iterators.md (V_elem_ch): New.
919 (VQH_type): Likewise.
920 * config/arm/arm.md (is_neon_type): New.
921 (conds): Use is_neon_type.
922 (anddi3_insn): Update type attribute.
923 (xordi3_insn): Likewise.
924 (one_cmpldi2): Likewise.
925 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
926 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
927 (*movmisalign<mode>_neon_store): Likewise.
928 (*movmisalign<mode>_neon_load): Likewise.
929 (vec_set<mode>_internal): Likewise.
930 (vec_set<mode>_internal): Likewise.
931 (vec_setv2di_internal): Likewise.
932 (vec_extract<mode>): Likewise.
933 (vec_extract<mode>): Likewise.
934 (vec_extractv2di): Likewise.
935 (*add<mode>3_neon): Likewise.
936 (adddi3_neon): Likewise.
937 (*sub<mode>3_neon): Likewise.
938 (subdi3_neon): Likewise.
939 (fma<VCVTF:mode>4): Likewise.
940 (fma<VCVTF:mode>4_intrinsic): Likewise.
941 (*fmsub<VCVTF:mode>4): Likewise.
942 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
943 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
944 (ior<mode>3): Likewise.
945 (and<mode>3): Likewise.
946 (orn<mode>3_neon): Likewise.
947 (orndi3_neon): Likewise.
948 (bic<mode>3_neon): Likewise.
949 (bicdi3_neon): Likewise.
950 (xor<mode>3): Likewise.
951 (one_cmpl<mode>2): Likewise.
952 (abs<mode>2): Likewise.
953 (neg<mode>2): Likewise.
954 (negdi2_neon): Likewise.
955 (*umin<mode>3_neon): Likewise.
956 (*umax<mode>3_neon): Likewise.
957 (*smin<mode>3_neon): Likewise.
958 (*smax<mode>3_neon): Likewise.
959 (vashl<mode>3): Likewise.
960 (vashr<mode>3_imm): Likewise.
961 (vlshr<mode>3_imm): Likewise.
962 (ashl<mode>3_signed): Likewise.
963 (ashl<mode>3_unsigned): Likewise.
964 (neon_load_count): Likewise.
965 (ashldi3_neon_noclobber): Likewise.
966 (ashldi3_neon): Likewise.
967 (signed_shift_di3_neon): Likewise.
968 (unsigned_shift_di3_neon): Likewise.
969 (ashrdi3_neon_imm_noclobber): Likewise.
970 (lshrdi3_neon_imm_noclobber): Likewise.
971 (<shift>di3_neon): Likewise.
972 (widen_ssum<mode>3): Likewise.
973 (widen_usum<mode>3): Likewise.
974 (quad_halves_<code>v4si): Likewise.
975 (quad_halves_<code>v4sf): Likewise.
976 (quad_halves_<code>v8hi): Likewise.
977 (quad_halves_<code>v16qi): Likewise.
978 (reduc_splus_v2di): Likewise.
979 (neon_vpadd_internal<mode>): Likewise.
980 (neon_vpsmin<mode>): Likewise.
981 (neon_vpsmax<mode>): Likewise.
982 (neon_vpumin<mode>): Likewise.
983 (neon_vpumax<mode>): Likewise.
984 (*ss_add<mode>_neon): Likewise.
985 (*us_add<mode>_neon): Likewise.
986 (*ss_sub<mode>_neon): Likewise.
987 (*us_sub<mode>_neon): Likewise.
988 (neon_vadd<mode>_unspec): Likewise.
989 (neon_vaddl<mode>): Likewise.
990 (neon_vaddw<mode>): Likewise.
991 (neon_vhadd<mode>): Likewise.
992 (neon_vqadd<mode>): Likewise.
993 (neon_vaddhn<mode>): Likewise.
994 (neon_vmul<mode>): Likewise.
995 (neon_vfms<VCVTF:mode>): Likewise.
996 (neon_vmlal<mode>): Likewise.
997 (neon_vmls<mode>): Likewise.
998 (neon_vmlsl<mode>): Likewise.
999 (neon_vqdmulh<mode>): Likewise.
1000 (neon_vqdmlal<mode>): Likewise.
1001 (neon_vqdmlsl<mode>): Likewise.
1002 (neon_vmull<mode>): Likewise.
1003 (neon_vqdmull<mode>): Likewise.
1004 (neon_vsub<mode>_unspec): Likewise.
1005 (neon_vsubl<mode>): Likewise.
1006 (neon_vsubw<mode>): Likewise.
1007 (neon_vqsub<mode>): Likewise.
1008 (neon_vhsub<mode>): Likewise.
1009 (neon_vsubhn<mode>): Likewise.
1010 (neon_vceq<mode>): Likewise.
1011 (neon_vcge<mode>): Likewise.
1012 (neon_vcgeu<mode>): Likewise.
1013 (neon_vcgt<mode>): Likewise.
1014 (neon_vcgtu<mode>): Likewise.
1015 (neon_vcle<mode>): Likewise.
1016 (neon_vclt<mode>): Likewise.
1017 (neon_vcage<mode>): Likewise.
1018 (neon_vcagt<mode>): Likewise.
1019 (neon_vtst<mode>): Likewise.
1020 (neon_vabd<mode>): Likewise.
1021 (neon_vabdl<mode>): Likewise.
1022 (neon_vaba<mode>): Likewise.
1023 (neon_vabal<mode>): Likewise.
1024 (neon_vmax<mode>): Likewise.
1025 (neon_vmin<mode>): Likewise.
1026 (neon_vpaddl<mode>): Likewise.
1027 (neon_vpadal<mode>): Likewise.
1028 (neon_vpmax<mode>): Likewise.
1029 (neon_vpmin<mode>): Likewise.
1030 (neon_vrecps<mode>): Likewise.
1031 (neon_vrsqrts<mode>): Likewise.
1032 (neon_vqabs<mode>): Likewise.
1033 (neon_vqneg<mode>): Likewise.
1034 (neon_vcls<mode>): Likewise.
1035 (clz<mode>2): Likewise.
1036 (popcount<mode>2): Likewise.
1037 (neon_vrecpe<mode>): Likewise.
1038 (neon_vrsqrte<mode>): Likewise.
1039 (neon_vget_lane<mode>_sext_internal): Likewise.
1040 (neon_vget_lane<mode>_zext_internal): Likewise.
1041 (neon_vdup_n<mode>): Likewise.
1042 (neon_vdup_n<mode>): Likewise.
1043 (neon_vdup_nv2di): Likewise.
1044 (neon_vdup_lane<mode>_interal): Likewise.
1045 (*neon_vswp<mode>): Likewise.
1046 (neon_vcombine<mode>): Likewise.
1047 (float<mode><V_cvtto>2): Likewise.
1048 (floatuns<mode><V_cvtto>2): Likewise.
1049 (fix_trunc<mode><V_cvtto>2): Likewise.
1050 (fixuns_trunc<mode><V_cvtto>2
1051 (neon_vcvt<mode>): Likewise.
1052 (neon_vcvt<mode>): Likewise.
1053 (neon_vcvtv4sfv4hf): Likewise.
1054 (neon_vcvtv4hfv4sf): Likewise.
1055 (neon_vcvt_n<mode>): Likewise.
1056 (neon_vcvt_n<mode>): Likewise.
1057 (neon_vmovn<mode>): Likewise.
1058 (neon_vqmovn<mode>): Likewise.
1059 (neon_vqmovun<mode>): Likewise.
1060 (neon_vmovl<mode>): Likewise.
1061 (neon_vmul_lane<mode>): Likewise.
1062 (neon_vmul_lane<mode>): Likewise.
1063 (neon_vmull_lane<mode>): Likewise.
1064 (neon_vqdmull_lane<mode>): Likewise.
1065 (neon_vqdmulh_lane<mode>): Likewise.
1066 (neon_vqdmulh_lane<mode>): Likewise.
1067 (neon_vmla_lane<mode>): Likewise.
1068 (neon_vmla_lane<mode>): Likewise.
1069 (neon_vmlal_lane<mode>): Likewise.
1070 (neon_vqdmlal_lane<mode>): Likewise.
1071 (neon_vmls_lane<mode>): Likewise.
1072 (neon_vmls_lane<mode>): Likewise.
1073 (neon_vmlsl_lane<mode>): Likewise.
1074 (neon_vqdmlsl_lane<mode>): Likewise.
1075 (neon_vext<mode>): Likewise.
1076 (neon_vrev64<mode>): Likewise.
1077 (neon_vrev32<mode>): Likewise.
1078 (neon_vrev16<mode>): Likewise.
1079 (neon_vbsl<mode>_internal): Likewise.
1080 (neon_vshl<mode>): Likewise.
1081 (neon_vqshl<mode>): Likewise.
1082 (neon_vshr_n<mode>): Likewise.
1083 (neon_vshrn_n<mode>): Likewise.
1084 (neon_vqshrn_n<mode>): Likewise.
1085 (neon_vqshrun_n<mode>): Likewise.
1086 (neon_vshl_n<mode>): Likewise.
1087 (neon_vqshl_n<mode>): Likewise.
1088 (neon_vqshlu_n<mode>): Likewise.
1089 (neon_vshll_n<mode>): Likewise.
1090 (neon_vsra_n<mode>): Likewise.
1091 (neon_vsri_n<mode>): Likewise.
1092 (neon_vsli_n<mode>): Likewise.
1093 (neon_vtbl1v8qi): Likewise.
1094 (neon_vtbl2v8qi): Likewise.
1095 (neon_vtbl3v8qi): Likewise.
1096 (neon_vtbl4v8qi): Likewise.
1097 (neon_vtbl1v16qi): Likewise.
1098 (neon_vtbl2v16qi): Likewise.
1099 (neon_vcombinev16qi): Likewise.
1100 (neon_vtbx1v8qi): Likewise.
1101 (neon_vtbx2v8qi): Likewise.
1102 (neon_vtbx3v8qi): Likewise.
1103 (neon_vtbx4v8qi): Likewise.
1104 (*neon_vtrn<mode>_insn): Likewise.
1105 (*neon_vzip<mode>_insn): Likewise.
1106 (*neon_vuzp<mode>_insn): Likewise.
1107 (neon_vld1<mode>): Likewise.
1108 (neon_vld1_lane<mode>): Likewise.
1109 (neon_vld1_lane<mode>): Likewise.
1110 (neon_vld1_dup<mode>): Likewise.
1111 (neon_vld1_dup<mode>): Likewise.
1112 (neon_vld1_dupv2di): Likewise.
1113 (neon_vst1<mode>): Likewise.
1114 (neon_vst1_lane<mode>): Likewise.
1115 (neon_vst1_lane<mode>): Likewise.
1116 (neon_vld2<mode>): Likewise.
1117 (neon_vld2<mode>): Likewise.
1118 (neon_vld2_lane<mode>): Likewise.
1119 (neon_vld2_lane<mode>): Likewise.
1120 (neon_vld2_dup<mode>): Likewise.
1121 (neon_vst2<mode>): Likewise.
1122 (neon_vst2<mode>): Likewise.
1123 (neon_vst2_lane<mode>): Likewise.
1124 (neon_vst2_lane<mode>): Likewise.
1125 (neon_vld3<mode>): Likewise.
1126 (neon_vld3qa<mode>): Likewise.
1127 (neon_vld3qb<mode>): Likewise.
1128 (neon_vld3_lane<mode>): Likewise.
1129 (neon_vld3_lane<mode>): Likewise.
1130 (neon_vld3_dup<mode>): Likewise.
1131 (neon_vst3<mode>): Likewise.
1132 (neon_vst3qa<mode>): Likewise.
1133 (neon_vst3qb<mode>): Likewise.
1134 (neon_vst3_lane<mode>): Likewise.
1135 (neon_vst3_lane<mode>): Likewise.
1136 (neon_vld4<mode>): Likewise.
1137 (neon_vld4qa<mode>): Likewise.
1138 (neon_vld4qb<mode>): Likewise.
1139 (neon_vld4_lane<mode>): Likewise.
1140 (neon_vld4_lane<mode>): Likewise.
1141 (neon_vld4_dup<mode>): Likewise.
1142 (neon_vst4<mode>): Likewise.
1143 (neon_vst4qa<mode>): Likewise.
1144 (neon_vst4qb<mode>): Likewise.
1145 (neon_vst4_lane<mode>): Likewise.
1146 (neon_vst4_lane<mode>): Likewise.
1147 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1148 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1149 (neon_vec_<US>mult_lo_<mode>): Likewise.
1150 (neon_vec_<US>mult_hi_<mode>): Likewise.
1151 (neon_vec_<US>shiftl_<mode>): Likewise.
1152 (neon_unpack<US>_<mode>): Likewise.
1153 (neon_vec_<US>mult_<mode>): Likewise.
1154 (vec_pack_trunc_<mode>): Likewise.
1155 (neon_vec_pack_trunc_<mode>): Likewise.
1156 (neon_vabd<mode>_2): Likewise.
1157 (neon_vabd<mode>_3): Likewise.
1159 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1161 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1162 (load_pair): Update type attribute.
1163 (store_pair): Update type attribute.
1164 * config/aarch64/iterators.md (q): New.
1166 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1168 * config/arm/types.md: Add new types for Neon insns.
1170 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1171 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1172 Sergey Lega <sergey.s.lega@intel.com>
1173 Anna Tikhonova <anna.tikhonova@intel.com>
1174 Ilya Tocar <ilya.tocar@intel.com>
1175 Andrey Turetskiy <andrey.turetskiy@intel.com>
1176 Ilya Verbin <ilya.verbin@intel.com>
1177 Kirill Yukhin <kirill.yukhin@intel.com>
1178 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1180 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1181 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1182 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1184 (srcp14<mode>): Ditto.
1185 (rsqrt14<mode>): Ditto.
1186 (rsqrt14<mode>): Ditto.
1187 (avx512f_vmscalef<mode>): Ditto.
1188 (avx512f_scalef<mode>): Ditto.
1189 (avx512f_getexp<mode>): Ditto.
1190 (avx512f_sgetexp<mode>): Ditto.
1191 (avx512f_fixupimm<mode>): Ditto.
1192 (avx512f_sfixupimm<mode>): Ditto.
1193 (avx512f_rndscale<mode>): Ditto.
1194 (*avx512er_exp2<mode>): Ditto.
1195 (*avx512er_rcp28<mode>): Ditto.
1196 (avx512er_rsqrt28<mode>): Ditto.
1197 (avx512f_getmant<mode>): Ditto.
1198 (avx512f_getmant<mode>): Ditto.
1199 (avx512f_rndscale<mode>): Fix formatting.
1201 2013-10-15 Martin Jambor <mjambor@suse.cz>
1203 * ipa-utils.h (ipa_edge_within_scc): Declare.
1204 * ipa-cp.c (edge_within_scc): Moved...
1205 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1207 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1208 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1209 Sergey Lega <sergey.s.lega@intel.com>
1210 Anna Tikhonova <anna.tikhonova@intel.com>
1211 Ilya Tocar <ilya.tocar@intel.com>
1212 Andrey Turetskiy <andrey.turetskiy@intel.com>
1213 Ilya Verbin <ilya.verbin@intel.com>
1214 Kirill Yukhin <kirill.yukhin@intel.com>
1215 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1217 * config/i386/predicates.md (const_8_to_15_operand): New.
1218 (const_16_to_31_operand): Ditto.
1219 * config/i386/sse.md (V8FI): New.
1221 (reduc_splus_v8df): Ditto.
1222 (reduc_splus_v16sf): Ditto.
1223 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1224 (vec_extract_hi_<mode>): Ditto.
1225 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1226 (vec_set_lo_<mode>): Ditto.
1227 (vec_set_hi_<mode>): Ditto.
1228 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1229 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1230 (avx512f_pshufd_1): Ditto.
1231 (avx512f_broadcast<mode>): Ditto.
1232 (avx512f_broadcast<mode>): Ditto.
1233 (define_split): Split vec_extract_lo into move.
1234 (ssequartermode): Ditto.
1235 (ssedoublemode): Extened with wider modes.
1236 (vec_extract_lo_<mode>): Ditto.
1238 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1239 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1240 Sergey Lega <sergey.s.lega@intel.com>
1241 Anna Tikhonova <anna.tikhonova@intel.com>
1242 Ilya Tocar <ilya.tocar@intel.com>
1243 Andrey Turetskiy <andrey.turetskiy@intel.com>
1244 Ilya Verbin <ilya.verbin@intel.com>
1245 Kirill Yukhin <kirill.yukhin@intel.com>
1246 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1248 * config/i386/predicates.md (register_or_constm1_operand): New.
1249 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1250 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1251 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1252 UNSPEC_SCATTER_PREFETCH
1254 (avx512f_ucmp<mode>3): Ditto.
1255 (avx512f_vternlog<mode>): Ditto.
1256 (avx512f_align<mode>): Ditto.
1257 (<shift_insn><mode>3): Ditto.
1258 (avx512f_<rotate>v<mode>): Ditto.
1259 (avx512f_<rotate><mode>): Ditto.
1260 (avx512f_eq<mode>3): Ditto.
1261 (avx512f_eq<mode>3_1): Ditto.
1262 (avx512f_gt<mode>3): Ditto.
1263 (avx512f_testm<mode>3): Ditto.
1264 (avx512f_testnm<mode>3): Ditto.
1265 (avx512pf_gatherpf<mode>): Ditto.
1266 (*avx512pf_gatherpf<mode>_mask): Ditto.
1267 (*avx512pf_gatherpf<mode>): Ditto.
1268 (avx512pf_scatterpf<mode>): Ditto.
1269 (*avx512pf_scatterpf<mode>_mask): Ditto.
1270 (*avx512pf_scatterpf<mode>): Ditto.
1271 (avx512f_vec_dup_gpr<mode>): Ditto.
1272 (clz<mode>2): Ditto.
1273 (conflict<mode>): Ditto.
1274 (REDUC_SMINMAX_MODE): Extened with wider modes.
1275 (reduc_<code>_<mode>): Ditto.
1276 (vlshr<mode>3): Ditto.
1277 (vashl<mode>3): Ditto.
1279 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1280 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1281 Sergey Lega <sergey.s.lega@intel.com>
1282 Anna Tikhonova <anna.tikhonova@intel.com>
1283 Ilya Tocar <ilya.tocar@intel.com>
1284 Andrey Turetskiy <andrey.turetskiy@intel.com>
1285 Ilya Verbin <ilya.verbin@intel.com>
1286 Kirill Yukhin <kirill.yukhin@intel.com>
1287 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1289 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1292 (avx512fmaskmode): Ditto.
1293 (bcstscalarsuff): Ditto.
1294 (avx512f_blendm<mode>): Ditto.
1295 (cmp_imm_predicate): Ditto.
1296 (avx512f_cmp<mode>3): Ditto.
1297 (avx512f_vec_dup<mode>): Ditto.
1298 (avx512f_vec_dup_mem<mode>): Ditto.
1299 (avx512f_vpermi2var<mode>3): Ditto.
1300 (avx512f_vpermt2var<mode>3): Ditto.
1301 (vec_init<mode>): Ditto.
1302 (avx512f_gathersi<mode>): Ditto.
1303 (*avx512f_gathersi<mode>): Ditto.
1304 (*avx512f_gathersi<mode>_2): Ditto.
1305 (avx512f_gatherdi<mode>): Ditto.
1306 (*avx512f_gatherdi<mode>): Ditto.
1307 (*avx512f_gatherdi<mode>_2): Ditto.
1308 (avx512f_scattersi<mode>): Ditto.
1309 (*avx512f_scattersi<mode>): Ditto.
1310 (avx512f_scatterdi<mode>): Ditto.
1311 (*avx512f_scatterdi<mode>): Ditto.
1312 (sseintprefix): Extened with wider modes.
1313 (VEC_GATHER_IDXSI): Ditto.
1314 (VEC_GATHER_IDXDI): Ditto.
1315 (VEC_GATHER_SRCDI): Ditto.
1317 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1318 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1320 * config/arm/t-aprofile: New file.
1321 * config.gcc: Handle --with-multilib-list option.
1323 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1325 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1328 2013-10-15 Richard Biener <rguenther@suse.de>
1330 * tree-loop-distribution.c (build_empty_rdg): Inline into
1332 (rdg_flag_vertex): Inline into single user.
1333 (rdg_flag_vertex_and_dependent): Likewise.
1334 (build_rdg_partition_for_vertex): Remove processed bitmap.
1335 (rdg_build_partitions): Simplify.
1337 2013-10-15 Richard Biener <rguenther@suse.de>
1339 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1340 Restructure forwarding through conversions and copies to
1341 avoid performing copy-propagation the wrong way. Adjust
1342 recursion invocations.
1343 (forward_propagate_addr_expr): Add argument stating if we
1344 are recursing from a single-use.
1345 (ssa_forward_propagate_and_combine): Adjust.
1347 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1349 * dumpfile.h (gcc::dump_manager): New class, to hold state
1350 relating to dumpfile management.
1351 (get_dump_file_name): Remove in favor of method of dump_manager.
1352 (dump_initialized_p): Likewise.
1353 (dump_start): Likewise.
1354 (dump_finish): Likewise.
1355 (dump_switch_p): Likewise.
1356 (dump_register): Likewise.
1357 (get_dump_file_info): Likewise.
1358 * context.c (gcc::context::context): Construct the dump_manager
1360 * context.h (gcc::context::get_dumps): New.
1361 (gcc::context::m_dumps): New.
1362 * coverage.c (coverage_init): Port to dump_manager API.
1363 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1364 (extra_dump_files_in_use): Likewise.
1365 (extra_dump_files_alloced): Likewise.
1366 (gcc::dump_manager::dump_manager): New.
1367 (dump_register): Convert to...
1368 (gcc::dump_manager::dump_register): ...method, replacing
1369 function-static next_dump with m_next_dump field.
1370 (get_dump_file_info): Convert to...
1371 (gcc::dump_manager::get_dump_file_info): ...method.
1372 (get_dump_file_name): Convert to...
1373 (gcc::dump_manager::get_dump_file_name): ...method.
1374 (dump_start): Convert to...
1375 (gcc::dump_manager::dump_start): ...method.
1376 (dump_finish): Convert to...
1377 (gcc::dump_manager::dump_finish): ...method.
1378 (dump_begin): Replace body with...
1379 (gcc::dump_manager::dump_begin): ...new method.
1380 (dump_phase_enabled_p): Convert to...
1381 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1382 (dump_phase_enabled_p): Convert to...
1383 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1384 (dump_initialized_p): Convert to...
1385 (gcc::dump_manager::dump_initialized_p): ...method.
1386 (dump_flag_name): Replace body with...
1387 (gcc::dump_manager::dump_flag_name): ...new method.
1388 (dump_enable_all): Convert to...
1389 (gcc::dump_manager::dump_enable_all): ...new method.
1390 (opt_info_enable_passes): Convert to...
1391 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1392 (dump_switch_p_1): Convert to...
1393 (gcc::dump_manager::dump_switch_p_1): ...new method.
1394 (dump_switch_p): Convert to...
1395 (gcc::dump_manager::dump_switch_p): ...new method.
1396 (opt_info_switch_p): Port to dump_manager API.
1397 (enable_rtl_dump_file): Likewise.
1398 * opts-global.c (handle_common_deferred_options): Port to new
1400 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1401 (pass_manager::register_one_dump_file): Likewise.
1402 (pass_manager::register_pass): Likewise.
1403 (pass_init_dump_file): Likewise.
1404 (pass_fini_dump_file): Likewise.
1405 * statistics.c (statistics_early_init): Likewise.
1407 2013-10-14 Richard Biener <rguenther@suse.de>
1409 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1410 iterative_hash_canonical_type, gimple_canonical_type_hash,
1411 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1412 gimple_register_canonical_type, print_gimple_types_stats,
1413 free_gimple_type_tables): Move to lto/lto.c
1414 (gt-gimple.h): Do not include.
1415 * gimple.h (gimple_register_canonical_type,
1416 print_gimple_types_stats, free_gimple_type_tables): Remove.
1417 * Makefile.in (GTFILES): Remove gimple.c.
1419 2013-10-14 Travis Snoozy <quandary@remstate.com>
1422 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1423 scanning for msp430x targets.
1425 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1428 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1429 (registers_ok_for_ldd_peep): Move around.
1430 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1431 * config/sparc/sparc.md (widening peepholes): Use it.
1433 2013-10-14 Richard Biener <rguenther@suse.de>
1437 * gimple.c (iterative_hash_canonical_type): Make sure to
1438 record the hash into the correct hashtable slot.
1440 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1442 PR rtl-optimization/58662
1443 * combine.c (try_combine): Take into account death nodes on I2 when
1444 splitting a PARALLEL of two independent SETs. Fix dump message.
1446 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1449 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1450 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1451 * config/sh/t-sh (sh_treg_combine.o): New entry.
1452 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1453 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1454 (register_sh_passes): New function. Register sh_treg_combine pass.
1455 (sh_option_override): Invoke it.
1456 (sh_canonicalize_comparison): Handle op0_preserve_value.
1457 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1458 opportunities. Canonicalize branch condition.
1459 (nott): Allow only if pseudos can be created for non-SH2A.
1461 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1464 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1465 (ix86_expand_movmem): Replace copy_addr_to_reg with
1466 ix86_copy_addr_to_reg.
1467 (ix86_expand_setmem): Likewise.
1469 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1471 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1472 provided by ix86_fp_compare_mode instead of CCFPUmode.
1474 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1476 * config/aarch64/arm_neon.h
1477 (vtbx<1,3>_<psu>8): Fix register constriants.
1479 2013-10-11 Jeff Law <law@redhat.com>
1481 PR tree-optimization/58640
1482 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1483 threading paths that cross over two loop entry points.
1485 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1487 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1488 handle vector float as well.
1489 (*vsx_le_perm_load_v4si): Likewise.
1490 (*vsx_le_perm_store_v2di): Likewise.
1491 (*vsx_le_perm_store_v4si): Likewise.
1493 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1495 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1496 directly to circumvent subtract from splat{31} workaround.
1497 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1499 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1500 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1501 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1502 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1503 separate big and little endian logic.
1504 (*altivec_vperm_<mode>_internal): New define_insn.
1505 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1506 separate big and little endian logic.
1507 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1508 (vec_permv16qi): Add little endian logic.
1510 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1512 * doc/extend.texi (returns_nonnull): Remove arguments.
1514 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1515 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1516 Sergey Lega <sergey.s.lega@intel.com>
1517 Anna Tikhonova <anna.tikhonova@intel.com>
1518 Ilya Tocar <ilya.tocar@intel.com>
1519 Andrey Turetskiy <andrey.turetskiy@intel.com>
1520 Ilya Verbin <ilya.verbin@intel.com>
1521 Kirill Yukhin <kirill.yukhin@intel.com>
1522 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1524 * config/i386/sse.md (VI48F_256_512): New.
1525 (avx2_permvar<mode>): Change to ...
1526 (<avx2_avx512f>_permvar<mode>): This.
1528 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1529 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1530 Sergey Lega <sergey.s.lega@intel.com>
1531 Anna Tikhonova <anna.tikhonova@intel.com>
1532 Ilya Tocar <ilya.tocar@intel.com>
1533 Andrey Turetskiy <andrey.turetskiy@intel.com>
1534 Ilya Verbin <ilya.verbin@intel.com>
1535 Kirill Yukhin <kirill.yukhin@intel.com>
1536 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1538 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1539 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1540 * config/i386/sse.md (VI4_AVX): New.
1541 (sf2simodelower): Ditto.
1542 (sse2_cvtps2dq): Change to ...
1543 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1545 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1546 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1547 Sergey Lega <sergey.s.lega@intel.com>
1548 Anna Tikhonova <anna.tikhonova@intel.com>
1549 Ilya Tocar <ilya.tocar@intel.com>
1550 Andrey Turetskiy <andrey.turetskiy@intel.com>
1551 Ilya Verbin <ilya.verbin@intel.com>
1552 Kirill Yukhin <kirill.yukhin@intel.com>
1553 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1555 * config/i386/sse.md (V_512): New.
1557 (vcond<V_512:mode><VF_512:mode>): Ditto.
1558 (vcond<V_512:mode><VI_512:mode>): Ditto.
1559 (vcondu<V_512:mode><VI_512:mode>): Ditto.
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/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1572 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1574 (fma<mode>4): Removed condition.
1575 (fms<mode>4): Ditto.
1576 (fnma<mode>4): Ditto.
1577 (fnms<mode>4): Ditto.
1578 (fma4i_fmadd_<mode>): Ditto.
1579 (*fma_fmadd_<mode>): Ditto.
1580 (*fma_fmsub_<mode>): Ditto.
1581 (*fma_fnmadd_<mode>): Ditto.
1582 (*fma_fnmsub_<mode>): Ditto.
1583 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1584 (*fma_fmaddsub_<mode>): Ditto.
1585 (*fma_fmsubadd_<mode>): Ditto.
1586 (*fmai_fmadd_<mode>): Ditto.
1587 (*fmai_fmsub_<mode>): Ditto.
1588 (*fmai_fnmadd_<mode>): Ditto.
1589 (*fmai_fnmsub_<mode>): Ditto.
1591 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1592 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1593 Sergey Lega <sergey.s.lega@intel.com>
1594 Anna Tikhonova <anna.tikhonova@intel.com>
1595 Ilya Tocar <ilya.tocar@intel.com>
1596 Andrey Turetskiy <andrey.turetskiy@intel.com>
1597 Ilya Verbin <ilya.verbin@intel.com>
1598 Kirill Yukhin <kirill.yukhin@intel.com>
1599 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1601 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1602 (VI124_256): Changed to ...
1603 (VI124_256_48_512): This.
1604 (ssepackmode): Extended with wider modes.
1605 (<code><mode>3): Changed iterator.
1606 (*avx2_<code><mode>3): Ditto.
1607 (vec_pack_trunc_<mode>): Ditto.
1609 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1610 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1611 Sergey Lega <sergey.s.lega@intel.com>
1612 Anna Tikhonova <anna.tikhonova@intel.com>
1613 Ilya Tocar <ilya.tocar@intel.com>
1614 Andrey Turetskiy <andrey.turetskiy@intel.com>
1615 Ilya Verbin <ilya.verbin@intel.com>
1616 Kirill Yukhin <kirill.yukhin@intel.com>
1617 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1619 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1620 (VI8F_256_512): Ditto.
1621 (abs<mode>2): Changed iterator.
1622 (avx2_perm<mode>): Changed to ...
1623 (<avx2_avx512f>_perm<mode>): This.
1624 (avx2_perm<mode>_1): Changed to ...
1625 (<avx2_avx512f>_perm<mode>_1): This.
1627 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1628 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1629 Sergey Lega <sergey.s.lega@intel.com>
1630 Anna Tikhonova <anna.tikhonova@intel.com>
1631 Ilya Tocar <ilya.tocar@intel.com>
1632 Andrey Turetskiy <andrey.turetskiy@intel.com>
1633 Ilya Verbin <ilya.verbin@intel.com>
1634 Kirill Yukhin <kirill.yukhin@intel.com>
1635 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1637 * config/i386/sse.md (VI48_AVX512F): New.
1638 (VI48_AVX2): Changed to ...
1639 (VI48_AVX2_48_AVX512F): This.
1640 (avx2_ashrv<mode>): Changed to ...
1641 (<avx2_avx512f>_ashrv<mode>): This.
1642 (avx2_<shift_insn>v<mode>): Changed to ...
1643 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1645 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1646 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1647 Sergey Lega <sergey.s.lega@intel.com>
1648 Anna Tikhonova <anna.tikhonova@intel.com>
1649 Ilya Tocar <ilya.tocar@intel.com>
1650 Andrey Turetskiy <andrey.turetskiy@intel.com>
1651 Ilya Verbin <ilya.verbin@intel.com>
1652 Kirill Yukhin <kirill.yukhin@intel.com>
1653 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1655 * config/i386/sse.md (VI4_AVX512F): New.
1656 (VI8_AVX2_AVX512F): Ditto.
1657 (mul<mode>3): Extended with wider modes.
1658 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1659 (mul<mode>3): Ditto.
1660 (vec_widen_<s>mult_odd_<mode>): Ditto.
1662 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1663 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1664 Sergey Lega <sergey.s.lega@intel.com>
1665 Anna Tikhonova <anna.tikhonova@intel.com>
1666 Ilya Tocar <ilya.tocar@intel.com>
1667 Andrey Turetskiy <andrey.turetskiy@intel.com>
1668 Ilya Verbin <ilya.verbin@intel.com>
1669 Kirill Yukhin <kirill.yukhin@intel.com>
1670 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1672 * config/i386/sse.md (VI2_AVX512F): New.
1673 (VI124_AVX512F): Ditto.
1674 (sseunpackmode): Extended with wider modes.
1675 (sseunpackfltmode): Ditto.
1676 (vec_unpacks_float_hi_<mode>): Ditto.
1677 (vec_unpacks_float_lo_<mode>): Ditto.
1678 (vec_unpacku_float_hi_<mode>): Ditto.
1679 (vec_unpacku_float_lo_<mode>): Ditto.
1680 (vec_unpacks_lo_<mode>): Ditto.
1681 (vec_unpacks_hi_<mode>): Ditto.
1682 (vec_unpacku_lo_<mode>): Ditto.
1683 (vec_unpacku_hi_<mode>): Ditto.
1685 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1686 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1687 Sergey Lega <sergey.s.lega@intel.com>
1688 Anna Tikhonova <anna.tikhonova@intel.com>
1689 Ilya Tocar <ilya.tocar@intel.com>
1690 Andrey Turetskiy <andrey.turetskiy@intel.com>
1691 Ilya Verbin <ilya.verbin@intel.com>
1692 Kirill Yukhin <kirill.yukhin@intel.com>
1693 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1695 * config/i386/i386.md (multdiv): New.
1696 (multdiv_mnemonic): Ditto.
1697 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1698 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1699 (<sse>_vmdiv<mode>3): Removed.
1701 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1702 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1703 Sergey Lega <sergey.s.lega@intel.com>
1704 Anna Tikhonova <anna.tikhonova@intel.com>
1705 Ilya Tocar <ilya.tocar@intel.com>
1706 Andrey Turetskiy <andrey.turetskiy@intel.com>
1707 Ilya Verbin <ilya.verbin@intel.com>
1708 Kirill Yukhin <kirill.yukhin@intel.com>
1709 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1711 * config/i386/sse.md (V): Extended with wider modes.
1713 (ssehalfvecmode): Ditto.
1715 (ssepackfltmode): Ditto.
1716 (avx_vec_concat<mode>): Ditto.
1717 (V_256_512): New iterator.
1718 (VF2_512_256): Ditto.
1719 (si2dfmode): New attribute.
1720 (si2dfmodelower): Ditto.
1722 (concat_tg_mode): Ditto.
1723 (floatv4siv4df2): Changed to ...
1724 (float<si2dfmodelower><mode>2): This.
1725 (avx_cvtps2pd256): Changed to ...
1726 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1727 (vec_pack_trunc_v4df): Changed to ...
1728 (vec_pack_trunc_<mode>): This.
1729 (avx_vpermil<mode>): Changed to ...
1730 (<sse2_avx_avx512f>_vpermil<mode>): This.
1731 (<fixsuffix>fix_truncv8dfv8si2): New.
1732 (vec_pack_sfix_trunc_v8df): Ditto.
1733 (avx512f_rndscale<mode>): Ditto.
1734 (avx512f_roundpd512): Ditto.
1735 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1737 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1738 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1739 Sergey Lega <sergey.s.lega@intel.com>
1740 Anna Tikhonova <anna.tikhonova@intel.com>
1741 Ilya Tocar <ilya.tocar@intel.com>
1742 Andrey Turetskiy <andrey.turetskiy@intel.com>
1743 Ilya Verbin <ilya.verbin@intel.com>
1744 Kirill Yukhin <kirill.yukhin@intel.com>
1745 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1747 * config/i386/i386.md (any_fix): New iterator.
1748 (fixsuffix): New attribute.
1749 * config/i386/sse.md (VF1): Extened with wider modes.
1753 (sseintvecmodelower): Ditto.
1754 (ssescalarmode): Ditto.
1755 (ssescalarnum): Ditto.
1757 (ssexmmmode): Ditto.
1758 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1759 (<sse>_rcp<mode>2): Change iterator.
1760 (rsqrt<mode>2): Ditto.
1761 (<sse>_rsqrt<mode>2): Ditto.
1762 (avx2_vec_dup<mode>): Ditto.
1763 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1764 (round<mode>2_sfix): Ditto.
1765 (avx2_pbroadcast<mode>): Ditto.
1766 (*andnot<mode>3): Handle XI mode.
1767 (*<code><mode>3): Ditto.
1768 (AVXTOSSEMODE): Removed.
1769 (avx_vpermil<mode>): Changed to ...
1770 (<sse2_avx_avx512f>_vpermil<mode>): This.
1772 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1773 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1774 Sergey Lega <sergey.s.lega@intel.com>
1775 Anna Tikhonova <anna.tikhonova@intel.com>
1776 Ilya Tocar <ilya.tocar@intel.com>
1777 Andrey Turetskiy <andrey.turetskiy@intel.com>
1778 Ilya Verbin <ilya.verbin@intel.com>
1779 Kirill Yukhin <kirill.yukhin@intel.com>
1780 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1782 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1783 (<sse>_comi): Ditto.
1784 (<sse>_ucomi): Ditto.
1785 (sse_cvtss2siq_2): Ditto.
1786 (sse2_cvtsd2si): Ditto.
1787 (sse2_cvtsd2siq): Ditto.
1788 (sse2_cvttsd2si): Ditto.
1789 (sse2_cvttsd2siq): Ditto.
1790 (<shift_insn><mode>3): Ditto.
1791 (sse2_cvtsi2sdq): Update constraint and prefix.
1792 (sse_cvtsi2ss): Update prefix.
1793 (sse_cvtsi2ssq): Ditto.
1795 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1797 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1798 ignore internal calls.
1800 2013-10-11 Richard Biener <rguenther@suse.de>
1802 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1803 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1804 TREE_OVERFLOW is set.
1806 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1808 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1810 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1811 * gimple.h: Likewise.
1813 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1815 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1818 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1820 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1821 * configure: Regenerate.
1823 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1825 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1826 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1827 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1828 similarly pp_right_paren.
1829 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1830 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1831 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1832 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1834 (omp_declare_simd_clauses_equal,
1835 omp_remove_redundant_declare_simd_attrs): New functions.
1836 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1837 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1838 * tree.h (OMP_LOOP_CHECK): Define.
1839 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1840 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1841 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1842 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1843 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1844 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1845 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1846 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1847 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1848 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1849 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1850 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1851 (omp_remove_redundant_declare_simd_attrs): New prototype.
1852 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1853 GIMPLE_OMP_TEAMS): New codes.
1854 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1855 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1857 (target_nesting_level): New variable.
1858 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1859 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1860 for collapse > 1 static schedule unless ordered.
1861 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1862 (determine_parallel_type): Adjust get_ws_args_for caller.
1863 (install_var_field): Handle mask & 4 for double indirection.
1864 (scan_sharing_clauses): Ignore shared clause on teams construct.
1865 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1866 (create_omp_child_function): If inside target or declare target
1867 constructs, set "omp declare target" attribute on the child function.
1868 (find_combined_for): New function.
1869 (scan_omp_parallel): Handle combined loops.
1870 (scan_omp_target, scan_omp_teams): New functions.
1871 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1872 restrictions and set ctx->cancellable for cancellable constructs.
1873 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1874 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1875 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1876 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1877 (omp_clause_aligned_alignment): New function.
1878 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
1879 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1880 on teams constructs. Handle user defined reductions and new
1882 (lower_reduction_clauses): Don't set placeholder to address of ref
1883 if it has already the right type.
1884 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1885 (expand_parallel_call): Use the new non-_start suffixed builtins,
1886 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1887 and GOMP_parallel_end after the call.
1888 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1889 (expand_omp_for_init_counts): Handle combined loops.
1890 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1892 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1893 end of cancellable loops.
1894 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1895 Likewise. Handle collapse > 1 loops.
1896 (expand_omp_simd): Handle combined loops.
1897 (expand_omp_for): Add inner_stmt argument, adjust callers of
1898 expand_omp_for* functions, use expand_omp_for_static*chunk even
1899 for collapse > 1 unless ordered.
1900 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1901 of cancellable sections.
1902 (expand_omp_single): Remove need_barrier variable, just rely on
1903 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1904 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1905 (expand_omp_atomic_load, expand_omp_atomic_store,
1906 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1907 (expand_omp_target): New function.
1908 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1909 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1910 (build_omp_regions_1): Immediately close region for
1911 GF_OMP_TARGET_KIND_UPDATE.
1912 (maybe_add_implicit_barrier_cancel): New function.
1913 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1915 (lower_omp_single): Likewise. Add clobber after the barrier.
1916 (lower_omp_taskgroup): New function.
1917 (lower_omp_for): Handle combined loops. Adjust
1918 lower_rec_input_clauses caller. Handle cancellation.
1919 (lower_depend_clauses): New function.
1920 (lower_omp_taskreg): Lower depend clauses. Adjust
1921 lower_rec_input_clauses caller. Add clobber after the call. Handle
1923 (lower_omp_target, lower_omp_teams): New functions.
1924 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1925 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1926 and GOMP_cancellation_point calls.
1927 (lower_omp): Fold stmts inside of target region.
1928 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1929 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1930 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1931 (BT_FN_VOID_OMPFN_PTR_UINT,
1932 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1933 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1934 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1935 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1936 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1937 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1938 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1939 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1940 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1941 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1942 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1943 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1944 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1945 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1946 BUILT_IN_GOMP_PARALLEL_END.
1947 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1948 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1949 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1950 GF_OMP_FOR_KIND_DISTRIBUTE.
1951 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1952 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1953 (dump_gimple_omp_return): Print lhs if it has any.
1954 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1955 gimple_omp_atomic_seq_cst_p.
1956 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1957 and GIMPLE_OMP_TEAMS.
1958 * langhooks.c (lhd_omp_mappable_type): New function.
1959 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1960 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1962 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1963 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1964 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
1965 (struct gimplify_omp_ctx): Add combined_loop field.
1966 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1967 on stmts inside of target region.
1968 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1969 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1970 ORT_TARGET and ORT_TARGET_DATA.
1971 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1973 (omp_notice_threadprivate_variable): Complain about threadprivate
1974 variables in target region.
1975 (omp_notice_variable): Complain about vars with non-mappable type
1976 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1977 (omp_check_private): Ignore ORT_TARGET* regions.
1978 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1979 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1980 (find_combined_omp_for): New function.
1981 (gimplify_omp_for): Handle gimplification of combined loops.
1982 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1984 (gimplify_omp_target_update): New function.
1985 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1986 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1987 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1988 (gimplify_body): If fndecl has "omp declare target" attribute, add
1989 implicit ORT_TARGET context around it.
1990 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1991 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1992 * tree-nested.c (convert_nonlocal_reference_stmt,
1993 convert_local_reference_stmt, convert_gimple_call): Handle
1994 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1995 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1996 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1997 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1998 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1999 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2000 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2001 BUILT_IN_GOMP_LOOP_END_CANCEL,
2002 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2003 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2004 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2005 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2006 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2007 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2008 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2009 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2010 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2011 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2012 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2013 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2014 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2015 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2016 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2017 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2018 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2019 gimple_build_omp_teams): New functions.
2020 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2021 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2022 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2023 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2024 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2025 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2026 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2027 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2028 GF_OMP_ATOMIC_SEQ_CST): New.
2029 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2030 gimple_build_omp_teams): New prototypes.
2031 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2032 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2033 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2034 GIMPLE_OMP_SINGLE as end of range.
2035 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2036 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2037 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2038 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2039 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2040 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2041 gimple_omp_target_kind, gimple_omp_target_set_kind,
2042 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2043 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2044 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2045 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2046 gimple_omp_teams_set_clauses): New inlines.
2047 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2048 and GIMPLE_OMP_TASKGROUP.
2049 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2050 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2051 enum omp_clause_proc_bind_kind): New.
2052 (union omp_clause_subcode): Add depend_kind, map_kind and
2053 proc_bind_kind fields.
2054 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2055 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2056 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2057 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2058 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2060 2013-10-10 Teresa Johnson <tejohnson@google.com>
2062 * predict.c (tree_estimate_probability): Add new parameter
2063 for estimate_bb_frequencies.
2064 (estimate_bb_frequencies): Add new parameter to force estimation.
2065 (rebuild_frequencies): When max frequency in function is small,
2066 recompute counts from frequencies.
2067 * predict.h (estimate_bb_frequencies): New parameter.
2069 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2071 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2072 optimizations are disabled.
2074 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2076 * coverage.c (coverage_finish): Fix leak of da_file_name.
2078 2013-10-10 Jan Hubicka <jh@suse.cz>
2080 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2081 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2084 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2087 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2088 if any labels are in FALLTHRU_BB, use a special label emitted
2089 immediately after the asm goto insn rather than label_rtx
2091 (expand_asm_stmt): Adjust caller.
2092 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2093 edge if the last insn in predecessor is a jump with single successor,
2094 but it isn't simplejump_p.
2096 2013-10-10 Richard Biener <rguenther@suse.de>
2098 PR tree-optimization/58656
2099 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2101 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2103 * gimplify.c: Include expr.h and tm_p.h for targets with special
2104 va-arg padding requirements.
2106 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2108 * tree-flow.h: Move some prototypes to gimple.h.
2109 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2110 * gimple.h: Relocate some prototypes from tree-flow.h
2111 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2113 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2114 (build_va_arg_indirect_ref): Relocate and make static.
2115 (std_gimplify_va_arg_expr): Relocate here.
2116 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2117 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2119 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2121 * doc/md.texi: Document the mnemonic attribute.
2123 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2126 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2127 (set_attr_alternative x ...) when searching for user defined
2130 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2132 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2134 2013-10-09 Easwaran Raman <eraman@google.com>
2136 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2137 * cfgexpand.c (defer_stack_allocation): ...use here
2138 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2140 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2142 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2143 (operand_equal_for_value_replacement): New function, extracted from
2144 value_replacement and enhanced to catch more cases.
2145 (value_replacement): Use operand_equal_for_value_replacement.
2147 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2149 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2150 get_max_loop_iterations.
2152 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2154 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2155 Place comment above function.
2157 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2159 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2160 are not related to tree-cfg.c.
2161 * tree-ssa-address.h: New file. Relocate prototypes.
2162 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2163 (addr_for_mem_ref): New. Combine call to get_address_description and
2164 return addr_for_mem_ref.
2165 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2166 * tree-ssa-live.h: Adjust prototypes.
2167 * passes.c: Include tree-ssa-live.h.
2168 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2169 * graphite.c (graphite_transform_loops): Make static.
2170 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2171 make_pass_graphite, pass_data_graphite_transforms,
2172 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2173 * ipa-pure-const.c (warn_function_noreturn): Make static.
2174 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2175 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2176 Relocate from tree-cfg.c
2177 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2179 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2180 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2181 Move to ipa-pure-const.c.
2182 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2183 Relocate from tree-optimize.c.
2184 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2185 make_pass_fixup_cfg): Move to tree-cfg.c.
2186 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2187 Relocate some prototypes.
2188 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2189 * tree-dump.c (dump_function_to_file): Remove prototype.
2190 Add tree-flow.h to the include file.
2191 * tree-dump.h: Remove prototype.
2192 * tree-parloops.h: New File. Add prototypes.
2193 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2194 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2195 from tree-ssa-loop.c.
2196 * tree-predcom.c (run_tree_predictive_commoning,
2197 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2198 Relocate here from tree-ssa-loop.c.
2199 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2200 ssa_name_values.release ().
2201 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2202 (ssa_name_values): Relocate from tree-flow.h.
2203 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2204 * tree-ssa-loop.c (run_tree_predictive_commoning,
2205 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2206 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2207 make_pass_graphite, pass_data_graphite_transforms,
2208 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2209 tree_parallelize_loops, pass_data_parallelize_loops,
2210 make_pass_parallelize_loops): Move to other files.
2211 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2213 * tree.h: Remove prototypes from tree-address.c.
2215 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2217 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2218 (struct int_tree_map): Move to tree-hasher.h
2219 (SCALE, LABEL, PERCENT): Move to gimple.h
2220 * tree-flow-inline.h: Delete. Move functions to other files.
2221 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2222 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2223 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2224 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2225 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2226 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2227 * tree-hasher.h: Don't include tree-flow.h.
2228 (struct int_tree_map): Relocate from tree-flow.h.
2229 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2230 tree-flow-inline.h and make static.
2231 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2233 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2234 tree-flow-inline.h and make static.
2235 * tree.h (is_global_var, may_be_aliased): Relocate from
2237 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2238 * value-prof.c: No longer include tree-flow-inline.h.
2239 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2241 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2243 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2244 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2245 (enum move_pos): Move to tree-ssa-loop-im.h
2246 * cfgloop.h: Move some prototypes.
2247 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2248 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2249 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2250 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2251 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2252 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2253 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2254 (enum move_pos): Relocate here.
2255 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2257 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2258 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2259 make_pass_lim): Relocate here from tree-ssa-loop.c.
2260 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2262 (loop_edge_to_cancel, unloop_loops): Make static.
2263 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2264 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2265 (tree_complete_unroll, gate_tree_complete_unroll,
2266 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2267 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2268 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2269 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2270 (stmt_invariant_in_loop_p): Remove unused function.
2271 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2272 * tree-ssa-loop-manip.h: New file. Add prototypes.
2273 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2274 (gcov_type_to_double_int): Move to cfgloop.h.
2275 (double_int_cmp, bound_index,
2276 estimate_numbers_of_iterations_loop): Make static.
2277 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2278 (max_loop_iterations): Factor out get_max_loop_iterations.
2279 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2281 * tree-ssa-loop-niter.h: New file. Add prototypes.
2282 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2283 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2284 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2285 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2286 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2287 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2288 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2289 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2290 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2291 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2292 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2293 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2294 pass_data_complete_unroll, make_pass_complete_unroll,
2295 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2296 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2297 tree-ssa-loop-ivcanon.c.
2298 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2299 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2300 tree-ssa-loop-prefetch.c.
2301 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2303 (get_lsm_tmp_name): Relocate and add suffix parameter.
2304 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2305 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2306 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2307 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2308 (get_estimated_loop_iterations): Factor out accessor from
2309 estimated_loop_iterations in tree-ssa-loop-niter.c.
2310 (get_max_loop_iterations): Factor out accessor from
2311 _max_loop_iterations in tree-ssa-niter.c.
2312 * loop-unroll.c (decide_unroll_constant_iterations,
2313 decide_unroll_runtime_iterations, decide_peel_simple,
2314 decide_unroll_stupid): Use new get_* accessors.
2316 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2318 PR tree-optimization/20318
2319 * doc/extend.texi (returns_nonnull): New function attribute.
2320 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2322 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2323 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2325 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2328 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2329 false if both components are bitfields.
2331 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2333 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2334 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2335 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2336 * config/aarch64/aarch64.h
2337 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2338 * config/aarch64/aarch64-simd-builtins.def
2339 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2341 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2343 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2345 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2347 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2349 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2351 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2352 (vneg_f64): New intrinsic.
2353 (vneg_s8): Asm replaced with C.
2354 (vneg_s16): Likewise.
2355 (vneg_s32): Likewise.
2356 (vneg_s64): New intrinsic.
2357 (vnegq_f32): Asm replaced with C.
2358 (vnegq_f64): Likewise.
2359 (vnegq_s8): Likewise.
2360 (vnegq_s16): Likewise.
2361 (vnegq_s32): Likewise.
2362 (vnegq_s64): Likewise.
2364 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2366 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2368 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2370 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2371 packed stack special handling.
2372 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2373 back to fixed stack slots for FPRs saved due to stdarg.
2375 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2377 * config/s390/s390.c (s390_frame_info): Restructure function.
2379 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2381 * config/s390/s390.c (struct s390_frame_layout): New field
2383 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2384 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2385 regs_ever_clobbered to char*.
2386 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2387 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2388 (s390_register_info): Enable FPR save slots. Move/Copy some
2389 functionality into ...
2390 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2391 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2393 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2394 not imply a stack frame.
2395 (s390_init_frame_layout): Remove variable clobbered_regs.
2396 (s390_update_register_info): Remove function.
2397 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2398 cannot be used for register renaming.
2399 (s390_hard_regno_scratch_ok): New function.
2400 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2401 (s390_initial_elimination_offset): Change offset calculation of
2402 the return address pointer.
2403 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2404 (restore_gprs): Set frame related flag.
2405 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2406 (s390_emit_prologue): Call s390_register_info instead of
2407 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2408 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2409 (s390_optimize_prologue): Call s390_optimize_register_info.
2410 Try to remove also FPR slot save/restore INSNs. Remove frame
2411 related flags from restore INSNs.
2413 2013-10-08 DJ Delorie <dj@redhat.com>
2415 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2418 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2419 avoid conflict with the MI use of %c.
2420 * config/rl78/rl78-real.md: change %c to %C throughout.
2421 * config/rl78/rl78-virt.md: Likewise.
2423 2013-10-08 Jan Hubicka <jh@suse.cz>
2425 * config/i386/i386.c (ix86_option_override_internal): Switch
2426 to SSE math for -ffast-math when target ISA supports SSE2.
2428 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2430 * tree-flow.h: Remove some prototypes.
2431 * tree.h: Remove some protypes, add a couple.
2432 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2433 using_eh_for_cleanups_p): Add interface routines for front ends.
2434 * tree-eh.h: New file. Add protoptyes.
2435 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2436 (add_stmt_to_eh_lp_fn): Make static.
2437 (lower_try_finally): Use new using_eh_for_cleanups_p.
2438 * emit-rtl.c: Include tree-eh.h.
2439 * gimple.h: Include tree-eh.h.
2441 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2443 PR tree-optimization/58480
2444 * tree-vrp.c (infer_nonnull_range): New function.
2445 (infer_value_range): Call infer_nonnull_range.
2447 2013-10-08 Dehao Chen <dehao@google.com>
2449 PR tree-optimization/58619
2450 * tree-inline.c (copy_phis_for_bb): Combine location data
2453 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2456 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2457 RTX_FRAME_RELATED_P on INSN.
2459 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2461 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2462 (altivec_expand_vec_perm_const): Call it.
2464 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2466 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2467 sequences for LE VSX loads and stores at expand time.
2468 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2470 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2471 (rs6000_gen_le_vsx_permute): New.
2472 (rs6000_gen_le_vsx_load): New.
2473 (rs6000_gen_le_vsx_store): New.
2474 (rs6000_gen_le_vsx_move): New.
2475 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2476 (*vsx_le_perm_load_v4si): New.
2477 (*vsx_le_perm_load_v8hi): New.
2478 (*vsx_le_perm_load_v16qi): New.
2479 (*vsx_le_perm_store_v2di): New.
2480 (*vsx_le_perm_store_v4si): New.
2481 (*vsx_le_perm_store_v8hi): New.
2482 (*vsx_le_perm_store_v16qi): New.
2483 (*vsx_xxpermdi2_le_<mode>): New.
2484 (*vsx_xxpermdi4_le_<mode>): New.
2485 (*vsx_xxpermdi8_le_V8HI): New.
2486 (*vsx_xxpermdi16_le_V16QI): New.
2487 (*vsx_lxvd2x2_le_<mode>): New.
2488 (*vsx_lxvd2x4_le_<mode>): New.
2489 (*vsx_lxvd2x8_le_V8HI): New.
2490 (*vsx_lxvd2x16_le_V16QI): New.
2491 (*vsx_stxvd2x2_le_<mode>): New.
2492 (*vsx_stxvd2x4_le_<mode>): New.
2493 (*vsx_stxvd2x8_le_V8HI): New.
2494 (*vsx_stxvd2x16_le_V16QI): New.
2496 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2498 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2500 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2502 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2503 loop to work also for 31bit ABI.
2504 Save the stack pointer for frame_size > 0.
2506 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2508 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2509 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2510 constraint letters from expanders.
2511 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2512 retry count to general_operand.
2513 ("tabort"): Give operand 0 a mode.
2514 ("tabort_1"): Add mode and constraint letter for operand 0.
2515 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2517 2013-10-04 Jeff Law <law@redhat.com>
2519 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2521 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2523 (thread_across_edge): Here. Call it.
2525 2013-10-04 Cary Coutant <ccoutant@google.com>
2527 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2528 discarding a location list expression (or a piece of one).
2530 2013-10-03 Jan Hubicka <jh@suse.cz>
2532 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2533 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2534 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2536 2013-10-03 Jan Hubicka <jh@suse.cz>
2538 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2539 accumulate-outgoing-args when producing unwind info.
2541 2013-10-03 Wei Mi <wmi@google.com>
2543 * lra-constraints.c (insert_move_for_subreg): New function
2544 extracted from simplify_operand_subreg.
2545 (simplify_operand_subreg): Add reload for paradoxical subreg.
2547 2013-10-03 Rong Xu <xur@google.com>
2549 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2550 the candidate of builtin_expect such that we should fix the
2551 size/time estimation.
2552 (estimate_function_body_sizes): Do the acutally size/time fix-up
2555 2013-10-03 Rong Xu <xur@google.com>
2557 * predict.c (tree_predict_by_opcode): Get the probability
2558 for builtin_expect from param builtin_expect_probability.
2559 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2560 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2561 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2563 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2566 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2567 * calls.c (alloca_call_p): Use get_callee_fndecl.
2568 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2569 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2571 (vrp_visit_stmt): Remove duplicated code.
2573 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2575 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2576 ceildf2, btruncdf2, instead of vsx_* name.
2578 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2579 iterators to only do V2DF and V4SF here. Move the DF code to
2580 rs6000.md where it is combined with SF mode. Replace <VSv> with
2581 just 'v' since only vector operations are handled with these insns
2582 after moving the DF support to rs6000.md.
2583 (vsx_sub<mode>3): Likewise.
2584 (vsx_mul<mode>3): Likewise.
2585 (vsx_div<mode>3): Likewise.
2586 (vsx_fre<mode>2): Likewise.
2587 (vsx_neg<mode>2): Likewise.
2588 (vsx_abs<mode>2): Likewise.
2589 (vsx_nabs<mode>2): Likewise.
2590 (vsx_smax<mode>3): Likewise.
2591 (vsx_smin<mode>3): Likewise.
2592 (vsx_sqrt<mode>2): Likewise.
2593 (vsx_rsqrte<mode>2): Likewise.
2594 (vsx_fms<mode>4): Likewise.
2595 (vsx_nfma<mode>4): Likewise.
2596 (vsx_copysign<mode>3): Likewise.
2597 (vsx_btrunc<mode>2): Likewise.
2598 (vsx_floor<mode>2): Likewise.
2599 (vsx_ceil<mode>2): Likewise.
2600 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2601 (vsx_sminsf3): Likewise.
2602 (vsx_fmadf4): Likewise.
2603 (vsx_fmsdf4): Likewise.
2604 (vsx_nfmadf4): Likewise.
2605 (vsx_nfmsdf4): Likewise.
2606 (vsx_cmpdf_internal1): Likewise.
2608 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2609 simpler to select whether a target has SPE or traditional floating
2610 point support in iterators.
2611 (TARGET_DF_SPE): Likewise.
2612 (TARGET_SF_FPR): Likewise.
2613 (TARGET_DF_FPR): Likewise.
2614 (TARGET_SF_INSN): Macros to say whether floating point support
2615 exists for a given operation for expanders.
2616 (TARGET_DF_INSN): Likewise.
2618 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2619 combining of SF/DF mode operations, using both traditional and VSX
2627 (abs<mode>2): Combine SF/DF modes using traditional floating point
2628 instructions. Add support for using the upper DF registers with
2629 VSX support, and SF registers with power8-vector support. Update
2630 expanders for operations supported by both the SPE and traditional
2631 floating point units.
2632 (abs<mode>2_fpr): Likewise.
2633 (nabs<mode>2): Likewise.
2634 (nabs<mode>2_fpr): Likewise.
2635 (neg<mode>2): Likewise.
2636 (neg<mode>2_fpr): Likewise.
2637 (add<mode>3): Likewise.
2638 (add<mode>3_fpr): Likewise.
2639 (sub<mode>3): Likewise.
2640 (sub<mode>3_fpr): Likewise.
2641 (mul<mode>3): Likewise.
2642 (mul<mode>3_fpr): Likewise.
2643 (div<mode>3): Likewise.
2644 (div<mode>3_fpr): Likewise.
2645 (sqrt<mode>3): Likewise.
2646 (sqrt<mode>3_fpr): Likewise.
2647 (fre<Fs>): Likewise.
2648 (rsqrt<mode>2): Likewise.
2649 (cmp<mode>_fpr): Likewise.
2650 (smax<mode>3): Likewise.
2651 (smin<mode>3): Likewise.
2652 (smax<mode>3_vsx): Likewise.
2653 (smin<mode>3_vsx): Likewise.
2654 (negsf2): Delete SF operations that are merged with DF.
2661 (fmasf4_fpr): Likewise.
2662 (fmssf4_fpr): Likewise.
2663 (nfmasf4_fpr): Likewise.
2664 (nfmssf4_fpr): Likewise.
2665 (sqrtsf2): Likewise.
2666 (rsqrtsf_internal1): Likewise.
2667 (smaxsf3): Likewise.
2668 (sminsf3): Likewise.
2669 (cmpsf_internal1): Likewise.
2670 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2671 (negdf2): Delete DF operations that are merged with SF.
2673 (nabsdf2): Likewise.
2679 (rsqrtdf_internal1): Likewise.
2680 (fmadf4_fpr): Likewise.
2681 (fmsdf4_fpr): Likewise.
2682 (nfmadf4_fpr): Likewise.
2683 (nfmsdf4_fpr): Likewise.
2684 (sqrtdf2): Likewise.
2685 (smaxdf3): Likewise.
2686 (smindf3): Likewise.
2687 (cmpdf_internal1): Likewise.
2688 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2689 (btrunc<mode>2): Delete separate expander, and combine with the
2690 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2691 (btrunc<mode>2_fpr): Likewise.
2692 (ceil<mode>2): Likewise.
2693 (ceil<mode>2_fpr): Likewise.
2694 (floor<mode>2): Likewise.
2695 (floor<mode>2_fpr): Likewise.
2696 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2697 Add support for using the upper registers with VSX and
2698 power8-vector. Move insns to be closer to the define_expands. On
2699 VSX systems, prefer the traditional form of FMA over the VSX
2700 version, since the traditional form allows the target not to
2701 overlap with the inputs.
2702 (fms<mode>4_fpr): Likewise.
2703 (nfma<mode>4_fpr): Likewise.
2704 (nfms<mode>4_fpr): Likewise.
2706 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2707 Richard Earnshaw <richard.earnshaw@arm.com>
2709 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2710 (struct mult_cost_table): Likewise.
2711 (struct mem_cost_table): Likewise.
2712 (struct fp_cost_table): Likewise.
2713 (struct vector_cost_table): Likewise.
2714 (cpu_cost_table): Likewise.
2715 * config/arm/arm.opt (mold-rts-costs): New option.
2716 (mnew-generic-costs): Likewise.
2717 * config/arm/arm.c (generic_extra_costs): New table.
2718 (cortexa15_extra_costs): Likewise.
2719 (arm_slowmul_tune): Use NULL as new costs.
2720 (arm_fastmul_tune): Likewise.
2721 (arm_strongarm_tune): Likewise.
2722 (arm_xscale_tune): Likewise.
2723 (arm_9e_tune): Likewise.
2724 (arm_v6t2_tune): Likewise.
2725 (arm_cortex_a5_tune): Likewise.
2726 (arm_cortex_a9_tune): Likewise.
2727 (arm_v6m_tune): Likewise.
2728 (arm_fa726te_tune): Likewise.
2729 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2730 (arm_cortex_tune): Use generict_extra_costs.
2731 (shifter_op_p): New function.
2732 (arm_unspec_cost): Likewise.
2733 (LIBCALL_COST): Define.
2734 (arm_new_rtx_costs): New function.
2735 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2736 table is available. Use old costs otherwise unless mnew-generic-costs
2738 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2739 (cpu_cost_table): Declare.
2741 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2744 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2745 (*subs_mul_imm_<mode>)
2746 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2747 (*sub_<shift>_<mode>)
2748 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2749 Remove k constraint.
2751 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2753 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2755 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2757 2013-10-02 Teresa Johnson <tejohnson@google.com>
2759 * predict.c (probably_never_executed): New function.
2760 (probably_never_executed_bb_p): Invoke probably_never_executed.
2761 (probably_never_executed_edge_p): Ditto.
2762 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2763 Treat profile insanities conservatively.
2765 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2767 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2769 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2771 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2772 use it. Use smaller increase for scratch. Don't increase reject
2773 for early clobber scratch.
2774 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2775 setting eliminated regs except setting fp from hfp.
2776 (lra_eliminate): Check lra_insn_recog_data on NULL.
2778 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2781 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2782 setting -mvsx-timode by default until the underlying problem is fixed.
2783 (RS6000_CPU, power7 defaults): Likewise.
2785 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2787 * config/x-linux (host-linux.o): Remove header dependencies.
2788 Use $(COMPILE) and $(POSTCOMPILE).
2789 * config/t-linux-android (linux-android.o): Ditto.
2791 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2793 * Makefile.in (expmed.o-warn): Remove.
2795 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2797 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2798 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2800 2013-10-02 Teresa Johnson <tejohnson@google.com>
2802 * dojump.c (do_jump_1): Divide probability between
2803 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2805 2013-10-02 Tom Tromey <tromey@redhat.com>
2807 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2809 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2811 * tree-flow.h: Remove some prototypes.
2812 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2813 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2814 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2815 mark_virtual_phi_result_for_renaming): Relocate here.
2816 * tree-into-ssa.h: Add prototypes.
2817 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2818 single_pred_before_succ_order.
2819 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2820 (nonfreeing_call_p) Move to gimple.c.
2821 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2823 * basic-block.h (single_pred_before_succ_order): Add prototype.
2824 * gimple.c (nonfreeing_call_p): Relocate here.
2825 * gimple.h: Add prototype.
2826 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2827 * tree-ssa-dom.h: New file. Relocate prototypes here.
2828 * tree-ssa.h: Include tree-ssa-dom.h.
2830 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2832 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2833 Use $(COMPILE) and $(POSTCOMPILE).
2835 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2837 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2839 * tree-flow.h: Remove some prototypes.
2840 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2841 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2842 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2843 propagate_tree_value*): Move from here to...
2844 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2845 propagate_tree_value*): Relocate here.
2846 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2847 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2848 * gimple-fold.c: Remove gimple-fold.h from include list.
2849 * tree-vrp.c: Remove gimple-fold.h from include list.
2850 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2851 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2852 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2853 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2854 * sese.c: Add tree-ssa-propagate.h to include list.
2856 2013-10-02 Richard Biener <rguenther@suse.de>
2858 * tree-loop-distribution.c: Include tree-vectorizer.h for
2860 (enum partition_kind): Remove PKIND_REDUCTION.
2861 (struct partition_s): Remove has_writes member, add reduction_p member.
2862 (partition_alloc): Adjust.
2863 (partition_builtin_p): Likewise.
2864 (partition_has_writes): Remove.
2865 (partition_reduction_p): New function.
2866 (partition_merge_into): Likewise.
2867 (generate_code_for_partition): Commonize builtin partition
2869 (rdg_cannot_recompute_vertex_p): Remove.
2870 (already_processed_vertex_p): Likewise.
2871 (rdg_flag_vertex): Do not set has_writes.
2872 (classify_partition): Adjust.
2873 (rdg_build_partitions): Do not set has_writes, treat all
2874 partitions as useful.
2875 (distribute_loop): Record number of library calls generated. Adjust.
2876 (tree_loop_distribution): Report number of loops and library
2877 calls generated as opt-info.
2879 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2881 * tree-flow.h: Include new .h files. Move prototypes.
2882 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2883 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2884 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2885 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2886 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2887 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2888 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2889 * tree.h (get_ref_base_and_extent): Move prototype out.
2890 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2892 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2893 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2894 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2895 * tree-scalar-evolution.c: Include tree.h.
2896 * sese.c: Include tree.h.
2897 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2898 * dwarf2out.c: Include tree-dfa.h.
2899 * tree-chrec.c: Include tree.h.
2900 * tree-data-ref.c: Include tree.h.
2902 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2904 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2907 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2909 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2910 Use $(COMPILE) and $(POSTCOMPILE).
2911 (sol2-cxx.o): Likewise.
2912 (sol2-stubs.o): Likewise.
2914 * config/x-solaris (host-solaris.o): Likewise.
2916 * config/sparc/t-sparc (sparc.o): Remove.
2917 (sparc-c.o): Remove header dependencies.
2918 Use $(COMPILE) and $(POSTCOMPILE).
2919 * config/sparc/x-sparc: Likewise.
2921 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2923 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2924 * config/arc/arc700.md: Likewise.
2925 * config/arc/arc-modes.def: Likewise.
2926 * config/arc/arc-simd.h: Likewise.
2927 * config/arc/t-arc-uClibc: Likewise.
2928 * config/arc/t-arc-newlib: Likewise.
2930 2013-10-02 Renlin Li <renlin.li@arm.com>
2932 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2934 (aarch64_expand_epilogue): Likewise.
2936 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2937 Yufeng Zhang <yufeng.zhang@arm.com>
2939 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2941 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2942 'base_in' represent a conversion and legal_cast_p_1 holds; set
2943 'base_in' with the returned value from get_unwidened.
2945 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2947 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2948 plus_constant is not used.
2950 2013-10-01 Wei Mi <wmi@google.com>
2952 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2953 * config/i386/i386.md: Add define_peephole2 to
2954 break partial reg stall for cvtss2sd/cvtsd2ss.
2956 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2958 * config/arc/arc.c (pass_arc_ifcvt::clone):
2959 Update for ctxt_ -> m_ctxt change.
2961 2013-10-01 Jeff Law <law@redhat.com>
2963 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2964 outgoing_edge and intermediate_edge fields. Instead store the path.
2965 (redirection_data::hash): Hash on the last edge's destination index.
2966 (redirection_data::equal): Check the entire thread path.
2967 (lookup_redirectio_data): Corresponding changes.
2968 (create_edge_and_update_destination_phis): Likewise.
2969 (thread_single_edge): Likewise.
2971 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2972 Diego Novillo <dnovillo@google.com>
2974 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2975 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2976 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2977 versions of these patterns.
2979 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2980 (__builtin_arc_aligned): Likewise.
2982 * config/arc/arc.md: Expand adc_0 comment stating the intended
2983 purpose and why it isn't ready.
2984 Replace commented out call_value_via_label_mixed with a
2985 plain comment about bl_s.
2987 * config/arc/arc.c (stdio.h): Don't include directly.
2988 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2989 Remove if (1) condition.
2990 (arc_encode_section_info): Fix comment.
2992 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2994 * config/arc/arc.c (arc_conditional_register_usage):
2995 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2996 Also set reg_alloc_order for DMA config regs.
2998 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2999 Jeremy Bennett <jeremy.bennett@embecosm.com>
3001 * doc/install.texi (--with-cpu): Mention ARC.
3002 (arc-*-elf32): New paragraph.
3003 (arc-linux-uclibc): Likewise.
3004 * doc/md.texi (Machine Constraints): Add ARC part.
3005 * doc/invoke.texi: (menu): Add ARC Options.
3006 (Machine Dependent Options) <ARC Options>: Add synopsis.
3007 (node ARC Options): Add.
3008 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3009 (ARC Built-in Functions): New section defining
3010 generic ARC built-in functions.
3011 (ARC SIMD Built-in Functions): New section defining SIMD specific
3013 (Declaring Attributes of Functions): Extended
3014 description of short_call and long_call attributes for ARC and
3015 added index entries.
3017 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3018 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3019 Joern Rennecke <joern.rennecke@embecosm.com>
3020 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3021 Brendan Kehoe <brendan@zen.org>
3022 Michael Eager <eager@eagercon.com>
3023 Simon Cook <simon.cook@embecosm.com>
3024 Jeremy Bennett <jeremy.bennett@embecosm.com>
3026 * config/arc, common/config/arc: New directories.
3028 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3029 Brendan Kehoe <brendan@zen.org>
3030 Simon Cook <simon.cook@embecosm.com>
3032 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3034 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3036 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3037 * tree-ssa-coalesce.h: New. Move prototype to here.
3038 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3039 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3040 (gimple_can_coalesce_p): Move to...
3041 * gimple.c (gimple_can_coalesce_p): Here.
3043 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3045 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3046 (dump_decl_set): Move to gimple.c.
3047 * gimple.h: Don't include tree-ssa-operands.h.
3048 (dump_decl_set): Add prototype.
3049 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3050 Move to gimple-ssa.h.
3051 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3052 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3053 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3054 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3055 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3057 (dump_decl_set): Relocate here.
3058 * gimple-ssa.h: New file.
3059 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3060 Relocate from gimple.h.
3061 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3062 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3063 swap_tree_operands and remove non-ssa path.
3064 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3065 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3067 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3068 vect_is_simple_reduction_1): Use swap_ssa_operands.
3069 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3070 (enum need_phi_state): Move to tree-into-ssa.c.
3071 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3072 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3073 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3074 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3075 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3076 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3077 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3078 num_imm_uses): Move to ssa-iterators.h.
3079 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3080 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3082 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3083 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3084 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3085 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3086 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3087 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3088 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3089 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3090 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3091 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3092 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3093 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3094 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3095 (set_phi_nodes): Move to tree-phinodes.h.
3096 * tree-ssa-operands.h (enum ssa_op_iter_type,
3097 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3098 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3099 (dump_decl_set): Remove prototype.
3100 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3101 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3102 (set_phi_nodes): Relocate from tree-flow-inline.h.
3103 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3105 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3106 include list. Temporarily add gimple.h to include list.
3107 * ssa-iterators.h: New file.
3108 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3109 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3110 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3111 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3112 Relocate from tree-ssa-operands.h.
3113 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3114 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3115 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3116 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3117 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3118 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3119 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3120 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3121 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3122 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3123 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3124 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3125 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3126 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3127 Relocate from tree-flow-inline.h.
3128 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3130 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3133 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3134 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3135 Insert '\t' to output template.
3136 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3137 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3138 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3139 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3141 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3143 * doc/install.texi (Host/target specific installation notes for GCC):
3144 Put @anchor before @heading.
3145 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3146 Use @email for email addresses.
3148 2013-10-01 Jeff Law <law@redhat.com>
3150 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3151 a vec. Only delete the path if we create one without successfully
3152 registering a jump thread.
3153 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3155 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3156 (paths): New vector of jump threading paths.
3157 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3158 (THREAD_PATH): New accessor macro for the entire thread path.
3159 (lookup_redirection_data): Get intermediate and final outgoing edge
3160 from the thread path.
3161 (create_edge_and_update_destination_phis): Copy the threading path.
3162 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3163 jump threading path.
3164 (ssa_redirect_edges): Get edges and block types from the jump threading
3165 path. Free the path vector.
3166 (thread_block): Get edges from the jump threading path. Look at the
3167 entire path to see if we thread to a loop exit. If we cancel a jump
3168 thread request, then free the path vector.
3169 (thread_single_edge): Get edges and block types from the jump threading
3170 path. Free the path vector.
3171 (thread_through_loop_header): Get edges and block types from the jump
3172 threading path. Free the path vector.
3173 (mark_threaded_blocks): Iterate over the vector of paths and store
3174 the path on the appropriate edge. Get edges and block types from the
3175 jump threading path.
3176 (mark_threaded_blocks): Get edges and block types from the jump
3177 threading path. Free the path vector.
3178 (thread_through_all_blocks): Use the vector of paths rather than
3179 a vector of 3-edge sets.
3180 (register_jump_thread): Accept pointer to a path vector rather
3181 than the path vector itself. Store the path vector for later use.
3184 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3185 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3188 * config/s390/s390.c (s390_split_branches): Modify check for table
3190 (s390_chunkify_start): Rearrange table jump insn check in order to
3191 deal with compare and branch insns correctly.
3193 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3197 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3198 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3199 define_insn_and_split.
3200 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3201 (shiftsi3_compare): New pattern.
3203 * config/arm/unspecs.md (UNSPEC_RRX): New.
3205 2013-10-01 Alan Modra <amodra@gmail.com>
3207 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3208 casing inout operands.
3210 2013-10-01 Richard Biener <rguenther@suse.de>
3212 PR tree-optimization/58553
3213 * tree-loop-distribution.c (struct partition_s): Add niter member.
3214 (classify_partition): Populate niter member for the partition
3215 and properly identify whether the relevant store happens before
3216 or after the loop exit.
3217 (generate_memset_builtin): Use niter member from the partition.
3218 (generate_memcpy_builtin): Likewise.
3220 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3222 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3223 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3225 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3227 * basic-block.h (edge_list): Prefix member names with "m_".
3228 * context.h (context): Likewise.
3229 * domwalk.h (dom_walker): Likewise.
3230 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3231 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3232 * hash-table.h (hash_table): Likewise.
3233 * machmode.h (bit_field_mode_iterator): Likewise.
3234 * pass_manager.h (pass_list): Likewise.
3235 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3236 * tree-pass.h (pass_data): Likewise.
3237 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3238 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3239 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3240 * asan.c (pass_data_asan): Update accordingly.
3241 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3242 (control_dependences::control_dependences): Likewise.
3243 (control_dependences::~control_dependences): Likewise.
3244 (control_dependences::~control_dependences): Likewise.
3245 (control_dependences::get_edges_dependent_on): Likewise.
3246 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3247 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3248 * context.c (gcc::context::context): Likewise.
3249 * cprop.c (pass_rtl_cprop::clone): Likewise.
3250 * domwalk.c (dom_walker::walk): Likewise.
3251 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3252 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3253 * mode-switching.c (pass_mode_switching::clone): Likewise.
3254 * passes.c (opt_pass::opt_pass): Likewise.
3255 (pass_manager::pass_manager): Likewise.
3256 * predict.c (pass_strip_predict_hints::clone): Likewise.
3257 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3258 (pass_split_all_insns::clone): Likewise.
3259 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3261 (bit_field_mode_iterator::next_mode): Likewise.
3262 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3263 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3264 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3265 * tree-complex.c (pass_lower_complex::clone): Likewise.
3266 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3267 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3268 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3269 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3270 (pass_fold_builtins::clone): Likewise.
3271 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3272 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3273 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3274 (pass_cd_dce::clone): Likewise.
3275 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3276 (pass_phi_only_cprop::clone): Likewise.
3277 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3278 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3279 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3280 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3281 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3282 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3283 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3284 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3285 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3286 * tree-vrp.c (pass_vrp::clone): Likewise.
3287 * tsan.c (pass_tsan::clone): Likewise.
3289 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3292 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3293 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3296 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3297 optimization, punt if sign_bit_p looked through any zero extension.
3299 2013-09-30 Teresa Johnson <tejohnson@google.com>
3301 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3302 Update redirected out edge count in joiner case.
3303 (ssa_redirect_edges): Common the joiner and non-joiner cases
3304 so that joiner case gets profile updates.
3306 2013-09-30 Richard Biener <rguenther@suse.de>
3308 PR tree-optimization/58554
3309 * tree-loop-distribution.c (classify_partition): Require
3310 unconditionally executed stores for memcpy and memset recognition.
3311 (tree_loop_distribution): Calculate dominance info.
3313 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3315 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3316 (NO_PROFILE_COUNTERS): Likewise.
3317 (PROFILE_HOOK): Likewise.
3318 (FUNCTION_PROFILER): Likewise.
3319 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3321 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3323 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3324 calls and defines in TARGET_MACHO conditional.
3325 (load_macho_picbase_di): Likewise.
3326 (reload_macho_picbase): Likewise.
3327 (reload_macho_picbase_si): Likewise.
3328 (reload_macho_picbase_di): Likewise.
3329 (nonlocal_goto_receiver): Likewise.
3331 2013-09-30 Nick Clifton <nickc@redhat.com>
3333 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3334 that use the MSP430X ISA.
3335 (msp430_option_override): Scan -mmcu command line option for any
3336 MCU name that supports the MSP430X ISA.
3337 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3338 -mmcu options which enable the MSP430X ISA.
3340 2013-09-30 Richard Biener <rguenther@suse.de>
3343 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3344 before looking for setjmp-like calls.
3346 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3349 * config/darwin-protos.h (machopic_get_function_picbase): New.
3350 * config/darwin.c (machopic_get_function_picbase): New.
3351 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3352 label for a new func. (load_macho_picbase_di): Likewise.
3353 (reload_macho_picbase): New expand.
3354 (reload_macho_picbase_si): New insn.
3355 (reload_macho_picbase_di): New insn.
3356 (nonlocal_goto_receiver): New define and split.
3357 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3358 (unspecv enum): Add UNSPECV_NLGR.
3360 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3362 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3363 that altivec registers are correctly sized on Darwin.
3365 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3367 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3368 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3369 * config/x-darwin (host-darwin.o): Likewise.
3370 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3371 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3372 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3374 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3376 * doc/invoke.texi: Fix usage of @tie{} command.
3378 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3380 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3382 2013-09-28 Jan Hubicka <jh@suse.cz>
3384 * config/i386/x86-tune.def: Add documentation for each of the options;
3387 2013-09-28 Jan Hubicka <jh@suse.cz>
3389 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3391 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3392 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3393 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3394 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3395 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3397 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3399 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3400 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3401 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3402 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3403 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3404 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3405 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3406 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3407 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3408 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3409 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3410 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3411 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3412 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3413 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3414 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3415 gimple.h, godump.c, graphite-clast-to-gimple.c,
3416 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3417 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3418 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3419 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3420 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3421 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3422 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3423 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3424 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3425 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3426 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3427 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3428 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3429 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3430 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3431 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3432 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3433 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3434 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3435 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3436 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3437 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3438 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3439 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3440 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3441 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3442 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3443 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3444 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3445 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3446 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3447 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3448 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3449 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3450 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3451 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3452 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3453 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3454 whitespace before "(".
3456 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3458 * expr.h (extract_bit_field): Remove packedp parameter.
3459 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3460 from forward declaration.
3461 (store_split_bit_field): Remove packedp arg from calls to
3462 extract_fixed_bit_field.
3463 (extract_bit_field_1): Remove packedp parameter and packedp
3464 argument from recursive calls and calls to extract_fixed_bit_field.
3465 (extract_bit_field): Remove packedp parameter and corresponding
3466 arg to extract_bit_field_1.
3467 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3469 (extract_split_bit_field): Remove packedp arg from call to
3470 extract_fixed_bit_field.
3471 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3472 (copy_blkmode_from_reg): Likewise.
3473 (copy_blkmode_to_reg): Likewise.
3474 (read_complex_part): Likewise.
3475 (store_field): Likewise.
3476 (expand_expr_real_1): Likewise.
3477 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3478 to extract_bit_field.
3479 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3480 call to extract_bit_field.
3481 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3482 call to extract_bit_field.
3483 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3484 and special packedp behavior from -fstrict-volatile-bitfields
3487 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3489 * lra-eliminations.c (init_elim_table): Guard value_p.
3491 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3493 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3494 DFmode, DImode, and SFmode in the upper VSX registers based on the
3495 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3496 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3497 of the VSX register handling.
3499 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3501 (zero_extendsidi2_lfiwzx): Add support for loading into the
3502 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3503 only do VSX memory options on Altivec registers.
3504 (extendsidi2_lfiwax): Likewise.
3505 (extendsfdf2_fpr): Likewise.
3506 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3507 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3508 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3509 (movdi_internal64): Likewise.
3511 2013-09-27 Xinliang David Li <davidxl@google.com>
3513 * opts.c (finish_options): Adjust parameters
3514 according to vect cost model.
3515 (common_handle_option): Set dynamic vect cost
3517 targhooks.c (default_add_stmt_cost): Compute stmt cost
3519 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3520 Use helper function.
3521 * tree-vectorizer.h (unlimited_cost_model): New function.
3522 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3523 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3525 (vect_enhance_data_refs_alignment): Ditto.
3526 * flag-types.h: New enum.
3527 * common/config/i386/i386-common.c (ix86_option_init_struct):
3528 No need to initialize vect_cost_model flag.
3529 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3532 2013-09-27 Diego Novillo <dnovillo@google.com>
3534 * gimple.h (enum ssa_mode): Remove.
3536 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3538 * cfgloop.h (number_of_loops): Fix typo in check for null.
3540 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3543 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3544 are children of outermost saved_cfun's loop, and set it up to
3545 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3546 if loop != loop0 and SESE region contains bbs that belong to loop0.
3548 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3550 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3551 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3552 (set_address_segment, set_address_base, set_address_index)
3553 (set_address_disp): Accept the argument unconditionally.
3554 (baseness): Remove must_be_base_p and must_be_index_p checks.
3555 (decompose_normal_address): Classify as much as possible in the
3558 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3560 * cse.c (count_reg_usage): Handle INT_LIST.
3561 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3562 * reginfo.c (reg_scan_mark_refs): Likewise.
3563 * reload1.c (eliminate_regs_1): Likewise.
3565 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3568 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3571 2013-09-27 Richard Biener <rguenther@suse.de>
3573 PR tree-optimization/58459
3574 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3575 restriction not propagating into loops.
3577 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3579 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3580 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3581 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3583 2013-09-26 Richard Biener <rguenther@suse.de>
3585 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3587 2013-09-26 Richard Biener <rguenther@suse.de>
3589 * alias.h (component_uses_parent_alias_set): Rename to ...
3590 (component_uses_parent_alias_set_from): ... this.
3591 * alias.c (component_uses_parent_alias_set): Rename to ...
3592 (component_uses_parent_alias_set_from): ... this and return
3594 (reference_alias_ptr_type_1): Use the result from
3595 component_uses_parent_alias_set_from instead of stripping
3596 components one at a time.
3597 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3599 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3601 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3602 Move prototypes to...
3603 * tree-ssa-ter.h: New File. Move prototypes here.
3604 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3605 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3606 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3607 from is_replaceable_p.
3608 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3609 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3610 (process_replaceable): Use ter_is_replaceable_p.
3611 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3612 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3613 newly refactored ssa_is_replaceable_p.
3614 * cfgexpand.c: Include tree-outof-ssa.h.
3615 * ssaexpand.h: Delete.
3617 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3619 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3620 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3622 (create_gimple_tmp): Delete.
3623 (get_expr_type, build_assign, build_type_cast): Move to...
3624 * gimple-builder.c: New File.
3625 (get_expr_type): Relocate from gimple.c.
3626 (build_assign, build_type_cast): Change to only create ssanames.
3627 * gimple.h: Move prototypes to...
3628 * gimple-builder.h: New File. Here.
3629 * tree-ssa.h: And here.
3630 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3631 count_uses_and_derefs): Relocate from gimple.c.
3632 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3633 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3634 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3635 gimple_replace_ssa_lhs.
3636 * asan.c: Include gimple-builder.h.
3637 * Makefile.in: Add gimple-builder.o.
3639 2013-09-26 Richard Biener <rguenther@suse.de>
3641 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3642 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3643 (loe_visit_block): Use gcc_checking_assert.
3644 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3645 gimple_assign_ssa_name_copy_p.
3646 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3649 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3651 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3652 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3654 2013-09-26 Richard Biener <rguenther@suse.de>
3656 PR tree-optimization/58539
3657 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3658 the fact that debug statements are not taking part in loop-closed
3661 2013-09-26 Nick Clifton <nickc@redhat.com>
3663 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3664 time warning message.
3665 (msp430_print_operand_raw): Delete unused letter parameter.
3666 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3667 (msp430_print_operand_address): New function.
3668 (msp430_print_operand): Move address printing code from here to
3670 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3672 (zero_extendpsisi2): Likewise.
3673 (extendpsisi2): New pattern.
3674 (andneghi3): New pattern.
3676 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3678 * config/aarch64/aarch64.opt (mlra): New option.
3679 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3680 (TARGET_LRA_P): Define.
3682 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3684 * expr.c (expand_assignment): Remove obsolete comment.
3686 2013-09-25 Jeff Law <law@redhat.com>
3688 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3689 tree-ssa-threadupdate.h.
3690 (register_jump_thread): Similarly.
3691 * tree-ssa-threadupdate.h: New header file.
3692 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3693 * tree-vrp.c: Likewise.
3694 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3695 (thread_around_empty_blocks): Change type of path vector argument to
3696 an edge,type pair from just an edge. Initialize both elements when
3697 appending to a jump threading path. Tweak references to elements
3699 (thread_across_edge): Similarly. Release memory for the elements
3701 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3702 (dump_jump_thread_path): New function broken out from
3703 register_jump_thread.
3704 (register_jump_thread): Use dump_jump_thread_path. Change type of
3705 path vector entries. Search the path for NULL edges and dump
3706 the path if one is found. Tweak the conversion of path to 3-edge
3707 form to use the block copy type information embedded in the path.
3709 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3711 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3713 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3714 Vladimir Makarov <vmakarov@redhat.com>
3716 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3717 from the least significant bit.
3718 (strip_address_mutations): Add bitfield operations handling.
3719 (must_be_index_p): Add shifting and rotate operations handling.
3720 (set_address_base): Use must_be_base_p predicate.
3721 (set_address_index): Use must_be_index_p predicate.
3723 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3724 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3725 Sergey Lega <sergey.s.lega@intel.com>
3726 Anna Tikhonova <anna.tikhonova@intel.com>
3727 Ilya Tocar <ilya.tocar@intel.com>
3728 Andrey Turetskiy <andrey.turetskiy@intel.com>
3729 Ilya Verbin <ilya.verbin@intel.com>
3730 Kirill Yukhin <kirill.yukhin@intel.com>
3731 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3733 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3735 (ix86_expand_vector_move_misalign): Support new unaligned load and
3736 stores and use new names.
3737 (CODE_FOR_sse2_storedqu): Rename to ...
3738 (CODE_FOR_sse2_storedquv16qi): ... this.
3739 (CODE_FOR_sse2_loaddqu): Rename to ...
3740 (CODE_FOR_sse2_loaddquv16qi): ... this.
3741 (CODE_FOR_avx_loaddqu256): Rename to ...
3742 (CODE_FOR_avx_loaddquv32qi): ... this.
3743 (CODE_FOR_avx_storedqu256): Rename to ...
3744 (CODE_FOR_avx_storedquv32qi): ... this.
3745 * config/i386/i386.md (fpint_logic): New.
3746 * config/i386/sse.md (VMOVE): Extend for AVX512.
3750 (VI_UNALIGNED_LOADSTORE): Ditto.
3751 (sse2_avx_avx512f): Ditto.
3752 (sse2_avx2): Extend for AVX512.
3753 (sse4_1_avx2): Ditto.
3754 (avx2_avx512f): New.
3755 (sse): Extend for AVX512.
3758 (avxsizesuffix): Ditto.
3759 (sseintvecmode): Ditto.
3761 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3762 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3763 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3764 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3765 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3766 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3767 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3768 (STORENT_MODE): Extend for AVX512.
3769 (*absneg<mode>2): Replace constraint "x" with "v".
3770 (*mul<mode>3): Ditto.
3771 (*ieee_smin<mode>3): Ditto.
3772 (*ieee_smax<mode>3): Ditto.
3773 (avx_cmp<mode>3): Replace VF with VF_128_256.
3774 (*<sse>_maskcmp<mode>3_comm): Ditto.
3775 (<sse>_maskcmp<mode>3): Ditto.
3776 (<sse>_andnot<mode>3): Extend for AVX512.
3777 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3778 (<code><mode>3, fpint_logic): New.
3779 (*<code><mode>3): Extend for AVX512.
3780 (avx512flogicsuff): New.
3781 (avx512f_<logic><mode>): Ditto.
3782 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3784 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3785 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3786 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3787 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3788 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3789 (xop_vpermil2<mode>3): Ditto.
3790 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3791 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3792 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3793 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3795 2013-09-25 Tom Tromey <tromey@redhat.com>
3797 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3798 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3799 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3800 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3801 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3802 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3803 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3804 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3805 (GRAPHITE_HTAB_H): Remove.
3807 2013-09-25 Tom Tromey <tromey@redhat.com>
3809 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3811 2013-09-25 Tom Tromey <tromey@redhat.com>
3813 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3815 2013-09-25 Tom Tromey <tromey@redhat.com>
3817 * config/i386/t-i386 (i386.o): Remove.
3818 (i386-c.o): Use COMPILE and POSTCOMPILE.
3820 2013-09-25 Tom Tromey <tromey@redhat.com>
3822 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3824 2013-09-25 Tom Tromey <tromey@redhat.com>
3826 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3827 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3828 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3829 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3830 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3831 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3832 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3833 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3834 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3835 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3836 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3837 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3838 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3839 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3840 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3841 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3842 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3843 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3844 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3845 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3846 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3847 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3848 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3849 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3850 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3851 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3852 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3853 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3854 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3855 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3856 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3857 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3858 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3859 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3860 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3861 (graphite-clast-to-gimple.o, graphite-dependences.o)
3862 (graphite-interchange.o, graphite-poly.o)
3863 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3864 (graphite-optimize-isl.o, tree-vect-loop.o)
3865 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3866 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3867 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3868 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3869 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3870 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3871 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3872 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3873 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3874 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3875 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3876 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3877 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3878 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3879 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3880 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3881 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3882 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3883 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3884 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3885 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3886 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3887 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3888 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3889 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3890 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3891 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3892 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3893 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3894 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3895 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3896 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3897 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3898 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3899 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3900 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3901 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3902 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3903 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3904 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3905 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3906 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3907 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3908 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3909 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3910 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3911 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3912 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3913 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3914 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3915 (gcov-dump.o): Remove.
3916 (default-c.o): Use COMPILE and POSTCOMPILE.
3917 (CFLAGS-gcc.o): New variable.
3918 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3920 2013-09-25 Tom Tromey <tromey@redhat.com>
3922 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3923 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3924 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3925 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3926 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3927 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3928 (c-family/array-notation-common.o, c-family/stub-objc.o)
3929 (c-family/c-ubsan.o): Remove.
3931 2013-09-25 Tom Tromey <tromey@redhat.com>
3933 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3935 2013-09-25 Tom Tromey <tromey@redhat.com>
3937 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3938 to add -DENABLE_SHARED_LIBGCC.
3939 (gcc.o): Don't use subshell.
3941 2013-09-25 Tom Tromey <tromey@redhat.com>
3943 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3944 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3945 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3946 * configure, config.in: Rebuild.
3948 2013-09-25 Tom Tromey <tromey@redhat.com>
3950 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3951 (COMPILE, POSTCOMPILE): New variables.
3952 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3953 (DEPFILES): New variable.
3954 Include ".Po" files.
3955 * configure.ac: Add checks for dependency checking.
3956 * configure, aclocal.m4: Regenerate.
3958 2013-09-25 Tom Tromey <tromey@redhat.com>
3960 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3961 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3963 2013-09-25 Tom Tromey <tromey@redhat.com>
3965 * Makefile.in (generated_files): Add options.h,
3966 target-hooks-def.h, insn-opinit.h,
3967 common/common-target-hooks-def.h, pass-instances.def,
3968 c-family/c-target-hooks-def.h.
3970 2013-09-25 Jeff Law <law@redhat.com>
3972 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3973 than foo[foo.length () - 1] to access last member in a vec.
3974 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3976 2013-09-25 Richard Biener <rguenther@suse.de>
3979 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3981 2013-09-25 Jan Hubicka <jh@suse.cz>
3983 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3986 2013-09-25 Marek Polacek <polacek@redhat.com>
3989 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3990 when determining the type name.
3992 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3994 * config/sh/sh.md: Fix formatting.
3996 2013-09-24 Xinliang David Li <davidxl@google.com>
3998 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3999 max peel iterations parameter.
4000 * param.def: New parameter.
4001 * doc/invoke.texi: Document New parameter.
4003 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4005 * gimple-pretty-print.c: Various whitespace tweaks.
4006 * tree-core.h: Likewise.
4007 * tree-pretty-print.c: Likewise.
4008 * tree-ssa-alias.c: Likewise.
4009 * tree-ssa-copy.c: Likewise.
4010 * tree-ssanames.c: Likewise.
4011 * tree-ssanames.h: Likewise.
4012 * tree-vrp.c: Likewise.
4014 2013-09-24 Alan Modra <amodra@gmail.com>
4018 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4019 for output operands that disallow regs. Don't use EXPAND_WRITE on
4022 2013-09-24 Richard Biener <rguenther@suse.de>
4025 * tree.c (reference_alias_ptr_type): Move ...
4026 * alias.c (reference_alias_ptr_type): ... here and implement
4027 in terms of the new reference_alias_ptr_type_1.
4028 (ref_all_alias_ptr_type_p): New helper.
4029 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4030 use ref_all_alias_ptr_type_p.
4031 (get_deref_alias_set): Add flag_strict_aliasing check here.
4032 (reference_alias_ptr_type_1): New function, split out from ...
4033 (get_alias_set): ... here.
4034 (alias_ptr_types_compatible_p): New function.
4035 * alias.h (reference_alias_ptr_type): Declare.
4036 (alias_ptr_types_compatible_p): Likewise.
4037 * tree.h (reference_alias_ptr_type): Remove.
4038 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4039 to compare MEM_REF alias types.
4041 2013-09-24 Richard Biener <rguenther@suse.de>
4043 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4045 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4047 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4048 reload helper function arrays into a single array reg_addr.
4049 (reload_fpr_gpr): Likewise.
4050 (reload_gpr_vsx): Likewise.
4051 (reload_vsx_gpr): Likewise.
4052 (struct rs6000_reg_addr): Likewise.
4053 (reg_addr): Likewise.
4054 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4055 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4056 (rs6000_init_hard_regno_mode_ok): Likewise.
4057 (rs6000_secondary_reload_direct_move): Likewise.
4058 (rs6000_secondary_reload): Likewise.
4060 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4061 constraints: wu, ww, and wy. Repurpose wv constraint added during
4062 power8 changes. Put wg constraint in alphabetical order.
4064 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4065 for future work to add ISA 2.07 VSX single precision support.
4066 (-mvsx-scalar-double): Change default from -1 to 1, update
4067 documentation comment.
4068 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4069 (-mupper-regs-df): New debug switch to control whether DF values
4070 can go in the traditional Altivec registers.
4071 (-mupper-regs-sf): New debug switch to control whether SF values
4072 can go in the traditional Altivec registers.
4074 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4076 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4077 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4078 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4079 (rs6000_debug_legitimate_address_p): Print if we are running
4080 before, during, or after reload.
4081 (rs6000_secondary_reload): Add a comment.
4082 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4084 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4085 constraints. Update documentation string.
4086 (wd constraint): Likewise.
4087 (wf constraint): Likewise.
4088 (wg constraint): Likewise.
4089 (wn constraint): Likewise.
4090 (ws constraint): Likewise.
4091 (wt constraint): Likewise.
4092 (wx constraint): Likewise.
4093 (wz constraint): Likewise.
4094 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4096 (ww constraint): Likewise.
4097 (wy constraint): Likewise.
4098 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4099 in the previous submissions.
4100 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4102 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4104 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4105 (REG_BR_PROB): Say that the probability is stored in an int_list.
4106 * reg-notes.def: Update commentary to mention INT_LIST.
4107 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4108 (INT_LIST): New rtx.
4109 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4110 * rtlanal.c (int_reg_note_p): New function.
4111 (alloc_reg_note): Assert that the note does not have an int argument.
4112 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4113 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4114 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4115 rather than an INSN_LIST. Handle INT_LIST.
4116 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4117 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4118 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4119 Manipulate them as ints rather than rtxes.
4120 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4121 * regmove.c (copy_src_to_dest): Likewise.
4122 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4124 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4125 into the cases that need it.
4126 * config/arm/arm.c (arm_unwind_emit): Likewise.
4128 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4129 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4130 * loop-doloop.c (add_test, doloop_modify): Likewise.
4131 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4132 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4133 * predict.c (combine_predictions_for_insn): Likewise.
4134 * print-rtl.c (print_rtx): Handle INT_LIST.
4135 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4136 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4137 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4138 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4139 (ix86_split_fp_branch, predict_jump): Likewise.
4140 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4141 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4142 * config/spu/spu.c (ea_load_store_inline): Likewise.
4144 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4145 value of a REG_BR_PROB note.
4146 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4147 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4148 * emit-rtl.c (try_split): Likewise.
4149 * predict.c (br_prob_note_reliable_p): Likewise.
4150 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4151 * reorg.c (mostly_true_jump): Likewise.
4152 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4153 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4154 * config/i386/i386.c (ix86_print_operand): Likewise.
4155 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4156 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4157 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4158 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4159 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4160 * config/sparc/sparc.c (output_cbranch): Likewise.
4161 * config/spu/spu.c (get_branch_target): Likewise.
4162 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4163 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4165 2013-09-23 Jan Hubicka <jh@suse.cz>
4167 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4169 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4170 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4172 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4173 (gimple_fold_call): Dump inconsistent devirtualizations; add
4174 sanity check for type based devirtualizations.
4175 * ipa-prop.c: Include ipa-utils.h
4176 (ipa_intraprocedural_devirtualization): Add sanity check.
4177 (try_make_edge_direct_virtual_call): Likewise.
4179 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4181 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4182 assignment statements.
4184 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4186 * gimple-pretty-print.c (dump_ssaname_info): New function.
4187 (dump_gimple_phi): Call it.
4188 (pp_gimple_stmt_1): Likewise.
4189 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4190 (range_info_def): Declare.
4191 * tree-pretty-print.c (pp_double_int): New function.
4192 (dump_generic_node): Call it.
4193 * tree-pretty-print.h (pp_double_int): Declare.
4194 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4195 * tree-ssanames.h (range_info_def): New structure.
4196 (value_range_type): Move definition here.
4197 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4199 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4201 (set_range_info): New function.
4202 (get_range_info): Likewise.
4203 (duplicate_ssa_name_range_info): Likewise.
4204 (duplicate_ssa_name_fn): Check pointer type and call
4205 duplicate_ssa_name_range_info.
4206 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4207 * tree-vrp.c (value_range_type): Remove definition, now in
4209 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4210 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4211 (SSA_NAME_RANGE_INFO): New macro.
4213 2013-09-23 Richard Biener <rguenther@suse.de>
4215 PR tree-optimization/58464
4216 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4217 (phi_trans_add): Change to add conditionally on being not present.
4218 (phi_translate_1): Remove recursion detection here.
4219 (phi_translate): Pre-seed the cache with NULL to catch
4220 recursion here in a more generic way.
4221 (bitmap_find_leader): Adjust comment.
4222 (get_representative_for): Dump value-numbers.
4223 (create_expression_by_pieces): Likewise.
4224 (insert_into_preds_of_block): Likewise.
4226 2013-09-23 Christian Bruel <christian.bruel@st.com>
4229 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4230 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4232 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4235 2013-09-20 Renlin Li <renlin.li@arm.com>
4237 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4239 (aarch64_expand_epilogue): Likewise.
4240 (aarch64_legitimize_reload_address): Likewise.
4242 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4244 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4245 NULL_TREE before pushing them onto the vector. Likewise for labels.
4247 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4249 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4250 comparison operators when -fno-trapping-math is in effect.
4251 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4252 comparison operators in TFmode and assert that unsupported operators
4254 (ia64_print_operand): Likewise.
4256 2013-09-21 Jan Hubicka <jh@suse.cz>
4258 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4259 (sse_typeless_stores): Enable for core
4260 (sse_load0_by_pxor): Likewise.
4261 (four_jump_limit): Disable for core.
4262 (pad_returns): Likewise.
4263 (avoid_vector_decode): Likewise.
4264 (fuse_cmp_and_branch): Enable for cores.
4265 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4267 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4270 * config/pa/pa.c (pa_option_override): Disable auto increment and
4271 decrement instructions until reload is completed.
4273 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4274 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4275 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4277 2013-09-20 DJ Delorie <dj@redhat.com>
4278 Nick Clifton <nickc@redhat.com>
4280 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4281 (need_to_save): Save bank 0 on interrupts.
4282 (characterize_address): Strip far address wrappers.
4283 (rl78_as_legitimate_address): Likewise.
4284 (transcode_memory_rtx): Likewise.
4285 (rl78_peep_movhi_p): Disable this peephole after devirt.
4286 (rl78_propogate_register_origins): Forget all origins when a
4288 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4289 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4291 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4292 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4293 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4295 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4297 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4299 * config/pa/pa.md: In "scc" insn patterns, change output template to
4300 handle const0_rtx in reg_or_0_operand operands.
4302 2013-09-20 Martin Husemann <martin@NetBSD.org>
4305 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4306 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4308 2013-09-20 Richard Biener <rguenther@suse.de>
4311 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4312 remembering SSA name version and block index.
4313 (new_scev_info_str): Adjust.
4314 (hash_scev_info): Likewise. Also hash the block index.
4315 (eq_scev_info): Adjust.
4316 (find_var_scev_info): Likewise.
4317 (struct instantiate_cache_entry): Remove.
4318 (struct instantiate_cache_type): Use a htab to map name, block
4320 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4321 (get_instantiated_value_entry): Likewise.
4322 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4323 (instantiate_scev_name): Adjust.
4325 2013-09-20 Jeff Law <law@redhat.com>
4327 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4329 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4331 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4332 Call aarch64_simd_expand_args to update op[argc].
4334 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4336 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4339 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4341 * gengtype.c (file_rules): Added rule for *.cc files.
4342 (get_output_file_with_visibility): Give fatal message when no
4345 2013-09-20 Renlin Li <renlin.li@arm.com>
4347 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4348 (aarch64_expand_epilogue): Likewise.
4349 (aarch64_legitimize_reload_address): Likewise.
4351 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4354 * expr.c (expand_assignment): Remove misalignp code path.
4356 2013-09-20 Marek Polacek <polacek@redhat.com>
4359 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4360 TYPE_PRECISION. Add asserts.
4362 2013-09-20 Richard Biener <rguenther@suse.de>
4364 PR tree-optimization/58453
4365 * tree-loop-distribution.c (distribute_loop): Apply the cost
4366 model for -ftree-loop-distribute-patterns, too.
4368 2013-09-20 Richard Biener <rguenther@suse.de>
4371 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4372 make type comparison less strict.
4374 2013-09-20 Alan Modra <amodra@gmail.com>
4376 * configure: Regenerate.
4377 * aclocal.m4: Regenerate.
4379 2013-09-20 Marek Polacek <polacek@redhat.com>
4382 * doc/extend.texi: Document that attribute used is meant to be used
4383 on variables with static storage duration.
4385 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4387 PR tree-optimization/58472
4388 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4389 simd_lane_access set inv_p = false.
4390 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4391 the simduid magic VAR_DECL.
4393 2013-09-19 Jan Hubicka <jh@suse.cz>
4395 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4397 2013-09-17 Jeff Law <law@redhat.com>
4399 * tree-ssa-dom.c (record_temporary_equivalences): New function
4400 split out of dom_opt_dom_walker::after_dom_children.
4401 (dom_opt_dom_walker::thread_across_edge): Move common code
4402 in here from dom_opt_dom_walker::after_dom_children.
4403 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4405 2013-09-19 Jan Hubicka <jh@suse.cz>
4407 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4408 (TARGET_GENERIC): Use PROCESOR_GENERIC
4409 (enum processor_type): Unify generic32 and 64.
4410 * i386.md (cpu): Likewise.
4411 * x86-tune.def (use_leave): Enable for generic32.
4412 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4413 * athlon.md: Change generic64 to generic in all occurences.
4414 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4415 (ix86_target_macros_internal): Likewise.
4416 * driver-i386.c (host_detect_local_cpu): Likewise.
4417 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4419 (generic_memcpy, generic_memset, generic_cost): This one.
4420 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4421 (m_GENERIC32, m_GENERIC64): Remove.
4422 (m_GENERIC): Turn into one flag.
4423 (processor_target): Unify generic tunnings.
4424 (ix86_option_override_internal): Replace generic32/64 by generic.
4425 (ix86_issue_rate): Likewise.
4426 (ix86_adjust_cost): Likewise.
4428 2013-09-19 Jan Hubicka <jh@suse.cz>
4430 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4431 of speculative flag.
4433 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4435 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4436 GOMP_sections_start, even if !exit_reachable.
4438 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4440 * lra-constraints.c (need_for_all_save_p): Use macro
4441 HARD_REGNO_CALL_PART_CLOBBERED.
4442 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4443 set up pseudo conflict hard regs.
4445 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4448 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4451 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4453 PR rtl-optimization/58438
4454 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4455 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4456 from previous subpasses.
4458 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4460 * arm.c (arm_get_frame_offsets): Validate architecture supports
4461 LDRD/STRD before accepting the tuning preference.
4462 (arm_expand_prologue): Likewise.
4463 (arm_expand_epilogue): Likewise.
4465 2013-09-18 Richard Biener <rguenther@suse.de>
4467 PR tree-optimization/58417
4468 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4469 have chrecs with symbols defined in the loop as operands.
4470 (chrec_fold_multiply): Likewise.
4471 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4472 parameters before folding binary operations.
4473 (struct instantiate_cache_entry_hasher): Remove.
4474 (struct instantiate_cache_type): Use a pointer-map.
4475 (instantiate_cache_type::instantiate_cache_type): New function.
4476 (instantiate_cache_type::get): Likewise.
4477 (instantiate_cache_type::set): Likewise.
4478 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4479 (get_instantiated_value_entry): Likewise.
4480 (global_cache): New global.
4481 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4482 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4483 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4484 (instantiate_scev_name): Adjust.
4485 (instantiate_scev): Construct global instead of local cache.
4486 (resolve_mixers): Likewise.
4488 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4489 Paolo Carlini <paolo.carlini@oracle.com>
4492 * doc/implement-cxx.texi: Fix references to the C++ standards.
4494 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4496 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4497 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4498 simd_lane_access drs, update also DR_ALIGNED_TO.
4500 2013-09-18 Marek Polacek <polacek@redhat.com>
4503 * doc/extend.texi: Document no_sanitize_undefined attribute.
4504 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4505 no_sanitize_undefined attribute.
4507 2013-09-18 Nick Clifton <nickc@redhat.com>
4509 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4510 (ASM_DECLARE_FUNCTION_NAME): Define.
4512 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4514 * compare-elim.c (find_comparison_dom_walker): New class
4515 (find_comparisons_in_bb): Rename to
4516 find_comparison_dom_walker::before_dom_children
4517 (find_comparisons): Adjust
4518 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4520 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4521 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4522 (init_walk_dominator_tree): Remove declaration.
4523 (fini_walk_dominator_tree): Remove declaration.
4524 * fwprop.c (single_def_use_dom_walker): New class
4525 (single_def_use_enter_block): Convert to
4526 single_def_use_dom_walker::before_dom_children.
4527 (single_def_use_leave_block): Convert to
4528 single_def_use_dom_walker::after_dom_children.
4529 (build_single_def_use_links): Adjust.
4530 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4532 (find_candidates_in_block): Convert to
4533 find_candidates_dom_walker::before_dom_children.
4534 (execute_strength_reduction): Adjust.
4535 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4536 (sese_dom_walker): New class.
4537 (sese_dom_walker::sese_dom_walker): New constructor.
4538 (sese_dom_walker::~sese_dom_walker): New destructor.
4539 (build_sese_conditions_before): Convert to
4540 sese_dom_walker::before_dom_children.
4541 (build_sese_conditions_after): Convert to
4542 sese_dom_walker::after_dom_children.
4543 (build_poly_scop): Adjust
4544 * tree-into-ssa.c (rewrite_dom_walker): New class
4545 (rewrite_enter_block): Convert to
4546 rewrite_dom_walker::before_dom_children.
4547 (rewrite_leave_block): Convert to
4548 rewrite_dom_walker::after_dom_children.
4549 (rewrite_update_dom_walker): New class.
4550 (rewrite_update_enter_block): Convert to
4551 rewrite_update_dom_walker::before_dom_children.
4552 (rewrite_update_leave_block): Convert to
4553 rewrite_update_dom_walker::after_dom_children.
4554 (rewrite_blocks, rewrite_into_ssa): Adjust.
4555 (mark_def_dom_walker): New class.
4556 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4557 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4558 (mark_def_sites_blocks): Convert to
4559 mark_def_dom_walker::before_dom_children.
4560 (mark_def_site_blocks): Remove.
4561 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4562 (tree_ssa_dominator_optimize): Adjust.
4563 (dom_thread_across_edge): Convert to method
4564 dom_opt_dom_walker::thread_across_edge.
4565 (dom_opt_enter_block): Convert to member function
4566 dom_opt_dom_walker::before_dom_children.
4567 (dom_opt_leave_block): Convert to member function
4568 dom_opt_dom_walker::after_dom_children.
4569 * tree-ssa-dse.c (dse_dom_walker): New class.
4570 (dse_enter_block): Convert to member function
4571 dse_dom_walker::before_dom_children.
4572 (tree_ssa_dse): Adjust.
4573 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4574 (determine_invariantness_stmt): Convert to method
4575 invariantness_dom_walker::before_dom_children.
4576 (determine_invariantness): Remove
4577 (move_computations_dom_walker): New class.
4578 (move_computations_stmt): Convert to method
4579 move_computations_dom_walker::before_dom_children.
4580 (move_computations, tree_ssa_lim): Adjust.
4581 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4582 (nt_init_block): Make method
4583 notrappping_dom_walker::before_dom_children.
4584 (nt_fini_block): Make
4585 method nontrapping_dom_walker::after_dom_children.
4586 (get_non_trapping): Adjust.
4587 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4588 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4589 (eliminate_leave_block): Make method.
4590 eliminate_dom_walker::after_dom_children.
4592 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4593 (strlen_enter_block): Make method
4594 strlen_dom_walker::before_dom_children.
4595 (strlen_leave_block): Make
4596 method strlen_dom_walker::after_dom_children.
4597 (tree_ssa_strlen): Adjust.
4598 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4599 (tree_ssa_uncprop): Adjust.
4600 (uncprop_leave_block): Make method
4601 uncprop_dom_walker::after_dom_children.
4602 (uncprop_leave_block): Make method
4603 uncprop_dom_walker::before_dom_children.
4605 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4607 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4608 branch in same basic block. Check both src and dest of the move insn.
4610 2013-09-17 Nick Clifton <nickc@redhat.com>
4612 * config/rl78/rl78-real.md (bf): New pattern.
4614 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4615 (rl78_print_operand): Do not put a # before a %B.
4616 * config/rl78/rl78.opt: Tweak doc strings.
4618 2013-09-17 DJ Delorie <dj@redhat.com>
4620 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4621 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4622 (rl78_option_override): Likewise, if -mallregs.
4623 (is_virtual_register): Likewise.
4624 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4625 (REGNO_OK_FOR_BASE_P): Likewise.
4626 * config/rl78/rl78.opt (-mallregs): New.
4628 2013-09-17 Nick Clifton <nickc@redhat.com>
4630 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4631 For interrupt functions: save all call clobbered registers if the
4632 interrupt handler is not a leaf function.
4633 (rl78_expand_prologue): Always recompute the frame information.
4634 For interrupt functions: only select bank 0 if one of the bank 0
4635 registers is going to be psuhed.
4637 2013-09-17 DJ Delorie <dj@redhat.com>
4639 * config/rl78/constraints.md: For each W* constraint, rename to C*
4640 and create a W* constraint that checks for an optional ES: prefix
4642 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4643 (es_addr): New. Used to wrap far addresses.
4644 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4645 (rl78_es_base): New.
4646 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4647 wrapped far addresses.
4648 (rl78_print_operand_1): Unwrap far addresses before processing.
4649 (rl78_lo16): Wrap far addresses in unspecs.
4650 (rl78_es_addr): New.
4651 (rl78_es_base): New.
4652 (insn_ok_now): Check for not-yet-wrapped far addresses.
4653 (transcode_memory_rtx): Properly re-wrap far addresses.
4655 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4657 * config/sparc/t-rtems: Add leon3 multilibs.
4659 2013-09-17 Cong Hou <congh@google.com>
4661 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4662 when checking the dot production pattern. The type of rhs operand
4663 of multiply is now checked correctly.
4665 2013-09-17 Jeff Law <law@redhat.com>
4667 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4668 edge implied equivalences into successor phis.
4669 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4670 here from tree-ssa-threadedge.c.
4671 (mark_threaded_blocks): When threading through a joiner, if both
4672 successors of the joiner's clone reach the same block, verify the
4673 PHI arguments are equal. If not, cancel the jump threading request.
4674 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4675 tree-ssa-threadupdate.c
4676 (thread_across_edge): Don't check PHI argument equality when
4677 threading through joiner block here.
4679 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4681 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4682 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4683 (warn_uninit, warn_uninitialized_vars,
4684 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4685 Move to tree-ssa-uninit.c.
4686 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4687 (has_undefined_value_p): New. Pass dependant parts of
4688 ssa_undefined_value_p.
4689 (uninit_undefined_value_p): Use has_undefined_value_p.
4690 (warn_uninit, warn_uninitialized_vars,
4691 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4692 Move from tree-ssa.c.
4693 * tree-ssa.h: Adjust prototypes.
4695 2013-09-17 Jan Hubicka <jh@suse.cz>
4698 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4699 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4700 FUNCTION_NOT_OPTIMIZED.
4701 * ipa-inline-analysis.c (compute_inline_parameters): Function
4702 not optimized is not inlinable unless it is alwaysinline.
4703 (inline_analyze_function): Force calls in not optimized
4704 function not inlinable.
4706 2013-09-17 Jan Hubicka <jh@suse.cz>
4709 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4711 * ipa.c (function_and_variable_visibility): Likewise.
4712 * ipa-profile.c (ipa_profile): Likewise.
4714 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4717 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4718 attribute, and don't inline the resolver in this case.
4720 2013-09-17 Teresa Johnson <tejohnson@google.com>
4722 * coverage.c (get_coverage_counts): Add missing newline.
4724 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4726 PR tree-optimization/58088
4727 * fold-const.c (mask_with_trailing_zeros): New function.
4728 (fold_binary_loc): Make sure we don't recurse infinitely
4729 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4730 Use mask_with_trailing_zeros where appropriate.
4732 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4734 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4735 of 'prev' var to get better distance estimation.
4737 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4739 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4740 * tree-inline.c (is_parameter_of): New predicate.
4741 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4742 a parameter has been remapped.
4743 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4744 (optimize_inline_calls): Initialize transform_parameter.
4745 (copy_gimple_seq_and_replace_locals): Likewise.
4746 (tree_function_versioning): Likewise.
4747 (maybe_inline_call_in_expr): Likewise.
4749 2013-09-17 Nick Clifton <nickc@redhat.com>
4751 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4752 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4754 (is_attr_func): New function.
4755 (msp430_is_interrupt_func): New function.
4756 (is_naked_func): New function.
4757 (is_reentrant_func): New function.
4758 (is_critical_func): New function.
4759 (msp430_start_function): Add annotations for function attributes.
4760 (msp430_attr): New function.
4761 (msp430_attribute_table): New.
4762 (msp430_function_section): New function.
4763 (TARGET_ASM_FUNCTION_SECTION): Define.
4764 (msp430_builtin): New enum.
4765 (msp430_init_builtins): New function.
4766 (msp430_builtin_devl): New function.
4767 (msp430_expand_builtin): New function.
4768 (TARGET_INIT_BUILTINS): Define.
4769 (TARGET_EXPAND_BUILTINS): Define.
4770 (TARGET_BUILTIN_DECL): Define.
4771 (msp430_expand_prologue): Add support for naked, interrupt,
4772 critical and reentrant functions.
4773 (msp430_expand_epilogue): Likewise.
4774 (msp430_print_operand): Handle 'O' character.
4775 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4777 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4778 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4779 (pushm): Use a 'n' rather than an 'i' constraint.
4780 (msp_return): Add generation of the interrupt return instruction.
4781 (disable_interrupts): New pattern.
4782 (enable_interrupts): New pattern.
4783 (push_intr_state): New pattern.
4784 (pop_intr_state): New pattern.
4785 (bic_SR): New pattern.
4786 (bis_SR): New pattern.
4787 * doc/extend.texi: Document MSP430 function attributes and builtin
4790 2013-09-17 Richard Biener <rguenther@suse.de>
4792 PR tree-optimization/58432
4793 * tree-loop-distribution.c (tree_loop_distribution): Also
4794 scan PHIs for outside loop uses and seed a partition from them.
4796 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4798 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4799 (restructure_reference): Call backtrace_base_for_ref.
4801 2013-09-17 Alan Modra <amodra@gmail.com>
4804 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4807 2013-09-16 DJ Delorie <dj@redhat.com>
4809 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4810 vregs location for RL78/G10.
4811 (rl78_expand_prologue): Avoid SEL on G10.
4812 (rl78_expand_epilogue): Likewise.
4813 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4814 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4815 __RL78_G10__ when appropriate.
4816 (ASM_SPEC): Pass -mg10 along to the assembler.
4817 * config/rl78/rl78.md (sel_rb): Disable for G10.
4818 * config/rl78/rl78.opt: Add -mg10 option.
4819 * config/rl78/t-rl78: Add -mg10 multilib.
4821 2013-09-16 Xinliang David Li <davidxl@google.com>
4823 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4824 * omp-low.c (omp_max_vf): Ditto.
4825 (expand_omp_simd): Ditto.
4826 * tree-vectorizer.c (vectorize_loops): Ditto.
4827 (gate_vect_slp): Ditto.
4828 (gate_increase_alignment): Ditto.
4829 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4830 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4831 (gate_tree_vectorize): Name change.
4832 (tree_vectorize): Ditto.
4833 (pass_vectorize::gate): Call new function.
4834 (pass_vectorize::execute): Ditto.
4835 * opts.c: O3 default setting change.
4836 (finish_options): Check new flag.
4837 * doc/invoke.texi: Document new flags.
4838 * common.opt: New flags.
4840 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4842 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4843 (Condition Code): Don't reference it.
4844 * doc/tm.texi: Regenerate.
4846 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4849 * lra-constraints.c (undo_optional_reloads): Consider all optional
4850 reload even if it did not get a hard reg.
4852 2013-09-16 Teresa Johnson <tejohnson@google.com>
4854 * dumpfile.c (dump_loc): Remove newline emission.
4855 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4856 emission to dump_printf_loc calls where missing.
4857 (vect_mark_for_runtime_alias_test): Ditto.
4858 (vect_analyze_data_ref_dependence): Ditto.
4859 (vect_analyze_data_ref_dependences): Ditto.
4860 (vect_slp_analyze_data_ref_dependence): Ditto.
4861 (vect_slp_analyze_data_ref_dependences): Ditto.
4862 (vect_compute_data_ref_alignment): Ditto.
4863 (vect_update_misalignment_for_peel): Ditto.
4864 (vect_verify_datarefs_alignment): Ditto.
4865 (vector_alignment_reachable_p): Ditto.
4866 (vect_get_data_access_cost): Ditto.
4867 (vect_enhance_data_refs_alignment): Ditto.
4868 (vect_find_same_alignment_drs): Ditto.
4869 (vect_analyze_data_refs_alignment): Ditto.
4870 (vect_analyze_group_access): Ditto.
4871 (vect_analyze_data_ref_access): Ditto.
4872 (vect_analyze_data_ref_accesses): Ditto.
4873 (vect_prune_runtime_alias_test_list): Ditto.
4874 (vect_analyze_data_refs): Ditto.
4875 (vect_create_addr_base_for_vector_ref): Ditto.
4876 (vect_create_data_ref_ptr): Ditto.
4877 (vect_grouped_store_supported): Ditto.
4878 (vect_grouped_load_supported): Ditto.
4879 * value-prof.c (check_counter): Ditto.
4880 (check_ic_target): Ditto.
4881 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4882 (vect_recog_widen_mult_pattern): Ditto.
4883 (vect_recog_widen_sum_pattern): Ditto.
4884 (vect_recog_over_widening_pattern): Ditto.
4885 (vect_recog_widen_shift_pattern): Ditto.
4886 (vect_recog_rotate_pattern): Ditto.
4887 (vect_recog_vector_vector_shift_pattern): Ditto.
4888 (vect_recog_divmod_pattern): Ditto.
4889 (vect_recog_mixed_size_cond_pattern): Ditto.
4890 (vect_recog_bool_pattern): Ditto.
4891 (vect_pattern_recog_1): Ditto.
4892 (vect_pattern_recog): Ditto.
4893 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4894 (vect_is_simple_iv_evolution): Ditto.
4895 (vect_analyze_scalar_cycles_1): Ditto.
4896 (vect_get_loop_niters): Ditto.
4897 (vect_analyze_loop_1): Ditto.
4898 (vect_analyze_loop_form): Ditto.
4899 (vect_analyze_loop_operations): Ditto.
4900 (vect_analyze_loop_2): Ditto.
4901 (vect_analyze_loop): Ditto.
4902 (report_vect_op): Ditto.
4903 (vect_is_slp_reduction): Ditto.
4904 (vect_is_simple_reduction_1): Ditto.
4905 (vect_get_known_peeling_cost): Ditto.
4906 (vect_estimate_min_profitable_iters): Ditto.
4907 (vect_model_reduction_cost): Ditto.
4908 (vect_model_induction_cost): Ditto.
4909 (get_initial_def_for_induction): Ditto.
4910 (vect_create_epilog_for_reduction): Ditto.
4911 (vectorizable_reduction): Ditto.
4912 (vectorizable_induction): Ditto.
4913 (vectorizable_live_operation): Ditto.
4914 (vect_loop_kill_debug_uses): Ditto.
4915 (vect_transform_loop): Ditto.
4916 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4917 (vect_stmt_relevant_p): Ditto.
4918 (process_use): Ditto.
4919 (vect_mark_stmts_to_be_vectorized): Ditto.
4920 (vect_model_simple_cost): Ditto.
4921 (vect_model_promotion_demotion_cost): Ditto.
4922 (vect_model_store_cost): Ditto.
4923 (vect_get_store_cost): Ditto.
4924 (vect_model_load_cost): Ditto.
4925 (vect_get_load_cost): Ditto.
4926 (vect_init_vector_1): Ditto.
4927 (vect_get_vec_def_for_operand): Ditto.
4928 (vect_finish_stmt_generation): Ditto.
4929 (vectorizable_call): Ditto.
4930 (vectorizable_conversion): Ditto.
4931 (vectorizable_assignment): Ditto.
4932 (vectorizable_shift): Ditto.
4933 (vectorizable_operation): Ditto.
4934 (vectorizable_store): Ditto.
4935 (vectorizable_load): Ditto.
4936 (vectorizable_condition): Ditto.
4937 (vect_analyze_stmt): Ditto.
4938 (vect_transform_stmt): Ditto.
4939 (vect_is_simple_use): Ditto.
4940 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4941 (vect_can_advance_ivs_p): Ditto.
4942 (vect_update_ivs_after_vectorizer): Ditto.
4943 (vect_do_peeling_for_loop_bound): Ditto.
4944 (vect_gen_niters_for_prolog_loop): Ditto.
4945 (vect_update_inits_of_drs): Ditto.
4946 (vect_create_cond_for_alias_checks): Ditto.
4947 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4948 (vect_build_slp_tree_1): Ditto.
4949 (vect_supported_load_permutation_p): Ditto.
4950 (vect_analyze_slp_instance): Ditto.
4951 (vect_analyze_slp): Ditto.
4952 (vect_make_slp_decision): Ditto.
4953 (vect_detect_hybrid_slp): Ditto.
4954 (vect_bb_vectorization_profitable_p): Ditto.
4955 (vect_slp_analyze_bb_1): Ditto.
4956 (vect_update_slp_costs_according_to_vf): Ditto.
4957 (vect_get_mask_element): Ditto.
4958 (vect_transform_slp_perm_load): Ditto.
4959 (vect_schedule_slp_instance): Ditto.
4960 (vect_schedule_slp): Ditto.
4961 (vect_slp_transform_bb): Ditto.
4962 * profile.c (read_profile_edge_counts): Ditto.
4963 (compute_branch_probabilities): Ditto.
4964 * coverage.c (get_coverage_counts): Ditto.
4966 2013-09-16 Diego Novillo <dnovillo@google.com>
4968 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4970 2013-09-16 Jan Hubicka <jh@suse.cz>
4972 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4973 abstract functions; for static functions check the presence of body.
4975 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4977 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4978 * config/aarch64/aarch64-simd.md
4979 (aarch64_mla_elt<mode>): New.
4980 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4981 (aarch64_mls_elt<mode>): Likewise.
4982 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4983 (aarch64_fma4_elt<mode>): Likewise.
4984 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4985 (aarch64_fma4_elt_to_128v2df): Likewise.
4986 (aarch64_fma4_elt_to_64df): Likewise.
4987 (fnma<mode>4): Likewise.
4988 (aarch64_fnma4_elt<mode>): Likewise.
4989 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4990 (aarch64_fnma4_elt_to_128v2df): Likewise.
4991 (aarch64_fnma4_elt_to_64df): Likewise.
4992 * config/aarch64/iterators.md (VDQSF): New.
4993 * config/aarch64/arm_neon.h
4994 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4995 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4997 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4999 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5000 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5001 (aarch64_mul3_elt_to_128df): Likewise.
5002 (aarch64_mul3_elt_to_64v2df): Likewise.
5003 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5005 (VMUL_CHANGE_NLANES) Likewise.
5008 * config/aarch64/arm_neon.h
5009 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5011 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5013 * config/aarch64/arm_neon.h
5014 (vcvtx_high_f32_f64): Fix parameters.
5016 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5017 Uros Bizjak <ubizjak@gmail.com>
5019 * config/alpha.c: Include tree-ssanames.h.
5021 2013-09-16 Richard Biener <rguenther@suse.de>
5023 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5024 (dot_rdg_1): Handle control_dd.
5025 (create_edge_for_control_dependence): New function.
5026 (create_rdg_edges): Add control dependences if asked for.
5027 (build_rdg): Likewise.
5028 (generate_loops_for_partition): If there are not necessary
5029 control stmts remove all their dependencies.
5030 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5031 (distribute_loop): Pass on control dependences.
5032 (tree_loop_distribution): Compute control dependences and remove
5033 restriction on number of loop nodes.
5035 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5037 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5040 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5042 * cse.c (try_const_anchors): Punt on CC modes.
5044 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5046 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5048 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5051 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5054 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5057 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5059 (darwin_objc1_section): Likewise.
5060 (darwin_file_end): Emit Image Info section when required.
5062 2013-09-14 Jan Hubicka <jh@suse.cz>
5064 * tree-into-ssa.c (gate_into_ssa): New.
5065 (pass_data_build_ssa): Use it.
5066 * cgraph.h (expand_thunk): Update prototype.
5067 * cgraphunit.c (analyze_function): Expand thunks early.
5068 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5069 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5070 set lowered flag; do not add new function.
5071 (assemble_thunks_and_aliases): Update.
5072 * tree-ssa.c (gate_init_datastructures): New gate.
5073 (pass_data_init_datastructures): Use it.
5075 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5078 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5079 xmm register set described in the psABI.
5081 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5083 * dwarf2out.c (should_emit_struct_debug): Add check
5084 for type_decl variable is not NULL.
5086 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5088 * config.gcc: Use new winnt-c.c target hooks
5089 * config/t-winnt: New file
5090 * config/winnt-c.c: New file
5091 * doc/tm.texi.in: Document new hook
5092 * doc/tm.texi: Regenerated
5094 2013-09-13 Jan Hubicka <jh@suse.cz>
5097 * ipa-inline.c (check_callers): New function.
5098 (check_caller_edge): Remove.
5099 (want_inline_function_to_all_callers_p): Also permit alises that are
5101 (inline_to_all_callers): Terminate the walk when devirtualization
5102 introduce new calls.
5104 2013-09-13 Jan Hubicka <jh@suse.cz>
5106 * ipa-inline-analysis.c (struct growth_data): Add node.
5107 (do_estimate_growth_1): Fix detection of recursion.
5109 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5111 PR tree-optimization/58392
5112 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5113 to avoid shadowing of outer loop variable. If
5114 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5115 replace_by_duplicate_decl simduid of loops that have it set and
5116 set dest_cfun->has_simduid_loops and/or
5117 dest_cfun->has_force_vect_loops.
5118 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5119 instead of maybe_lookup_decl.
5120 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5121 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5122 call. Copy over force_vect and copy and remap simduid. Set
5123 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5124 (copy_cfg_body): Remove blocks_to_copy argument. Use
5125 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5126 caller. Don't set cfun->has_simduid_loops and/or
5127 cfun->has_force_vect_loops here.
5128 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5130 (expand_call_inline, tree_function_versioning): Adjust copy_body
5133 2013-09-13 Martin Jambor <mjambor@suse.cz>
5136 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5137 the assert if the edge was a speculative one.
5139 2013-09-13 Richard Biener <rguenther@suse.de>
5141 * tree-data-ref.h (known_dependences_p): Move here ...
5142 * tree-loop-distribution.c (known_dependences_p): ... from here.
5143 (dump_rdg_component, debug_rdg_component): Remove.
5145 (generate_loops_for_partition): Use gimple_uid instead of
5146 relying on matching stmt visit order.
5147 (rdg_build_partitions): Take starting stmt vector.
5148 (ldist_gen): Merge into ...
5149 (distribute_loop): ... this function. Do not compute starting
5151 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5153 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5155 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5156 Set type attribute correctly. Set predicable_short_it attribute.
5157 (cmpsi_shiftsi): Remove %? from output template.
5159 2013-09-13 Richard Biener <rguenther@suse.de>
5161 * tree-loop-distribution.c (struct rdg_component,
5162 rdg_defs_used_in_other_loops_p, free_rdg_components,
5163 rdg_build_components): Remove.
5164 (stmts_from_loop): Do not record virtual PHIs.
5165 (generate_loops_for_partition): Skip virtual PHIs.
5166 (build_rdg_partition_for_component): Rename to ...
5167 (build_rdg_partition_for_vertex): ... this and adjust.
5168 (rdg_build_partitions): Take a vector of starting vertices
5169 instead of components. Remove unnecessary leftover handling.
5170 (ldist_gen): Do not build components or record other stores.
5171 (distribute_loop): Do not distribute loops containing stmts
5174 2013-09-13 Christian Bruel <christian.bruel@st.com>
5177 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5179 2013-09-13 Kai Tietz <ktietz@redhat.com>
5181 * config.gcc: Separate cases for mingw and cygwin targets,
5182 and add 64-bit cygwin target case.
5184 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5185 dll-export inline-functions.
5186 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5189 2013-09-13 Jeff Law <law@redhat.com>
5193 2013-09-06 Jeff Law <law@redhat.com>
5195 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5196 edge implied equivalences into successor phis.
5198 2013-09-12 DJ Delorie <dj@redhat.com>
5200 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5203 2013-09-12 Brooks Moses <bmoses@google.com>
5206 * Makefile.in: Do not install driver binaries in $(target)/bin.
5208 2013-09-12 DJ Delorie <dj@redhat.com>
5210 * config/rl78/rl78.opt (mrelax): New.
5211 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5212 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5214 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5216 (rl78_expand_epilogue): Likewise.
5218 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5221 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5222 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5223 pointer elimination with using remove_reg_equal_offset_note.
5225 2013-09-12 DJ Delorie <dj@redhat.com>
5227 * config/msp430/: New port.
5228 * config.gcc (msp430): Added.
5229 * doc/invoke.texi: Document MSP430 options.
5230 * doc/install.texi: Document msp430-elf
5231 * doc/md.texi: Document msp430-elf
5232 * doc/contrib.texi: Document msp430-elf
5234 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5237 2013-09-12 Martin Jambor <mjambor@suse.cz>
5240 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5241 reference descriptor is NULL.
5242 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5245 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5247 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5248 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5249 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5250 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5251 make_temp_ssa_name): move to tree-ssanames.h
5252 * tree-ssa-alias.h: Move prototype.
5253 * tree-ssa.h: Include tree-ssanames.h.
5254 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5255 * tree-ssanames.h: New. Move items from tree-flow*.h
5256 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5258 2013-09-12 Richard Biener <rguenther@suse.de>
5260 PR tree-optimization/58404
5261 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5262 propagate non-invariant addresses into dereferences wrapped
5263 in component references.
5265 2013-09-12 Richard Biener <rguenther@suse.de>
5267 PR tree-optimization/58402
5268 * passes.def: Move pass_late_warn_uninitialized later.
5270 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5272 * tree-ssa.h: New. Move content from tree-flow.h and
5274 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5275 Move prototypes belonging to tree-ssa.c.
5276 * tree-flow-inline.h (redirect_edge_var_map_def,
5277 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5279 * gimple.h: Adjust prototypes.
5280 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5282 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5283 * tree.h: Move prototype to tree-ssa.h.
5284 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5285 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5286 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5287 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5288 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5289 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5290 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5291 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5292 graphite-blocking.c, graphite-clast-to-gimple.c,
5293 graphite-dependences.c, graphite-interchange.c,
5294 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5295 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5296 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5297 ipa-reference.c, ipa-split.c, ipa-utils.c,
5298 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5299 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5300 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5301 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5302 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5303 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5304 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5305 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5306 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5307 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5308 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5309 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5310 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5311 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5312 value-prof.c, var-tracking.c,
5313 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5315 2013-09-12 Richard Biener <rguenther@suse.de>
5317 PR tree-optimization/58396
5318 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5319 (build_rdg): Take a loop-nest parameter, fix memleaks.
5320 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5322 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5324 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5327 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5329 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5330 The FMA instruction names should have a 'v' prefix.
5332 2013-09-12 Richard Biener <rguenther@suse.de>
5334 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5335 (dot_rdg): Use popen instead of system in optional code.
5336 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5337 (already_processed_vertex_p): Adjust.
5338 (has_anti_or_output_dependence, predecessor_has_mem_write,
5339 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5340 rdg_flag_uses): Remove.
5341 (rdg_flag_vertex): Simplify.
5342 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5344 (build_rdg_partition_for_component): Process the first vertex
5345 of a component only.
5346 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5348 2013-09-12 Alan Modra <amodra@gmail.com>
5350 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5352 2013-09-11 DJ Delorie <dj@redhat.com>
5353 Nick Clifton <nickc@redhat.com>
5355 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5356 (rl78_stack_based_mem): New.
5357 * config/rl78/constraints.md (Iv08): New.
5366 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5368 (movsi): Change from expand to insn-and-split.
5369 (ashrsi3): Clobber AX.
5373 * config/rl78/rl78.md (CC_REG): Fix.
5374 (addsi3): Allow memory and immediate operands.
5375 (addsi3_internal): Split into...
5376 (addsi3_internal_virt): ...new, and ...
5377 (addsi3_internal_real): ...new.
5379 (subsi3_internal_virt): New.
5380 (subsi3_internal_real): New.
5381 (mulsi3): Add memory operand.
5382 (mulsi3_rl78): Likewise.
5383 (mulsi3_g13): Likewise.
5384 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5385 (cbranchqi4_real): Add more constraint options.
5386 (cbranchhi4_real): Expand pattern.
5387 (cbranchhi4_real_signed): New.
5388 (cbranchhi4_real_inverted): New.
5389 (cbranchsi4_real_lt): New.
5390 (cbranchsi4_real_ge): New.
5391 (cbranchsi4_real_signed): New.
5392 (cbranchsi4_real): New.
5394 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5396 (lshrsi3_virt): Likewise.
5397 (ashlsi3_virt): Likewise.
5398 (cbranchqi4_virt_signed): New.
5399 (cbranchhi4_virt_signed): New.
5400 (cbranchsi4_virt): New.
5401 * config/rl78/rl78.c: Whitespace fixes throughout.
5402 (move_elim_pass): New.
5403 (pass_data_rl78_move_elim): New.
5404 (pass_rl78_move_elim): New.
5405 (make_pass_rl78_move_elim): New.
5406 (rl78_devirt_info): Run devirt earlier.
5407 (rl78_move_elim_info): New.
5408 (rl78_asm_file_start): Register it.
5409 (rl78_split_movsi): New.
5410 (rl78_as_legitimate_address): Allow virtual base registers when
5412 (rl78_addr_space_convert): Remove spurious debug stuff.
5413 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5414 (rl78_print_operand): More cases for not printing '#'.
5415 (rl78_expand_compare): Remove most of the logic.
5416 (content_memory): New.
5417 (clear_content_memory): New.
5418 (get_content_index): New.
5419 (get_content_name): New.
5420 (display_content_memory): New.
5421 (update_content): New.
5422 (record_content): New.
5423 (already_contains): New.
5424 (insn_ok_now): Re-recog insns with virtual registers.
5425 (add_postponed_content_update): New.
5426 (process_postponed_content_update): New.
5427 (gen_and_emit_move): New.
5428 (transcode_memory_rtx): Record new location content.
5429 Use gen_and_emit_move.
5430 (force_into_acc): New.
5431 (move_to_acc): Use gen_and_emit_move.
5432 (move_from_acc): Likewise.
5433 (move_acc_to_reg): Likewise.
5434 (move_to_x): Likewise.
5435 (move_to_hl): Likewise.
5436 (move_to_de): Likewise.
5437 (rl78_alloc_physical_registers_op1): Record location content.
5438 (has_constraint): New.
5439 (rl78_alloc_physical_registers_op2): Record location content.
5441 (rl78_alloc_physical_registers_ro1): Likewise.
5442 (rl78_alloc_physical_registers_cmp): Likewise.
5443 (rl78_alloc_physical_registers_umul): Likewise.
5444 (rl78_alloc_address_registers_macax): New.
5445 (rl78_alloc_physical_registers): Initialize and set location
5446 content memory as needed.
5447 (rl78_reorg): Make sure split2 is called.
5448 (rl78_rtx_costs): New.
5450 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5452 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5453 for (not (neg ...)) and (neg (not ...)) cases.
5455 2013-09-11 Richard Biener <rguenther@suse.de>
5458 * passes.def: Split critical edges before late uninit warning passes.
5459 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5461 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5463 PR tree-optimization/58385
5464 * fold-const.c (build_range_check): If both low and high are NULL,
5465 use omit_one_operand_loc to preserve exp side-effects.
5467 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5469 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5471 2013-09-11 Richard Biener <rguenther@suse.de>
5473 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5474 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5475 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5476 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5477 stmts_from_loop, known_dependences_p, build_empty_rdg,
5478 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5479 * tree-loop-distribution.c: ... here.
5480 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5481 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5482 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5483 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5484 * tree-loop-distribution.c: ... here.
5485 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5486 (struct partition_s): Add loops member.
5487 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5488 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5489 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5491 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5492 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5493 Sergey Lega <sergey.s.lega@intel.com>
5494 Anna Tikhonova <anna.tikhonova@intel.com>
5495 Ilya Tocar <ilya.tocar@intel.com>
5496 Andrey Turetskiy <andrey.turetskiy@intel.com>
5497 Ilya Verbin <ilya.verbin@intel.com>
5498 Kirill Yukhin <kirill.yukhin@intel.com>
5499 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5501 * config/i386/constraints.md (k): New.
5503 * config/i386/i386.c (const regclass_map): Add new mask registers.
5504 (dbx_register_map): Ditto.
5505 (dbx64_register_map): Ditto.
5506 (svr4_dbx_register_map): Ditto.
5507 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5509 (ix86_preferred_reload_class): Disable constants for mask registers.
5510 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5511 (ix86_hard_regno_mode_ok): Support new mask registers.
5512 (x86_order_regs_for_local_alloc): Ditto.
5513 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5514 (FIXED_REGISTERS): Add new mask registers.
5515 (CALL_USED_REGISTERS): Ditto.
5516 (REG_ALLOC_ORDER): Ditto.
5517 (VALID_MASK_REG_MODE): New.
5518 (FIRST_MASK_REG): Ditto.
5519 (LAST_MASK_REG): Ditto.
5520 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5521 (MAYBE_MASK_CLASS_P): New.
5522 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5523 (REG_CLASS_CONTENTS): Ditto.
5524 (MASK_REGNO_P): New.
5525 (ANY_MASK_REG_P): Ditto.
5526 (HI_REGISTER_NAMES): Add new mask registers.
5527 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5528 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5530 (attribute "type"): Add mskmov, msklog.
5531 (attribute "length_immediate"): Support them.
5532 (attribute "memory"): Ditto.
5533 (attribute "prefix_0f"): Ditto.
5534 (*movhi_internal): Support new mask registers.
5535 (*movqi_internal): Ditto.
5536 (define_split): Split out clobber pattern is a logic
5537 insn on mask registers.
5538 (*k<logic><mode>): New.
5539 (*andhi_1): Extend to support mask regs.
5540 (*andqi_1): Extend to support mask regs.
5542 (define_split): Split and-not to and and not if operands
5544 (*<code><mode>_1): Separate HI mode to new pattern...
5545 (*<code>hi_1): This.
5546 (*<code>qi_1): Extend to support mask regs.
5548 (kortestzhi): Ditto.
5549 (kortestchi): Ditto.
5551 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5552 (*one_cmplhi2_1): ...Here, now with mask registers support.
5553 (*one_cmplqi2_1): Support new mask registers.
5554 (HI/QImode arithmetics splitter): Don't split if mask registers
5556 (HI/QImode not splitter): Ditto.
5557 * config/i386/predicated.md (mask_reg_operand): New.
5558 (general_reg_operand): Ditto.
5560 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5562 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5563 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5565 2013-09-10 Jeff Law <law@redhat.com>
5567 PR tree-optimization/58380
5568 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5569 of threading through a buried loop header.
5571 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5572 return value for single successor case.
5574 2013-09-10 Jan Hubicka <jh@suse.cz>
5576 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5578 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5581 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5582 support conditional execution.
5583 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5585 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5587 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5589 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5590 Check destination too to check move insn.
5591 (undo_optional_reloads): Add check that the original peudo did not
5592 changed its allocation and the optional reload was inherited on last
5593 inheritance pass. Break loop after deciding to keep optional reload.
5594 (lra_undo_inheritance): Add check that inherited pseudo still in
5597 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5599 * config/aarch64/aarch64.md (generic_sched): New.
5600 * config/aarch64/aarch64-generic.md (load): Make conditional
5601 on generic_sched attribute.
5602 (nonload): Likewise.
5604 2013-09-10 Jan Hubicka <jh@suse.cz>
5606 * lto-cgraph.c: Include ipa-utils.h.
5607 (compute_ltrans_boundary): Also add possible targets into the boundary.
5609 2013-09-10 Jan Hubicka <jh@suse.cz>
5611 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5612 VAR_DECL of vtable rather than full expression.
5614 2013-09-10 Jan Hubicka <jh@suse.cz>
5615 Paolo Carlini <paolo.carlini@oracle.com>
5617 * cgraphunit.c (analyze_functions): Save input_location, set it
5618 to UNKNOWN_LOCATION and restore it at the end.
5620 2013-09-10 Martin Jambor <mjambor@suse.cz>
5622 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5623 represented by a thunk.
5625 2013-09-10 Jeff Law <law@redhat.com>
5627 PR tree-optimization/58343
5628 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5629 jump threading requests through loop headers buried in the middle
5630 of a jump threading path.
5632 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5633 in return value/type.
5635 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5637 PR rtl-optimization/58365
5638 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5639 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5642 2013-09-10 Richard Biener <rguenther@suse.de>
5644 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5645 * tree-data-ref.c (create_rdg_vertices): Collect all data
5646 references, signal failure to the caller, use data-ref API.
5647 (build_rdg): Compute data references only once. Maintain lifetime
5648 of data references and data dependences from within RDG.
5649 (free_rdg): Free dependence relations.
5650 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5651 inventing extra dependences.
5652 (distribute_loop): Update for RDG API changes.
5654 2013-09-10 Kai Tietz <ktietz@redhat.com>
5656 * doc/invoke.texi (fms-extensions): Document changed
5657 behavior for ms-abi targets.
5658 * config/i386/i386.c (ix86_option_override_internal):
5659 Set default value of option -fms-extension for ms-abi targets.
5661 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5663 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5665 2013-09-10 Alan Modra <amodra@gmail.com>
5668 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5670 2013-09-10 Alan Modra <amodra@gmail.com>
5672 * config/rs6000/predicates.md (add_cint_operand): New.
5673 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5674 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5675 using add_cint_operand.
5676 (largetoc_high_plus_aix): Likewise.
5678 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5680 PR tree-optimization/58364
5681 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5682 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5683 the current range can't be an unconditional true or false.
5685 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5687 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5689 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5691 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5693 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5696 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5698 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5700 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5701 NO_REGS for immediate that can't be moved directly into FP_REGS.
5703 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5705 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5706 comparison with negated operand.
5707 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5710 2013-09-09 Richard Biener <rguenther@suse.de>
5713 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5714 of a subloop record all its block as affecting loop-closed SSA form.
5716 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5718 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5719 of an rtx/bitpos pair.
5720 (store_fixed_bit_field): Update accordingly.
5722 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5724 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5726 * builtins.c (expand_errno_check): Likewise.
5727 * dwarf2cfi.c (init_return_column_size): Likewise.
5728 * except.c (sjlj_mark_call_sites): Likewise.
5729 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5730 * lra-constraints.c (emit_inc): Likewise.
5731 * ree.c (combine_set_extension): Likewise.
5732 * regmove.c (fixup_match_2): Likewise.
5733 * reload1.c (inc_for_reload): Likewise.
5735 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5737 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5738 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5739 Use gen_int_mode with the mode of the associated simplify_* call.
5740 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5741 * expmed.c (expand_shift_1): Likewise.
5742 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5743 * loop-iv.c (iv_number_of_iterations): Likewise.
5744 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5745 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5747 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5749 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5750 of the associated expand_* call.
5751 (asan_emit_stack_protection): Likewise.
5752 * builtins.c (round_trampoline_addr): Likewise.
5753 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5754 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5755 (emit_store_flag): Likewise.
5756 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5757 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5759 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5760 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5761 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5762 * modulo-sched.c (generate_prolog_epilog): Likewise.
5763 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5764 (expand_ctz, expand_ffs, expand_unop): Likewise.
5766 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5768 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5769 of the associated gen_rtx_* call.
5770 * caller-save.c (init_caller_save): Likewise.
5771 * combine.c (find_split_point, make_extraction): Likewise.
5772 (make_compound_operation): Likewise.
5773 * dwarf2out.c (mem_loc_descriptor): Likewise.
5774 * explow.c (plus_constant, probe_stack_range): Likewise.
5775 * expmed.c (expand_mult_const): Likewise.
5776 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5777 * reload1.c (init_reload): Likewise.
5778 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5779 * var-tracking.c (adjust_mems): Likewise.
5780 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5781 rather than gen_rtx_fmt_ee.
5783 2013-09-09 Jan Hubicka <jh@suse.cz>
5786 * value-prof.c (gimple_ic): Copy also abnormal edges.
5788 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5790 * asan.c (asan_shadow_cst): Use gen_int_mode.
5792 2013-09-08 Jan Hubicka <jh@suse.cz>
5794 * ipa-profile.c: Add toplevel comment.
5795 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5796 (contains_hot_call_p): New function.
5797 (ipa_propagate_frequency): Set frequencies based on counts when
5799 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5800 profile; do not tamper with profile after inlining if it is read.
5802 2013-09-08 Jan Hubicka <jh@suse.cz>
5804 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5807 2013-09-08 Jan Hubicka <jh@suse.cz>
5809 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5812 2013-09-08 Jeff Law <law@redhat.com>
5815 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5818 2013-09-08 Andi Kleen <ak@linux.intel.com>
5820 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5822 2013-09-08 Jan Hubicka <jh@suse.cz>
5824 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5826 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5828 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5829 for non-debug insns.
5830 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5831 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5832 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5835 2013-09-08 Jan Hubicka <jh@suse.cz>
5837 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5838 targets and devirtualize to BUILT_IN_UNREACHABLE.
5839 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5840 * ipa.c (walk_polymorphic_call_targets): New function.
5841 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5842 functions; use the new timevar.
5843 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5844 was removed from the program.
5845 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5846 not consider it in the walk when its vtable is dead.
5847 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5849 (devirt_variable_node_removal_hook): New function.
5850 (possible_polymorphic_call_targets): Also register
5851 devirt_variable_node_removal_hook.
5852 (ipa_devirt): Do not do non-speculative devirtualization.
5853 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5855 2013-09-08 Jan Hubicka <jh@suse.cz>
5857 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5858 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5859 varpool_remove_variable_insertion_hook): Declare.
5860 * varpool.c (varpool_node_hook_list): New structure.
5861 (first_varpool_node_removal_hook,
5862 first_varpool_variable_insertion_hook): New variables.
5863 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5864 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5865 varpool_remove_variable_insertion_hook,
5866 varpool_call_variable_insertion_hooks): New functions.
5867 (varpool_remove_node): Use it.
5869 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5872 * diagnostic.c (diagnostic_build_prefix): When s.file is
5873 "<built-in>" don't output line and column numbers.
5875 2013-09-06 Jan Hubicka <jh@suse.cz>
5877 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5878 * lto-streamer-out.c: Stream thunks, too.
5879 * lto-streamer-in.c (input_function): Pop cfun here
5880 (lto_read_body): Instead of here.
5882 2013-09-06 Caroline Tice <cmtice@google.com>
5884 * doc/install.texi: Add documentation for the --enable-vtable-verify
5885 and the --disable-libvtv configure options.
5887 2013-09-06 Jeff Law <law@redhat.com>
5889 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5890 edge implied equivalences into successor phis.
5892 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5894 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5896 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5898 * resource.c (mark_target_live_regs): Compute resources taking
5899 into account if a call is predicated or not.
5901 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5903 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5904 the compiler in the function names.
5906 2013-09-06 Jan Hubicka <jh@suse.cz>
5909 * ipa-inline.c (has_caller_p): New function.
5910 (want_inline_function_to_all_callers_p): Use it.
5911 (sum_callers, inline_to_all_callers): Break out from ...
5912 (ipa_inline): ... here.
5914 2013-09-06 Jan Hubicka <jh@suse.cz>
5916 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5917 only when AVX is enabled.
5919 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5921 * config/aarch64/aarch64.md
5922 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5923 is fpsimd_<load/store>2.
5924 (load_pair<mode>): Likewise.
5925 (store_pair<mode>): Likewise.
5927 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5929 * config/arm/types.md (type): Add "mrs" type.
5930 * config/aarch64/aarch64.md
5931 (aarch64_load_tp_hard): Make type "mrs".
5933 (load_tp_hard): Make type "mrs".
5934 * config/arm/cortex-a15.md: Update with new attributes.
5935 * config/arm/cortex-a5.md: Update with new attributes.
5936 * config/arm/cortex-a53.md: Update with new attributes.
5937 * config/arm/cortex-a7.md: Update with new attributes.
5938 * config/arm/cortex-a8.md: Update with new attributes.
5939 * config/arm/cortex-a9.md: Update with new attributes.
5940 * config/arm/cortex-m4.md: Update with new attributes.
5941 * config/arm/cortex-r4.md: Update with new attributes.
5942 * config/arm/fa526.md: Update with new attributes.
5943 * config/arm/fa606te.md: Update with new attributes.
5944 * config/arm/fa626te.md: Update with new attributes.
5945 * config/arm/fa726te.md: Update with new attributes.
5947 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5949 * config/aarch64/aarch64.md
5950 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5951 (*movtf_aarch64): Likewise.
5953 (thumb1_movdi_insn): Use "multiple" for type where more than one
5954 instruction is used for a move.
5955 (*arm32_movhf): Likewise.
5956 (*thumb_movdf_insn): Likewise.
5958 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5960 * config/arm/types.md (type): Rename fcpys to fmov.
5962 (*arm_movsi_vfp): Rename type fcpys as fmov.
5963 (*thumb2_movsi_vfp): Likewise
5964 (*movhf_vfp_neon): Likewise
5965 (*movhf_vfp): Likewise
5966 (*movsf_vfp): Likewise
5967 (*thumb2_movsf_vfp): Likewise
5968 (*movsfcc_vfp): Likewise
5969 (*thumb2_movsfcc_vfp): Likewise
5970 * config/aarch64/aarch64-simd.md
5971 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5972 * config/aarch64/aarch64.md
5973 (*movsi_aarch64): Replace type mov_reg with fmovs.
5974 (*movdi_aarch64): Likewise
5975 (*movsf_aarch64): Likewise
5976 (*movdf_aarch64): Likewise
5978 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5979 * config/arm/iwmmxt.md
5980 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5981 * config/arm/arm1020e.md: Update with new attributes.
5982 * config/arm/cortex-a15-neon.md: Update with new attributes.
5983 * config/arm/cortex-a5.md: Update with new attributes.
5984 * config/arm/cortex-a53.md: Update with new attributes.
5985 * config/arm/cortex-a7.md: Update with new attributes.
5986 * config/arm/cortex-a8-neon.md: Update with new attributes.
5987 * config/arm/cortex-a9.md: Update with new attributes.
5988 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5989 * config/arm/cortex-r4f.md: Update with new attributes.
5990 * config/arm/marvell-pj4.md: Update with new attributes.
5991 * config/arm/vfp11.md: Update with new attributes.
5993 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5995 * config/aarch64/aarch64.md
5996 (*madd<mode>): Fix type attribute.
5997 (*maddsi_uxtw): Likewise.
5998 (*msub<mode>): Likewise.
5999 (*msubsi_uxtw): Likewise.
6000 (<su_optab>maddsidi4): Likewise.
6001 (<su_optab>msubsidi4): Likewise.
6003 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6005 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6006 * config/arm/arm.md (core_cycles): Remove fdiv.
6007 * config/arm/vfp.md:
6008 (*sqrtsf2_vfp): Update for attribute changes.
6009 (*sqrtdf2_vfp): Likewise.
6010 * config/aarch64/aarch64.md:
6011 (sqrt<mode>2): Update for attribute changes.
6012 * config/arm/arm1020e.md: Update with new attributes.
6013 * config/arm/cortex-a15-neon.md: Update with new attributes.
6014 * config/arm/cortex-a5.md: Update with new attributes.
6015 * config/arm/cortex-a53.md: Update with new attributes.
6016 * config/arm/cortex-a7.md: Update with new attributes.
6017 * config/arm/cortex-a8-neon.md: Update with new attributes.
6018 * config/arm/cortex-a9.md: Update with new attributes.
6019 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6020 * config/arm/cortex-r4f.md: Update with new attributes.
6021 * config/arm/marvell-pj4.md: Update with new attributes.
6022 * config/arm/vfp11.md: Update with new attributes.
6024 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6026 * config/arm/types.md
6027 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6028 * config/aarch64/aarch64.md
6029 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6031 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6032 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6033 (float<GPI:mode><GPF:mode>2): Likewise.
6035 (*truncsisf2_vfp): Update with new attributes.
6036 (*truncsidf2_vfp): Likewise.
6037 (fixuns_truncsfsi2): Likewise.
6038 (fixuns_truncdfsi2): Likewise.
6039 (*floatsisf2_vfp): Likewise.
6040 (*floatsidf2_vfp): Likewise.
6041 (floatunssisf2): Likewise.
6042 (floatunssidf2): Likewise.
6043 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6044 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6045 * config/arm/arm1020e.md: Update with new attributes.
6046 * config/arm/cortex-a15-neon.md: Update with new attributes.
6047 * config/arm/cortex-a5.md: Update with new attributes.
6048 * config/arm/cortex-a53.md: Update with new attributes.
6049 * config/arm/cortex-a7.md: Update with new attributes.
6050 * config/arm/cortex-a8-neon.md: Update with new attributes.
6051 * config/arm/cortex-a9.md: Update with new attributes.
6052 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6053 * config/arm/cortex-r4f.md: Update with new attributes.
6054 * config/arm/marvell-pj4.md: Update with new attributes.
6055 * config/arm/vfp11.md: Update with new attributes.
6057 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6059 * config/aarch64/arm_neon.h
6060 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6061 (vqtbx<1,2,3,4><q>_s8): Likewise.
6063 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6065 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6066 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6067 (add<mode>3): Add type attribute.
6068 (add<mode>3): Likewise.
6069 (usadd<mode>3): Likewise.
6070 (ssadd<mode>3): Likewise.
6071 (sub<mode>3): Likewise.
6072 (sub<mode>3): Likewise.
6073 (ussub<mode>3): Likewise.
6074 (sssub<mode>3): Likewise.
6075 (ssmulsa3): Likewise.
6076 (usmulusa3): Likewise.
6077 (arm_usatsihi): Likewise.
6079 (*movdi_vfp): Add types for all instructions.
6080 (*movdi_vfp_cortexa8): Likewise.
6081 (*movhf_vfp_neon): Likewise.
6082 (*movhf_vfp): Likewise.
6083 (*movdf_vfp): Likewise.
6084 (*thumb2_movdf_vfp): Likewise.
6085 (*thumb2_movdfcc_vfp): Likewise.
6086 * config/arm/arm.md: Add type attribute to all insn patterns.
6087 (*thumb1_adddi3): Add type attribute.
6088 (*arm_adddi3): Likewise.
6089 (*adddi_sesidi_di): Likewise.
6090 (*adddi_zesidi_di): Likewise.
6091 (*thumb1_addsi3): Likewise.
6092 (addsi3_compare0): Likewise.
6093 (*addsi3_compare0_scratch): Likewise.
6094 (*compare_negsi_si): Likewise.
6095 (cmpsi2_addneg): Likewise.
6096 (*addsi3_carryin_<optab>): Likewise.
6097 (*addsi3_carryin_alt2_<optab>): Likewise.
6098 (*addsi3_carryin_clobercc_<optab>): Likewise.
6099 (*subsi3_carryin): Likewise.
6100 (*subsi3_carryin_const): Likewise.
6101 (*subsi3_carryin_compare): Likewise.
6102 (*subsi3_carryin_compare_const): Likewise.
6103 (*arm_subdi3): Likewise.
6104 (*thumb_subdi3): Likewise.
6105 (*subdi_di_zesidi): Likewise.
6106 (*subdi_di_sesidi): Likewise.
6107 (*subdi_zesidi_di): Likewise.
6108 (*subdi_sesidi_di): Likewise.
6109 (*subdi_zesidi_ze): Likewise.
6110 (thumb1_subsi3_insn): Likewise.
6111 (*arm_subsi3_insn): Likewise.
6112 (*anddi3_insn): Likewise.
6113 (*anddi_zesidi_di): Likewise.
6114 (*anddi_sesdi_di): Likewise.
6115 (*ne_zeroextracts): Likewise.
6116 (*ne_zeroextracts): Likewise.
6117 (*ite_ne_zeroextr): Likewise.
6118 (*ite_ne_zeroextr): Likewise.
6119 (*anddi_notdi_di): Likewise.
6120 (*anddi_notzesidi): Likewise.
6121 (*anddi_notsesidi): Likewise.
6122 (andsi_notsi_si): Likewise.
6123 (thumb1_bicsi3): Likewise.
6124 (*iordi3_insn): Likewise.
6125 (*iordi_zesidi_di): Likewise.
6126 (*iordi_sesidi_di): Likewise.
6127 (*thumb1_iorsi3_insn): Likewise.
6128 (*xordi3_insn): Likewise.
6129 (*xordi_zesidi_di): Likewise.
6130 (*xordi_sesidi_di): Likewise.
6131 (*arm_xorsi3): Likewise.
6132 (*andsi_iorsi3_no): Likewise.
6133 (*smax_0): Likewise.
6134 (*smax_m1): Likewise.
6135 (*arm_smax_insn): Likewise.
6136 (*smin_0): Likewise.
6137 (*arm_smin_insn): Likewise.
6138 (*arm_umaxsi3): Likewise.
6139 (*arm_uminsi3): Likewise.
6140 (*minmax_arithsi): Likewise.
6141 (*minmax_arithsi_): Likewise.
6142 (*satsi_<SAT:code>): Likewise.
6143 (arm_ashldi3_1bit): Likewise.
6144 (arm_ashrdi3_1bit): Likewise.
6145 (arm_lshrdi3_1bit): Likewise.
6146 (*arm_negdi2): Likewise.
6147 (*thumb1_negdi2): Likewise.
6148 (*arm_negsi2): Likewise.
6149 (*thumb1_negsi2): Likewise.
6150 (*negdi_extendsid): Likewise.
6151 (*negdi_zero_extend): Likewise.
6152 (*arm_abssi2): Likewise.
6153 (*thumb1_abssi2): Likewise.
6154 (*arm_neg_abssi2): Likewise.
6155 (*thumb1_neg_abss): Likewise.
6156 (one_cmpldi2): Likewise.
6157 (extend<mode>di2): Likewise.
6158 (*compareqi_eq0): Likewise.
6159 (*arm_extendhisi2addsi): Likewise.
6160 (*arm_movdi): Likewise.
6161 (*thumb1_movdi_insn): Likewise.
6162 (*arm_movt): Likewise.
6163 (*thumb1_movsi_insn): Likewise.
6164 (pic_add_dot_plus_four): Likewise.
6165 (pic_add_dot_plus_eight): Likewise.
6166 (tls_load_dot_plus_eight): Likewise.
6167 (*thumb1_movhi_insn): Likewise.
6168 (*thumb1_movsf_insn): Likewise.
6169 (*movdf_soft_insn): Likewise.
6170 (*thumb_movdf_insn): Likewise.
6171 (cbranchsi4_insn): Likewise.
6172 (cbranchsi4_scratch): Likewise.
6173 (*negated_cbranchsi4): Likewise.
6174 (*tbit_cbranch): Likewise.
6175 (*tlobits_cbranch): Likewise.
6176 (*tstsi3_cbranch): Likewise.
6177 (*cbranchne_decr1): Likewise.
6178 (*addsi3_cbranch): Likewise.
6179 (*addsi3_cbranch_scratch): Likewise.
6180 (*arm_cmpdi_insn): Likewise.
6181 (*arm_cmpdi_unsig): Likewise.
6182 (*arm_cmpdi_zero): Likewise.
6183 (*thumb_cmpdi_zero): Likewise.
6184 (*deleted_compare): Likewise.
6185 (*mov_scc): Likewise.
6186 (*mov_negscc): Likewise.
6187 (*mov_notscc): Likewise.
6188 (*cstoresi_eq0_thumb1_insn): Likewise.
6189 (cstoresi_nltu_thumb1): Likewise.
6190 (cstoresi_ltu_thu): Likewise.
6191 (thumb1_addsi3_addgeu): Likewise.
6192 (*arm_jump): Likewise.
6193 (*thumb_jump): Likewise.
6194 (*check_arch2): Likewise.
6195 (arm_casesi_internal): Likewise.
6196 (thumb1_casesi_dispatch): Likewise.
6197 (*arm_indirect_jump): Likewise.
6198 (*thumb1_indirect_jump): Likewise.
6200 (*and_scc): Likewise.
6201 (*ior_scc): Likewise.
6202 (*compare_scc): Likewise.
6203 (*cond_move): Likewise.
6204 (*cond_arith): Likewise.
6205 (*cond_sub): Likewise.
6206 (*cmp_ite0): Likewise.
6207 (*cmp_ite1): Likewise.
6208 (*cmp_and): Likewise.
6209 (*cmp_ior): Likewise.
6210 (*ior_scc_scc): Likewise.
6211 (*ior_scc_scc_cmp): Likewise.
6212 (*and_scc_scc): Likewise.
6213 (*and_scc_scc_cmp): Likewise.
6214 (*and_scc_scc_nod): Likewise.
6215 (*negscc): Likewise.
6216 (movcond_addsi): Likewise.
6217 (movcond): Likewise.
6218 (*ifcompare_plus_move): Likewise.
6219 (*if_plus_move): Likewise.
6220 (*ifcompare_move_plus): Likewise.
6221 (*if_move_plus): Likewise.
6222 (*ifcompare_arith_arith): Likewise.
6223 (*if_arith_arith): Likewise.
6224 (*ifcompare_arith_move): Likewise.
6225 (*if_arith_move): Likewise.
6226 (*ifcompare_move_arith): Likewise.
6227 (*if_move_arith): Likewise.
6228 (*ifcompare_move_not): Likewise.
6229 (*if_move_not): Likewise.
6230 (*ifcompare_not_move): Likewise.
6231 (*if_not_move): Likewise.
6232 (*ifcompare_shift_move): Likewise.
6233 (*if_shift_move): Likewise.
6234 (*ifcompare_move_shift): Likewise.
6235 (*if_move_shift): Likewise.
6236 (*ifcompare_shift_shift): Likewise.
6237 (*ifcompare_not_arith): Likewise.
6238 (*ifcompare_arith_not): Likewise.
6239 (*if_arith_not): Likewise.
6240 (*ifcompare_neg_move): Likewise.
6241 (*if_neg_move): Likewise.
6242 (*ifcompare_move_neg): Likewise.
6243 (*if_move_neg): Likewise.
6244 (prologue_thumb1_interwork): Likewise.
6245 (*cond_move_not): Likewise.
6246 (*sign_extract_onebit): Likewise.
6247 (*not_signextract_onebit): Likewise.
6248 (stack_tie): Likewise.
6249 (align_4): Likewise.
6250 (align_8): Likewise.
6251 (consttable_end): Likewise.
6252 (consttable_1): Likewise.
6253 (consttable_2): Likewise.
6254 (consttable_4): Likewise.
6255 (consttable_8): Likewise.
6256 (consttable_16): Likewise.
6257 (*thumb1_tablejump): Likewise.
6258 (prefetch): Likewise.
6259 (force_register_use): Likewise.
6260 (thumb_eh_return): Likewise.
6261 (load_tp_hard): Likewise.
6262 (load_tp_soft): Likewise.
6263 (tlscall): Likewise.
6264 (*arm_movtas_ze): Likewise.
6265 (*arm_rev): Likewise.
6266 (*arm_revsh): Likewise.
6267 (*arm_rev16): Likewise.
6268 * config/arm/thumb2.md
6269 (*thumb2_smaxsi3): Likewise.
6270 (*thumb2_sminsi3): Likewise.
6271 (*thumb32_umaxsi3): Likewise.
6272 (*thumb2_uminsi3): Likewise.
6273 (*thumb2_negdi2): Likewise.
6274 (*thumb2_abssi2): Likewise.
6275 (*thumb2_neg_abss): Likewise.
6276 (*thumb2_movsi_insn): Likewise.
6277 (tls_load_dot_plus_four): Likewise.
6278 (*thumb2_movhi_insn): Likewise.
6279 (*thumb2_mov_scc): Likewise.
6280 (*thumb2_mov_negs): Likewise.
6281 (*thumb2_mov_negs): Likewise.
6282 (*thumb2_mov_nots): Likewise.
6283 (*thumb2_mov_nots): Likewise.
6284 (*thumb2_movsicc_): Likewise.
6285 (*thumb2_movsfcc_soft_insn): Likewise.
6286 (*thumb2_indirect_jump): Likewise.
6287 (*thumb2_and_scc): Likewise.
6288 (*thumb2_ior_scc): Likewise.
6289 (*thumb2_ior_scc_strict_it): Likewise.
6290 (*thumb2_cond_move): Likewise.
6291 (*thumb2_cond_arith): Likewise.
6292 (*thumb2_cond_ari): Likewise.
6293 (*thumb2_cond_sub): Likewise.
6294 (*thumb2_negscc): Likewise.
6295 (*thumb2_movcond): Likewise.
6296 (thumb2_casesi_internal): Likewise.
6297 (thumb2_casesi_internal_pic): Likewise.
6298 (*thumb2_alusi3_short): Likewise.
6299 (*thumb2_mov<mode>_shortim): Likewise.
6300 (*thumb2_addsi_short): Likewise.
6301 (*thumb2_subsi_short): Likewise.
6302 (thumb2_addsi3_compare0): Likewise.
6303 (*thumb2_cbz): Likewise.
6304 (*thumb2_cbnz): Likewise.
6305 (*thumb2_one_cmplsi2_short): Likewise.
6306 (*thumb2_negsi2_short): Likewise.
6307 (*orsi_notsi_si): Likewise.
6308 * config/arm/arm1020e.md: Update with new attributes.
6309 * config/arm/arm1026ejs.md: Update with new attributes.
6310 * config/arm/arm1136jfs.md: Update with new attributes.
6311 * config/arm/arm926ejs.md: Update with new attributes.
6312 * config/arm/cortex-a15.md: Update with new attributes.
6313 * config/arm/cortex-a5.md: Update with new attributes.
6314 * config/arm/cortex-a53.md: Update with new attributes.
6315 * config/arm/cortex-a7.md: Update with new attributes.
6316 * config/arm/cortex-a8.md: Update with new attributes.
6317 * config/arm/cortex-a9.md: Update with new attributes.
6318 * config/arm/cortex-m4.md: Update with new attributes.
6319 * config/arm/cortex-r4.md: Update with new attributes.
6320 * config/arm/fa526.md: Update with new attributes.
6321 * config/arm/fa606te.md: Update with new attributes.
6322 * config/arm/fa626te.md: Update with new attributes.
6323 * config/arm/fa726te.md: Update with new attributes.
6325 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6327 * config/aarch64/aarch64-simd.md
6328 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6329 <vwx> iterator to ensure correct register choice.
6330 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6331 (aarch64_sqdmull_n<mode>): Likewise.
6332 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6333 * config/aarch64/arm_neon.h
6334 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6335 (vml<as><q>_n_<su>16): Likewise.
6336 (vml<as>l_high_lane<q>_<su>16): Likewise.
6337 (vml<as>l_high_n_<su>16): Likewise.
6338 (vml<as>l_lane<q>_<su>16): Likewise.
6339 (vml<as>l_n_<su>16): Likewise.
6340 (vmul<q>_lane<q>_<su>16): Likewise.
6341 (vmul<q>_n_<su>16): Likewise.
6342 (vmull_lane<q>_<su>16): Likewise.
6343 (vmull_n_<su>16): Likewise.
6344 (vmull_high_lane<q>_<su>16): Likewise.
6345 (vmull_high_n_<su>16): Likewise.
6346 (vqrdmulh<q>_n_s16): Likewise.
6348 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6350 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6351 have the correct lane parameter.
6353 2013-09-06 Richard Biener <rguenther@suse.de>
6355 * cfganal.c (control_dependences::~control_dependences):
6356 Properly free all of the vector.
6358 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6361 * config/i386/i386.c (ix86_conditional_register_usage):
6362 Proper initialize extended SSE registers.
6364 2013-09-06 Jan Hubicka <jh@suse.cz>
6366 PR tree-optimization/58311
6367 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6369 2013-09-06 Jan Hubicka <jh@suse.cz>
6371 * Makefile.in (tree-sra.o): Update dependencies.
6372 * tree-sra.c: Include ipa-utils.h
6373 (scan_function): Use recursive_call_p.
6374 (has_caller_p): New function.
6375 (cgraph_for_node_and_aliases): Count also callers of aliases.
6377 2013-09-06 Jan Hubicka <jh@suse.cz>
6380 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6381 * tree-tailcall.c: Include ipa-utils.h.
6382 (find_tail_calls): Use it.
6383 * ipa-pure-const.c (check_call): Likewise.
6384 * ipa-utils.c (recursive_call_p): New function.
6385 * ipa-utils.h (recursive_call_p): Dclare.
6386 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6387 (symtab_semantically_equivalent_p): New function.
6388 * Makefile.in (tree-tailcall.o): Update dependencies.
6390 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6392 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6395 2013-09-06 Richard Biener <rguenther@suse.de>
6397 * lto-streamer.h (lto_global_var_decls): Remove.
6398 * Makefile.in (OBJS): Remove lto-symtab.o.
6399 (lto-symtab.o): Remove.
6400 (GTFILES): Remove lto-symtab.c
6401 * lto-symtab.c: Move to lto/
6403 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6405 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6406 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6407 (UNSPEC_FPINT_RINT): New constant definitions.
6408 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6409 definition with 2 attributes.
6410 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6411 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6414 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6416 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6417 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6418 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6419 attribute to "z196_alone".
6420 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6423 2013-09-06 Richard Biener <rguenther@suse.de>
6425 * basic-block.h (class control_dependences): New.
6426 * tree-ssa-dce.c (control_dependence_map): Remove.
6428 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6429 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6430 find_pdom, find_control_dependence, find_all_control_dependences):
6432 (mark_control_dependent_edges_necessary,
6433 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6434 tree_dce_done, perform_tree_ssa_dce): Adjust.
6435 * cfganal.c (set_control_dependence_map_bit,
6436 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6437 find_all_control_dependences): Move from tree-ssa-dce.c and
6438 implement as methods of control_dependences class.
6439 (control_dependences::control_dependences): New.
6440 (control_dependences::~control_dependences): Likewise.
6441 (control_dependences::get_edges_dependent_on): Likewise.
6442 (control_dependences::get_edge): Likewise.
6444 2013-09-04 Jan Hubicka <jh@suse.cz>
6446 * tree.c (types_same_for_odr): Drop overactive check.
6447 * ipa-devirt.c (hash_type_name): Likewise.
6449 2013-09-04 Jan Hubicka <jh@suse.cz>
6451 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6452 (analyze_functions): ... here.
6454 2013-09-04 Jan Hubicka <jh@suse.cz>
6457 * cgraphunit.c (analyze_functions): Clear AUX fields
6458 after processing; initialize assembler name has.
6460 2013-09-05 Jeff Law <law@redhat.com>
6462 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6463 from thread_around_empty_block. Record threading path into PATH.
6464 Recurse if threading through the initial block is successful.
6465 (thread_across_edge): Corresponding changes to slightly simplify.
6467 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6469 * config/aarch64/aarch64.md
6470 (type): Remove frecpe, frecps, frecpx.
6471 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6472 fix to be a TARGET_SIMD instruction.
6473 (aarch64_frecps): Remove.
6474 * config/aarch64/aarch64-simd.md
6475 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6476 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6478 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6479 Sofiane Naci <sofiane.naci@arm.com>
6481 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6482 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6483 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6484 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6485 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6486 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6487 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6488 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6489 into "clz, "rbit". Rename "shift" to "shift_imm".
6490 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6491 changes. Update for attribute changes all occurrences of arlo_* and
6493 * config/arm/arm-fixed.md: Update for attribute changes
6494 all occurrences of arlo_* types.
6495 * config/arm/thumb2.md: Update for attribute changes all occurrences
6497 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6498 (cortexa7_older_only): Likewise.
6499 (cortexa7_younger): Likewise.
6500 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6501 (1020alu_shift_op): Likewise.
6502 (1020alu_shift_reg_op): Likewise.
6503 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6504 (alu_shift_op): Likewise.
6505 (alu_shift_reg_op): Likewise.
6506 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6507 (11_alu_shift_op): Likewise.
6508 (11_alu_shift_reg_op): Likewise.
6509 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6510 (9_alu_shift_reg_op): Likewise.
6511 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6513 (cortex_a15_alu_shift): Likewise.
6514 (cortex_a15_alu_shift_reg): Likewise.
6515 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6517 (cortex_a5_alu_shift): Likewise.
6518 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6520 (cortex_a53_alu_shift): Likewise.
6521 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6523 (cortex_a7_alu_reg): Likewise.
6524 (cortex_a7_alu_shift): Likewise.
6525 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6527 (cortex_a8_alu_shift): Likewise.
6528 (cortex_a8_alu_shift_reg): Likewise.
6529 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6530 (cortex_a9_dp_shift): Likewise.
6531 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6533 * config/arm/cortex-r4.md
6534 (cortex_r4_alu): Update for attribute changes.
6535 (cortex_r4_mov): Likewise.
6536 (cortex_r4_alu_shift_reg): Likewise.
6537 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6538 (526_alu_shift_op): Likewise.
6539 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6540 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6541 (626te_alu_shift_op): Likewise.
6542 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6543 (726te_alu_shift_op): Likewise.
6544 (726te_alu_shift_reg_op): Likewise.
6545 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6546 (mp626_alu_shift_op): Likewise.
6547 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6548 (pj4_alu_conds): Likewise.
6549 (pj4_shift): Likewise.
6550 (pj4_shift_conds): Likewise.
6551 (pj4_alu_shift): Likewise.
6552 (pj4_alu_shift_conds): Likewise.
6553 * config/aarch64/aarch64.md: Update for attribute change
6554 all occurrences of arlo_* and shift* types.
6556 2013-09-05 Mike Stump <mikestump@comcast.net>
6558 * tree.h: Move documentation for tree_function_decl to tree-core.h
6559 with the declaration.
6561 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6564 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6565 looking for widest mode.
6567 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6569 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6571 2013-09-05 Richard Biener <rguenther@suse.de>
6573 PR tree-optimization/58137
6574 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6575 Do not create vectors of pointers.
6576 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6577 types for the components of the vector initializer.
6578 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6579 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6581 2013-09-05 Martin Jambor <mjambor@suse.cz>
6583 * ipa-prop.c (remove_described_reference): Accept missing references,
6584 return false if that hppens, otherwise return true.
6585 (cgraph_node_for_jfunc): New function.
6586 (try_decrement_rdesc_refcount): Likewise.
6587 (try_make_edge_direct_simple_call): Use them.
6588 (ipa_edge_removal_hook): Remove references from rdescs.
6589 (ipa_edge_duplication_hook): Clone rdescs and their references
6590 when the new edge has the same caller as the old one.
6591 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6592 reference before removing any edges.
6594 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6596 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6598 * thumb2.md (thumb2_storewb_parisi): New pattern.
6600 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6602 * config/aarch64/aarch64-option-extensions.def: Add
6603 AARCH64_OPT_EXTENSION of 'crc'.
6604 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6605 (AARCH64_ISA_CRC): Ditto.
6606 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6607 description of the CRC extension.
6609 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6611 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6612 * config/rs6000/linux.h: Ditto.
6613 * alpha/linux.h: Ditto.
6614 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6615 no_c99_libc_has_function.
6616 * config/c6x/uclinux-elf.h: Ditto.
6617 * config/lm32/uclinux-elf.h: Ditto.
6618 * config/m68k/uclinux.h: Ditto.
6619 * config/moxie/uclinux.h: Ditto.
6620 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6621 (crisv32-*-linux*, cris-*-linux*): Ditto.
6622 * config/bfin/bfin.c: Include "tm_p.h".
6624 2013-09-05 Richard Biener <rguenther@suse.de>
6626 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6627 check for a definition without a basic-block.
6629 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6630 Sofiane Naci <sofiane.naci@arm.com>
6632 * config/aarch64/aarch64.md
6633 (*movti_aarch64): Rename r_2_f and f_2_r.
6634 (*movsf_aarch64): Likewise.
6635 (*movdf_aarch64): Likewise.
6636 (*movtf_aarch64): Likewise.
6637 (aarch64_movdi_<mode>low): Likewise.
6638 (aarch64_movdi_<mode>high): Likewise.
6639 (aarch64_mov<mode>high_di): Likewise.
6640 (aarch64_mov<mode>low_di): Likewise.
6641 (aarch64_movtilow_tilow): Likewise.
6642 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6643 values to config/arm/types.md
6644 (attribute "conds"): Update for attribute change.
6645 (anddi3_insn): Likewise.
6646 (iordi3_insn): Likewise.
6647 (xordi3_insn): Likewise.
6648 (one_cmpldi2): Likewise.
6649 * config/arm/types.md (type): Add Neon types.
6650 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6651 use "type" attribute.
6652 (movmisalign<mode>_neon_store): Likewise.
6653 (movmisalign<mode>_neon_load): Likewise.
6654 (vec_set<mode>_internal): Likewise.
6655 (vec_setv2di_internal): Likewise.
6656 (vec_extract<mode>): Likewise.
6657 (vec_extractv2di): Likewise.
6658 (add<mode>3_neon): Likewise.
6659 (adddi3_neon): Likewise.
6660 (sub<mode>3_neon): Likewise.
6661 (subdi3_neon): Likewise.
6662 (mul<mode>3_neon): Likewise.
6663 (mul<mode>3add<mode>_neon): Likewise.
6664 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6665 (fma<VCVTF:mode>4)): Likewise.
6666 (fma<VCVTF:mode>4_intrinsic): Likewise.
6667 (fmsub<VCVTF:mode>4)): Likewise.
6668 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6669 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6670 (ior<mode>3): Likewise.
6671 (and<mode>3): Likewise.
6672 (anddi3_neon): Likewise.
6673 (orn<mode>3_neon): Likewise.
6674 (orndi3_neon): Likewise.
6675 (bic<mode>3_neon): Likewise.
6676 (bicdi3_neon): Likewise.
6677 (xor<mode>3): Likewise.
6678 (one_cmpl<mode>2): Likewise.
6679 (abs<mode>2): Likewise.
6680 (neg<mode>2): Likewise.
6681 (umin<mode>3_neon): Likewise.
6682 (umax<mode>3_neon): Likewise.
6683 (smin<mode>3_neon): Likewise.
6684 (smax<mode>3_neon): Likewise.
6685 (vashl<mode>3): Likewise.
6686 (vashr<mode>3_imm): Likewise.
6687 (vlshr<mode>3_imm): Likewise.
6688 (ashl<mode>3_signed): Likewise.
6689 (ashl<mode>3_unsigned): Likewise.
6690 (neon_load_count): Likewise.
6691 (ashldi3_neon_noclobber): Likewise.
6692 (signed_shift_di3_neon): Likewise.
6693 (unsigned_shift_di3_neon): Likewise.
6694 (ashrdi3_neon_imm_noclobber): Likewise.
6695 (lshrdi3_neon_imm_noclobber): Likewise.
6696 (widen_ssum<mode>3): Likewise.
6697 (widen_usum<mode>3): Likewise.
6698 (quad_halves_<code>v4si): Likewise.
6699 (quad_halves_<code>v4sf): Likewise.
6700 (quad_halves_<code>v8hi): Likewise.
6701 (quad_halves_<code>v16qi): Likewise.
6702 (reduc_splus_v2di): Likewise.
6703 (neon_vpadd_internal<mode>): Likewise.
6704 (neon_vpsmin<mode>): Likewise.
6705 (neon_vpsmax<mode>): Likewise.
6706 (neon_vpumin<mode>): Likewise.
6707 (neon_vpumax<mode>): Likewise.
6708 (ss_add<mode>_neon): Likewise.
6709 (us_add<mode>_neon): Likewise.
6710 (ss_sub<mode>_neon): Likewise.
6711 (us_sub<mode>_neon): Likewise.
6712 (neon_vadd<mode>_unspec): Likewise.
6713 (neon_vaddl<mode>): Likewise.
6714 (neon_vaddw<mode>): Likewise.
6715 (neon_vhadd<mode>): Likewise.
6716 (neon_vqadd<mode>): Likewise.
6717 (neon_vaddhn<mode>): Likewise.
6718 (neon_vmul<mode>): Likewise.
6719 (neon_vmla<mode>): Likewise.
6720 (neon_vmlal<mode>): Likewise.
6721 (neon_vmls<mode>): Likewise.
6722 (neon_vmlsl<mode>): Likewise.
6723 (neon_vqdmulh<mode>): Likewise.
6724 (neon_vqdmlal<mode>): Likewise.
6725 (neon_vqdmlsl<mode>): Likewise.
6726 (neon_vmull<mode>): Likewise.
6727 (neon_vqdmull<mode>): Likewise.
6728 (neon_vsub<mode>_unspec): Likewise.
6729 (neon_vsubl<mode>): Likewise.
6730 (neon_vsubw<mode>): Likewise.
6731 (neon_vqsub<mode>): Likewise.
6732 (neon_vhsub<mode>): Likewise.
6733 (neon_vsubhn<mode>): Likewise.
6734 (neon_vceq<mode>): Likewise.
6735 (neon_vcge<mode>): Likewise.
6736 (neon_vcgeu<mode>): Likewise.
6737 (neon_vcgt<mode>): Likewise.
6738 (neon_vcgtu<mode>): Likewise.
6739 (neon_vcle<mode>): Likewise.
6740 (neon_vclt<mode>): Likewise.
6741 (neon_vcage<mode>): Likewise.
6742 (neon_vcagt<mode>): Likewise.
6743 (neon_vtst<mode>): Likewise.
6744 (neon_vabd<mode>): Likewise.
6745 (neon_vabdl<mode>): Likewise.
6746 (neon_vaba<mode>): Likewise.
6747 (neon_vabal<mode>): Likewise.
6748 (neon_vmax<mode>): Likewise.
6749 (neon_vmin<mode>): Likewise.
6750 (neon_vpaddl<mode>): Likewise.
6751 (neon_vpadal<mode>): Likewise.
6752 (neon_vpmax<mode>): Likewise.
6753 (neon_vpmin<mode>): Likewise.
6754 (neon_vrecps<mode>): Likewise.
6755 (neon_vrsqrts<mode>): Likewise.
6756 (neon_vqabs<mode>): Likewise.
6757 (neon_vqneg<mode>): Likewise.
6758 (neon_vcls<mode>): Likewise.
6759 (clz<mode>2): Likewise.
6760 (popcount<mode>2): Likewise.
6761 (neon_vrecpe): Likewise.
6762 (neon_vrsqrte): Likewise.
6763 (neon_vget_lane<mode>_sext_internal): Likewise.
6764 (neon_vget_lane<mode>_zext_internal): Likewise.
6765 (neon_vdup_n<mode>): Likewise.
6766 (neon_vdup_nv2di): Likewise.
6767 (neon_vdpu_lane<mode>_internal): Likewise.
6768 (neon_vswp<mode>): Likewise.
6769 (float<mode><V_cvtto>2): Likewise.
6770 (floatuns<mode><V_cvtto>2): Likewise.
6771 (fix_trunc<mode><V_cvtto>)2): Likewise
6772 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6773 (neon_vcvt<mode>): Likewise.
6774 (neon_vcvtv4sfv4hf): Likewise.
6775 (neon_vcvtv4hfv4sf): Likewise.
6776 (neon_vcvt_n<mode>): Likewise.
6777 (neon_vmovn<mode>): Likewise.
6778 (neon_vqmovn<mode>): Likewise.
6779 (neon_vqmovun<mode>): Likewise.
6780 (neon_vmovl<mode>): Likewise.
6781 (neon_vmul_lane<mode>): Likewise.
6782 (neon_vmull_lane<mode>): Likewise.
6783 (neon_vqdmull_lane<mode>): Likewise.
6784 (neon_vqdmulh_lane<mode>): Likewise.
6785 (neon_vmla_lane<mode>): Likewise.
6786 (neon_vmlal_lane<mode>): Likewise.
6787 (neon_vqdmlal_lane<mode>): Likewise.
6788 (neon_vmls_lane<mode>): Likewise.
6789 (neon_vmlsl_lane<mode>): Likewise.
6790 (neon_vqdmlsl_lane<mode>): Likewise.
6791 (neon_vext<mode>): Likewise.
6792 (neon_vrev64<mode>): Likewise.
6793 (neon_vrev32<mode>): Likewise.
6794 (neon_vrev16<mode>): Likewise.
6795 (neon_vbsl<mode>_internal): Likewise.
6796 (neon_vshl<mode>): Likewise.
6797 (neon_vqshl<mode>): Likewise.
6798 (neon_vshr_n<mode>): Likewise.
6799 (neon_vshrn_n<mode>): Likewise.
6800 (neon_vqshrn_n<mode>): Likewise.
6801 (neon_vqshrun_n<mode>): Likewise.
6802 (neon_vshl_n<mode>): Likewise.
6803 (neon_vqshl_n<mode>): Likewise.
6804 (neon_vqshlu_n<mode>): Likewise.
6805 (neon_vshll_n<mode>): Likewise.
6806 (neon_vsra_n<mode>): Likewise.
6807 (neon_vsri_n<mode>): Likewise.
6808 (neon_vsli_n<mode>): Likewise.
6809 (neon_vtbl1v8qi): Likewise.
6810 (neon_vtbl2v8qi): Likewise.
6811 (neon_vtbl3v8qi): Likewise.
6812 (neon_vtbl4v8qi): Likewise.
6813 (neon_vtbx1v8qi): Likewise.
6814 (neon_vtbx2v8qi): Likewise.
6815 (neon_vtbx3v8qi): Likewise.
6816 (neon_vtbx4v8qi): Likewise.
6817 (neon_vtrn<mode>_internal): Likewise.
6818 (neon_vzip<mode>_internal): Likewise.
6819 (neon_vuzp<mode>_internal): Likewise.
6820 (neon_vld1<mode>): Likewise.
6821 (neon_vld1_lane<mode>): Likewise.
6822 (neon_vld1_dup<mode>): Likewise.
6823 (neon_vld1_dupv2di): Likewise.
6824 (neon_vst1<mode>): Likewise.
6825 (neon_vst1_lane<mode>): Likewise.
6826 (neon_vld2<mode>): Likewise.
6827 (neon_vld2_lane<mode>): Likewise.
6828 (neon_vld2_dup<mode>): Likewise.
6829 (neon_vst2<mode>): Likewise.
6830 (neon_vst2_lane<mode>): Likewise.
6831 (neon_vld3<mode>): Likewise.
6832 (neon_vld3qa<mode>): Likewise.
6833 (neon_vld3qb<mode>): Likewise.
6834 (neon_vld3_lane<mode>): Likewise.
6835 (neon_vld3_dup<mode>): Likewise.
6836 (neon_vst3<mode>): Likewise.
6837 (neon_vst3qa<mode>): Likewise.
6838 (neon_vst3qb<mode>): Likewise.
6839 (neon_vst3_lane<mode>): Likewise.
6840 (neon_vld4<mode>): Likewise.
6841 (neon_vld4qa<mode>): Likewise.
6842 (neon_vld4qb<mode>): Likewise.
6843 (neon_vld4_lane<mode>): Likewise.
6844 (neon_vld4_dup<mode>): Likewise.
6845 (neon_vst4<mode>): Likewise.
6846 (neon_vst4qa<mode>): Likewise.
6847 (neon_vst4qb<mode>): Likewise.
6848 (neon_vst4_lane<mode>): Likewise.
6849 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6850 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6851 (neon_vec_<US>mult_lo_<mode>): Likewise.
6852 (neon_vec_<US>mult_hi_<mode>): Likewise.
6853 (neon_vec_<US>shiftl_<mode>): Likewise.
6854 (neon_unpack<US>_<mode>): Likewise.
6855 (neon_vec_<US>mult_<mode>): Likewise.
6856 (vec_pack_trunc_<mode>): Likewise.
6857 (neon_vec_pack_trunk_<mode>): Likewise.
6858 (neon_vabd<mode>_2): Likewise.
6859 (neon_vabd<mode>_3): Likewise.
6860 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6861 (thumb2_movsi_vfp): Likewise.
6862 (movdi_vfp): Likewise.
6863 (movdi_vfp_cortexa8): Likewise.
6864 (movhf_vfp_neon): Likewise.
6865 (movhf_vfp): Likewiwse.
6866 (movsf_vfp): Likewiwse.
6867 (thumb2_movsf_vfp): Likewiwse.
6868 (movdf_vfp): Likewise.
6869 (thumb2_movdf_vfp): Likewise.
6870 (movsfcc_vfp): Likewise.
6871 (thumb2_movsfcc_vfp): Likewise.
6872 (movdfcc_vfp): Likewise.
6873 (thumb2_movdfcc_vfp): Likewise.
6874 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6875 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6876 (v10_v2c): Likewise.
6877 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6879 (cortex_a15_neon_int_2): Likewise.
6880 (cortex_a15_neon_int_3): Likewise.
6881 (cortex_a15_neon_int_4): Likewise.
6882 (cortex_a15_neon_int_5): Likewise.
6883 (cortex_a15_neon_vqneg_vqabs): Likewise.
6884 (cortex_a15_neon_vmov): Likewise.
6885 (cortex_a15_neon_vaba): Likewise.
6886 (cortex_a15_neon_vaba_qqq): Likewise.
6887 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6888 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6889 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6890 scalar_64_32_long_scalar): Likewise.
6891 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6892 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6893 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6894 lotype_qdd_64_32_long): Likewise.
6895 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6896 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6897 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6898 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6899 (cortex_a15_neon_shift_1): Likewise.
6900 (cortex_a15_neon_shift_2): Likewise.
6901 (cortex_a15_neon_shift_3): Likewise.
6902 (cortex_a15_neon_vshl_ddd): Likewise.
6903 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6904 (cortex_a15_neon_vsra_vrsra): Likewise.
6905 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6906 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6907 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6908 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6909 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6910 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6911 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6912 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6913 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6914 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6915 (cortex_a15_neon_bp_simple): Likewise.
6916 (cortex_a15_neon_bp_2cycle): Likewise.
6917 (cortex_a15_neon_bp_3cycle): Likewise.
6918 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6919 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6920 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6921 (cortex_a15_neon_vld2_4_regs): Likewise.
6922 (cortex_a15_neon_vld3_vld4): Likewise.
6923 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6924 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6925 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6926 (cortex_a15_neon_vst3_vst4): Likewise.
6927 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6928 (cortex_a15_neon_vld3_vld4_lane" 10
6929 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6930 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6931 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6932 (cortex_a15_neon_ldm_2): Likewise.0
6933 (cortex_a15_neon_stm_2): Likewise.
6934 (cortex_a15_neon_mcr): Likewise.
6935 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6936 (cortex_a15_neon_mrc): Likewise.
6937 (cortex_a15_neon_mrrc): Likewise.
6938 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6940 (cortex_a15_alu_shift): Likewise.
6941 (cortex_a15_alu_shift_reg): Likewise.
6942 (cortex_a15_mult32): Likewise.
6943 (cortex_a15_mult64): Likewise.
6944 (cortex_a15_block): Likewise.
6945 (cortex_a15_branch): Likewise.
6946 (cortex_a15_load1): Likewise.
6947 (cortex_a15_load3): Likewise.
6948 (cortex_a15_store1): Likewise.
6949 (cortex_a15_store3): Likewise.
6950 (cortex_a15_call): Likewise.
6951 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6952 (cortex_a5_f2r): Likewise.
6953 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6955 (cortex_a53_f2r): Likewise.
6956 * config/arm/cortex-a7.md
6957 (cortex_a7_branch): Update for attribute change.
6958 (cortex_a7_call): Likewise.
6959 (cortex_a7_alu_imm): Likewise.
6960 (cortex_a7_alu_reg): Likewise.
6961 (cortex_a7_alu_shift): Likewise.
6962 (cortex_a7_mul): Likewise.
6963 (cortex_a7_load1): Likewise.
6964 (cortex_a7_store1): Likewise.
6965 (cortex_a7_load2): Likewise.
6966 (cortex_a7_store2): Likewise.
6967 (cortex_a7_load3): Likewise.
6968 (cortex_a7_store3): Likewise.
6969 (cortex_a7_load4): Likewise.
6970 (cortex_a7_store4): Likewise.
6971 (cortex_a7_fpalu): Likewise.
6972 (cortex_a7_fconst): Likewise.
6973 (cortex_a7_fpmuls): Likewise.
6974 (cortex_a7_neon_mul): Likewise.
6975 (cortex_a7_fpmacs): Likewise.
6976 (cortex_a7_neon_mla: Likewise.
6977 (cortex_a7_fpmuld: Likewise.
6978 (cortex_a7_fpmacd: Likewise.
6979 (cortex_a7_fpfmad: Likewise.
6980 (cortex_a7_fdivs: Likewise.
6981 (cortex_a7_fdivd: Likewise.
6982 (cortex_a7_r2f: Likewise.
6983 (cortex_a7_f2r: Likewise.
6984 (cortex_a7_f_flags: Likewise.
6985 (cortex_a7_f_loads: Likewise.
6986 (cortex_a7_f_loadd: Likewise.
6987 (cortex_a7_f_stores: Likewise.
6988 (cortex_a7_f_stored: Likewise.
6989 (cortex_a7_neon): Likewise.
6990 * config/arm/cortex-a8-neon.md
6991 (cortex_a8_neon_mrc): Update for attribute change.
6992 (cortex_a8_neon_mrrc): Likewise.
6993 (cortex_a8_neon_int_1): Likewise.
6994 (cortex_a8_neon_int_2): Likewise.
6995 (cortex_a8_neon_int_3): Likewise.
6996 (cortex_a8_neon_int_4): Likewise.
6997 (cortex_a8_neon_int_5): Likewise.
6998 (cortex_a8_neon_vqneg_vqabs): Likewise.
6999 (cortex_a8_neon_vmov): Likewise.
7000 (cortex_a8_neon_vaba): Likewise.
7001 (cortex_a8_neon_vaba_qqq): Likewise.
7002 (cortex_a8_neon_vsma): Likewise.
7003 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7004 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7005 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7006 long_scalar): Likewise.
7007 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7008 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7009 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7010 long_scalar_qdd_64_32_long): Likewise.
7011 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7012 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7013 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7014 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7015 (cortex_a8_neon_shift_1): Likewise.
7016 (cortex_a8_neon_shift_2): Likewise.
7017 (cortex_a8_neon_shift_3): Likewise.
7018 (cortex_a8_neon_vshl_ddd): Likewise.
7019 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7020 (cortex_a8_neon_vsra_vrsra): Likewise.
7021 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7022 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7023 (cortex_a8_neon_fp_vsum): Likewise.
7024 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7025 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7026 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7027 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7028 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7029 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7030 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7031 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7032 (cortex_a8_neon_bp_simple): Likewise.
7033 (cortex_a8_neon_bp_2cycle): Likewise.
7034 (cortex_a8_neon_bp_3cycle): Likewise.
7035 (cortex_a8_neon_ldr): Likewise.
7036 (cortex_a8_neon_str): Likewise.
7037 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7038 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7039 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7040 (cortex_a8_neon_vld2_4_regs): Likewise.
7041 (cortex_a8_neon_vld3_vld4): Likewise.
7042 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7043 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7044 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7045 (cortex_a8_neon_vst3_vst4): Likewise.
7046 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7047 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7048 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7049 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7050 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7051 (cortex_a8_neon_mcr): Likewise.
7052 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7053 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7054 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7056 (ca9_neon_mrrc): Likewise.
7057 (cortex_a9_neon_int_1): Likewise.
7058 (cortex_a9_neon_int_2): Likewise.
7059 (cortex_a9_neon_int_3): Likewise.
7060 (cortex_a9_neon_int_4): Likewise.
7061 (cortex_a9_neon_int_5): Likewise.
7062 (cortex_a9_neon_vqneg_vqabs): Likewise.
7063 (cortex_a9_neon_vmov): Likewise.
7064 (cortex_a9_neon_vaba): Likewise.
7065 (cortex_a9_neon_vaba_qqq): Likewise.
7066 (cortex_a9_neon_vsma): Likewise.
7067 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7068 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7069 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7070 long_scalar): Likewise.
7071 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7072 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7073 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7074 long_scalar_qdd_64_32_long): Likewise.
7075 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7076 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7077 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7078 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7079 (cortex_a9_neon_shift_1): Likewise.
7080 (cortex_a9_neon_shift_2): Likewise.
7081 (cortex_a9_neon_shift_3): Likewise.
7082 (cortex_a9_neon_vshl_ddd): Likewise.
7083 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7084 (cortex_a9_neon_vsra_vrsra): Likewise.
7085 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7086 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7087 (cortex_a9_neon_fp_vsum): Likewise.
7088 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7089 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7090 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7091 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7092 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7093 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7094 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7095 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7096 (cortex_a9_neon_bp_simple): Likewise.
7097 (cortex_a9_neon_bp_2cycle): Likewise.
7098 (cortex_a9_neon_bp_3cycle): Likewise.
7099 (cortex_a9_neon_ldr): Likewise.
7100 (cortex_a9_neon_str): Likewise.
7101 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7102 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7103 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7104 (cortex_a9_neon_vld2_4_regs): Likewise.
7105 (cortex_a9_neon_vld3_vld4): Likewise.
7106 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7107 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7108 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7109 (cortex_a9_neon_vst3_vst4): Likewise.
7110 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7111 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7112 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7113 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7114 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7115 (cortex_a9_neon_mcr): Likewise.
7116 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7117 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7118 (cortex_a9_fps): Likewise.
7119 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7121 (cortex_m4_fmuls): Likewise.
7122 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7124 (cortex_r4_mrc): Likewise.
7125 * config/arm/iterators.md: Update comment referring to neon_type.
7126 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7127 (iwmmxt_movsi_insn): Likewise.
7128 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7130 (pj4_core_to_vfp): Likewise.
7131 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7133 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7134 (vfp_fstore): Likewise.
7135 * doc/md.texi: Change references to neon_type to refer to type.
7137 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7139 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7141 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7143 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7146 2013-09-04 Easwaran Raman <eraman@google.com>
7149 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7150 (build_and_add_sum): Use it.
7151 (appears_later_in_bb): Simplify code.
7153 2013-09-04 Teresa Johnson <tejohnson@google.com>
7155 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7157 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7159 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7161 2013-09-04 Jan Hubicka <jh@suse.cz>
7163 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7164 * ipa-devirt.c: Include diganostic.h
7165 (odr_type_d): Add types and types_set.
7166 (hash_type_name): Work for types with vtables during LTO.
7167 (odr_hasher::remove): Fix comment; destroy types_set.
7168 (add_type_duplicate): New function,
7169 (get_odr_type): Use it.
7170 (dump_type_inheritance_graph): Dump type duplicates.
7171 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7173 * tree.c (types_same_for_odr): Give exact answers on types with
7176 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7178 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7179 explaining their differences.
7181 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7183 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7185 2013-09-03 Jeff Law <law@redhat.com>
7187 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7188 when not threading through a joiner block. Pass joiner/no joiner
7189 state to register_jump_thread.
7190 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7191 state from argument rather than implying on path length.
7192 Dump the entire jump thread path into debugging dump.
7193 * tree-flow.h (register_jump_thread): Update prototype.
7195 2013-08-29 Xinliang David Li <davidxl@google.com>
7197 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7198 Remove a trivial gcc_assert.
7200 2013-08-29 Xinliang David Li <davidxl@google.com>
7202 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7203 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7204 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7205 Delay base decl alignment adjustment.
7206 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7207 * tree-vectorizer.h: New data structure.
7208 (set_dr_misalignment): New function.
7209 (dr_misalignment): Ditto.
7210 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7211 (vectorizable_load): Ditto.
7212 (ensure_base_align): New function.
7213 (vectorize_loops): Add dbg_cnt support.
7214 (execute_vect_slp): Ditto.
7215 * dbgcnt.def: New debug counter.
7216 * Makefile: New dependency.
7218 2013-09-03 Meador Inge <meadori@codesourcery.com>
7222 2013-08-30 Meador Inge <meadori@codesourcery.com>
7224 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7226 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7228 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7229 function descriptor.
7231 2013-09-03 Richard Biener <rguenther@suse.de>
7233 * tree-affine.c (add_elt_to_tree): Fix association issue,
7234 avoid useless converts and make sure to always return a
7235 properly typed result.
7237 2013-09-03 Richard Biener <rguenther@suse.de>
7240 * fold-const.c (negate_expr_p): Fix division case.
7241 (negate_expr): Likewise.
7243 2013-09-03 Richard Biener <rguenther@suse.de>
7246 * tree-streamer-out.c: Include tm.h.
7247 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7249 2013-09-03 Jan Hubicka <jh@suse.cz>
7251 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7253 2013-09-03 Alan Modra <amodra@gmail.com>
7255 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7256 multilibs to big-endian target and vice versa.
7257 * config/rs6000/t-linux64: Use := assignment on all vars.
7258 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7259 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7260 * config/rs6000/t-linux64le: New file.
7261 * config/rs6000/t-linux64bele: New file.
7262 * config/rs6000/t-linux64lebe: New file.
7264 2013-09-02 Jan Hubicka <jh@suse.cz>
7266 * ipa-inline-transform.c (inline_transform): Do not
7267 optimize_inline_calls when not optimizing.
7269 2013-09-02 Jan Hubicka <jh@suse.cz>
7271 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7272 duplicated nodes for assembler names.
7273 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7274 hard registers from assembler name hash.
7276 2013-09-02 Jan Hubicka <jh@suse.cz>
7278 * ipa-split.c (execute_split_functions): Split externally visible
7279 functions called once.
7281 2013-09-02 Martin Jambor <mjambor@suse.cz>
7284 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7285 linked list. When finding the correct duplicate, also consider also
7286 the caller in additon to its inlined_to node.
7288 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7290 * config/aarch64/aarch64-simd-builtins.def
7291 (dup_lane_scalar): Remove.
7292 * config/aarch64/aarch64-simd.md
7293 (aarch64_simd_dup): Add 'w->w' alternative.
7294 (aarch64_dup_lane<mode>): Allow for VALL.
7295 (aarch64_dup_lane_scalar<mode>): Remove.
7296 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7297 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7298 (aarch64_get_lane_unsigned<mode>): Likewise.
7299 (aarch64_get_lane<mode>): Likewise.
7300 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7301 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7302 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7303 (VCON): Change container of V2SF.
7304 (vswap_width_name): Likewise.
7305 * config/aarch64/arm_neon.h
7306 (__aarch64_vdup_lane_any): New.
7307 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7308 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7309 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7311 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7314 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7315 the source or the destination is a single hard register.
7317 2013-09-02 Richard Biener <rguenther@suse.de>
7320 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7323 2013-09-02 Richard Biener <rguenther@suse.de>
7325 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7326 arithmetic to sizetype.
7328 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7330 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7331 Find auto-increment use both before and after candidate.
7333 2013-09-02 Marek Polacek <polacek@redhat.com>
7335 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7337 2013-09-01 Jan Hubicka <jh@suse.cz>
7339 * Makefile.in: Add ipa-profile.o
7340 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7341 * cgraph.c (struct cgraph_propagate_frequency_data,
7342 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7343 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7344 * cgraph.h (cgraph_propagate_frequency): Remove.
7345 * ipa-inline-analysis.c: Include ipa-utils.h;
7346 drop duplicated cfgloop.h.
7347 (inline_update_callee_summaries): Update.
7348 * ipa-profile.c: New file.
7349 * ipa-utils.h (ipa_propagate_frequency): Declare.
7350 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7351 data-streamer.h, value-prof.h.
7352 (symtab_remove_unreachable_nodes): Update profile.
7353 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7354 account_time_size, cmp_counts, dump_histogram,
7355 ipa_profile_generate_summary, ipa_profile_write_summary,
7356 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7357 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7358 Move to ipa-profile.c.
7360 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7362 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7364 2013-09-01 Jan Hubicka <jh@suse.cz>
7366 * common.opt (fdevirtualize-speculatively): New function.
7367 * invoke.texi (fdevirtualize-speculatively): Document.
7368 * ipa-devirt.c: Include ipa-inline.h
7369 (likely_target_p): New function.
7370 (ipa_devirt): New function.
7371 (gate_ipa_devirt): New function.
7372 (pass_data_ipa_devirt): New static var.
7373 (pass_ipa_devirt): Likewise.
7374 (make_pass_ipa_devirt): New function.
7375 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7376 (common_handle_option): Disable devirtualization when
7377 value range profiling is available.
7378 * passes.def (pass_ipa_devirt): Add.
7379 * timever.def (TV_IPA_DEVIRT): New timevar.
7380 * tree-pass.h (make_pass_ipa_devirt):
7382 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7384 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7385 include sanitize(undefined).
7387 2013-08-31 Diego Novillo <dnovillo@google.com>
7389 * Makefile.in (TREE_CORE_H): Define.
7391 (GTFILES): Add tree-core.h.
7392 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7394 * tree-core.h: New file.
7395 Move all data structures, enum, typedefs, global
7396 declarations and constants from ...
7399 2013-08-31 Jan Hubicka <jh@suse.cz>
7401 * bulitins.c (expand_builtin): Do not early exit for gcov
7402 instrumented functions.
7404 2013-08-31 Marek Polacek <polacek@redhat.com>
7406 * ubsan.c: Include tm_p.h.
7408 2013-08-31 Jan Hubicka <jh@suse.cz>
7410 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7413 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7414 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7415 * tree-cfg.c (verify_gimple_label): ... here.
7416 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7417 (ipa_merge_profiles): New function.
7418 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7419 (lto_input_function_body): Likewise.
7420 * ipa-utils.h (ipa_merge_profiles): Declare.
7421 * lto-streamer.h (lto_input_function_body): Update prototype.
7422 (emit_label_in_global_context_p): Remove.
7423 * lto-symtab.c: Include ipa-utils.h
7424 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7426 2013-08-31 Jan Hubicka <jh@suse.cz>
7428 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7430 2013-08-31 Jan Hubicka <jh@suse.cz>
7432 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7434 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7436 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7437 "cmp" RTX before signed_comparison_operator check to account
7440 2013-08-30 Jan Hubicka <jh@suse.cz>
7442 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7443 (detect_type_change_1): Rename to ...
7444 (detect_type_change): ... this one; early return on non-polymorphic
7446 (detect_type_change_ssa): Add comp_type parameter; update
7447 use of detect_type_change.
7448 (compute_complex_assign_jump_func): Add param_type parameter;
7449 update use of detect_type_change_ssa.
7450 (compute_complex_ancestor_jump_func): Likewise.
7451 (ipa_get_callee_param_type): New function.
7452 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7453 update calls to the jump function computation functions.
7455 2013-08-30 Teresa Johnson <tejohnson@google.com>
7456 Steven Bosscher <steven@gcc.gnu.org>
7458 * cfgrtl.c (fixup_new_cold_bb): New routine.
7459 (commit_edge_insertions): Invoke fixup_partitions.
7460 (find_partition_fixes): New routine.
7461 (fixup_partitions): Ditto.
7462 (verify_hot_cold_block_grouping): Update comments.
7463 (rtl_verify_edges): Invoke find_partition_fixes.
7464 (rtl_verify_bb_pointers): Update comments.
7465 (rtl_verify_bb_layout): Ditto.
7466 * basic-block.h (probably_never_executed_edge_p): Declare.
7467 (fixup_partitions): Ditto.
7468 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7469 * bb-reorder.c (sanitize_hot_paths): New function.
7470 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7472 * predict.c (probably_never_executed_edge_p): New routine.
7473 * cfg.c (check_bb_profile): Add partition insanity warnings.
7475 2013-08-30 Meador Inge <meadori@codesourcery.com>
7477 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7479 2013-08-30 Marek Polacek <polacek@redhat.com>
7481 * Makefile.in (ubsan.o): Add.
7482 (c-family/c-ubsan.o): Add.
7483 (builtins.o): Add ubsan.h dependency.
7484 * ubsan.h: New file.
7485 * ubsan.c: New file.
7486 * common.opt: Add -fsanitize=undefined option.
7487 (flag_sanitize): Add variable.
7488 (fsanitize=): Add option. Add Driver.
7489 (fsanitize=thread): Remove option.
7490 (fsanitize=address): Likewise.
7491 (static-libubsan): New option.
7492 * doc/invoke.texi: Document the new flag and -static-libubsan.
7493 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7494 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7495 * builtin-attrs.def (ATTR_COLD): Define.
7496 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7497 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7498 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7499 * flag-types.h (sanitize_code): New enum.
7500 * opts.c (common_handle_option): Parse command line arguments
7501 of -fsanitize=. Add -fsanitize=unreachable option.
7502 * varasm.c (get_variable_section): Adjust.
7503 (assemble_noswitch_variable): Likewise.
7504 (assemble_variable): Likewise.
7505 (output_constant_def_contents): Likewise.
7506 (categorize_decl_for_section): Likewise.
7507 (place_block_symbol): Likewise.
7508 (output_object_block): Likewise.
7509 * builtins.def: Likewise.
7510 * toplev.c (compile_file): Likewise.
7511 (process_options): Likewise.
7512 * cppbuiltin.c: Likewise.
7513 * tsan.c (tsan_pass): Likewise.
7514 (tsan_gate): Likewise.
7515 (tsan_gate_O0): Likewise.
7516 * cfgexpand.c (partition_stack_vars): Likewise.
7517 (expand_stack_vars): Likewise.
7518 (defer_stack_allocation): Likewise.
7519 (expand_used_vars): Likewise.
7520 * cfgcleanup.c (old_insns_match_p): Likewise.
7521 * asan.c (asan_finish_file): Likewise.
7522 (asan_instrument): Likewise.
7523 (gate_asan): Likewise.
7524 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7525 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7526 (asan_global_struct): Use pointer_sized_int_node instead
7527 calling build_nonstandard_integer_type.
7528 (initialize_sanitizer_builtins): Likewise.
7529 (asan_finish_file): Likewise.
7530 * gcc.c: Document %{%:function(args):X}.
7531 (static_spec_functions): Add sanitize.
7532 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7533 store funcval != NULL there.
7534 (do_spec_1): Adjust handle_spec_function caller.
7535 (handle_braces): Allow %:function(args) as condition.
7536 (sanitize_spec_function): New function.
7537 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7538 (LIBUBSAN_SPEC): Likewise.
7539 (LIBUBSAN_EARLY_SPEC): Likewise.
7540 (SANITIZER_SPEC): Handle libubsan.
7541 (SANITIZER_EARLY_SPEC): Likewise.
7542 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7543 instead of fsanitize=address.
7544 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7545 instead of fsanitize=address*.
7546 * builtins.c: Include ubsan.h.
7547 (fold_builtin_0): Instrument __builtin_unreachable.
7548 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7549 instead of flag_asan.
7550 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7551 (pointer_sized_int_node): Define.
7552 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7554 2013-08-30 Mike Stump <mikestump@comcast.net>
7556 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7559 2013-08-29 Jan Hubicka <jh@suse.cz>
7561 * cgraph.c (cgraph_function_body_availability): Handle weakref
7563 * passes.def: Remove pass_fixup_cfg.
7564 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7565 track when we need to remove functions.
7566 (gate_ipa_inline): Execute inlining always; add comment why.
7567 (pass_data_ipa_inline): Remove TODO_remove_functions.
7568 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7569 do not produce summaries.
7570 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7571 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7572 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7573 constant pool and vtable.
7575 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7577 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7578 New arm_neon.h's internal macros to specify 64-bit constants.
7579 Avoid using stdint.h's macros.
7581 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7583 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7584 before applying REG_P.
7586 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7588 PR tree-optimization/58277
7589 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7590 after seeing too many stmts with vdef in between dombb and current
7591 bb, invalidate everything.
7593 2013-08-30 Richard Biener <rguenther@suse.de>
7595 * fold-const.c (fold_single_bit_test): Fix overflow test.
7597 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7599 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7600 and which can live in a register, always retrieve the value on entry.
7601 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7602 passed by invisible reference specially.
7603 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7604 (vt_add_function_parameter): Correctly deal with a parameter passed by
7605 invisible reference.
7607 2013-08-30 Jan Hubicka <jh@suse.cz>
7609 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7611 2013-08-30 Richard Biener <rguenther@suse.de>
7613 PR tree-optimization/58228
7614 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7615 allow invariant loads in nested loop vectorization.
7617 2013-08-30 Richard Biener <rguenther@suse.de>
7619 PR tree-optimization/58223
7620 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7621 (has_anti_or_output_dependence): ... this and adjust to also
7622 look for output dependences.
7623 (mark_nodes_having_upstream_mem_writes): Adjust.
7624 (rdg_flag_uses): Likewise.
7626 2013-08-30 Richard Biener <rguenther@suse.de>
7628 PR tree-optimization/58010
7629 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7630 assert that we have a loop-closed PHI.
7632 2013-08-29 Jan Hubicka <jh@suse.cz>
7634 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7635 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7636 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7637 (lto_free_function_in_decl_state_for_node): New function.
7639 2013-08-29 Xinliang David Li <davidxl@google.com>
7641 * loop-unroll.c (report_unroll_peel): Minor message change.
7642 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7643 Emit alignment peeling message with default -fopt-info.
7644 (vect_loop_versioning): Emit loop version info message.
7645 * tree-vectorizer.c (vectorize_loops): Minor message change.
7646 (execute_vect_slp): Ditto.
7648 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7650 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7651 of the clone from the DECL_NAME of the original function.
7653 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7655 * passes.c (register_pass): Add overload.
7656 * tree-pass.h (register_pass): Forward declare it. Add comment.
7658 2013-08-29 Jan Hubicka <jh@suse.cz>
7660 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7661 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7662 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7663 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7664 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7665 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7666 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7667 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7669 2013-08-29 Teresa Johnson <tejohnson@google.com>
7671 * dumpfile.c (dump_loc): Output column number.
7672 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7673 * doc/invoke.texi: Document optall -fopt-info flag.
7674 * profile.c (read_profile_edge_counts): Use new dump framework.
7675 (compute_branch_probabilities): Ditto.
7676 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7677 when pass not in any opt group.
7678 * pass_manager.h (pass_manager::get_pass_profile): New method.
7679 * value-prof.c (check_counter): Use new dump framework.
7680 (check_ic_target): Ditto.
7681 * coverage.c (get_coverage_counts): Ditto.
7682 (coverage_init): Setup new dump framework.
7684 2013-08-29 Richard Biener <rguenther@suse.de>
7686 PR tree-optimization/58246
7687 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7688 handle the dominance check inside a basic-block.
7690 2013-08-29 Richard Biener <rguenther@suse.de>
7693 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7694 of default defs that appear in abnormal PHI nodes.
7696 2013-08-29 Richard Biener <rguenther@suse.de>
7698 PR tree-optimization/57685
7699 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7700 single-use operands to avoid exponential complexity.
7702 2013-08-28 Dehao Chen <dehao@google.com>
7704 * ipa-inline.c (edge_badness): Fix integer underflow.
7706 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7708 * gtm-builtins.def (_ITM_free): Declare leaf.
7710 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7713 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7714 (*tls_local_dynamic_base_64_largepic): Likewise.
7715 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7716 Remove predicate from call operand.
7717 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7718 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7720 2013-08-28 Jeff Law <law@redhat.com>
7722 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7723 checks for the number of predecessors and successors allowed.
7724 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7725 which require copying a joiner block if there is a request which
7726 is a subpath that requires no joiner block copying.
7728 2013-08-28 Jan Hubicka <jh@suse.cz>
7730 * lto-streamer-out.c (DFS_write_tree_body): Drop
7731 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7732 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7733 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7734 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7735 Do not read DECL_ERROR_ISSUED.
7736 (unpack_ts_decl_with_vis_value_fields): Do not read
7738 (lto_input_ts_binfo_tree_pointers): Do not read
7739 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7740 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7741 write DECL_ERROR_ISSUED..
7742 (pack_ts_decl_with_vis_value_fields): Do not write
7744 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7745 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7746 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7747 * tree.h (tree_decl_common): Update comment.
7748 (DECL_ERROR_ISSUED): Remove.
7750 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7753 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7755 2013-08-28 Jan Hubicka <jh@suse.cz>
7757 * builtins.def (free): Declare leaf.
7759 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7762 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7763 * configure: Regenerate.
7765 2013-08-27 Martin Jambor <mjambor@suse.cz>
7767 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7768 (ipa_ancestor_jf_data): Likewise.
7769 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7770 (ipa_get_jf_pass_through_type_preserved): New function.
7771 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7772 (ipa_get_jf_ancestor_type_preserved): New function.
7773 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7774 (ipa_get_jf_ancestor_result): Likewise.
7775 (propagate_vals_accross_pass_through): Use
7776 ipa_get_jf_pass_through_result to do all the value mappings.
7777 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7778 type_preserved flag.
7779 (ipa_set_jf_cst_copy): New function.
7780 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7781 (ipa_set_jf_arith_pass_through): Likewise.
7782 (ipa_set_ancestor_jf): Likewise.
7783 (compute_complex_assign_jump_func): Set type_preserved instead of
7785 (ipa_compute_jump_functions_for_edge): Likewise.
7786 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7787 (update_jump_functions_after_inlining): Update type_preserved.
7788 Explicitely create jump functions when combining one with pass_through.
7789 (ipa_write_jump_function): Stream the type_preserved flags.
7790 (ipa_read_jump_function): Likewise.
7792 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7793 Aldy Hernandez <aldyh@redhat.com>
7795 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7796 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7797 * function.h (struct function): Add has_force_vect_loops and
7799 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7800 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7802 * gimple.def: Update CLAUSES comments.
7803 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7804 (gimple_build_omp_for): Add argument to prototype.
7805 (gimple_omp_for_kind): New.
7806 (gimple_omp_for_set_kind): New.
7807 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7808 GOVD_DATA_SHARE_CLASS.
7809 (enum omp_region_type): Add ORT_SIMD.
7810 (gimple_add_tmp_var): Handle ORT_SIMD.
7811 (gimplify_var_or_parm_decl): Same.
7812 (is_gimple_stmt): Same.
7813 (omp_firstprivatize_variable): Same.
7814 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7815 (omp_notice_variable): Handle ORT_SIMD.
7816 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7817 (omp_check_private): Handle ORT_SIMD.
7818 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7820 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7821 Handle OMP_CLAUSE_LASTPRIVATE.
7822 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7824 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7825 (gimplify_expr): Handle OMP_SIMD.
7826 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7827 (expand_GOMP_SIMD_VF): New.
7828 (expand_GOMP_SIMD_LAST_LANE): New.
7829 * internal-fn.def (GOMP_SIMD_LANE): New.
7830 (GOMP_SIMD_VF): New.
7831 (GOMP_SIMD_LAST_LANE): New.
7832 * omp-low.c: Include target.h.
7833 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7835 (check_omp_nesting_restrictions): Same.
7837 (lower_rec_simd_input_clauses): New.
7838 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7840 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7841 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7842 (expand_omp_build_assign): New.
7843 (expand_omp_for_init_counts): New.
7844 (expand_omp_for_init_vars): New.
7845 (extract_omp_for_update_vars): New.
7846 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7847 and rewrite accordingly.
7848 (expand_omp_simd): New.
7849 (expand_omp_for): Use expand_omp_simd.
7850 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7851 (lower_omp_for): Do not lower the body.
7852 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7854 * tree-flow.h (find_omp_clause): Remove prototype.
7855 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7856 forcing vectorization of the loop, or if flag_tree_vectorize.
7857 (gate_tree_if_conversion): Similarly.
7858 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7859 gimple_build_omp_for.
7860 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7861 * tree-parloops (create_parallel_loop): Pass kind argument to
7862 gimple_build_omp_for.
7863 * tree-pretty-print.c (dump_omp_clause): Add cases for
7864 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7865 OMP_CLAUSE__SIMDUID_.
7866 (dump_generic_node): Handle OMP_SIMD.
7867 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7868 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7869 unroll OMP_SIMD loops here.
7870 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7871 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7873 (vect_analyze_data_refs): Handle simd loops.
7874 * tree-vect-loop.c (vectorizable_live_operation): Handle
7876 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7877 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7878 (vectorizable_load): Same.
7879 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7880 (struct simduid_to_vf): New.
7881 (simduid_to_vf::hash): New.
7882 (simduid_to-vf::equal): New.
7883 (struct simd_array_to_simduid): New.
7884 (simd_array_to_simduid::hash): New.
7885 (simd_array_to_simduid::equal): New.
7886 (adjust_simduid_builtins): New.
7887 (struct note_simd_array_uses_struct): New.
7888 (note_simd_array_uses_cb): New.
7889 (note_simd_array_uses): New.
7890 (vectorize_loops): Handle simd hints and adjust simd builtins
7892 * tree-vectorizer.h (struct _stmt_vec_info): Add
7893 simd_lane_access_p field.
7894 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7895 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7896 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7897 (omp_clause_code_name): Same.
7898 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7899 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7900 * tree.def (OMP_SIMD): New entry.
7901 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7902 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7903 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7904 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7905 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7906 (OMP_CLAUSE_LINEAR_STEP): New.
7907 (OMP_CLAUSE_SAFELEN_EXPR): New.
7908 (OMP_CLAUSE__SIMDUID__DECL): New.
7909 (find_omp_clause): New prototype.
7911 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7913 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7914 Haswell processor detection.
7916 2013-08-27 Christian Widmer <shadow@umbrox.de>
7919 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7920 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7921 AVX2 capable processors.
7923 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7925 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7926 of vget_low_* with implementations in terms of other intrinsics.
7928 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7931 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7932 values to -1, 0 and 1.
7934 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7936 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7937 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7938 (<optab><mode>3_insn): Remove.
7939 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7940 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7941 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7942 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7943 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7944 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7945 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7946 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7947 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7948 (ror<mode>3_insn): Likewise.
7949 * config/aarch64/predicates.md (aarch64_simd_register): New.
7951 2013-08-27 Richard Biener <rguenther@suse.de>
7953 PR tree-optimization/57521
7954 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7955 one edge is non-critical.
7956 (find_phi_replacement_condition): Make sure to use a non-critical
7957 edge. Cleanup and remove old bug workarounds.
7958 (bb_postdominates_preds): Remove.
7959 (if_convertible_loop_p_1): Do not compute post-dominators.
7960 (combine_blocks): Do not free post-dominators.
7961 (main_tree_if_conversion): Likewise.
7962 (pass_data_if_conversion): Add TODO_verify_ssa.
7964 2013-08-27 DJ Delorie <dj@redhat.com>
7966 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7968 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7970 * function.c (assign_parm_find_data_types): Set passed_mode and
7971 nominal_mode to the TYPE_MODE of nominal_type for the built
7972 pointer type in case of the struct-pass-by-reference.
7974 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7976 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7977 (UINT16_TYPE): Change default to "unsigned int".
7979 * config/avr/avr.opt (mfract-convert-truncate): New option.
7980 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7981 is set, round negative fractional integers according to n1169
7982 when converting to integer types.
7984 2013-08-26 Jan Hubicka <jh@suse.cz>
7986 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7987 methods can not be called indirectly when their address is not taken.
7989 2013-08-26 Jan Hubicka <jh@suse.cz>
7991 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7994 2013-08-26 Jan Hubicka <jh@suse.cz>
7996 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7999 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8001 * reload.c (find_valid_class): Allow classes that do not include
8002 FIRST_PSEUDO_REGISTER - 1.
8004 2013-08-26 Jan Hubicka <jh@suse.cz>
8006 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8007 fix edge count/frequency when speculation failed; fix type check
8008 for the direct call.
8010 2013-08-26 Jan Hubicka <jh@suse.cz>
8012 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8014 2013-08-26 Jan Hubicka <jh@suse.cz>
8016 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8017 to be changed by edge redirection.
8019 2013-08-26 Jan Hubicka <jh@suse.cz>
8021 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8022 formating; add sanity check.
8023 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8024 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8025 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8026 (resolve_noninline_speculation): Update callee keys, too.
8028 2013-08-26 Jan Hubicka <jh@suse.cz>
8030 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8031 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8033 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8035 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8038 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8040 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8041 (x86_64_elf_select_section): Ditto.
8042 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8044 (ix86_pass_by_reference): Ditto.
8045 (output_set_got): Ditto.
8046 (ix86_unary_operator_ok): Ditto.
8047 (ix86_expand_builtin): Ditto.
8049 2013-08-23 Jan Hubicka <jh@suse.cz>
8051 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8053 2013-08-23 Jan Hubicka <jh@suse.cz>
8055 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8056 (tree_decl_with_vis): Add FINAL field.
8058 2013-08-23 Jeff Law <law@redhat.com>
8060 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8061 the debugging dump when the expression is fully redundant.
8063 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8065 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8066 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8067 * pretty-print.c (pp_formatted_text_data): Likewise.
8068 (pp_write_text_to_stream): Likewise.
8069 (pp_write_text_as_dot_label_to_stream): Likewise.
8070 (pp_append_r): Likewise.
8071 (pp_format): Likewise.
8072 (pp_flush): Likewise.
8073 (pp_clear_output_area): Likewise.
8074 (pp_append_text): Likewise.
8075 (pp_formatted_text): Likewise.
8076 (pp_remaining_character_count_for_line): Likewise.
8077 (pp_newline): Likewise.
8078 (pp_character): Likewise.
8079 (output_buffer::~output_buffer): Define.
8080 (pretty_printer::~pretty_printer): Destruct output buffer.
8081 * pretty-print.h (output_buffer::~output_buffer): Declare.
8082 (pretty_printer::~pretty_printer): Declare virtual.
8084 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8087 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8088 HOST_WIDE_INT_M1U): New macros.
8089 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8090 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8091 unsigned -1 for lshift.
8092 * cse.c (cse_insn): Likewise.
8093 * double-int.c (rshift_double, lshift_double): Likewise.
8094 * builtins.c (fold_builtin_bitop): Likewise.
8095 * combine.c (force_to_mode): Likewise.
8096 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8097 * simplify-rtx.c (simplify_const_unary_operation,
8098 simplify_const_binary_operation): Likewise.
8099 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8100 check_va_list_escapes): Likewise.
8101 * rtlanal.c (nonzero_bits1): Likewise.
8102 * expmed.c (expand_smod_pow2): Likewise.
8103 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8105 2013-08-23 Jan Hubicka <jh@suse.cz>
8107 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8108 as having address taken.
8110 2013-08-23 Jan Hubicka <jh@suse.cz>
8112 * ipa-utils.h (method_class_type): Declare.
8113 * ipa-devirt.c (method_class_type): Export.
8115 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8116 do not walk base classes of anonymous types.
8118 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8120 PR rtl-optimization/58220
8122 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8123 handle SEQUENCE insns properly.
8125 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8127 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8129 (pp_newline_and_indent): Likewise.
8130 (pp_separate_with): Likewise.
8131 * pretty-print.c (pp_newline_and_flush): Define.
8132 (pp_newline_and_indent): Likewise.
8133 (pp_separate_with): Likewise.
8135 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8138 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8139 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8141 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8143 * config/i386/predicates.md (ext_sse_reg_operand): New.
8144 * config/i386/i386.md (*movti_internal): Use
8145 predicate to determine if EVEX is needed.
8146 (*movsi_internal): Ditto.
8147 (*movdf_internal): Ditto.
8148 (*movsf_internal): Ditto.
8149 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8151 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8153 PR tree-optimization/58209
8154 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8155 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8156 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8157 emit POINTER_PLUS_EXPR.
8158 (create_tailcall_accumulator): For pointer result type accumulate in
8161 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8163 * configure.ac: Add backslashes missing from the last change.
8164 * configure: Regenerate.
8166 2013-08-22 Jan Hubicka <jh@susue.cz>
8168 * ipa.c (function_and_variable_visibility): First remember function
8169 was global and then make it local.
8171 2013-08-22 Julian Brown <julian@codesourcery.com>
8173 * configure.ac: Add aarch64 to list of arches which use "nop" in
8175 * configure: Regenerate.
8177 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8179 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8180 gnu_libc_has_function.
8181 * config/s390/tpf.h: Likewise.
8183 2013-08-22 Jan Hubicka <jh@susue.cz>
8185 * timevar.c (validate_phases): Add cast.
8187 2013-08-22 Jan Hubicka <jh@susue.cz>
8189 * timevar.c (validate_phases): Use size_t for memory.
8190 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8192 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8194 * pretty-print.h (output_buffer::output_buffer): Declare.
8195 (pretty_printer::pretty_printer): Likewise.
8196 (pp_construct): Remove.
8197 * pretty-print.c (output_buffer::output_buffer): Define.
8198 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8199 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8200 (print_gimple_expr): Likewise.
8201 (print_gimple_seq): Likewise.
8202 (gimple_dump_bb): Likewise.
8203 * sched-vis.c (dump_value_slim): Likewise.
8204 (dump_insn_slim): Likewise.
8205 (dump_rtl_slim): Likewise.
8206 (str_pattern_slim): Likewise.
8207 * tree-mudflap.c (mf_varname_tree): Likewise.
8208 * graph.c (print_graph_cfg): Likewise.
8209 (start_graph_dump): Likewise.
8210 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8212 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8213 pretty printer initialization.
8214 * coretypes.h (diagnostic_context): Remove superflous type alias
8216 (pretty_printer): Likewise. Declare directly as a class.
8217 (pretty_print_info): Remove declaration as class.
8218 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8219 and pp_clear_output_area.
8220 (asan_add_global): Likewise.
8222 2013-08-22 Jan Hubicka <jh@suse.cz>
8224 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8225 * ipa-utils.h (update_type_inheritance_graph): Declare.
8226 (possible_polymorphic_call_target_p): Declare.
8227 (possible_polymorphic_call_target_p): New.
8228 * ipa-devirt.c: Update toplevel comments.
8229 (cached_polymorphic_call_targets): Move up.
8230 (odr_type_d): Move ID down.
8231 (polymorphic_type_binfo_p): Update comment.
8232 (odr_hasher::remove): Likewise;
8233 (get_odr_type): Set anonymous_namespace.
8234 (dump_odr_type): Dump it.
8235 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8236 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8237 (record_binfo): Add comment.
8238 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8240 (devirt_node_removal_hook): Do not iCE when cache is freed.
8241 (possible_polymorphic_call_target_p): New predicate.
8242 (update_type_inheritance_graph): New function.
8244 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8245 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8246 Sergey Lega <sergey.s.lega@intel.com>
8247 Anna Tikhonova <anna.tikhonova@intel.com>
8248 Ilya Tocar <ilya.tocar@intel.com>
8249 Andrey Turetskiy <andrey.turetskiy@intel.com>
8250 Ilya Verbin <ilya.verbin@intel.com>
8251 Kirill Yukhin <kirill.yukhin@intel.com>
8252 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8254 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8255 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8256 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8257 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8258 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8259 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8260 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8261 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8262 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8263 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8264 OPT_mavx512pf, OPT_mavx512er cases.
8265 * config/i386/constraints.md (v): New constraint.
8266 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8267 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8268 (bit_AVX512CD): New.
8269 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8270 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8271 * config/i386/i386-c.c (ix86_target_macros_internal):
8272 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8274 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8275 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8276 * config/i386/i386.c (regclass_map, dbx_register_map)
8277 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8278 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8279 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8281 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8282 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8283 -mavx512pf options. Fix formatting.
8284 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8285 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8286 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8287 -mavx512cd, -mavx512pf options.
8288 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8289 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8290 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8292 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8293 (ix86_set_reg_reg_cost): Ditto.
8294 (x86_order_regs_for_local_alloc): Ditto.
8295 (MAX_VECT_LEN): Extend to 64-byte.
8296 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8297 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8298 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8299 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8300 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8301 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8302 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8303 (SSE_REG_MODE_P): Support new modes.
8304 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8305 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8306 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8307 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8308 (REG_CLASS_CONTENTS): Add new registers.
8309 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8310 (EXT_REX_SSE_REGNO_P): New.
8311 (HI_REGISTER_NAMES): Add new registers.
8312 * config/i386/i386.md: Define constants for new registers.
8313 (mode): Add new 512-bit modes.
8314 (prefix): Support evex prefix.
8315 (isa): Support avx512f, noavx512f, fma_avx512f.
8316 (ssemodesuffix): Add new 512-bit modes.
8318 (*movxi_internal_avx512f): Ditto.
8319 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8321 (*movsi_internal): Likewise.
8322 (*movdf_internal): Likewise.
8323 (*movsf_internal): Likewise.
8324 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8325 (<code><mode>3): Likewise.
8326 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8328 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8329 with the new constraint "v".
8330 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8332 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8333 with the new constraint "v".
8334 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8335 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8336 (*<plusminus_insn><mode>3): Likewise.
8337 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8338 (*mul<mode>3): Likewise.
8339 (<sse>_vmmul<mode>3): Likewise.
8340 (<sse>_div<mode>3): Likewise.
8341 (<sse>_vmdiv<mode>3): Likewise.
8342 (<sse>_sqrt<mode>2): Likewise.
8343 (<sse>_vmsqrt<mode>2): Likewise.
8344 (*<code><mode>3_finite): Likewise.
8345 (*<code><mode>3) <smaxmin>: Likewise.
8346 (<sse>_vm<code><mode>3): Likewise.
8347 (*<code><mode>3) <any_logic>: Likewise.
8348 (*fma_fmadd_<mode>): Likewise.
8349 (*fma_fmsub_<mode>): Likewise.
8350 (*fma_fnmadd_<mode>): Likewise.
8351 (*fma_fnmsub_<mode>): Likewise.
8352 (*fma_fmaddsub_<mode>): Likewise.
8353 (*fma_fmsubadd_<mode>): Likewise.
8354 (*fmai_fmadd_<mode>): Likewise.
8355 (*fmai_fmsub_<mode>): Likewise.
8356 (*fmai_fnmadd_<mode>): Likewise.
8357 (*fmai_fnmsub_<mode>): Likewise.
8358 (sse_cvtsi2ss): Likewise.
8359 (sse_cvtsi2ssq): Likewise.
8360 (sse_cvtss2si): Likewise.
8361 (sse_cvtss2si_2): Likewise.
8362 (sse_cvtss2siq): Likewise.
8363 (sse_cvtss2siq_2): Likewise.
8364 (sse_cvttss2si): Likewise.
8365 (sse_cvtss2siq_2): Likewise.
8366 (float<sseintvecmodelower><mode>2): Likewise.
8367 (sse2_cvtsd2si_2): Likewise.
8368 (sse2_cvtsd2siq_2): Likewise.
8369 (*<plusminus_insn><mode>3): Likewise.
8370 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8371 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8372 (ashr<mode>3): Likewise.
8373 (<shift_insn><mode>3): Likewise.
8374 (avx2_<code><mode>3): Likewise.
8375 (*avx2_<code><mode>3): Likewise.
8376 (*andnot<mode>3): Likewise.
8377 (*<code><mode>3) <any_logic>: Likewise.
8378 (abs<mode>2): Likewise.
8379 (avx2_permvar<mode>): Likewise.
8380 (avx2_perm<mode>_1): Likewise.
8381 (*avx_vpermilp<mode>): Likewise.
8382 (avx_vpermilvar<mode>3): Likewise.
8383 (avx2_ashrv<mode>): Likewise.
8384 (avx2_<shift_insn>v<mode>): Likewise.
8385 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8387 * doc/rtl.texi: Document XImode.
8389 2013-08-21 Jeff Law <law@redhat.com>
8391 * tree-flow.h (register_jump_thread): Pass vector of edges
8392 instead of each important edge.
8393 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8394 thread path into a vector and pass that to register_jump_thread.
8395 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8396 passed in edge vector to the current 3-edge form.
8399 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8401 * dce.c (fini_dce): Call df_analyze again just in case
8402 delete_unmarked_insns removed anything.
8404 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8406 * reload.h (struct reg_equivs): Rename to ..
8407 (struct reg_equivs_s): .. this.
8409 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8411 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8413 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8415 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8417 2013-08-21 Jeff Law <law@redhat.com>
8419 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8420 simplify assignments too. If the RHS collapses to a singleton
8421 range, then return the value for the range.
8423 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8425 * config/i386/sse.md (V16): Rename to...
8427 (mov<mode>): Update iterator name.
8428 (*mov<mode>_internal): Ditto.
8429 (push<mode>1): Ditto.
8430 (movmisalign<mode>): Ditto.
8432 2013-08-20 Jan Hubicka <jh@suse.cz>
8435 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8436 entry for direct edges.
8437 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8439 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8441 Revert my last two changes, r201865 and r201864:
8444 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8446 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8447 instances can own GC refs.
8449 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8450 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8451 (gcc::context::gt_pch_nx): Likewise.
8452 (gcc::context::gt_pch_nx): Likewise.
8453 * ggc.h (gt_ggc_mx <T>): New.
8454 (gt_pch_nx_with_op <T>): New.
8455 (gt_pch_nx <T>): New.
8456 * passes.c (opt_pass::gt_ggc_mx): New.
8457 (opt_pass::gt_pch_nx): New.
8458 (opt_pass::gt_pch_nx_with_op): New.
8459 (pass_manager::gt_ggc_mx): New.
8460 (pass_manager::gt_pch_nx): New.
8461 (pass_manager::gt_pch_nx_with_op): New.
8462 (pass_manager::operator new): Use
8463 ggc_internal_cleared_alloc_stat rather than xcalloc.
8464 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8465 (pass_manager::gt_ggc_mx): New.
8466 (pass_manager::gt_pch_nx): New.
8467 (pass_manager::gt_pch_nx_with_op): New.
8468 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8469 (opt_pass::operator new): New.
8470 (opt_pass::gt_ggc_mx): New.
8471 (opt_pass::gt_pch_nx): New.
8472 (opt_pass::gt_pch_nx_with_op): New.
8475 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8477 * Makefile.in (GTFILES): Add context.h.
8478 * context.c (gcc::context::operator new): New.
8479 (gcc::context::gt_ggc_mx): New.
8480 (gcc::context::gt_pch_nx): New.
8481 (gcc::context::gt_pch_nx): New.
8482 * context.h (gcc::context): Add GTY((user)) marking.
8483 (gcc::context::operator new): New.
8484 (gcc::context::gt_ggc_mx): New.
8485 (gcc::context::gt_pch_nx): New.
8486 (gcc::context::gt_pch_nx): New.
8487 (g): Add GTY marking.
8488 (gt_ggc_mx (gcc::context *)): New.
8489 (gt_pch_nx (gcc::context *)): New.
8490 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8491 void *cookie)): New.
8492 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8494 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8496 * dce.c (fini_dce): Call df_analyze again just in case
8497 delete_unmarked_insns removed anything.
8499 2013-08-20 Teresa Johnson <tejohnson@google.com>
8501 PR rtl-optimizations/57451
8502 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8503 from crossing split section boundaries.
8505 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8507 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8508 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8509 with MULTLIB_DEFAULTS.
8511 2013-08-20 Nick Clifton <nickc@redhat.com>
8513 * target.def (narrow_volatile_bitfield): Note that the default
8514 value is false, not !TARGET_STRICT_ALIGN.
8515 * doc/tm.texi: Regenerate.
8517 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8519 Fix LIB_SPEC for systems without libpthread.
8521 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8522 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8524 * config/i386/linux-common.h: Likewise.
8525 * config/mips/linux-common.h: Likewise.
8527 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8529 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8532 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8534 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8535 instances can own GC refs.
8537 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8538 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8539 (gcc::context::gt_pch_nx): Likewise.
8540 (gcc::context::gt_pch_nx): Likewise.
8541 * ggc.h (gt_ggc_mx <T>): New.
8542 (gt_pch_nx_with_op <T>): New.
8543 (gt_pch_nx <T>): New.
8544 * passes.c (opt_pass::gt_ggc_mx): New.
8545 (opt_pass::gt_pch_nx): New.
8546 (opt_pass::gt_pch_nx_with_op): New.
8547 (pass_manager::gt_ggc_mx): New.
8548 (pass_manager::gt_pch_nx): New.
8549 (pass_manager::gt_pch_nx_with_op): New.
8550 (pass_manager::operator new): Use
8551 ggc_internal_cleared_alloc_stat rather than xcalloc.
8552 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8553 (pass_manager::gt_ggc_mx): New.
8554 (pass_manager::gt_pch_nx): New.
8555 (pass_manager::gt_pch_nx_with_op): New.
8556 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8557 (opt_pass::operator new): New.
8558 (opt_pass::gt_ggc_mx): New.
8559 (opt_pass::gt_pch_nx): New.
8560 (opt_pass::gt_pch_nx_with_op): New.
8562 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8564 * Makefile.in (GTFILES): Add context.h.
8565 * context.c (gcc::context::operator new): New.
8566 (gcc::context::gt_ggc_mx): New.
8567 (gcc::context::gt_pch_nx): New.
8568 (gcc::context::gt_pch_nx): New.
8569 * context.h (gcc::context): Add GTY((user)) marking.
8570 (gcc::context::operator new): New.
8571 (gcc::context::gt_ggc_mx): New.
8572 (gcc::context::gt_pch_nx): New.
8573 (gcc::context::gt_pch_nx): New.
8574 (g): Add GTY marking.
8575 (gt_ggc_mx (gcc::context *)): New.
8576 (gt_pch_nx (gcc::context *)): New.
8577 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8578 void *cookie)): New.
8579 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8581 2013-08-20 Alan Modra <amodra@gmail.com>
8584 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8585 (rs6000_emit_epilogue): Likewise.
8587 2013-08-19 Dehao Chen <dehao@google.com>
8589 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8591 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8592 Jakub Jelinek <jakub@redhat.com>
8594 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8595 (BUILT_IN_FABSD64): Likewise.
8596 (BUILT_IN_FABSD128): Likewise.
8597 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8598 (fold_builtin_1): Likewise.
8599 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8600 destination and source operands.
8601 (*abstd2_fpr): Likewise.
8602 (*nabstd2_fpr): Likewise.
8604 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8606 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8609 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8611 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8614 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8616 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8617 * builtins.c (default_libc_has_function): New.
8618 (gnu_libc_has_function): Ditto.
8619 (no_c99_libc_has_function): Ditto.
8620 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8621 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8622 (fold_builtin_sincos): Likewise.
8623 (fold_builtin_cexp): Likewise.
8624 * builtins.def (DEF_C94_BUILTIN): Likewise.
8625 (DEF_C99_BUILTIN): Likewise.
8626 (DEF_C99_C90RES_BUILTIN): Likewise.
8627 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8628 definitions to using this define.
8629 * config/darwin-protos.h (darwin_libc_has_function): New.
8630 * config/darwin.c (darwin_libc_has_function): Ditto.
8631 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8632 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8633 * config/darwin.h: Ditto.
8634 * config/elfos.h: Ditto.
8635 * config/freebsd.h: Ditto.
8636 * config/i386/cygming.h: Ditto.
8637 * config/i386/djgpp.h: Ditto.
8638 * config/i386/i386-interix.h: Ditto.
8639 * config/microblaze/microblaze.h: Ditto.
8640 * config/mmix/mmix.h: Ditto.
8641 * config/gnu-user.h: Ditto.
8642 * config/ia64/hpux.h: Ditto.
8643 * config/pa/pa-hpux.h: Ditto.
8644 * config/pdp11/pdp11.h: Ditto.
8645 * config/picochip/picochip.h: Ditto.
8646 * config/linux.h: Ditto.
8647 * config/netbsd.h: Ditto.
8648 * config/openbsd.h: Ditto.
8649 * config/rs6000/aix43.h: Ditto.
8650 * config/rs6000/aix51.h: Ditto.
8651 * config/rs6000/aix52.h: Ditto.
8652 * config/rs6000/aix53.h: Ditto.
8653 * config/rs6000/aix61.h: Ditto.
8654 * config/rs6000/darwin.h: Ditto.
8655 * config/rs6000/linux.h: Ditto.
8656 * config/rs6000/linux64.h: Ditto.
8657 * config/s390/tpf.h: Ditto.
8658 * config/sol2-10.h: Ditto.
8659 * config/sol2.h: Ditto.
8660 * config/vms/vms.h: Ditto.
8661 * config/vxworks.h: Ditto.
8662 * config/linux-android.c (linux_android_libc_has_function):
8663 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8664 * config/linux-protos.h (linux_android_libc_has_function):
8666 * config/i386/i386.c (ix86_libc_has_function): New.
8667 * config/i386/i386-protos.h
8668 (ix86_libc_has_function): New declaration.
8669 * config/i386/i386.md
8670 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8671 ("isinf<mode>2): Likewise.
8672 * convert.c (convert_to_integer): Using new target hook
8673 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8674 TARGET_C99_FUNCTIONS.
8675 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8676 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8677 * coretypes.h (function_class): New enum for different
8678 classes of functions.
8679 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8680 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8681 (TARGET_HAS_SINCOS): Likewise.
8682 (TARGET_LIBC_HAS_FUNCTION): New.
8683 * doc/tm.texi: Regenerated.
8684 * targhooks.h (default_libc_has_function): New declaration.
8685 (no_c99_libc_has_function): Ditto.
8686 (gnu_libc_has_function): Ditto.
8687 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8688 and TARGET_HAS_SINCOS.
8690 2013-08-18 Jan Hubicka <jh@suse.cz>
8692 * Makeifle-in (ipa-devirt.o): New.
8693 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8694 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8695 (analyze_functions): Look into possible targets of polymorphic call.
8696 * dumpfile.c (dump_files): Add type-inheritance dump.
8697 * dumpfile.h (TDI_inheritance): New.
8698 * ipa-devirt.c: New file.
8699 * ipa-utils.h (odr_type_d): Forward declare.
8700 (odr_type): New type.
8701 (build_type_inheritance_graph): Declare.
8702 (possible_polymorphic_call_targets): Declare and introduce inline
8703 variant when only edge is pased.
8704 (dump_possible_polymorphic_call_targets): Likewise.
8705 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8706 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8707 (types_same_for_odr): ... here.
8708 * tree.h (type_in_anonymous_namespace_p): Declare.
8710 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8712 PR tree-optimization/58006
8713 * tree-parloops.c (take_address_of): Don't ICE if get_name
8715 (eliminate_local_variables_stmt): Remove clobber stmts.
8717 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8719 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8720 error message is issued for an alias to undefined symbol.
8722 2013-08-18 Jan Hubicka <jh@suse.cz>
8724 * cgraph.c (cgraph_create_indirect_edge): Discover
8725 polymorphic calls and record basic info into indirect_info.
8726 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8727 devirtualization, ignore objc function calls.
8728 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8729 call with no parm index info.
8730 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8731 * tree.c (virtual_method_call_p): New function.
8732 * tree.h (virtual_method_call_p): Declare.
8734 2013-08-16 Jan Hubicka <jh@suse.cz>
8737 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8739 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8741 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8744 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8746 * gengtype.c (type_for_name): Add special-case support for
8747 locating types within the "gcc::" namespace.
8748 (open_base_files): Emit a "using namespace gcc" directive.
8750 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8753 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8754 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8756 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8757 array instead of each individual operand as a separate argument.
8758 (emit_fusion_gpr_load): Likewise.
8759 (expand_fusion_gpr_load): Add new function declaration.
8761 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8762 signature to have the operands passed as an array, instead of as
8763 separate arguments. Allow ZERO_EXTEND to be in the memory
8764 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8765 depend on the register live/dead flags when peepholes are run.
8766 (expand_fusion_gpr_load): New function to be called from the
8767 peephole2 pass, to change the register that addis sets to be the
8769 (emit_fusion_gpr_load): Change the calling signature to have the
8770 operands passed as an array, instead of as separate arguments.
8771 Allow ZERO_EXTEND to be in the memory address, and also
8772 SIGN_EXTEND if -mpower8-fusion-sign.
8774 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8776 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8777 adjust the register addis loads up in the peephole2 pass. Do not
8778 depend on the register live/dead state when the peephole pass is done.
8780 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8782 * gengtype.c (create_user_defined_type): Ensure that the kind
8783 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8784 declaration is seen before the GTY((user)) marking.
8786 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8789 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8791 2013-08-16 Jan Hubicka <jh@suse.cz>
8793 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8794 arugment expected_type.
8795 (gimple_fold_call): Use it.
8796 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8797 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8798 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8799 (try_make_edge_direct_virtual_call): Likewise.
8800 * tree.c (obj_type_ref_class): New.
8801 * tree.h (obj_type_ref_class): Use it.
8803 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8805 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8806 (rtl_slim_pp): Likewise.
8807 (init_rtl_slim_pretty_print): Likewise.
8808 (dump_value_slim): Don't call it. Use local pretty printer.
8809 (dump_insn_slim): Likewise.
8810 (dump_rtl_slim): Likewise.
8811 (str_pattern_slim): Likewise.
8812 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8815 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8817 PR tree-optimization/58164
8818 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8819 walk gimple_goto_dest of GIMPLE_GOTO.
8821 PR tree-optimization/58165
8822 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8823 bi_call must be the last stmt in a bb, don't split_block, instead
8824 use fallthru edge from it and give up if there is none.
8825 Release conds vector when returning early.
8827 2013-08-14 Xinliang David Li <davidxl@google.com>
8829 * config/i386/i386.c (ix86_option_override_internal):
8830 Remove unused variable and field.
8832 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8835 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8836 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8837 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8838 and Linux, correct BLKmode alignment when 128-bit alignment is
8839 required and compatibility flag is not set.
8840 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8841 for zero-size arguments when compatibility flag is not set.
8843 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8845 PR tree-optimization/58145
8846 * tree-sra.c (build_ref_for_offset): If prev_base has
8847 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8849 2013-08-14 Xinliang David Li <davidxl@google.com>
8851 * config/i386/i386.c (ix86_option_override_internal):
8852 Fix uninitialized variable error.
8854 2013-08-14 Xinliang David Li <davidxl@google.com>
8856 * config/i386/i386.opt: Define two new options.
8857 * config/i386/x86-tune.def: Add arch selector field in macros.
8858 * config/i386/i386.h: Adjust macro definition.
8859 * config/i386/i386.c (ix86_option_override_internal):
8861 (parse_mtune_ctrl_str): New function.
8862 (set_ix86_tune_features): New function.
8863 (ix86_function_specific_restore): Call the new helper function.
8865 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8867 PR rtl-optimization/57662
8868 * sel-sched.c (code_motion_process_successors): When the current insn
8869 is removed after the recursive traversal, break from the loop.
8870 Add comments and debug printouts.
8872 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8873 Alexandre Oliva <aoliva@redhat.com>
8876 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8877 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8878 there also UNSPEC_PLTOFF.
8880 2013-08-14 Marek Polacek <polacek@redhat.com>
8882 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8883 NUM_CONDITIONS bit positions.
8885 2013-08-13 Cary Coutant <ccoutant@google.com>
8887 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8888 (attr_checksum): Hash vector contents instead of pointer.
8889 (attr_checksum_ordered): Likewise.
8891 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8893 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8894 when Pmode != word_mode. Add length_address attribute.
8895 (sse3_monitor_<mode>): Merge from sse3_monitor and
8896 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8897 Pmode != word_mode. Update insn length attribute.
8898 * config/i386/i386.c (ix86_option_override_internal): Update
8899 ix86_gen_monitor selection for merged sse3_monitor insn.
8901 2013-08-13 Julian Brown <julian@codesourcery.com>
8903 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8904 perform invalid legitimization on greater-than-word-size modes for
8907 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8909 * ira.c (setup_class_translate_array): Use aclass instead of cl
8910 for classes not fully covered by allocno classes.
8912 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8914 PR tree-optimization/57661
8915 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8916 * tree-inline.c (tree_function_versioning): Initialize it.
8917 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8918 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8919 that is not being copied.
8922 * asan.c (instrument_strlen_call): Fix typo in comment.
8923 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8925 2013-08-13 Steve Ellcey <sellcey@mips.com>
8927 * config/mips/mips.md (prefetch): Use lw instead of ld on
8928 loongson in 32bit mode.
8930 2013-08-13 Nick Clifton <nickc@redhat.com>
8932 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8934 2013-08-13 Jan Hubicka <jh@suse.cz>
8936 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8937 introduced edge; fix typo in sanity check.
8938 (cgraph_resolve_speculation): Export; improve diagnostic.
8939 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8940 speculation at type mismatch.
8941 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8942 (cgraph_resolve_speculation): Declare.
8943 (symtab_can_be_discarded): New function.
8944 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8945 * ipa-inline-transform.c (speculation_removed): New global var.
8946 (clone_inlined_nodes): See if speculation can be removed.
8947 (inline_call): If speculations was removed, we growths may not match.
8948 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8949 (speculation_useful_p): New function.
8950 (resolve_noninline_speculation): New function.
8951 (inline_small_functions): Resolve useless speculations.
8952 * ipa-inline.h (speculation_useful_p): Declare
8953 * ipa.c (can_replace_by_local_alias): Simplify.
8954 (ipa_profile): Produce speculative calls in non-lto, too;
8955 add simple cost model; produce local aliases.
8957 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8959 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8962 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8964 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8966 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8968 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8969 iterator instead of X87MODEF.
8971 2013-08-12 Perez Read <netfirewall@gmail.com>
8974 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8975 operand 0 for intel asm alternative.
8976 (*movabs<mode>_2): Ditto for operand 1.
8978 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8980 * config/aarch64/arm_none.h
8981 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8983 2013-08-12 Nick Clifton <nickc@redhat.com>
8985 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8987 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8989 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8990 expand for QI/HImode operand to produce more effictive code for
8991 unsigned char(short) --> float(double) conversion.
8993 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8995 * doc/invoke.texi: Mention that -ftls-model does not force the final
8998 2013-08-12 Marek Polacek <polacek@redhat.com>
8999 Marc Glisse <marc.glisse@inria.fr>
9001 PR tree-optimization/57980
9002 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9003 when creating -1 constant.
9005 2013-08-10 Jan Hubicka <jh@suse.cz>
9007 Workaround binutils PR14342.
9008 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9009 (gimple_init_edge_profiler): Likewise.
9010 (gimple_gen_ic_func_profiler): Likewise.
9012 2013-08-09 Jan Hubicka <jh@suse.cz>
9014 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9016 2013-08-09 Xinliang David Li <davidxl@google.com>
9018 * config/i386/stringop.def: New file.
9019 * config/i386/stringop.opt: New file.
9020 * config/i386/i386-opts.h: Include stringopt.def.
9021 * config/i386/i386.opt: Include stringopt.opt.
9022 * config/i386/i386.c (ix86_option_override_internal):
9023 Override default size based stringop inline strategies with options.
9024 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9027 2013-08-09 Jan Hubicka <jh@suse.cz>
9029 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9031 2013-08-09 Jan Hubicka <jh@suse.cz>
9033 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9035 (dump_cgraph_node): Dump profile-id.
9036 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9037 and common_target_probability.
9038 * lto-cgraph.c (lto_output_edge): Stream common targets.
9039 (lto_output_node): Stream profile ids.
9040 (input_node): Stream profile ids.
9041 (input_edge): Stream common targets.
9042 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9043 * ipa.c: Include value-prof.h
9044 (ipa_profile_generate_summary): Turn indirect call statement histograms
9045 into common targets.
9046 (ipa_profile): Turn common targets into speculative edges.
9048 2013-08-09 Jan Hubicka <jh@suse.cz>
9050 * cgraph.h (cgraph_node): Add profile_id.
9051 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9052 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9053 (del_node_map): Update.
9054 (find_func_by_funcdef_no): Replace by ...
9055 (find_func_by_profile_id): ... this one.
9056 (gimple_ic_transform): Do not remove useful histograms when
9057 speculation is not done; dump info when indirect call removal
9059 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9060 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9061 (__gcov_indirect_call_profiler_v2): .. this one.
9062 * profile.h (init_node_map): Update.
9063 * coverage.c (coverage_compute_profile_id): New function.
9064 * coverage.h (coverage_compute_profile_id): Declare.
9065 * tree-profile.c (init_ic_make_global_vars): Make
9066 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9067 (gimple_init_edge_profiler): Update prototype of
9068 __gcov_indirect_call_profiler.
9069 (gimple_gen_ic_func_profiler): Simplify.
9070 (tree_profiling): Use init_node_map
9072 2013-08-09 Jan Hubicka <jh@suse.cz>
9074 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9075 non-speculative refs.
9076 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9077 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9078 (cgraph_set_call_stmt): Likewise.
9079 (cgraph_create_edge_1): Fix release checking compilatoin;
9081 (cgraph_free_edge): Free indirect info.
9082 (cgraph_turn_edge_to_speculative): New function.
9083 (cgraph_speculative_call_info): New function.
9084 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9085 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9086 (dump_cgraph_node): Dump speculation.
9087 (verify_edge_count_and_frequency): Accept speculative edges.
9088 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9089 (verify_cgraph_node): Handle speculation.
9090 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9091 (cgraph_set_call_stmt): Update prototype.
9092 (cgraph_make_edge_direct): Update prototype.
9093 (cgraph_speculative_call_info): Declare.
9094 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9095 to change; update call of ipa_find_references.
9096 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9097 lto_stmt_uid and speculative flags.
9098 (ipa_dump_references): Dump speculation.
9099 (ipa_clone_references): Clone speculative flag.
9100 (ipa_clone_referring): Likewise.
9101 (ipa_clone_ref): New function.
9102 (ipa_find_reference): Look into lto_stmt_uids
9103 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9104 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9105 (ipa_find_reference): Update declaration.
9106 (ipa_clone_ref): Declare.
9107 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9108 stream speculative flag.
9109 (lto_output_ref): Stream statements uids and speculation.
9110 (input_ref): Likewise.
9111 (input_edge): Stream speuclation.
9112 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9113 (cgraph_set_call_stmt_including_clones): Handle speculation.
9114 * ipa-inline.c (heap_edge_removal_hook): New function.
9115 (inline_small_functions): Register it.
9116 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9117 also initialize refs.
9118 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9120 (try_make_edge_direct_simple_call): Likewise.
9121 (try_make_edge_direct_simple_call): Likewise.
9122 (update_indirect_edges_after_inlining): Likewise.
9123 (remove_described_reference): Look proper lto_stmt_uid.
9124 (propagate_controlled_uses): Likewise.
9125 (propagate_controlled_uses): Liekwise.
9126 * tree-inline.c (copy_bb): Copy speculative edges.
9127 (redirect_all_calls): New function.
9128 (copy_cfg_body): Do redirection after loop info is updated.
9129 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9131 2013-08-09 Jan Hubicka <jh@suse.cz>
9133 * lto-streamer-out.c (output_function): Renumber PHIs.
9134 * lto-streamer-in.c (input_function): Likewise.
9136 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9138 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9139 (get_lane_unsigned): Likewise.
9140 (dup_lane_scalar): Likewise.
9141 (get_lane): enable for VALL.
9142 * config/aarch64/aarch64-simd.md
9143 (aarch64_dup_lane_scalar<mode>): Remove.
9144 (aarch64_get_lane_signed<mode>): Likewise.
9145 (aarch64_get_lane_unsigned<mode>): Likewise.
9146 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9147 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9148 (aarch64_get_lane<mode>): Enable for all vector modes.
9149 (aarch64_get_lanedi): Remove misleading constraints.
9150 * config/aarch64/arm_neon.h
9151 (__aarch64_vget_lane_any): Define.
9152 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9153 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9154 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9155 * config/aarch64/iterators.md (VDQQH): New.
9159 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9161 * configure.ac: Add GAS check for LEON instructions on SPARC.
9162 * configure: Regenerate.
9163 * config.in: Likewise.
9164 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9166 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9167 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9169 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9170 (AS_LEON_FLAG): New macro.
9171 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9172 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9173 Deal with LEON and LEON3 for the memory model.
9174 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9175 (atomic_compare_and_swap<mode>_1): Likewise.
9176 (*atomic_compare_and_swap<mode>_1): Likewise.
9178 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9180 * config/arm/neon.md (vcond): Fix floating-point vector
9181 comparisons against 0.
9183 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9185 * lra-constraints.c (emit_spill_move): Remove assert.
9186 (process_alt_operands): Add more debugging
9187 output. Increase reject for spilling into memory. Decrease
9188 reject for reloading scratch.
9189 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9191 2013-08-08 Steve Ellcey <sellcey@mips.com>
9193 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9194 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9195 micromips incompatible. Add nan2008.
9196 (MULTILIB_DIRNAMES): Add nan2008.
9197 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9198 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9199 and micromips incompatible. Add nan2008.
9200 (MULTILIB_DIRNAMES): Add nan2008.
9201 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9203 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9205 PR rtl-optimization/58079
9206 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9207 simplify_comparison has widened a comparison with an integer.
9209 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9211 * config/arm/neon.md (movmisalign<mode>): Disable when we
9212 don't allow unaligned accesses.
9213 (*movmisalign<mode>_neon_store): Likewise.
9214 (*movmisalign<mode>_neon_load): Likewise.
9215 (*movmisalign<mode>_neon_store): Likewise.
9216 (*movmisalign<mode>_neon_load): Likewise.
9218 2013-08-08 Jan Hubicka <jh@suse.cz>
9220 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9221 (make_pass_rebuild_cgraph_edges): Also clear references.
9222 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9223 * ipa-inline-transform.c (inline_transform): Remove all references
9225 * cgraphunit.c (expand_function): Remove all references after
9227 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9228 (ipa_find_reference): Rewrite to iterator.
9229 (remove_stmt_references): Likewise.
9230 (ipa_clear_stmts_in_references): New function.
9231 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9232 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9234 * ipa-split.c (split_function): Remove references in split function.
9236 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9239 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9240 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9242 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9245 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9246 suggested mode for the assignment isn't compatible with the
9249 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9252 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9254 2013-08-07 Xinliang David Li <davidxl@google.com>
9256 * config/i386/i386.opt: New option -mtune-ctrl=.
9257 * config/i386/x86-tune.def: New file.
9258 * config/i386/i386.h: include x86-tune.def.
9259 * config/i386/i386.c (ix86_option_override_internal):
9260 Parsing -mtune-ctrl= option and set tune features.
9262 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9265 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9266 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9269 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9271 * rtl.h (update_alignments): Declare.
9272 * final.c (grow_label_align): New function extracted from...
9273 (shorten_branches): ...here. Call it.
9274 (update_alignments): New function.
9275 * reorg.c (sibling_labels): New variable.
9276 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9277 the new label along with it onto the sibling_labels vector.
9278 (fill_simple_delay_slots): Adjust call to get_label_before.
9279 (fill_slots_from_thread): Likewise.
9280 (relax_delay_slots): Likewise.
9281 (make_return_insns): Likewise.
9282 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9284 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9286 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9287 document its semantics.
9288 (diagnostic_report_diagnostic): Adjust accordingly.
9290 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9292 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9293 (sparc_option_override): ...and port to new C++ pass API.
9294 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9296 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9298 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9300 2013-08-06 Caroline Tice <cmtice@google.com>
9302 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9303 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9304 * tree-pass.h: Add pass_vtable_verify.
9305 * varasm.c (assemble_variable): Add code to properly set the comdat
9306 section and name for the .vtable_map_vars section.
9307 (assemble_vtyv_preinit_initializer): New function.
9308 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9310 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9311 * vtable-verify.c: New file.
9312 * vtable-verify.h: New file.
9313 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9314 initialiation levels.
9315 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9316 * passes.def: Insert pass_vtable_verify.
9317 * aclocal.m4: Reorder includes.
9318 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9319 -fvtv-counts options.
9320 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9321 as appropriate, if -fvtable-verify=... is used.
9322 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9323 -fvtable-verify=... is used.
9324 * Makefile.in (OBJS): Add vtable-verify.o to list.
9325 (vtable-verify.o): Add new build rule.
9326 (GTFILES): Add vtable-verify.c to list.
9327 * common.opt (fvtable-verify=): New flag.
9328 (vtv_priority): Values for fvtable-verify= flag.
9329 (fvtv-counts): New flag.
9330 (fvtv-debug): New flag.
9331 * tree.h (save_vtable_map_decl): New extern function decl.
9333 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9335 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9336 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9337 (pass_data_rl78_devirt): ...new pass_data instance and...
9338 (make_pass_rl78_devirt): ...new function.
9339 (rl78_asm_file_start): Port pass registration to new C++ API.
9341 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9343 * coretypes.h (rtl_opt_pass): Add.
9344 (gcc::context): Add.
9345 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9346 (epiphany_init): Port to new C++ pass API.
9347 (epiphany_optimize_mode_switching): Likewise.
9348 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9349 (pass_manager::get_pass_mode_switching): New.
9350 (pass_manager::get_pass_peephole2): New.
9351 * mode-switching.c (pass_mode_switching): Add clone method.
9352 * recog.c (pass_peephole2): Add clone method.
9353 (pass_split_all_insns): Add clone method.
9355 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9357 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9358 (mips_option_override): ...here, porting to new C++ API for passes.
9360 2013-08-06 Jan Hubicka <jh@suse.cz>
9362 * cgraph.c (cgraph_get_body): New function based on lto.c
9364 * cgraph.h (cgraph_get_body): Declare.
9365 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9367 * cgraphunit.c (expand_function): Get body prior expanding.
9368 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9369 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9371 * passes.c (do_per_function_toporder): Get body.
9372 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9373 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9375 2013-08-06 Martin Jambor <mjambor@suse.cz>
9378 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9379 is not re-finalized. Rename second parameter to no_collect.
9381 2013-08-06 Martin Jambor <mjambor@suse.cz>
9384 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9385 MEM_REF has proper alignment information.
9387 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9390 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9392 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9393 rtx (*) (rtx, ...) with insn_gen_fn.
9394 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9395 insn_gen_fn::stored_funcptr. Add initializer braces.
9397 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9399 Rewrite how instances of passes are cloned to remove assumptions
9400 about their sizes (thus allowing pass subclasses to have
9401 additional data fields, albeit non-GC-managed ones at this point).
9403 * passes.c (make_pass_instance): Now that passes have clone
9404 methods, rewrite this function to eliminate XNEW and memcpy
9405 calls that used hardcoded sizes. Since this function no longer
9406 creates pass instances, rename it to...
9407 (add_pass_instance): ...this. Document the old way that passes were
9408 numbered and flagged, and rework this function to continue using it.
9409 (next_pass_1): Add an initial_pass argument for use by
9411 (position_pass): When adding multiple instances of a pass, use
9412 the pass's clone method, rather than relying on the XNEW/memcpy
9413 within the former make_pass_instance (now add_pass_instance).
9414 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9415 the initial instance of the current pass within the pass manager.
9417 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9419 This is the automated part of the conversion of passes from C
9420 structs to C++ classes.
9422 Patch autogenerated by refactor_passes.py from
9423 https://github.com/davidmalcolm/gcc-refactoring-scripts
9424 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9426 * asan.c (pass_asan): Convert from a global struct to a subclass of
9427 gimple_opt_pass along with...
9428 (pass_data_asan): ...new pass_data instance and...
9429 (make_pass_asan): ...new function.
9430 (pass_asan_O0): Convert from a global struct to a subclass of
9431 gimple_opt_pass along with...
9432 (pass_data_asan_O0): ...new pass_data instance and...
9433 (make_pass_asan_O0): ...new function.
9434 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9435 subclass of rtl_opt_pass along with...
9436 (pass_data_inc_dec): ...new pass_data instance and...
9437 (make_pass_inc_dec): ...new function.
9438 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9439 a subclass of rtl_opt_pass along with...
9440 (pass_data_reorder_blocks): ...new pass_data instance and...
9441 (make_pass_reorder_blocks): ...new function.
9442 (pass_duplicate_computed_gotos): Convert from a global struct to a
9443 subclass of rtl_opt_pass along with...
9444 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9445 (make_pass_duplicate_computed_gotos): ...new function.
9446 (pass_partition_blocks): Convert from a global struct to a subclass of
9447 rtl_opt_pass along with...
9448 (pass_data_partition_blocks): ...new pass_data instance and...
9449 (make_pass_partition_blocks): ...new function.
9450 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9451 struct to a subclass of rtl_opt_pass along with...
9452 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9454 (make_pass_branch_target_load_optimize1): ...new function.
9455 (pass_branch_target_load_optimize2): Convert from a global struct to a
9456 subclass of rtl_opt_pass along with...
9457 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9459 (make_pass_branch_target_load_optimize2): ...new function.
9460 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9461 of rtl_opt_pass along with...
9462 (pass_data_jump): ...new pass_data instance and...
9463 (make_pass_jump): ...new function.
9464 (pass_jump2): Convert from a global struct to a subclass of
9465 rtl_opt_pass along with...
9466 (pass_data_jump2): ...new pass_data instance and...
9467 (make_pass_jump2): ...new function.
9468 * cfgexpand.c (pass_expand): Convert from a global struct to a
9469 subclass of rtl_opt_pass along with...
9470 (pass_data_expand): ...new pass_data instance and...
9471 (make_pass_expand): ...new function.
9472 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9473 of rtl_opt_pass along with...
9474 (pass_data_free_cfg): ...new pass_data instance and...
9475 (make_pass_free_cfg): ...new function.
9476 (pass_into_cfg_layout_mode): Convert from a global struct to a
9477 subclass of rtl_opt_pass along with...
9478 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9479 (make_pass_into_cfg_layout_mode): ...new function.
9480 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9481 subclass of rtl_opt_pass along with...
9482 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9483 (make_pass_outof_cfg_layout_mode): ...new function.
9484 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9485 struct to a subclass of gimple_opt_pass along with...
9486 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9487 (make_pass_build_cgraph_edges): ...new function.
9488 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9489 subclass of gimple_opt_pass along with...
9490 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9491 (make_pass_rebuild_cgraph_edges): ...new function.
9492 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9493 subclass of gimple_opt_pass along with...
9494 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9496 (make_pass_remove_cgraph_callee_edges): ...new function.
9497 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9498 struct to a subclass of rtl_opt_pass along with...
9499 (pass_data_stack_adjustments): ...new pass_data instance and...
9500 (make_pass_stack_adjustments): ...new function.
9501 * combine.c (pass_combine): Convert from a global struct to a subclass
9502 of rtl_opt_pass along with...
9503 (pass_data_combine): ...new pass_data instance and...
9504 (make_pass_combine): ...new function.
9505 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9506 global struct to a subclass of rtl_opt_pass along with...
9507 (pass_data_compare_elim_after_reload): ...new pass_data instance
9509 (make_pass_compare_elim_after_reload): ...new function.
9510 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9511 of rtl_opt_pass along with...
9512 (pass_data_rtl_cprop): ...new pass_data instance and...
9513 (make_pass_rtl_cprop): ...new function.
9514 * cse.c (pass_cse): Convert from a global struct to a subclass of
9515 rtl_opt_pass along with...
9516 (pass_data_cse): ...new pass_data instance and...
9517 (make_pass_cse): ...new function.
9518 (pass_cse2): Convert from a global struct to a subclass of
9519 rtl_opt_pass along with...
9520 (pass_data_cse2): ...new pass_data instance and...
9521 (make_pass_cse2): ...new function.
9522 (pass_cse_after_global_opts): Convert from a global struct to a
9523 subclass of rtl_opt_pass along with...
9524 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9525 (make_pass_cse_after_global_opts): ...new function.
9526 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9527 of rtl_opt_pass along with...
9528 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9529 (make_pass_ud_rtl_dce): ...new function.
9530 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9531 rtl_opt_pass along with...
9532 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9533 (make_pass_fast_rtl_dce): ...new function.
9534 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9535 a subclass of rtl_opt_pass along with...
9536 (pass_data_df_initialize_opt): ...new pass_data instance and...
9537 (make_pass_df_initialize_opt): ...new function.
9538 (pass_df_initialize_no_opt): Convert from a global struct to a
9539 subclass of rtl_opt_pass along with...
9540 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9541 (make_pass_df_initialize_no_opt): ...new function.
9542 (pass_df_finish): Convert from a global struct to a subclass of
9543 rtl_opt_pass along with...
9544 (pass_data_df_finish): ...new pass_data instance and...
9545 (make_pass_df_finish): ...new function.
9546 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9547 rtl_opt_pass along with...
9548 (pass_data_rtl_dse1): ...new pass_data instance and...
9549 (make_pass_rtl_dse1): ...new function.
9550 (pass_rtl_dse2): Convert from a global struct to a subclass of
9551 rtl_opt_pass along with...
9552 (pass_data_rtl_dse2): ...new pass_data instance and...
9553 (make_pass_rtl_dse2): ...new function.
9554 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9555 subclass of rtl_opt_pass along with...
9556 (pass_data_dwarf2_frame): ...new pass_data instance and...
9557 (make_pass_dwarf2_frame): ...new function.
9558 * except.c (pass_set_nothrow_function_flags): Convert from a global
9559 struct to a subclass of rtl_opt_pass along with...
9560 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9562 (make_pass_set_nothrow_function_flags): ...new function.
9563 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9564 subclass of rtl_opt_pass along with...
9565 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9567 (make_pass_convert_to_eh_region_ranges): ...new function.
9568 * final.c (pass_compute_alignments): Convert from a global struct to a
9569 subclass of rtl_opt_pass along with...
9570 (pass_data_compute_alignments): ...new pass_data instance and...
9571 (make_pass_compute_alignments): ...new function.
9572 (pass_final): Convert from a global struct to a subclass of
9573 rtl_opt_pass along with...
9574 (pass_data_final): ...new pass_data instance and...
9575 (make_pass_final): ...new function.
9576 (pass_shorten_branches): Convert from a global struct to a subclass of
9577 rtl_opt_pass along with...
9578 (pass_data_shorten_branches): ...new pass_data instance and...
9579 (make_pass_shorten_branches): ...new function.
9580 (pass_clean_state): Convert from a global struct to a subclass of
9581 rtl_opt_pass along with...
9582 (pass_data_clean_state): ...new pass_data instance and...
9583 (make_pass_clean_state): ...new function.
9584 * function.c (pass_instantiate_virtual_regs): Convert from a global
9585 struct to a subclass of rtl_opt_pass along with...
9586 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9587 (make_pass_instantiate_virtual_regs): ...new function.
9588 (pass_leaf_regs): Convert from a global struct to a subclass of
9589 rtl_opt_pass along with...
9590 (pass_data_leaf_regs): ...new pass_data instance and...
9591 (make_pass_leaf_regs): ...new function.
9592 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9593 subclass of rtl_opt_pass along with...
9594 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9596 (make_pass_thread_prologue_and_epilogue): ...new function.
9597 (pass_match_asm_constraints): Convert from a global struct to a
9598 subclass of rtl_opt_pass along with...
9599 (pass_data_match_asm_constraints): ...new pass_data instance and...
9600 (make_pass_match_asm_constraints): ...new function.
9601 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9602 subclass of rtl_opt_pass along with...
9603 (pass_data_rtl_fwprop): ...new pass_data instance and...
9604 (make_pass_rtl_fwprop): ...new function.
9605 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9606 rtl_opt_pass along with...
9607 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9608 (make_pass_rtl_fwprop_addr): ...new function.
9609 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9610 rtl_opt_pass along with...
9611 (pass_data_rtl_pre): ...new pass_data instance and...
9612 (make_pass_rtl_pre): ...new function.
9613 (pass_rtl_hoist): Convert from a global struct to a subclass of
9614 rtl_opt_pass along with...
9615 (pass_data_rtl_hoist): ...new pass_data instance and...
9616 (make_pass_rtl_hoist): ...new function.
9617 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9618 subclass of gimple_opt_pass along with...
9619 (pass_data_lower_cf): ...new pass_data instance and...
9620 (make_pass_lower_cf): ...new function.
9621 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9622 from a global struct to a subclass of gimple_opt_pass along with...
9623 (pass_data_strength_reduction): ...new pass_data instance and...
9624 (make_pass_strength_reduction): ...new function.
9625 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9626 of rtl_opt_pass along with...
9627 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9628 (make_pass_rtl_ifcvt): ...new function.
9629 (pass_if_after_combine): Convert from a global struct to a subclass of
9630 rtl_opt_pass along with...
9631 (pass_data_if_after_combine): ...new pass_data instance and...
9632 (make_pass_if_after_combine): ...new function.
9633 (pass_if_after_reload): Convert from a global struct to a subclass of
9634 rtl_opt_pass along with...
9635 (pass_data_if_after_reload): ...new pass_data instance and...
9636 (make_pass_if_after_reload): ...new function.
9637 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9638 a subclass of rtl_opt_pass along with...
9639 (pass_data_initialize_regs): ...new pass_data instance and...
9640 (make_pass_initialize_regs): ...new function.
9641 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9642 of ipa_opt_pass_d along with...
9643 (pass_data_ipa_cp): ...new pass_data instance and...
9644 (make_pass_ipa_cp): ...new function.
9645 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9646 global struct to a subclass of gimple_opt_pass along with...
9647 (pass_data_inline_parameters): ...new pass_data instance and...
9648 (make_pass_inline_parameters): ...new function.
9649 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9650 subclass of gimple_opt_pass along with...
9651 (pass_data_early_inline): ...new pass_data instance and...
9652 (make_pass_early_inline): ...new function.
9653 (pass_ipa_inline): Convert from a global struct to a subclass of
9654 ipa_opt_pass_d along with...
9655 (pass_data_ipa_inline): ...new pass_data instance and...
9656 (make_pass_ipa_inline): ...new function.
9657 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9658 struct to a subclass of gimple_opt_pass along with...
9659 (pass_data_local_pure_const): ...new pass_data instance and...
9660 (make_pass_local_pure_const): ...new function.
9661 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9662 ipa_opt_pass_d along with...
9663 (pass_data_ipa_pure_const): ...new pass_data instance and...
9664 (make_pass_ipa_pure_const): ...new function.
9665 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9666 to a subclass of ipa_opt_pass_d along with...
9667 (pass_data_ipa_reference): ...new pass_data instance and...
9668 (make_pass_ipa_reference): ...new function.
9669 * ipa-split.c (pass_split_functions): Convert from a global struct to
9670 a subclass of gimple_opt_pass along with...
9671 (pass_data_split_functions): ...new pass_data instance and...
9672 (make_pass_split_functions): ...new function.
9673 (pass_feedback_split_functions): Convert from a global struct to a
9674 subclass of gimple_opt_pass along with...
9675 (pass_data_feedback_split_functions): ...new pass_data instance and...
9676 (make_pass_feedback_split_functions): ...new function.
9677 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9678 global struct to a subclass of simple_ipa_opt_pass along with...
9679 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9681 (make_pass_ipa_function_and_variable_visibility): ...new function.
9682 (pass_ipa_free_inline_summary): Convert from a global struct to a
9683 subclass of simple_ipa_opt_pass along with...
9684 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9685 (make_pass_ipa_free_inline_summary): ...new function.
9686 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9687 subclass of ipa_opt_pass_d along with...
9688 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9690 (make_pass_ipa_whole_program_visibility): ...new function.
9691 (pass_ipa_profile): Convert from a global struct to a subclass of
9692 ipa_opt_pass_d along with...
9693 (pass_data_ipa_profile): ...new pass_data instance and...
9694 (make_pass_ipa_profile): ...new function.
9695 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9696 ipa_opt_pass_d along with...
9697 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9698 (make_pass_ipa_cdtor_merge): ...new function.
9699 * ira.c (pass_ira): Convert from a global struct to a subclass of
9700 rtl_opt_pass along with...
9701 (pass_data_ira): ...new pass_data instance and...
9702 (make_pass_ira): ...new function.
9703 (pass_reload): Convert from a global struct to a subclass of
9704 rtl_opt_pass along with...
9705 (pass_data_reload): ...new pass_data instance and...
9706 (make_pass_reload): ...new function.
9707 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9708 subclass of rtl_opt_pass along with...
9709 (pass_data_cleanup_barriers): ...new pass_data instance and...
9710 (make_pass_cleanup_barriers): ...new function.
9711 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9712 of rtl_opt_pass along with...
9713 (pass_data_loop2): ...new pass_data instance and...
9714 (make_pass_loop2): ...new function.
9715 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9716 rtl_opt_pass along with...
9717 (pass_data_rtl_loop_init): ...new pass_data instance and...
9718 (make_pass_rtl_loop_init): ...new function.
9719 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9720 rtl_opt_pass along with...
9721 (pass_data_rtl_loop_done): ...new pass_data instance and...
9722 (make_pass_rtl_loop_done): ...new function.
9723 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9724 subclass of rtl_opt_pass along with...
9725 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9726 (make_pass_rtl_move_loop_invariants): ...new function.
9727 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9728 rtl_opt_pass along with...
9729 (pass_data_rtl_unswitch): ...new pass_data instance and...
9730 (make_pass_rtl_unswitch): ...new function.
9731 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9732 subclass of rtl_opt_pass along with...
9733 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9735 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9736 (pass_rtl_doloop): Convert from a global struct to a subclass of
9737 rtl_opt_pass along with...
9738 (pass_data_rtl_doloop): ...new pass_data instance and...
9739 (make_pass_rtl_doloop): ...new function.
9740 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9741 a subclass of rtl_opt_pass along with...
9742 (pass_data_lower_subreg): ...new pass_data instance and...
9743 (make_pass_lower_subreg): ...new function.
9744 (pass_lower_subreg2): Convert from a global struct to a subclass of
9745 rtl_opt_pass along with...
9746 (pass_data_lower_subreg2): ...new pass_data instance and...
9747 (make_pass_lower_subreg2): ...new function.
9748 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9749 struct to a subclass of ipa_opt_pass_d along with...
9750 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9751 (make_pass_ipa_lto_gimple_out): ...new function.
9752 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9753 of ipa_opt_pass_d along with...
9754 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9755 (make_pass_ipa_lto_finish_out): ...new function.
9756 * mode-switching.c (pass_mode_switching): Convert from a global struct
9757 to a subclass of rtl_opt_pass along with...
9758 (pass_data_mode_switching): ...new pass_data instance and...
9759 (make_pass_mode_switching): ...new function.
9760 * modulo-sched.c (pass_sms): Convert from a global struct to a
9761 subclass of rtl_opt_pass along with...
9762 (pass_data_sms): ...new pass_data instance and...
9763 (make_pass_sms): ...new function.
9764 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9765 subclass of gimple_opt_pass along with...
9766 (pass_data_expand_omp): ...new pass_data instance and...
9767 (make_pass_expand_omp): ...new function.
9768 (pass_lower_omp): Convert from a global struct to a subclass of
9769 gimple_opt_pass along with...
9770 (pass_data_lower_omp): ...new pass_data instance and...
9771 (make_pass_lower_omp): ...new function.
9772 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9773 of gimple_opt_pass along with...
9774 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9775 (make_pass_diagnose_omp_blocks): ...new function.
9776 * passes.c (pass_early_local_passes): Convert from a global struct to
9777 a subclass of simple_ipa_opt_pass along with...
9778 (pass_data_early_local_passes): ...new pass_data instance and...
9779 (make_pass_early_local_passes): ...new function.
9780 (pass_all_early_optimizations): Convert from a global struct to a
9781 subclass of gimple_opt_pass along with...
9782 (pass_data_all_early_optimizations): ...new pass_data instance and...
9783 (make_pass_all_early_optimizations): ...new function.
9784 (pass_all_optimizations): Convert from a global struct to a subclass
9785 of gimple_opt_pass along with...
9786 (pass_data_all_optimizations): ...new pass_data instance and...
9787 (make_pass_all_optimizations): ...new function.
9788 (pass_all_optimizations_g): Convert from a global struct to a subclass
9789 of gimple_opt_pass along with...
9790 (pass_data_all_optimizations_g): ...new pass_data instance and...
9791 (make_pass_all_optimizations_g): ...new function.
9792 (pass_rest_of_compilation): Convert from a global struct to a subclass
9793 of rtl_opt_pass along with...
9794 (pass_data_rest_of_compilation): ...new pass_data instance and...
9795 (make_pass_rest_of_compilation): ...new function.
9796 (pass_postreload): Convert from a global struct to a subclass of
9797 rtl_opt_pass along with...
9798 (pass_data_postreload): ...new pass_data instance and...
9799 (make_pass_postreload): ...new function.
9800 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9801 subclass of rtl_opt_pass along with...
9802 (pass_data_gcse2): ...new pass_data instance and...
9803 (make_pass_gcse2): ...new function.
9804 * postreload.c (pass_postreload_cse): Convert from a global struct to
9805 a subclass of rtl_opt_pass along with...
9806 (pass_data_postreload_cse): ...new pass_data instance and...
9807 (make_pass_postreload_cse): ...new function.
9808 * predict.c (pass_profile): Convert from a global struct to a subclass
9809 of gimple_opt_pass along with...
9810 (pass_data_profile): ...new pass_data instance and...
9811 (make_pass_profile): ...new function.
9812 (pass_strip_predict_hints): Convert from a global struct to a subclass
9813 of gimple_opt_pass along with...
9814 (pass_data_strip_predict_hints): ...new pass_data instance and...
9815 (make_pass_strip_predict_hints): ...new function.
9816 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9817 of rtl_opt_pass along with...
9818 (pass_data_peephole2): ...new pass_data instance and...
9819 (make_pass_peephole2): ...new function.
9820 (pass_split_all_insns): Convert from a global struct to a subclass of
9821 rtl_opt_pass along with...
9822 (pass_data_split_all_insns): ...new pass_data instance and...
9823 (make_pass_split_all_insns): ...new function.
9824 (pass_split_after_reload): Convert from a global struct to a subclass
9825 of rtl_opt_pass along with...
9826 (pass_data_split_after_reload): ...new pass_data instance and...
9827 (make_pass_split_after_reload): ...new function.
9828 (pass_split_before_regstack): Convert from a global struct to a
9829 subclass of rtl_opt_pass along with...
9830 (pass_data_split_before_regstack): ...new pass_data instance and...
9831 (make_pass_split_before_regstack): ...new function.
9832 (pass_split_before_sched2): Convert from a global struct to a subclass
9833 of rtl_opt_pass along with...
9834 (pass_data_split_before_sched2): ...new pass_data instance and...
9835 (make_pass_split_before_sched2): ...new function.
9836 (pass_split_for_shorten_branches): Convert from a global struct to a
9837 subclass of rtl_opt_pass along with...
9838 (pass_data_split_for_shorten_branches): ...new pass_data instance
9840 (make_pass_split_for_shorten_branches): ...new function.
9841 * ree.c (pass_ree): Convert from a global struct to a subclass of
9842 rtl_opt_pass along with...
9843 (pass_data_ree): ...new pass_data instance and...
9844 (make_pass_ree): ...new function.
9845 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9846 subclass of rtl_opt_pass along with...
9847 (pass_data_stack_regs): ...new pass_data instance and...
9848 (make_pass_stack_regs): ...new function.
9849 (pass_stack_regs_run): Convert from a global struct to a subclass of
9850 rtl_opt_pass along with...
9851 (pass_data_stack_regs_run): ...new pass_data instance and...
9852 (make_pass_stack_regs_run): ...new function.
9853 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9854 subclass of rtl_opt_pass along with...
9855 (pass_data_cprop_hardreg): ...new pass_data instance and...
9856 (make_pass_cprop_hardreg): ...new function.
9857 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9858 subclass of rtl_opt_pass along with...
9859 (pass_data_reginfo_init): ...new pass_data instance and...
9860 (make_pass_reginfo_init): ...new function.
9861 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9862 of rtl_opt_pass along with...
9863 (pass_data_regmove): ...new pass_data instance and...
9864 (make_pass_regmove): ...new function.
9865 * regrename.c (pass_regrename): Convert from a global struct to a
9866 subclass of rtl_opt_pass along with...
9867 (pass_data_regrename): ...new pass_data instance and...
9868 (make_pass_regrename): ...new function.
9869 * reorg.c (pass_delay_slots): Convert from a global struct to a
9870 subclass of rtl_opt_pass along with...
9871 (pass_data_delay_slots): ...new pass_data instance and...
9872 (make_pass_delay_slots): ...new function.
9873 (pass_machine_reorg): Convert from a global struct to a subclass of
9874 rtl_opt_pass along with...
9875 (pass_data_machine_reorg): ...new pass_data instance and...
9876 (make_pass_machine_reorg): ...new function.
9877 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9878 of rtl_opt_pass along with...
9879 (pass_data_sched): ...new pass_data instance and...
9880 (make_pass_sched): ...new function.
9881 (pass_sched2): Convert from a global struct to a subclass of
9882 rtl_opt_pass along with...
9883 (pass_data_sched2): ...new pass_data instance and...
9884 (make_pass_sched2): ...new function.
9885 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9886 to a subclass of rtl_opt_pass along with...
9887 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9888 (make_pass_stack_ptr_mod): ...new function.
9889 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9890 to a subclass of rtl_opt_pass along with...
9891 (pass_data_rtl_store_motion): ...new pass_data instance and...
9892 (make_pass_rtl_store_motion): ...new function.
9893 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9894 of gimple_opt_pass along with...
9895 (pass_data_tracer): ...new pass_data instance and...
9896 (make_pass_tracer): ...new function.
9897 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9898 to a subclass of gimple_opt_pass along with...
9899 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9900 (make_pass_diagnose_tm_blocks): ...new function.
9901 (pass_lower_tm): Convert from a global struct to a subclass of
9902 gimple_opt_pass along with...
9903 (pass_data_lower_tm): ...new pass_data instance and...
9904 (make_pass_lower_tm): ...new function.
9905 (pass_tm_init): Convert from a global struct to a subclass of
9906 gimple_opt_pass along with...
9907 (pass_data_tm_init): ...new pass_data instance and...
9908 (make_pass_tm_init): ...new function.
9909 (pass_tm_mark): Convert from a global struct to a subclass of
9910 gimple_opt_pass along with...
9911 (pass_data_tm_mark): ...new pass_data instance and...
9912 (make_pass_tm_mark): ...new function.
9913 (pass_tm_edges): Convert from a global struct to a subclass of
9914 gimple_opt_pass along with...
9915 (pass_data_tm_edges): ...new pass_data instance and...
9916 (make_pass_tm_edges): ...new function.
9917 (pass_tm_memopt): Convert from a global struct to a subclass of
9918 gimple_opt_pass along with...
9919 (pass_data_tm_memopt): ...new pass_data instance and...
9920 (make_pass_tm_memopt): ...new function.
9921 (pass_ipa_tm): Convert from a global struct to a subclass of
9922 simple_ipa_opt_pass along with...
9923 (pass_data_ipa_tm): ...new pass_data instance and...
9924 (make_pass_ipa_tm): ...new function.
9925 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9926 subclass of gimple_opt_pass along with...
9927 (pass_data_call_cdce): ...new pass_data instance and...
9928 (make_pass_call_cdce): ...new function.
9929 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9930 subclass of gimple_opt_pass along with...
9931 (pass_data_build_cfg): ...new pass_data instance and...
9932 (make_pass_build_cfg): ...new function.
9933 (pass_split_crit_edges): Convert from a global struct to a subclass of
9934 gimple_opt_pass along with...
9935 (pass_data_split_crit_edges): ...new pass_data instance and...
9936 (make_pass_split_crit_edges): ...new function.
9937 (pass_warn_function_return): Convert from a global struct to a
9938 subclass of gimple_opt_pass along with...
9939 (pass_data_warn_function_return): ...new pass_data instance and...
9940 (make_pass_warn_function_return): ...new function.
9941 (pass_warn_function_noreturn): Convert from a global struct to a
9942 subclass of gimple_opt_pass along with...
9943 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9944 (make_pass_warn_function_noreturn): ...new function.
9945 (pass_warn_unused_result): Convert from a global struct to a subclass
9946 of gimple_opt_pass along with...
9947 (pass_data_warn_unused_result): ...new pass_data instance and...
9948 (make_pass_warn_unused_result): ...new function.
9949 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9950 a subclass of gimple_opt_pass along with...
9951 (pass_data_merge_phi): ...new pass_data instance and...
9952 (make_pass_merge_phi): ...new function.
9953 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9954 a subclass of gimple_opt_pass along with...
9955 (pass_data_lower_complex): ...new pass_data instance and...
9956 (make_pass_lower_complex): ...new function.
9957 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9958 gimple_opt_pass along with...
9959 (pass_data_lower_complex_O0): ...new pass_data instance and...
9960 (make_pass_lower_complex_O0): ...new function.
9961 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9962 subclass of gimple_opt_pass along with...
9963 (pass_data_lower_eh): ...new pass_data instance and...
9964 (make_pass_lower_eh): ...new function.
9965 (pass_refactor_eh): Convert from a global struct to a subclass of
9966 gimple_opt_pass along with...
9967 (pass_data_refactor_eh): ...new pass_data instance and...
9968 (make_pass_refactor_eh): ...new function.
9969 (pass_lower_resx): Convert from a global struct to a subclass of
9970 gimple_opt_pass along with...
9971 (pass_data_lower_resx): ...new pass_data instance and...
9972 (make_pass_lower_resx): ...new function.
9973 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9974 of gimple_opt_pass along with...
9975 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9976 (make_pass_lower_eh_dispatch): ...new function.
9977 (pass_cleanup_eh): Convert from a global struct to a subclass of
9978 gimple_opt_pass along with...
9979 (pass_data_cleanup_eh): ...new pass_data instance and...
9980 (make_pass_cleanup_eh): ...new function.
9981 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9982 to a subclass of simple_ipa_opt_pass along with...
9983 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9984 (make_pass_ipa_lower_emutls): ...new function.
9985 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9986 a subclass of gimple_opt_pass along with...
9987 (pass_data_if_conversion): ...new pass_data instance and...
9988 (make_pass_if_conversion): ...new function.
9989 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9990 subclass of gimple_opt_pass along with...
9991 (pass_data_build_ssa): ...new pass_data instance and...
9992 (make_pass_build_ssa): ...new function.
9993 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9994 global struct to a subclass of gimple_opt_pass along with...
9995 (pass_data_loop_distribution): ...new pass_data instance and...
9996 (make_pass_loop_distribution): ...new function.
9997 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9998 subclass of gimple_opt_pass along with...
9999 (pass_data_mudflap_1): ...new pass_data instance and...
10000 (make_pass_mudflap_1): ...new function.
10001 (pass_mudflap_2): Convert from a global struct to a subclass of
10002 gimple_opt_pass along with...
10003 (pass_data_mudflap_2): ...new pass_data instance and...
10004 (make_pass_mudflap_2): ...new function.
10005 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10006 subclass of gimple_opt_pass along with...
10007 (pass_data_mudflap_1): ...new pass_data instance and...
10008 (make_pass_mudflap_1): ...new function.
10009 (pass_mudflap_2): Convert from a global struct to a subclass of
10010 gimple_opt_pass along with...
10011 (pass_data_mudflap_2): ...new pass_data instance and...
10012 (make_pass_mudflap_2): ...new function.
10013 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10014 gimple_opt_pass along with...
10015 (pass_data_nrv): ...new pass_data instance and...
10016 (make_pass_nrv): ...new function.
10017 (pass_return_slot): Convert from a global struct to a subclass of
10018 gimple_opt_pass along with...
10019 (pass_data_return_slot): ...new pass_data instance and...
10020 (make_pass_return_slot): ...new function.
10021 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10022 to a subclass of gimple_opt_pass along with...
10023 (pass_data_object_sizes): ...new pass_data instance and...
10024 (make_pass_object_sizes): ...new function.
10025 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10026 global struct to a subclass of gimple_opt_pass along with...
10027 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10029 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10030 (pass_fixup_cfg): Convert from a global struct to a subclass of
10031 gimple_opt_pass along with...
10032 (pass_data_fixup_cfg): ...new pass_data instance and...
10033 (make_pass_fixup_cfg): ...new function.
10034 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10035 (make_pass_mudflap_1): ...new function.
10036 (pass_mudflap_2): Replace declaration with that of...
10037 (make_pass_mudflap_2): ...new function.
10038 (pass_asan): Replace declaration with that of...
10039 (make_pass_asan): ...new function.
10040 (pass_asan_O0): Replace declaration with that of...
10041 (make_pass_asan_O0): ...new function.
10042 (pass_tsan): Replace declaration with that of...
10043 (make_pass_tsan): ...new function.
10044 (pass_tsan_O0): Replace declaration with that of...
10045 (make_pass_tsan_O0): ...new function.
10046 (pass_lower_cf): Replace declaration with that of...
10047 (make_pass_lower_cf): ...new function.
10048 (pass_refactor_eh): Replace declaration with that of...
10049 (make_pass_refactor_eh): ...new function.
10050 (pass_lower_eh): Replace declaration with that of...
10051 (make_pass_lower_eh): ...new function.
10052 (pass_lower_eh_dispatch): Replace declaration with that of...
10053 (make_pass_lower_eh_dispatch): ...new function.
10054 (pass_lower_resx): Replace declaration with that of...
10055 (make_pass_lower_resx): ...new function.
10056 (pass_build_cfg): Replace declaration with that of...
10057 (make_pass_build_cfg): ...new function.
10058 (pass_early_tree_profile): Replace declaration with that of...
10059 (make_pass_early_tree_profile): ...new function.
10060 (pass_cleanup_eh): Replace declaration with that of...
10061 (make_pass_cleanup_eh): ...new function.
10062 (pass_sra): Replace declaration with that of...
10063 (make_pass_sra): ...new function.
10064 (pass_sra_early): Replace declaration with that of...
10065 (make_pass_sra_early): ...new function.
10066 (pass_early_ipa_sra): Replace declaration with that of...
10067 (make_pass_early_ipa_sra): ...new function.
10068 (pass_tail_recursion): Replace declaration with that of...
10069 (make_pass_tail_recursion): ...new function.
10070 (pass_tail_calls): Replace declaration with that of...
10071 (make_pass_tail_calls): ...new function.
10072 (pass_tree_loop): Replace declaration with that of...
10073 (make_pass_tree_loop): ...new function.
10074 (pass_tree_loop_init): Replace declaration with that of...
10075 (make_pass_tree_loop_init): ...new function.
10076 (pass_lim): Replace declaration with that of...
10077 (make_pass_lim): ...new function.
10078 (pass_tree_unswitch): Replace declaration with that of...
10079 (make_pass_tree_unswitch): ...new function.
10080 (pass_predcom): Replace declaration with that of...
10081 (make_pass_predcom): ...new function.
10082 (pass_iv_canon): Replace declaration with that of...
10083 (make_pass_iv_canon): ...new function.
10084 (pass_scev_cprop): Replace declaration with that of...
10085 (make_pass_scev_cprop): ...new function.
10086 (pass_empty_loop): Replace declaration with that of...
10087 (make_pass_empty_loop): ...new function.
10088 (pass_record_bounds): Replace declaration with that of...
10089 (make_pass_record_bounds): ...new function.
10090 (pass_graphite): Replace declaration with that of...
10091 (make_pass_graphite): ...new function.
10092 (pass_graphite_transforms): Replace declaration with that of...
10093 (make_pass_graphite_transforms): ...new function.
10094 (pass_if_conversion): Replace declaration with that of...
10095 (make_pass_if_conversion): ...new function.
10096 (pass_loop_distribution): Replace declaration with that of...
10097 (make_pass_loop_distribution): ...new function.
10098 (pass_vectorize): Replace declaration with that of...
10099 (make_pass_vectorize): ...new function.
10100 (pass_slp_vectorize): Replace declaration with that of...
10101 (make_pass_slp_vectorize): ...new function.
10102 (pass_complete_unroll): Replace declaration with that of...
10103 (make_pass_complete_unroll): ...new function.
10104 (pass_complete_unrolli): Replace declaration with that of...
10105 (make_pass_complete_unrolli): ...new function.
10106 (pass_parallelize_loops): Replace declaration with that of...
10107 (make_pass_parallelize_loops): ...new function.
10108 (pass_loop_prefetch): Replace declaration with that of...
10109 (make_pass_loop_prefetch): ...new function.
10110 (pass_iv_optimize): Replace declaration with that of...
10111 (make_pass_iv_optimize): ...new function.
10112 (pass_tree_loop_done): Replace declaration with that of...
10113 (make_pass_tree_loop_done): ...new function.
10114 (pass_ch): Replace declaration with that of...
10115 (make_pass_ch): ...new function.
10116 (pass_ccp): Replace declaration with that of...
10117 (make_pass_ccp): ...new function.
10118 (pass_phi_only_cprop): Replace declaration with that of...
10119 (make_pass_phi_only_cprop): ...new function.
10120 (pass_build_ssa): Replace declaration with that of...
10121 (make_pass_build_ssa): ...new function.
10122 (pass_build_alias): Replace declaration with that of...
10123 (make_pass_build_alias): ...new function.
10124 (pass_build_ealias): Replace declaration with that of...
10125 (make_pass_build_ealias): ...new function.
10126 (pass_dominator): Replace declaration with that of...
10127 (make_pass_dominator): ...new function.
10128 (pass_dce): Replace declaration with that of...
10129 (make_pass_dce): ...new function.
10130 (pass_dce_loop): Replace declaration with that of...
10131 (make_pass_dce_loop): ...new function.
10132 (pass_cd_dce): Replace declaration with that of...
10133 (make_pass_cd_dce): ...new function.
10134 (pass_call_cdce): Replace declaration with that of...
10135 (make_pass_call_cdce): ...new function.
10136 (pass_merge_phi): Replace declaration with that of...
10137 (make_pass_merge_phi): ...new function.
10138 (pass_split_crit_edges): Replace declaration with that of...
10139 (make_pass_split_crit_edges): ...new function.
10140 (pass_pre): Replace declaration with that of...
10141 (make_pass_pre): ...new function.
10142 (pass_profile): Replace declaration with that of...
10143 (make_pass_profile): ...new function.
10144 (pass_strip_predict_hints): Replace declaration with that of...
10145 (make_pass_strip_predict_hints): ...new function.
10146 (pass_lower_complex_O0): Replace declaration with that of...
10147 (make_pass_lower_complex_O0): ...new function.
10148 (pass_lower_complex): Replace declaration with that of...
10149 (make_pass_lower_complex): ...new function.
10150 (pass_lower_vector): Replace declaration with that of...
10151 (make_pass_lower_vector): ...new function.
10152 (pass_lower_vector_ssa): Replace declaration with that of...
10153 (make_pass_lower_vector_ssa): ...new function.
10154 (pass_lower_omp): Replace declaration with that of...
10155 (make_pass_lower_omp): ...new function.
10156 (pass_diagnose_omp_blocks): Replace declaration with that of...
10157 (make_pass_diagnose_omp_blocks): ...new function.
10158 (pass_expand_omp): Replace declaration with that of...
10159 (make_pass_expand_omp): ...new function.
10160 (pass_expand_omp_ssa): Replace declaration with that of...
10161 (make_pass_expand_omp_ssa): ...new function.
10162 (pass_object_sizes): Replace declaration with that of...
10163 (make_pass_object_sizes): ...new function.
10164 (pass_strlen): Replace declaration with that of...
10165 (make_pass_strlen): ...new function.
10166 (pass_fold_builtins): Replace declaration with that of...
10167 (make_pass_fold_builtins): ...new function.
10168 (pass_stdarg): Replace declaration with that of...
10169 (make_pass_stdarg): ...new function.
10170 (pass_early_warn_uninitialized): Replace declaration with that of...
10171 (make_pass_early_warn_uninitialized): ...new function.
10172 (pass_late_warn_uninitialized): Replace declaration with that of...
10173 (make_pass_late_warn_uninitialized): ...new function.
10174 (pass_cse_reciprocals): Replace declaration with that of...
10175 (make_pass_cse_reciprocals): ...new function.
10176 (pass_cse_sincos): Replace declaration with that of...
10177 (make_pass_cse_sincos): ...new function.
10178 (pass_optimize_bswap): Replace declaration with that of...
10179 (make_pass_optimize_bswap): ...new function.
10180 (pass_optimize_widening_mul): Replace declaration with that of...
10181 (make_pass_optimize_widening_mul): ...new function.
10182 (pass_warn_function_return): Replace declaration with that of...
10183 (make_pass_warn_function_return): ...new function.
10184 (pass_warn_function_noreturn): Replace declaration with that of...
10185 (make_pass_warn_function_noreturn): ...new function.
10186 (pass_cselim): Replace declaration with that of...
10187 (make_pass_cselim): ...new function.
10188 (pass_phiopt): Replace declaration with that of...
10189 (make_pass_phiopt): ...new function.
10190 (pass_forwprop): Replace declaration with that of...
10191 (make_pass_forwprop): ...new function.
10192 (pass_phiprop): Replace declaration with that of...
10193 (make_pass_phiprop): ...new function.
10194 (pass_tree_ifcombine): Replace declaration with that of...
10195 (make_pass_tree_ifcombine): ...new function.
10196 (pass_dse): Replace declaration with that of...
10197 (make_pass_dse): ...new function.
10198 (pass_nrv): Replace declaration with that of...
10199 (make_pass_nrv): ...new function.
10200 (pass_rename_ssa_copies): Replace declaration with that of...
10201 (make_pass_rename_ssa_copies): ...new function.
10202 (pass_sink_code): Replace declaration with that of...
10203 (make_pass_sink_code): ...new function.
10204 (pass_fre): Replace declaration with that of...
10205 (make_pass_fre): ...new function.
10206 (pass_check_data_deps): Replace declaration with that of...
10207 (make_pass_check_data_deps): ...new function.
10208 (pass_copy_prop): Replace declaration with that of...
10209 (make_pass_copy_prop): ...new function.
10210 (pass_vrp): Replace declaration with that of...
10211 (make_pass_vrp): ...new function.
10212 (pass_uncprop): Replace declaration with that of...
10213 (make_pass_uncprop): ...new function.
10214 (pass_return_slot): Replace declaration with that of...
10215 (make_pass_return_slot): ...new function.
10216 (pass_reassoc): Replace declaration with that of...
10217 (make_pass_reassoc): ...new function.
10218 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10219 (make_pass_rebuild_cgraph_edges): ...new function.
10220 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10221 (make_pass_remove_cgraph_callee_edges): ...new function.
10222 (pass_build_cgraph_edges): Replace declaration with that of...
10223 (make_pass_build_cgraph_edges): ...new function.
10224 (pass_local_pure_const): Replace declaration with that of...
10225 (make_pass_local_pure_const): ...new function.
10226 (pass_tracer): Replace declaration with that of...
10227 (make_pass_tracer): ...new function.
10228 (pass_warn_unused_result): Replace declaration with that of...
10229 (make_pass_warn_unused_result): ...new function.
10230 (pass_diagnose_tm_blocks): Replace declaration with that of...
10231 (make_pass_diagnose_tm_blocks): ...new function.
10232 (pass_lower_tm): Replace declaration with that of...
10233 (make_pass_lower_tm): ...new function.
10234 (pass_tm_init): Replace declaration with that of...
10235 (make_pass_tm_init): ...new function.
10236 (pass_tm_mark): Replace declaration with that of...
10237 (make_pass_tm_mark): ...new function.
10238 (pass_tm_memopt): Replace declaration with that of...
10239 (make_pass_tm_memopt): ...new function.
10240 (pass_tm_edges): Replace declaration with that of...
10241 (make_pass_tm_edges): ...new function.
10242 (pass_split_functions): Replace declaration with that of...
10243 (make_pass_split_functions): ...new function.
10244 (pass_feedback_split_functions): Replace declaration with that of...
10245 (make_pass_feedback_split_functions): ...new function.
10246 (pass_strength_reduction): Replace declaration with that of...
10247 (make_pass_strength_reduction): ...new function.
10248 (pass_ipa_lower_emutls): Replace declaration with that of...
10249 (make_pass_ipa_lower_emutls): ...new function.
10250 (pass_ipa_function_and_variable_visibility): Replace declaration with
10252 (make_pass_ipa_function_and_variable_visibility): ...new function.
10253 (pass_ipa_tree_profile): Replace declaration with that of...
10254 (make_pass_ipa_tree_profile): ...new function.
10255 (pass_early_local_passes): Replace declaration with that of...
10256 (make_pass_early_local_passes): ...new function.
10257 (pass_ipa_whole_program_visibility): Replace declaration with that
10259 (make_pass_ipa_whole_program_visibility): ...new function.
10260 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10261 (make_pass_ipa_lto_gimple_out): ...new function.
10262 (pass_ipa_increase_alignment): Replace declaration with that of...
10263 (make_pass_ipa_increase_alignment): ...new function.
10264 (pass_ipa_inline): Replace declaration with that of...
10265 (make_pass_ipa_inline): ...new function.
10266 (pass_ipa_free_lang_data): Replace declaration with that of...
10267 (make_pass_ipa_free_lang_data): ...new function.
10268 (pass_ipa_free_inline_summary): Replace declaration with that of...
10269 (make_pass_ipa_free_inline_summary): ...new function.
10270 (pass_ipa_cp): Replace declaration with that of...
10271 (make_pass_ipa_cp): ...new function.
10272 (pass_ipa_reference): Replace declaration with that of...
10273 (make_pass_ipa_reference): ...new function.
10274 (pass_ipa_pure_const): Replace declaration with that of...
10275 (make_pass_ipa_pure_const): ...new function.
10276 (pass_ipa_pta): Replace declaration with that of...
10277 (make_pass_ipa_pta): ...new function.
10278 (pass_ipa_lto_finish_out): Replace declaration with that of...
10279 (make_pass_ipa_lto_finish_out): ...new function.
10280 (pass_ipa_tm): Replace declaration with that of...
10281 (make_pass_ipa_tm): ...new function.
10282 (pass_ipa_profile): Replace declaration with that of...
10283 (make_pass_ipa_profile): ...new function.
10284 (pass_ipa_cdtor_merge): Replace declaration with that of...
10285 (make_pass_ipa_cdtor_merge): ...new function.
10286 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10288 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10289 (pass_init_datastructures): Replace declaration with that of...
10290 (make_pass_init_datastructures): ...new function.
10291 (pass_fixup_cfg): Replace declaration with that of...
10292 (make_pass_fixup_cfg): ...new function.
10293 (pass_expand): Replace declaration with that of...
10294 (make_pass_expand): ...new function.
10295 (pass_instantiate_virtual_regs): Replace declaration with that of...
10296 (make_pass_instantiate_virtual_regs): ...new function.
10297 (pass_rtl_fwprop): Replace declaration with that of...
10298 (make_pass_rtl_fwprop): ...new function.
10299 (pass_rtl_fwprop_addr): Replace declaration with that of...
10300 (make_pass_rtl_fwprop_addr): ...new function.
10301 (pass_jump): Replace declaration with that of...
10302 (make_pass_jump): ...new function.
10303 (pass_jump2): Replace declaration with that of...
10304 (make_pass_jump2): ...new function.
10305 (pass_lower_subreg): Replace declaration with that of...
10306 (make_pass_lower_subreg): ...new function.
10307 (pass_cse): Replace declaration with that of...
10308 (make_pass_cse): ...new function.
10309 (pass_fast_rtl_dce): Replace declaration with that of...
10310 (make_pass_fast_rtl_dce): ...new function.
10311 (pass_ud_rtl_dce): Replace declaration with that of...
10312 (make_pass_ud_rtl_dce): ...new function.
10313 (pass_rtl_dce): Replace declaration with that of...
10314 (make_pass_rtl_dce): ...new function.
10315 (pass_rtl_dse1): Replace declaration with that of...
10316 (make_pass_rtl_dse1): ...new function.
10317 (pass_rtl_dse2): Replace declaration with that of...
10318 (make_pass_rtl_dse2): ...new function.
10319 (pass_rtl_dse3): Replace declaration with that of...
10320 (make_pass_rtl_dse3): ...new function.
10321 (pass_rtl_cprop): Replace declaration with that of...
10322 (make_pass_rtl_cprop): ...new function.
10323 (pass_rtl_pre): Replace declaration with that of...
10324 (make_pass_rtl_pre): ...new function.
10325 (pass_rtl_hoist): Replace declaration with that of...
10326 (make_pass_rtl_hoist): ...new function.
10327 (pass_rtl_store_motion): Replace declaration with that of...
10328 (make_pass_rtl_store_motion): ...new function.
10329 (pass_cse_after_global_opts): Replace declaration with that of...
10330 (make_pass_cse_after_global_opts): ...new function.
10331 (pass_rtl_ifcvt): Replace declaration with that of...
10332 (make_pass_rtl_ifcvt): ...new function.
10333 (pass_into_cfg_layout_mode): Replace declaration with that of...
10334 (make_pass_into_cfg_layout_mode): ...new function.
10335 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10336 (make_pass_outof_cfg_layout_mode): ...new function.
10337 (pass_loop2): Replace declaration with that of...
10338 (make_pass_loop2): ...new function.
10339 (pass_rtl_loop_init): Replace declaration with that of...
10340 (make_pass_rtl_loop_init): ...new function.
10341 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10342 (make_pass_rtl_move_loop_invariants): ...new function.
10343 (pass_rtl_unswitch): Replace declaration with that of...
10344 (make_pass_rtl_unswitch): ...new function.
10345 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10346 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10347 (pass_rtl_doloop): Replace declaration with that of...
10348 (make_pass_rtl_doloop): ...new function.
10349 (pass_rtl_loop_done): Replace declaration with that of...
10350 (make_pass_rtl_loop_done): ...new function.
10351 (pass_web): Replace declaration with that of...
10352 (make_pass_web): ...new function.
10353 (pass_cse2): Replace declaration with that of...
10354 (make_pass_cse2): ...new function.
10355 (pass_df_initialize_opt): Replace declaration with that of...
10356 (make_pass_df_initialize_opt): ...new function.
10357 (pass_df_initialize_no_opt): Replace declaration with that of...
10358 (make_pass_df_initialize_no_opt): ...new function.
10359 (pass_reginfo_init): Replace declaration with that of...
10360 (make_pass_reginfo_init): ...new function.
10361 (pass_inc_dec): Replace declaration with that of...
10362 (make_pass_inc_dec): ...new function.
10363 (pass_stack_ptr_mod): Replace declaration with that of...
10364 (make_pass_stack_ptr_mod): ...new function.
10365 (pass_initialize_regs): Replace declaration with that of...
10366 (make_pass_initialize_regs): ...new function.
10367 (pass_combine): Replace declaration with that of...
10368 (make_pass_combine): ...new function.
10369 (pass_if_after_combine): Replace declaration with that of...
10370 (make_pass_if_after_combine): ...new function.
10371 (pass_ree): Replace declaration with that of...
10372 (make_pass_ree): ...new function.
10373 (pass_partition_blocks): Replace declaration with that of...
10374 (make_pass_partition_blocks): ...new function.
10375 (pass_match_asm_constraints): Replace declaration with that of...
10376 (make_pass_match_asm_constraints): ...new function.
10377 (pass_regmove): Replace declaration with that of...
10378 (make_pass_regmove): ...new function.
10379 (pass_split_all_insns): Replace declaration with that of...
10380 (make_pass_split_all_insns): ...new function.
10381 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10382 (make_pass_fast_rtl_byte_dce): ...new function.
10383 (pass_lower_subreg2): Replace declaration with that of...
10384 (make_pass_lower_subreg2): ...new function.
10385 (pass_mode_switching): Replace declaration with that of...
10386 (make_pass_mode_switching): ...new function.
10387 (pass_sms): Replace declaration with that of...
10388 (make_pass_sms): ...new function.
10389 (pass_sched): Replace declaration with that of...
10390 (make_pass_sched): ...new function.
10391 (pass_ira): Replace declaration with that of...
10392 (make_pass_ira): ...new function.
10393 (pass_reload): Replace declaration with that of...
10394 (make_pass_reload): ...new function.
10395 (pass_clean_state): Replace declaration with that of...
10396 (make_pass_clean_state): ...new function.
10397 (pass_branch_prob): Replace declaration with that of...
10398 (make_pass_branch_prob): ...new function.
10399 (pass_value_profile_transformations): Replace declaration with that
10401 (make_pass_value_profile_transformations): ...new function.
10402 (pass_postreload_cse): Replace declaration with that of...
10403 (make_pass_postreload_cse): ...new function.
10404 (pass_gcse2): Replace declaration with that of...
10405 (make_pass_gcse2): ...new function.
10406 (pass_split_after_reload): Replace declaration with that of...
10407 (make_pass_split_after_reload): ...new function.
10408 (pass_branch_target_load_optimize1): Replace declaration with that
10410 (make_pass_branch_target_load_optimize1): ...new function.
10411 (pass_thread_prologue_and_epilogue): Replace declaration with that
10413 (make_pass_thread_prologue_and_epilogue): ...new function.
10414 (pass_stack_adjustments): Replace declaration with that of...
10415 (make_pass_stack_adjustments): ...new function.
10416 (pass_peephole2): Replace declaration with that of...
10417 (make_pass_peephole2): ...new function.
10418 (pass_if_after_reload): Replace declaration with that of...
10419 (make_pass_if_after_reload): ...new function.
10420 (pass_regrename): Replace declaration with that of...
10421 (make_pass_regrename): ...new function.
10422 (pass_cprop_hardreg): Replace declaration with that of...
10423 (make_pass_cprop_hardreg): ...new function.
10424 (pass_reorder_blocks): Replace declaration with that of...
10425 (make_pass_reorder_blocks): ...new function.
10426 (pass_branch_target_load_optimize2): Replace declaration with that
10428 (make_pass_branch_target_load_optimize2): ...new function.
10429 (pass_leaf_regs): Replace declaration with that of...
10430 (make_pass_leaf_regs): ...new function.
10431 (pass_split_before_sched2): Replace declaration with that of...
10432 (make_pass_split_before_sched2): ...new function.
10433 (pass_compare_elim_after_reload): Replace declaration with that of...
10434 (make_pass_compare_elim_after_reload): ...new function.
10435 (pass_sched2): Replace declaration with that of...
10436 (make_pass_sched2): ...new function.
10437 (pass_stack_regs): Replace declaration with that of...
10438 (make_pass_stack_regs): ...new function.
10439 (pass_stack_regs_run): Replace declaration with that of...
10440 (make_pass_stack_regs_run): ...new function.
10441 (pass_df_finish): Replace declaration with that of...
10442 (make_pass_df_finish): ...new function.
10443 (pass_compute_alignments): Replace declaration with that of...
10444 (make_pass_compute_alignments): ...new function.
10445 (pass_duplicate_computed_gotos): Replace declaration with that of...
10446 (make_pass_duplicate_computed_gotos): ...new function.
10447 (pass_variable_tracking): Replace declaration with that of...
10448 (make_pass_variable_tracking): ...new function.
10449 (pass_free_cfg): Replace declaration with that of...
10450 (make_pass_free_cfg): ...new function.
10451 (pass_machine_reorg): Replace declaration with that of...
10452 (make_pass_machine_reorg): ...new function.
10453 (pass_cleanup_barriers): Replace declaration with that of...
10454 (make_pass_cleanup_barriers): ...new function.
10455 (pass_delay_slots): Replace declaration with that of...
10456 (make_pass_delay_slots): ...new function.
10457 (pass_split_for_shorten_branches): Replace declaration with that of...
10458 (make_pass_split_for_shorten_branches): ...new function.
10459 (pass_split_before_regstack): Replace declaration with that of...
10460 (make_pass_split_before_regstack): ...new function.
10461 (pass_convert_to_eh_region_ranges): Replace declaration with that
10463 (make_pass_convert_to_eh_region_ranges): ...new function.
10464 (pass_shorten_branches): Replace declaration with that of...
10465 (make_pass_shorten_branches): ...new function.
10466 (pass_set_nothrow_function_flags): Replace declaration with that of...
10467 (make_pass_set_nothrow_function_flags): ...new function.
10468 (pass_dwarf2_frame): Replace declaration with that of...
10469 (make_pass_dwarf2_frame): ...new function.
10470 (pass_final): Replace declaration with that of...
10471 (make_pass_final): ...new function.
10472 (pass_rtl_seqabstr): Replace declaration with that of...
10473 (make_pass_rtl_seqabstr): ...new function.
10474 (pass_release_ssa_names): Replace declaration with that of...
10475 (make_pass_release_ssa_names): ...new function.
10476 (pass_early_inline): Replace declaration with that of...
10477 (make_pass_early_inline): ...new function.
10478 (pass_inline_parameters): Replace declaration with that of...
10479 (make_pass_inline_parameters): ...new function.
10480 (pass_update_address_taken): Replace declaration with that of...
10481 (make_pass_update_address_taken): ...new function.
10482 (pass_convert_switch): Replace declaration with that of...
10483 (make_pass_convert_switch): ...new function.
10484 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10485 to a subclass of simple_ipa_opt_pass along with...
10486 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10487 (make_pass_ipa_tree_profile): ...new function.
10488 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10489 subclass of gimple_opt_pass along with...
10490 (pass_data_sra_early): ...new pass_data instance and...
10491 (make_pass_sra_early): ...new function.
10492 (pass_sra): Convert from a global struct to a subclass of
10493 gimple_opt_pass along with...
10494 (pass_data_sra): ...new pass_data instance and...
10495 (make_pass_sra): ...new function.
10496 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10497 gimple_opt_pass along with...
10498 (pass_data_early_ipa_sra): ...new pass_data instance and...
10499 (make_pass_early_ipa_sra): ...new function.
10500 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10501 subclass of gimple_opt_pass along with...
10502 (pass_data_ccp): ...new pass_data instance and...
10503 (make_pass_ccp): ...new function.
10504 (pass_fold_builtins): Convert from a global struct to a subclass of
10505 gimple_opt_pass along with...
10506 (pass_data_fold_builtins): ...new pass_data instance and...
10507 (make_pass_fold_builtins): ...new function.
10508 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10509 subclass of gimple_opt_pass along with...
10510 (pass_data_copy_prop): ...new pass_data instance and...
10511 (make_pass_copy_prop): ...new function.
10512 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10513 global struct to a subclass of gimple_opt_pass along with...
10514 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10515 (make_pass_rename_ssa_copies): ...new function.
10516 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10517 subclass of gimple_opt_pass along with...
10518 (pass_data_dce): ...new pass_data instance and...
10519 (make_pass_dce): ...new function.
10520 (pass_dce_loop): Convert from a global struct to a subclass of
10521 gimple_opt_pass along with...
10522 (pass_data_dce_loop): ...new pass_data instance and...
10523 (make_pass_dce_loop): ...new function.
10524 (pass_cd_dce): Convert from a global struct to a subclass of
10525 gimple_opt_pass along with...
10526 (pass_data_cd_dce): ...new pass_data instance and...
10527 (make_pass_cd_dce): ...new function.
10528 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10529 subclass of gimple_opt_pass along with...
10530 (pass_data_dominator): ...new pass_data instance and...
10531 (make_pass_dominator): ...new function.
10532 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10533 gimple_opt_pass along with...
10534 (pass_data_phi_only_cprop): ...new pass_data instance and...
10535 (make_pass_phi_only_cprop): ...new function.
10536 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10537 subclass of gimple_opt_pass along with...
10538 (pass_data_dse): ...new pass_data instance and...
10539 (make_pass_dse): ...new function.
10540 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10541 a subclass of gimple_opt_pass along with...
10542 (pass_data_forwprop): ...new pass_data instance and...
10543 (make_pass_forwprop): ...new function.
10544 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10545 struct to a subclass of gimple_opt_pass along with...
10546 (pass_data_tree_ifcombine): ...new pass_data instance and...
10547 (make_pass_tree_ifcombine): ...new function.
10548 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10549 subclass of gimple_opt_pass along with...
10550 (pass_data_ch): ...new pass_data instance and...
10551 (make_pass_ch): ...new function.
10552 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10553 subclass of gimple_opt_pass along with...
10554 (pass_data_tree_loop): ...new pass_data instance and...
10555 (make_pass_tree_loop): ...new function.
10556 (pass_tree_loop_init): Convert from a global struct to a subclass of
10557 gimple_opt_pass along with...
10558 (pass_data_tree_loop_init): ...new pass_data instance and...
10559 (make_pass_tree_loop_init): ...new function.
10560 (pass_lim): Convert from a global struct to a subclass of
10561 gimple_opt_pass along with...
10562 (pass_data_lim): ...new pass_data instance and...
10563 (make_pass_lim): ...new function.
10564 (pass_tree_unswitch): Convert from a global struct to a subclass of
10565 gimple_opt_pass along with...
10566 (pass_data_tree_unswitch): ...new pass_data instance and...
10567 (make_pass_tree_unswitch): ...new function.
10568 (pass_predcom): Convert from a global struct to a subclass of
10569 gimple_opt_pass along with...
10570 (pass_data_predcom): ...new pass_data instance and...
10571 (make_pass_predcom): ...new function.
10572 (pass_vectorize): Convert from a global struct to a subclass of
10573 gimple_opt_pass along with...
10574 (pass_data_vectorize): ...new pass_data instance and...
10575 (make_pass_vectorize): ...new function.
10576 (pass_graphite): Convert from a global struct to a subclass of
10577 gimple_opt_pass along with...
10578 (pass_data_graphite): ...new pass_data instance and...
10579 (make_pass_graphite): ...new function.
10580 (pass_graphite_transforms): Convert from a global struct to a subclass
10581 of gimple_opt_pass along with...
10582 (pass_data_graphite_transforms): ...new pass_data instance and...
10583 (make_pass_graphite_transforms): ...new function.
10584 (pass_check_data_deps): Convert from a global struct to a subclass of
10585 gimple_opt_pass along with...
10586 (pass_data_check_data_deps): ...new pass_data instance and...
10587 (make_pass_check_data_deps): ...new function.
10588 (pass_iv_canon): Convert from a global struct to a subclass of
10589 gimple_opt_pass along with...
10590 (pass_data_iv_canon): ...new pass_data instance and...
10591 (make_pass_iv_canon): ...new function.
10592 (pass_scev_cprop): Convert from a global struct to a subclass of
10593 gimple_opt_pass along with...
10594 (pass_data_scev_cprop): ...new pass_data instance and...
10595 (make_pass_scev_cprop): ...new function.
10596 (pass_record_bounds): Convert from a global struct to a subclass of
10597 gimple_opt_pass along with...
10598 (pass_data_record_bounds): ...new pass_data instance and...
10599 (make_pass_record_bounds): ...new function.
10600 (pass_complete_unroll): Convert from a global struct to a subclass of
10601 gimple_opt_pass along with...
10602 (pass_data_complete_unroll): ...new pass_data instance and...
10603 (make_pass_complete_unroll): ...new function.
10604 (pass_complete_unrolli): Convert from a global struct to a subclass of
10605 gimple_opt_pass along with...
10606 (pass_data_complete_unrolli): ...new pass_data instance and...
10607 (make_pass_complete_unrolli): ...new function.
10608 (pass_parallelize_loops): Convert from a global struct to a subclass
10609 of gimple_opt_pass along with...
10610 (pass_data_parallelize_loops): ...new pass_data instance and...
10611 (make_pass_parallelize_loops): ...new function.
10612 (pass_loop_prefetch): Convert from a global struct to a subclass of
10613 gimple_opt_pass along with...
10614 (pass_data_loop_prefetch): ...new pass_data instance and...
10615 (make_pass_loop_prefetch): ...new function.
10616 (pass_iv_optimize): Convert from a global struct to a subclass of
10617 gimple_opt_pass along with...
10618 (pass_data_iv_optimize): ...new pass_data instance and...
10619 (make_pass_iv_optimize): ...new function.
10620 (pass_tree_loop_done): Convert from a global struct to a subclass of
10621 gimple_opt_pass along with...
10622 (pass_data_tree_loop_done): ...new pass_data instance and...
10623 (make_pass_tree_loop_done): ...new function.
10624 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10625 struct to a subclass of gimple_opt_pass along with...
10626 (pass_data_cse_reciprocals): ...new pass_data instance and...
10627 (make_pass_cse_reciprocals): ...new function.
10628 (pass_cse_sincos): Convert from a global struct to a subclass of
10629 gimple_opt_pass along with...
10630 (pass_data_cse_sincos): ...new pass_data instance and...
10631 (make_pass_cse_sincos): ...new function.
10632 (pass_optimize_bswap): Convert from a global struct to a subclass of
10633 gimple_opt_pass along with...
10634 (pass_data_optimize_bswap): ...new pass_data instance and...
10635 (make_pass_optimize_bswap): ...new function.
10636 (pass_optimize_widening_mul): Convert from a global struct to a
10637 subclass of gimple_opt_pass along with...
10638 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10639 (make_pass_optimize_widening_mul): ...new function.
10640 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10641 subclass of gimple_opt_pass along with...
10642 (pass_data_phiopt): ...new pass_data instance and...
10643 (make_pass_phiopt): ...new function.
10644 (pass_cselim): Convert from a global struct to a subclass of
10645 gimple_opt_pass along with...
10646 (pass_data_cselim): ...new pass_data instance and...
10647 (make_pass_cselim): ...new function.
10648 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10649 subclass of gimple_opt_pass along with...
10650 (pass_data_phiprop): ...new pass_data instance and...
10651 (make_pass_phiprop): ...new function.
10652 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10653 subclass of gimple_opt_pass along with...
10654 (pass_data_pre): ...new pass_data instance and...
10655 (make_pass_pre): ...new function.
10656 (pass_fre): Convert from a global struct to a subclass of
10657 gimple_opt_pass along with...
10658 (pass_data_fre): ...new pass_data instance and...
10659 (make_pass_fre): ...new function.
10660 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10661 subclass of gimple_opt_pass along with...
10662 (pass_data_reassoc): ...new pass_data instance and...
10663 (make_pass_reassoc): ...new function.
10664 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10665 subclass of gimple_opt_pass along with...
10666 (pass_data_sink_code): ...new pass_data instance and...
10667 (make_pass_sink_code): ...new function.
10668 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10669 subclass of gimple_opt_pass along with...
10670 (pass_data_strlen): ...new pass_data instance and...
10671 (make_pass_strlen): ...new function.
10672 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10673 struct to a subclass of gimple_opt_pass along with...
10674 (pass_data_build_alias): ...new pass_data instance and...
10675 (make_pass_build_alias): ...new function.
10676 (pass_build_ealias): Convert from a global struct to a subclass of
10677 gimple_opt_pass along with...
10678 (pass_data_build_ealias): ...new pass_data instance and...
10679 (make_pass_build_ealias): ...new function.
10680 (pass_ipa_pta): Convert from a global struct to a subclass of
10681 simple_ipa_opt_pass along with...
10682 (pass_data_ipa_pta): ...new pass_data instance and...
10683 (make_pass_ipa_pta): ...new function.
10684 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10685 subclass of gimple_opt_pass along with...
10686 (pass_data_uncprop): ...new pass_data instance and...
10687 (make_pass_uncprop): ...new function.
10688 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10689 global struct to a subclass of gimple_opt_pass along with...
10690 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10691 (make_pass_late_warn_uninitialized): ...new function.
10692 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10693 to a subclass of gimple_opt_pass along with...
10694 (pass_data_init_datastructures): ...new pass_data instance and...
10695 (make_pass_init_datastructures): ...new function.
10696 (pass_early_warn_uninitialized): Convert from a global struct to a
10697 subclass of gimple_opt_pass along with...
10698 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10699 (make_pass_early_warn_uninitialized): ...new function.
10700 (pass_update_address_taken): Convert from a global struct to a
10701 subclass of gimple_opt_pass along with...
10702 (pass_data_update_address_taken): ...new pass_data instance and...
10703 (make_pass_update_address_taken): ...new function.
10704 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10705 struct to a subclass of gimple_opt_pass along with...
10706 (pass_data_release_ssa_names): ...new pass_data instance and...
10707 (make_pass_release_ssa_names): ...new function.
10708 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10709 subclass of gimple_opt_pass along with...
10710 (pass_data_stdarg): ...new pass_data instance and...
10711 (make_pass_stdarg): ...new function.
10712 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10713 global struct to a subclass of gimple_opt_pass along with...
10714 (pass_data_convert_switch): ...new pass_data instance and...
10715 (make_pass_convert_switch): ...new function.
10716 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10717 to a subclass of gimple_opt_pass along with...
10718 (pass_data_tail_recursion): ...new pass_data instance and...
10719 (make_pass_tail_recursion): ...new function.
10720 (pass_tail_calls): Convert from a global struct to a subclass of
10721 gimple_opt_pass along with...
10722 (pass_data_tail_calls): ...new pass_data instance and...
10723 (make_pass_tail_calls): ...new function.
10724 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10725 struct to a subclass of gimple_opt_pass along with...
10726 (pass_data_lower_vector): ...new pass_data instance and...
10727 (make_pass_lower_vector): ...new function.
10728 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10729 gimple_opt_pass along with...
10730 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10731 (make_pass_lower_vector_ssa): ...new function.
10732 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10733 to a subclass of gimple_opt_pass along with...
10734 (pass_data_slp_vectorize): ...new pass_data instance and...
10735 (make_pass_slp_vectorize): ...new function.
10736 (pass_ipa_increase_alignment): Convert from a global struct to a
10737 subclass of simple_ipa_opt_pass along with...
10738 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10739 (make_pass_ipa_increase_alignment): ...new function.
10740 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10741 gimple_opt_pass along with...
10742 (pass_data_vrp): ...new pass_data instance and...
10743 (make_pass_vrp): ...new function.
10744 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10745 subclass of simple_ipa_opt_pass along with...
10746 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10747 (make_pass_ipa_free_lang_data): ...new function.
10748 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10749 gimple_opt_pass along with...
10750 (pass_data_tsan): ...new pass_data instance and...
10751 (make_pass_tsan): ...new function.
10752 (pass_tsan_O0): Convert from a global struct to a subclass of
10753 gimple_opt_pass along with...
10754 (pass_data_tsan_O0): ...new pass_data instance and...
10755 (make_pass_tsan_O0): ...new function.
10756 * var-tracking.c (pass_variable_tracking): Convert from a global
10757 struct to a subclass of rtl_opt_pass along with...
10758 (pass_data_variable_tracking): ...new pass_data instance and...
10759 (make_pass_variable_tracking): ...new function.
10760 * web.c (pass_web): Convert from a global struct to a subclass of
10761 rtl_opt_pass along with...
10762 (pass_data_web): ...new pass_data instance and...
10763 (make_pass_web): ...new function.
10764 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10765 declaration with that of...
10766 (make_pass_mode_switch_use): ...new function.
10767 (pass_resolve_sw_modes): Replace declaration with that of...
10768 (make_pass_resolve_sw_modes): ...new function.
10769 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10770 from a global struct to a subclass of rtl_opt_pass along with...
10771 (pass_data_mode_switch_use): ...new pass_data instance and...
10772 (make_pass_mode_switch_use): ...new function.
10773 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10774 from a global struct to a subclass of rtl_opt_pass along with...
10775 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10776 (make_pass_resolve_sw_modes): ...new function.
10777 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10778 struct to a subclass of rtl_opt_pass along with...
10779 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10780 (make_pass_insert_vzeroupper): ...new function.
10781 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10782 global struct to a subclass of rtl_opt_pass along with...
10783 (pass_data_work_around_errata): ...new pass_data instance and...
10784 (make_pass_work_around_errata): ...new function.
10785 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10786 struct to a subclass of rtl_opt_pass along with...
10787 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10788 (make_pass_mips_machine_reorg2): ...new function.
10790 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10792 * passes.c (pass_manager::operator new): New.
10794 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10796 Handwritten part of conversion of passes to C++ classes.
10798 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10799 (toplev.o): Add dep on PASS_MANAGER_H.
10800 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10801 of early local pases to reflect this moving from a global to a
10802 member of gcc::pass_manager.
10803 (cgraph_add_new_function): Likewise.
10804 * lto-cgraph.c (lto_output_node): Update for conversion of
10805 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10806 * passes.c (opt_pass::clone): New.
10807 (opt_pass::gate): New.
10808 (opt_pass::execute): New.
10809 (opt_pass::opt_pass): New.
10810 (pass_manager::execute_early_local_passes): New.
10811 (pass_manager::execute_pass_mode_switching): new.
10812 (finish_optimization_passes): Convert to...
10813 (pass_manager::finish_optimization_passes): ...this.
10814 (finish_optimization_passes): Update for conversion of passes to
10816 (register_dump_files_1): Use has_gate since we cannot portably
10817 check a vtable entry against NULL.
10818 (dump_one_pass): Likewise.
10819 (ipa_write_summaries_2): Likewise.
10820 (ipa_write_optimization_summaries_1): Likewise.
10821 (ipa_read_summaries_1): Likewise.
10822 (ipa_read_optimization_summaries_1): Likewise.
10823 (execute_ipa_stmt_fixups): Likewise.
10824 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10825 pass-creation functions rather than wiring up globals, and
10826 storing the results in fields of pass_manager generated using
10827 pass-instances.def.
10828 (pass_manager::dump_profile_report): Update for conversion of
10829 passes to C++ classes.
10830 (pass_manager::execute_ipa_summary_passes): Likewise.
10831 (execute_one_ipa_transform_pass): Likewise.
10832 (execute_one_pass): Use has_gate and has_execute since we cannot
10833 portably check a vtable entry against NULL.
10834 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10835 (pass_manager): Use pass-instances.def to add fields for the
10836 various pass instances.
10837 * toplev.c (finalize): Update for move of
10838 finish_optimization_passes to a method of gcc::pass_manager.
10839 * toplev.h (finish_optimization_passes): Move to method of class
10841 * tree-pass.h (struct pass_data): New.
10842 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10843 (opt_pass::gate): Convert to virtual function.
10844 (opt_pass::~opt_pass): New.
10845 (opt_pass::clone): New.
10846 (opt_pass::execute): Convert to virtual function.
10847 (opt_pass::opt_pass): New.
10848 (opt_pass::ctxt_): new.
10849 (gimple_opt_pass): Convert to subclass of opt_pass.
10850 (gimple_opt_pass::gimple_opt_pass): New.
10851 (rtl_opt_pass): Convert to subclass of opt_pass.
10852 (rtl_opt_pass::rtl_opt_pass): New.
10853 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10854 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10855 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10856 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10857 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10858 invocation of pass_mode_switching to reflect this moving from a
10859 global to a member of gcc::pass_manager.
10860 (ix86_option_override): Rework how pass_insert_vzeroupper is
10861 added to the pass_manager to reflect autogenerated changes.
10862 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10864 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10866 PR rtl-optimization/57708
10867 * recog.c (peep2_find_free_register): Validate all regs in a
10870 2013-08-05 Jan Hubicka <jh@suse.cz>
10873 * cgraph.c (verify_cgraph_node): Accept local flags from other
10875 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10876 (function_and_variable_visibility): Likewise.
10877 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10879 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10881 * graph.c (init_graph_slim_pretty_print): Remove.
10882 (print_graph_cfg): Do not call it. Use local pretty printer.
10883 (start_graph_dump): Likewise.
10885 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10887 * gimple-pretty-print.c (buffer): Remove.
10888 (initialized): Likewise.
10889 (maybe_init_pretty_print): Likewise.
10890 (print_gimple_stmt): Do not call it. Use non-static local
10891 pretty_printer variable.
10892 (print_gimple_expr): Likewise.
10893 (print_gimple_seq): Likewise.
10894 (gimple_dump_bb): Likewise.
10896 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10898 * asan.c (asan_pp): Remove.
10899 (asan_pp_initialized): Likewise.
10900 (asan_pp_initialize): Likewise.
10901 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10902 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10903 (asan_add_global): Likewise.
10905 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10907 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10908 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10909 * pretty-print.c (pp_formatted_text_data): Likewise.
10910 (pp_indent): Rename from pp_base_indent.
10911 (pp_format): Rename from pp_base_format.
10912 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10913 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10914 (pp_flush): Rename from pp_base_flush.
10915 (pp_set_line_maximum_length): Rename from
10916 pp_base_set_line_maximum_length.
10917 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10918 (pp_set_prefix): Rename from pp_base_set_prefix.
10919 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10920 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10921 (pp_append_text): Rename from pp_base_append_text.
10922 (pp_formatted_text): Rename from pp_base_formatted_text.
10923 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10924 (pp_remaining_character_count_for_line): Rename from
10925 pp_base_remaining_character_count_for_line.
10926 (pp_newline): Rename from pp_base_newline.
10927 (pp_character): Rename from pp_base_character.
10928 (pp_string): Rename from pp_base_string.
10929 (pp_maybe_space): Rename from pp_base_maybe_space.
10930 * asan.c (asan_pp_string): Adjust.
10931 (asan_emit_stack_protection): Likewise.
10932 (asan_add_global): Likewise.
10933 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10934 * tree-mudflap.c (mf_varname_tree): Likewise.
10935 * tree-pretty-print.c (pp_tree_identifier): Rename from
10936 pp_base_tree_identifier.
10937 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10938 Declare as function.
10940 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10942 * pretty-print.h (pp_bar_bar): New.
10943 (pp_ampersand_ampersand): Likewise.
10944 (pp_less_equal): Likewise.
10945 (pp_greater_equal): Likewise.
10946 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10947 printer functions instead of pp_string or operators and punctuators.
10948 (dump_gimple_call): Likewise.
10949 (dump_gimple_omp_for): Likewise.
10950 (dump_gimple_transaction): Likewise.
10951 (dump_gimple_phi): Likewise.
10952 (pp_gimple_stmt_1): Likewise.
10953 * sched-vis.c (print_insn): Likewise.
10954 * tree-mudflap.c (mf_varname_tree): Likewise.
10955 * tree-pretty-print.c (dump_block_node): Likewise.
10956 (dump_generic_node): Likewise.
10958 2013-08-02 Jan Hubicka <jh@suse.cz>
10960 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10962 * lto-streamer-out.c (tree_is_indexable): Results decls and
10963 parm decls are not indexable.
10964 (DFS_write_tree_body): Do not follow args and results.
10965 (hash_tree): Likewise.
10966 (output_functions): Rearrange so struct function is needed
10967 only when real body is output; be able to also ouptut abstract
10968 functions; output DECL_ARGUMENTS and DECL_RESULT.
10969 (lto_output): When not in WPA, ale store abstract functions.
10970 (write_symbol): Do not care about RESULT_DECL.
10971 (output_symbol_p): Handle correctly sbtract decls.
10972 * lto-streamer-in.c (input_function): Rearrange so struct
10973 function can be NULL at entry; allow streaming of
10974 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10975 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10976 sanity check during LTO.
10977 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10978 RESULT_DECl and DECL_ARGUMENTS.
10979 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10982 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10984 * pretty-print.h (pp_underscore): New.
10986 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10987 printer functions instead of pp_character.
10988 (dump_binary_rhs): Likewise.
10989 (dump_ternary_rhs): Likewise.
10990 (dump_gimple_call_args): Likewise.
10991 (pp_points_to_solution): Likewise.
10992 (dump_gimple_call): Likewise.
10993 (dump_gimple_switch): Likewise.
10994 (dump_gimple_cond): Likewise.
10995 (dump_gimple_bind): Likewise.
10996 (dump_gimple_try): Likewise.
10997 (dump_gimple_omp_for): Likewise.
10998 (dump_gimple_omp_continue): Likewise.
10999 (dump_gimple_omp_single): Likewise.
11000 (dump_gimple_omp_sections): Likewise.
11001 (dump_gimple_omp_block): Likewise.
11002 (dump_gimple_omp_critical): Likewise.
11003 (dump_gimple_transaction): Likewise.
11004 (dump_gimple_asm): Likewise.
11005 (dump_gimple_phi): Likewise.
11006 (dump_gimple_omp_parallel): Likewise.
11007 (dump_gimple_omp_task): Likewise.
11008 (dump_gimple_omp_atomic_load): Likewise.
11009 (dump_gimple_omp_atomic_store): Likewise.
11010 (dump_gimple_mem_ops): Likewise.
11011 (pp_gimple_stmt_1): Likewise.
11012 (pp_cfg_jump): Likewise.
11013 (dump_implicit_edges): Likewise.
11014 (gimple_dump_bb_for_graph): Likewise.
11015 * graph.c (draw_cfg_node): Likewise.
11016 * langhooks.c (lhd_print_error_function): Likewise.
11017 * sched-vis.c (print_exp): Likewise.
11018 (print_value): Likewise.
11019 (print_pattern): Likewise.
11020 (print_insn): Likewise.
11021 (rtl_dump_bb_for_graph): Likewise.
11022 * tree-pretty-print.c (dump_function_declaration): Likewise.
11023 (dump_array_domain): Likewise.
11024 (dump_omp_clause): Likewise.
11025 (dump_location): Likewise.
11026 (dump_generic_node): Likewise.
11027 (print_struct_decl): Likewise.
11028 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11030 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11032 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11033 candidate table when replacing a candidate statement.
11034 (replace_rhs_if_not_dup): Likewise.
11035 (replace_one_candidate): Likewise.
11037 2013-08-02 Jan Hubicka <jh@suse.cz>
11038 Martin Liska <marxin.liska@gmail.com>
11040 * cgraphunit.c (add_new_function): Fix logic when adding from
11042 (assemble_thunk): Rename to ...
11043 (expand_thunk); .. this one; export; get it working with
11044 general functions; make produced gimple valid.
11045 * cgraph.h (expand_thunk): Declare.
11047 2013-08-02 Jan Hubicka <jh@suse.cz>
11049 * ipa-cp.c (gather_context_independent_values): Use
11050 ipa_get_param_move_cost.
11051 (get_replacement_map): Remove PARAM; move parameter folding
11053 (create_specialized_node): Update.
11054 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11055 assert that we have gimple body; update move_cost.
11056 (count_formal_params): Assert that we have gimple body.
11057 (ipa_dump_param): New function.
11058 (ipa_alloc_node_params): Break out from ...
11059 (ipa_initialize_node_params): ... here.
11060 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11061 (ipa_write_node_info): Stream move costs.
11062 (ipa_read_node_info): Read move costs.
11063 (ipa_update_after_lto_read): Do not recompute node params.
11064 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11065 (ipa_get_param): Check we are not in WPA.
11066 (ipa_get_param_move_cost): New.
11067 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11068 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11069 parm numbers to be present.
11071 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11073 PR rtl-optimization/58048
11074 * lra-constraints.c (process_alt_operands): Don't check asm
11075 operand on register.
11077 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11079 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11080 the implied StoreLoad barrier for atomic operations if before.
11082 2013-08-02 Jan Hubicka <jh@suse.cz>
11083 Martin Liska <marxin.liska@gmail.com>
11085 * cgraph.c (cgraph_function_body_availability): Do not check
11087 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11088 symtab_node_availability): Declare.
11089 * ipa.c (can_replace_by_local_alias): New.
11090 (function_and_variable_visibility): Use it.
11091 * symtab.c (symtab_for_node_and_aliases,
11092 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11094 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11096 PR rtl-optimization/57963
11097 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11098 (lra_constraints): Use them.
11100 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11102 * config/arm/types.md (define_attr "type"): Add "load_acq"
11104 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11106 (cortex_a53_store1): Likewise.
11108 2013-08-01 Jan Hubicka <jh@suse.cz>
11110 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11111 partitions are not needed.
11113 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11115 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11116 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11117 MAYBE_NON_Q_CLASS_P where appropriate.
11119 2013-08-01 Jan Hubicka <jh@suse.cz>
11121 * cgraph.h (release_function_body): Declare.
11122 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11123 of unused delcarations.
11125 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11127 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11128 RTL form when subtracting a constant.
11130 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11132 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11133 Generate canonical plus rtx with negated immediate instead of minus
11135 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11137 2013-08-01 Jan Hubicka <jh@suse.cz>
11139 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11140 (cgraph_release_function_body): Likewise.
11141 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11142 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11143 to used_as_abstract_origin.
11144 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11145 symbols used as abstract origins.
11146 * cgraphunit.c (analyze_functions): Update.
11147 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11148 used_as_abstract_origin.
11149 * tree-inline.c (tree_function_versioning): Update
11150 used_as_abstract_origin; be ready for DECL_RESULT and
11151 DECL_ARGUMENTS to be NULL.
11153 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11154 for abstract functions.
11155 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11158 2013-08-01 Jan Hubicka <jh@suse.cz>
11160 * profile.c (compute_value_histograms): Fix thinko.
11162 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11164 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11165 aarch-common-protos.h to extra_headers.
11166 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11167 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11168 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11170 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11172 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11173 Include "../arm/types.md". Define "type" attribute for all patterns.
11174 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11177 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11179 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11180 to support power8 load fusion.
11181 (fusion_gpr_mem_load): Likewise.
11183 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11185 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11186 declarations for power8 load fusion.
11187 (emit_fusion_gpr_load): Likewise.
11189 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11190 tuning for power8, turn on fusion mode by default. Turn on sign
11191 extending fusion mode if normal fusion mode is on, and we are at
11193 (fusion_gpr_load_p): New function, return true if we can fuse an
11194 addis instruction with a dependent load to a GPR.
11195 (emit_fusion_gpr_load): Emit the instructions for power8 load
11198 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11199 (VSX load fusion peepholes): New peepholes to fuse together an
11200 addi instruction with a VSX load instruction.
11202 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11203 peepholes to fuse an addis instruction with a load to a GPR base
11204 register. If we are supporting sign extending fusions, convert
11205 sign extending loads to zero extending loads and add an explicit
11208 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11210 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11211 aarch-common-protos.h to extra_headers.
11212 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11213 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11214 (arm_early_store_addr_dep): Likewise.
11215 (arm_no_early_alu_shift_dep): Likewise.
11216 (arm_no_early_alu_shift_value_dep): Likewise.
11217 (arm_no_early_mul_dep): Likewise.
11218 (arm_no_early_store_addr_dep): Likewise.
11219 (arm_mac_accumulator_is_mul_result): Likewise.
11220 (arm_mac_accumulator_is_result): Likewise.
11221 * config/arm/aarch-common.c: ... here. New file.
11222 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11224 (arm_early_store_addr_dep): Likewise.
11225 (arm_no_early_alu_shift_dep): Likewise.
11226 (arm_no_early_alu_shift_value_dep): Likewise.
11227 (arm_no_early_mul_dep): Likewise.
11228 (arm_no_early_store_addr_dep): Likewise.
11229 (arm_mac_accumulator_is_mul_result): Likewise.
11230 (arm_mac_accumulator_is_result): Likewise.
11231 * config/arm/aarch-common-protos.h: ... here. New file.
11232 * config/arm/t-arm (aarch-common.o): Define.
11234 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11236 * config/arm/arm.md: Include new file "types.md".
11237 (define_attr "type"): Move from here to ...
11238 (define_attr "mul32"): Likewise.
11239 (define_attr "mul64"): Likewise.
11240 * config/arm/types.md: ... here. New file.
11242 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11244 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11245 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11247 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11249 * gen-pass-instances.awk: Fix offset of substr().
11251 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11253 * Makefile.in (pass-instances.def): New.
11254 (passes.o): Replace dependency on passes.def with one on
11257 * gen-pass-instances.awk: New.
11259 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11260 rather than passes.def, updating local definition of NEXT_PASS
11261 macro to add an extra NUM parameter (currently unused).
11263 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11265 * Makefile.in (PASS_MANAGER_H): New.
11266 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11267 (passes.o): Likewise.
11268 (statistics.o): Likewise.
11269 (cgraphunit.o): Likewise.
11270 (context.o): Depend on PASS_MANAGER_H.
11272 * pass_manager.h: New.
11274 * cgraphunit.c (cgraph_add_new_function): Update for moves
11275 of globals to fields of pass_manager.
11276 (analyze_function): Likewise.
11277 (expand_function): Likewise.
11278 (ipa_passes): Likewise.
11279 (compile): Likewise.
11281 * context.c (context::context): New.
11282 * context.h (context::context): New.
11283 (context::get_passes): New.
11284 (context::passes_): New.
11286 * lto-cgraph.c (input_node): Update for moves of globals to
11287 fields of pass_manager.
11289 * passes.c (all_passes): Remove, in favor of a field of the
11290 same name within the new class pass_manager.
11291 (all_small_ipa_passes): Likewise.
11292 (all_lowering_passes): Likewise.
11293 (all_regular_ipa_passes): Likewise.
11294 (all_late_ipa_passes): Likewise.
11295 (all_lto_gen_passes): Likewise.
11296 (passes_by_id): Likewise.
11297 (passes_by_id_size): Likewise.
11298 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11299 the new class pass_manager.
11300 (set_pass_for_id): Convert to...
11301 (pass_manager::set_pass_for_id): ...method.
11302 (get_pass_for_id): Convert to...
11303 (pass_manager::get_pass_for_id): ...method.
11304 (register_one_dump_file): Move body of implementation into...
11305 (pass_manager::register_one_dump_file): ...here.
11306 (register_dump_files_1): Convert to...
11307 (pass_manager::register_dump_files_1): ...method.
11308 (register_dump_files): Convert to...
11309 (pass_manager::register_dump_files): ...method.
11310 (create_pass_tab): Update for moves of globals to fields of
11312 (dump_passes): Move body of implementation into...
11313 (pass_manager::dump_passes): ...here.
11314 (register_pass): Move body of implementation into...
11315 (pass_manager::register_pass): ...here.
11316 (init_optimization_passes): Convert into...
11317 (pass_manager::pass_manager): ...constructor for new
11318 pass_manager class, and initialize the pass_lists array.
11319 (check_profile_consistency): Update for moves of globals to
11320 fields of pass_manager.
11321 (dump_profile_report): Move body of implementation into...
11322 (pass_manager::dump_profile_report): ...here.
11323 (ipa_write_summaries_1): Update for moves of pass lists from
11324 being globals to fields of pass_manager.
11325 (ipa_write_optimization_summaries): Likewise.
11326 (ipa_read_summaries): Likewise.
11327 (ipa_read_optimization_summaries): Likewise.
11328 (execute_all_ipa_stmt_fixups): Likewise.
11330 * statistics.c (statistics_fini): Update for moves of globals to
11331 fields of pass_manager.
11333 * toplev.c (general_init): Replace call to
11334 init_optimization_passes with construction of the pass_manager
11337 * tree-pass.h (all_passes): Remove, in favor of a field of the
11338 same name within the new class pass_manager.
11339 (all_small_ipa_passes): Likewise.
11340 (all_lowering_passes): Likewise.
11341 (all_regular_ipa_passes): Likewise.
11342 (all_lto_gen_passes): Likewise.
11343 (all_late_ipa_passes): Likewise.
11344 (passes_by_id): Likewise.
11345 (passes_by_id_size): Likewise.
11346 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11347 the new class pass_manager.
11348 (get_pass_for_id): Remove.
11350 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11352 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11355 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11357 * arm.md (mulhi3): New expand pattern.
11359 2013-07-30 Jan Hubicka <jh@suse.cz>
11360 Martin Liska <marxin.liska@gmail.com>
11362 * profile.c (compute_value_histograms): Do not ICE when
11363 there is mismatch only on some counters.
11365 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11367 PR rtl-optimization/57637
11368 * function.c (move_insn_for_shrink_wrap): Also check the
11369 GEN set of the LIVE problem for the liveness analysis
11370 if it exists, otherwise give up.
11372 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11374 PR tree-optimization/57993
11375 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11376 replaced statement in the candidate table.
11377 (phi_add_costs): Return infinite cost when the hidden basis does
11378 not dominate all phis on which the candidate is dependent.
11379 (replace_one_candidate): Record replaced statement in the
11382 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11384 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11385 (ashlv2si3): New expander.
11386 (*ashlv2si3_i): New define_insn_and_split.
11387 * predicates.md (float_operation): Allow patterns with three
11388 basic sub-patterns.
11390 PR rtl-optimization/58021
11391 * mode-switching.c (create_pre_exit): Always split off preceding
11392 insns if we are not at the basic block head.
11394 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11396 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11397 (UCLIBC_DYNAMIC_LINKER): New macro.
11398 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11400 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11401 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11403 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11404 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11405 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11406 for SF and DF modes. Use ieee_quad_format for TF mode.
11407 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11408 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11409 (mips_option_override): Handle `-mnan=legacy'.
11410 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11411 `-mabs=2008' and `-mnan=2008'.
11412 (OPTION_DEFAULT_SPECS): Add "nan" default.
11413 (ASM_SPEC): Handle `-mnan='.
11414 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11415 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11416 comment accordingly.
11417 (neg<mode>2): Likewise.
11418 * config/mips/mips.opt (mabs, mnan): New options.
11419 * doc/install.texi (Configuration): Document `--with-nan=' option.
11420 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11422 (MIPS Options): Document them.
11423 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11424 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11425 * configure: Regenerate.
11426 * config.in: Regenerate.
11428 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11430 * config/i386/i386.md (float post-reload splitters): Do not check
11431 for subregs of SSE registers.
11433 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11434 H.J. Lu <hongjiu.lu@intel.com>
11438 * config/i386/i386.md (post-reload splitter
11439 to avoid partial SSE reg dependency stalls): New pattern.
11441 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11443 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11445 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11447 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11448 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11450 2013-07-26 Cary Coutant <ccoutant@google.com>
11452 * dwarf2out.c (die_checksum_ordered): Don't include template
11453 instantiations in signature.
11454 (is_template_parameter): New function.
11455 (is_template_instantiation): New function.
11456 (generate_skeleton_bottom_up): Don't include template instantiations
11458 (generate_skeleton): Likewise.
11459 (break_out_comdat_types): Move recursive call to break out nested
11461 (prune_unused_types_mark_generic_parms_dies): Call
11462 is_template_parameter.
11464 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11466 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11467 uses vector registers.
11468 * config/aarch64/iterators.md: Add attributes rtn and vas.
11470 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11471 Richard Earnshaw <richard.earnshaw@arm.com>
11473 * combine.c (simplify_comparison): Re-canonicalize operands
11475 * config/arm/arm.md (movcond_addsi): New splitter.
11477 2013-07-25 Sterling Augustine <saugustine@google.com>
11479 * dwarf2out.c (size_of_pubnames): Move code to...
11480 (include_pubname_in_output): ...here. New.
11481 (want_pubnames): Rearrange.
11482 (output_pubnames): Call include_pubname_in_output. Move assertion.
11484 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11486 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11488 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11491 * doc/extend.texi: Remove obsolete builtins. Fix
11492 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11494 2013-07-25 Jan Hubicka <jh@suse.cz>
11496 * cgraph.c (release_function_body): Break out from ...
11497 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11498 and DECL_ARGUMENTS.
11499 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11500 old_tree in the map.
11501 (create_specialized_node): Update.
11502 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11504 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11505 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11506 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11507 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11510 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11512 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11513 addsi3_carryin_alt2_<optab>): Correct output template.
11515 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11517 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11518 Adjust for arm_restrict_it.
11519 Remove trailing whitespace.
11521 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11523 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11524 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11526 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11528 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11530 PR rtl-optimization/57960
11531 * lra-constraints.c (process_alt_operands): Use the right mode
11532 when checking strict_low.
11534 2013-07-25 Jan Hubicka <jh@suse.cz>
11536 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11537 * cgraph.c (cgraph_remove_node): Do not release function body
11538 when in cgraph streaming.
11539 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11540 in other partitions are not considered reachable; fix handling of
11543 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11545 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11547 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11552 * config/arm/arm.md ("*sibcall_insn): Replace use of
11553 Ss with US. Adjust output for v5 and v4t.
11554 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11556 * config/arm/constraints.md ("Ss"): Rename to US.
11558 2013-07-25 Terry Guo <terry.guo@arm.com>
11560 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11561 shift_add/shift_sub0/shift_sub1 RTXs.
11563 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11564 Anton Blanchard <anton@au1.ibm.com>
11566 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11567 (altivec_vpks<VI_char>ss): Likewise.
11568 (altivec_vpks<VI_char>us): Likewise.
11569 (altivec_vpku<VI_char>us): Likewise.
11570 (altivec_vpku<VI_char>um): Likewise.
11572 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11574 Introduce context class.
11576 * Makefile.in (CONTEXT_H): New.
11577 (OBJS): Add context.o.
11578 (toplev.o): Add CONTEXT_H to dependencies.
11581 * toplev.c (general_init): Create the singleton gcc::context instance.
11587 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11589 PR rtl-optimization/57968
11590 * mode-switching.c (create_pre_exit): Allow instructions that
11591 don't set a return register to need a non-exit mode.
11593 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11594 Anton Blanchard <anton@au1.ibm.com>
11596 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11597 operands to vperm for little endian.
11598 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11599 of lvsl to create the control mask for a vperm for little endian.
11601 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11602 Anton Blanchard <anton@au1.ibm.com>
11604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11605 two operands for little-endian.
11607 2013-07-23 Steve Ellcey <sellcey@mips.com>
11609 * config/mips/mips.c (mips_case_values_threshold): New.
11610 (TARGET_CASE_VALUES_THRESHOLD): Define.
11612 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11613 Anton Blanchard <anton@au1.ibm.com>
11615 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11616 selection of field for vector splat in little endian mode.
11618 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11620 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11621 expanders to rs6000.md.
11622 (ior<mode>3): Likewise.
11623 (and<mode>3): Likewise.
11624 (one_cmpl<mode>2): Likewise.
11625 (nor<mode>3): Likewise.
11626 (andc<mode>3): Likewise.
11627 (eqv<mode>3): Likewise.
11628 (nand<mode>3): Likewise.
11629 (orc<mode>3): Likewise.
11631 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11634 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11635 to split multi-word logical operations.
11636 (rs6000_split_logical_di): Likewise.
11637 (rs6000_split_logical): Likewise.
11639 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11640 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11641 and allow TImode operations in 32-bit.
11642 (vsx_and<mode>3_64bit): Likewise.
11643 (vsx_ior<mode>3_32bit): Likewise.
11644 (vsx_ior<mode>3_64bit): Likewise.
11645 (vsx_xor<mode>3_32bit): Likewise.
11646 (vsx_xor<mode>3_64bit): Likewise.
11647 (vsx_one_cmpl<mode>2_32bit): Likewise.
11648 (vsx_one_cmpl<mode>2_64bit): Likewise.
11649 (vsx_nor<mode>3_32bit): Likewise.
11650 (vsx_nor<mode>3_64bit): Likewise.
11651 (vsx_andc<mode>3_32bit): Likewise.
11652 (vsx_andc<mode>3_64bit): Likewise.
11653 (vsx_eqv<mode>3_32bit): Likewise.
11654 (vsx_eqv<mode>3_64bit): Likewise.
11655 (vsx_nand<mode>3_32bit): Likewise.
11656 (vsx_nand<mode>3_64bit): Likewise.
11657 (vsx_orc<mode>3_32bit): Likewise.
11658 (vsx_orc<mode>3_64bit): Likewise.
11660 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11661 logical types in GPRs.
11663 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11664 logical insns to rs6000.md, and allow TImode operations in
11666 (altivec_ior<mode>3): Likewise.
11667 (altivec_xor<mode>3): Likewise.
11668 (altivec_one_cmpl<mode>2): Likewise.
11669 (altivec_nor<mode>3): Likewise.
11670 (altivec_andc<mode>3): Likewise.
11672 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11673 attributes for moving the 128-bit logical operations into
11675 (BOOL_REGS_OUTPUT): Likewise.
11676 (BOOL_REGS_OP1): Likewise.
11677 (BOOL_REGS_OP2): Likewise.
11678 (BOOL_REGS_UNARY): Likewise.
11679 (BOOL_REGS_AND_CR0): Likewise.
11680 (one_cmpl<mode>2): Add support for DI logical operations on
11681 32-bit, splitting the operations to 32-bit.
11682 (anddi3): Likewise.
11683 (iordi3): Likewise.
11684 (xordi3): Likewise.
11685 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11686 changes to combine the 32/64-bit code, allow logical operations on
11687 TI mode in 32-bit, and to use similar match_operator patterns like
11688 scalar mode uses. Combine the Altivec and VSX code for logical
11689 operations, and move it here.
11690 (ior<mode>3, 128-bit types): Likewise.
11691 (xor<mode>3, 128-bit types): Likewise.
11692 (one_cmpl<mode>3, 128-bit types): Likewise.
11693 (nor<mode>3, 128-bit types): Likewise.
11694 (andc<mode>3, 128-bit types): Likewise.
11695 (eqv<mode>3, 128-bit types): Likewise.
11696 (nand<mode>3, 128-bit types): Likewise.
11697 (orc<mode>3, 128-bit types): Likewise.
11698 (and<mode>3_internal): Likewise.
11699 (bool<mode>3_internal): Likewise.
11700 (boolc<mode>3_internal1): Likewise.
11701 (boolc<mode>3_internal2): Likewise.
11702 (boolcc<mode>3_internal1): Likewise.
11703 (boolcc<mode>3_internal2): Likewise.
11704 (eqv<mode>3_internal1): Likewise.
11705 (eqv<mode>3_internal2): Likewise.
11706 (one_cmpl1<mode>3_internal): Likewise.
11708 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11710 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11711 Rename flag_stack_usage to flag_stack_usage_info.
11713 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11715 * config/microblaze/sync.md: New file.
11716 * config/microblaze/microblaze.md: Include sync.md
11717 * config/microblaze/microblaze.c: Add print_operand 'y'.
11718 * config/microblaze/constraints.md: Add memory_contraint
11719 'Q' which is a single register.
11721 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11723 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11725 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11727 * reload.c (find_reloads): Exit loop once we find this operand
11728 cannot be reloaded somehow for this alternative.
11730 * reload.c (find_reloads): Exit loop once we find a hard register.
11732 * rtlanal.c (computed_jump_p): Exit loop once we find label
11735 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11737 * cfgloopmanip.c (remove_path): Exit loop after setting
11740 * gensupport.c (subst_dup): Avoid loop if code is not
11741 MATCH_DUP nor MATCH_OP_DUP.
11743 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11745 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11747 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11749 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11750 true for SP_REGNUM if mode == ptr_mode.
11751 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11752 with value R0_REGNUM + 31.
11754 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11756 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11757 pad pointer-typed argument downward.
11759 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11761 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11762 and __ILP32__ when the ILP32 model is in use.
11764 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11766 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11767 (aarch64_load_symref_appropriately): In the case of
11768 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11769 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11770 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11771 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11772 if 'mode' doesn't equal to 'ptr_mode'.
11773 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11774 'vcall_offset'; change to call aarch64_emit_move differently depending
11775 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11776 to calculate the upper bound of 'vcall_offset'.
11777 (aarch64_cannot_force_const_mem): Change to also return true if
11779 (aarch64_legitimize_reload_address): In the case of large
11780 displacements, add new local variable 'xmode' and an assertion
11781 based on it; change to use 'xmode' to generate the new rtx and
11783 (aarch64_asm_trampoline_template): Change to generate the template
11784 differently depending on TARGET_ILP32 or not; change to use
11785 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11786 (aarch64_trampoline_size): Removed.
11787 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11788 and replace immediate literals with it. Change to use 'ptr_mode'
11789 instead of 'DImode' and call convert_memory_address if the mode
11790 of 'fnaddr' doesn't equal to 'ptr_mode'.
11791 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11793 (aarch64_elf_asm_destructor): Likewise.
11794 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11795 on TARGET_ILP32 instead of aarch64_trampoline_size.
11796 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11797 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11798 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11799 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11800 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11801 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11802 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11803 depending on the value of 'mode'.
11804 (add_losym_<mode>): New.
11805 (ldr_got_small_<mode>): New, based on ldr_got_small.
11806 (ldr_got_small): Remove.
11807 (ldr_got_small_sidi): New.
11808 * config/aarch64/iterators.md (P): New.
11809 (PTR): Change to 'ptr_mode' in the condition.
11811 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11813 * config.gcc (aarch64*-*-*): Support --with-abi.
11814 (aarch64*-*-elf): Support --with-multilib-list.
11815 (aarch64*-*-linux*): Likewise.
11816 (supported_defaults): Add abi to aarch64*-*-*.
11817 * configure.ac: Mention AArch64 for --with-multilib-list.
11818 * configure: Re-generated.
11819 * config/aarch64/biarchilp32.h: New file.
11820 * config/aarch64/biarchlp64.h: New file.
11821 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11823 (MULTILIB_DEFAULTS): Ditto.
11824 (DRIVER_SELF_SPECS): Ditto.
11825 (ASM_SPEC): Update to also substitute -mabi.
11826 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11827 file whose name depends on -mabi= and -mbig-endian.
11828 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11830 (POINTER_SIZE): New define.
11831 (POINTERS_EXTEND_UNSIGNED): Ditto.
11832 (enum aarch64_abi_type): New enumeration tag.
11833 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11834 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11835 (TARGET_ILP32): New define.
11836 * config/aarch64/aarch64.opt (mabi): New.
11837 (aarch64_abi): New.
11838 (ilp32, lp64): New values for -mabi.
11839 * config/aarch64/t-aarch64 (comma): New define.
11840 (MULTILIB_OPTIONS): Ditto.
11841 (MULTILIB_DIRNAMES): Ditto.
11842 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11843 * doc/invoke.texi: Document -mabi for AArch64.
11845 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11847 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11849 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11850 Anton Blanchard <anton@au1.ibm.com>
11852 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11853 endianness when selecting field to splat.
11855 2013-07-22 Eric Christopher <echristo@gmail.com>
11857 * dwarf2out.c (die_odr_checksum): New function to use
11858 CHECKSUM_ macros and ULEB128 for DIE tag.
11859 (generate_type_signature): Use.
11861 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11863 * config.gcc (sparc*-*-*): Accept leon3 processor.
11864 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11865 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11866 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11867 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11868 (mfix-ut699): Adjust comment.
11869 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11870 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11871 (CPP_CPU_SPEC): Likewise.
11872 (ASM_CPU_SPEC): Likewise.
11873 * config/sparc/sparc.c (leon3_cost): New constant.
11874 (sparc_option_override): Add leon3 support.
11875 (mem_ref): New function.
11876 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11877 (sparc_do_work_around_errata): Look into the instruction in the delay
11878 slot and adjust accordingly. Add fix for the data cache nullify issues
11879 of the UT699. Change insertion position for the NOP.
11880 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11881 (leon3_load): New reservation.
11882 (leon_store): Bump latency to 2.
11883 (grfpu): New automaton.
11884 (grfpu_alu): New unit.
11885 (grfpu_ds): Likewise.
11886 (leon_fp_alu): Adjust.
11887 (leon_fp_mult): Delete.
11888 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11889 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11890 * config/sparc/sparc.md (cpu): Add leon3.
11891 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11892 (swapsi): Likewise.
11893 (atomic_test_and_set): Likewise.
11894 (ldstub): Likewise.
11896 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11898 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11899 default for R5900 targets.
11900 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11901 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11902 * config/mips/mips.c (mips_option_override): Report an error for
11903 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11904 for -march=r5900 -mhard-float.
11906 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11908 * df-problems.c (can_move_insns_across): Exit loop once we
11909 find a non-fixed, non-global register.
11911 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11914 * omega.c (omega_eliminate_red): Break after setting red_found.
11915 (omega_problem_has_red_equations): Similarly after setting found.
11916 (omega_query_variable): Similarly after setting coupled.
11918 2013-07-22 Marek Polacek <polacek@redhat.com>
11920 * gimplify.c: Don't include gimple.h twice.
11922 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11924 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11925 instead of TARGET_THUMB1.
11926 (Pz): New constraint.
11927 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11929 (compare_negsi_si): Likewise.
11930 (compare_addsi2_op0): Likewise.
11931 (compare_addsi2_op1): Likewise.
11932 (addsi3_carryin_<optab>): Likewise.
11933 (addsi3_carryin_alt2_<optab>): Likewise.
11934 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11935 for arm_restrict_it.
11936 (subsi3_carryin): Likewise.
11937 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11938 (minmax_arithsi): Disable for arm_restrict_it.
11939 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11940 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11941 (satsi_<SAT:code>_shift): Likewise.
11942 (arm_shiftsi3): Add alternative for 16-bit encoding.
11943 (arm32_movhf): Disable for arm_restrict_it.
11944 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11945 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11947 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11949 * config/arm/arm.md (attribute "insn"): Delete.
11950 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11951 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11952 (not_shiftsi): Update for attribute change.
11953 (not_shiftsi_compare0): Likewise.
11954 (not_shiftsi_compare0_scratch): Likewise.
11955 (arm_one_cmplsi2): Likewise.
11956 (thumb1_one_cmplsi2): Likewise.
11957 (notsi_compare0): Likewise.
11958 (notsi_compare0_scratch): Likewise.
11959 (thumb1_movdi_insn): Likewise.
11960 (arm_movsi_insn): Likewise.
11961 (movhi_insn_arch4): Likewise.
11962 (movhi_bytes): Likewise.
11963 (arm_movqi_insn): Likewise.
11964 (thumb1_movqi_insn): Likewise.
11965 (arm32_movhf): Likewise.
11966 (thumb1_movhf): Likewise.
11967 (arm_movsf_soft_insn): Likewise.
11968 (thumb1_movsf_insn): Likewise.
11969 (thumb_movdf_insn): Likewise.
11970 (movsicc_insn): Likewise.
11971 (movsfcc_soft_insn): Likewise.
11972 (and_scc): Likewise.
11973 (cond_move): Likewise.
11974 (if_move_not): Likewise.
11975 (if_not_move): Likewise.
11976 (if_shift_move): Likewise.
11977 (if_move_shift): Likewise.
11978 (if_shift_shift): Likewise.
11979 (if_not_arith): Likewise.
11980 (if_arith_not): Likewise.
11981 (cond_move_not): Likewise.
11982 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11983 (neon_mov<mode>): Likewise.
11984 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11985 (thumb2_movsi_vfp): Likewise.
11986 (movsf_vfp): Likewise.
11987 (thumb2_movsf_vfp): Likewise.
11988 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11990 (cortexa7_older_only): Likewise.
11991 (cortexa7_younger): Likewise.
11992 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11993 (1020alu_shift_op): Likewise.
11994 (1020alu_shift_reg_op): Likewise.
11995 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11996 (alu_shift_op): Likewise.
11997 (alu_shift_reg_op): Likewise.
11998 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11999 (11_alu_shift_op): Likewise.
12000 (11_alu_shift_reg_op): Likewise.
12001 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12002 (9_alu_shift_reg_op): Likewise.
12003 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12005 (cortex_a15_alu_shift): Likewise.
12006 (cortex_a15_alu_shift_reg): Likewise.
12007 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12008 (cortex_a5_alu_shift): Likewise.
12009 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12011 (cortex_a53_alu_shift): Likewise.
12012 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12014 (cortex_a7_alu_reg): Likewise.
12015 (cortex_a7_alu_shift): Likewise.
12016 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12017 (cortex_a8_alu_shift): Likewise.
12018 (cortex_a8_alu_shift_reg): Likewise.
12019 (cortex_a8_mov): Likewise.
12020 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12021 (cortex_a9_dp_shift): Likewise.
12022 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12023 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12024 (cortex_r4_mov): Likewise.
12025 (cortex_r4_alu_shift): Likewise.
12026 (cortex_r4_alu_shift_reg): Likewise.
12027 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12028 (526_alu_shift_op): Likewise.
12029 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12030 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12031 (626te_alu_shift_op): Likewise.
12032 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12033 (726te_alu_op): Likewise.
12034 (726te_alu_shift_op): Likewise.
12035 (726te_alu_shift_reg_op): Likewise.
12036 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12037 (mp626_alu_shift_op): Likewise.
12038 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12039 (pj4_alu_e1_conds): Likewise.
12040 (pj4_alu): Likewise.
12041 (pj4_alu_conds): Likewise.
12042 (pj4_shift): Likewise.
12043 (pj4_shift_conds): Likewise.
12044 (pj4_alu_shift): Likewise.
12045 (pj4_alu_shift_conds): Likewise.
12047 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12049 * config/arm/predicates.md (shiftable_operator_strict_it):
12051 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12052 Disable cond_exec version for arm_restrict_it.
12053 (thumb2_smaxsi3): Convert to generate cond_exec.
12054 (thumb2_sminsi3): Likewise.
12055 (thumb32_umaxsi3): Likewise.
12056 (thumb2_uminsi3): Likewise.
12057 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12058 (thumb2_neg_abssi2): Likewise.
12059 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12060 (thumb2_movsicc_insn): Adjust alternatives.
12061 (thumb2_mov_negscc): Disable for arm_restrict_it.
12062 (thumb2_mov_negscc_strict_it): New pattern.
12063 (thumb2_mov_notscc_strict_it): New pattern.
12064 (thumb2_mov_notscc): Disable for arm_restrict_it.
12065 (thumb2_ior_scc): Likewise.
12066 (thumb2_ior_scc_strict_it): New pattern.
12067 (thumb2_cond_move): Adjust for arm_restrict_it.
12068 (thumb2_cond_arith): Disable for arm_restrict_it.
12069 (thumb2_cond_arith_strict_it): New pattern.
12070 (thumb2_cond_sub): Adjust for arm_restrict_it.
12071 (thumb2_movcond): Likewise.
12072 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12073 (thumb2_zero_extendhisi2_v6): Likewise.
12074 (thumb2_zero_extendqisi2_v6): Likewise.
12075 (orsi_notsi_si): Likewise.
12076 (orsi_not_shiftsi_si): Likewise.
12078 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12080 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12081 instruction sequence is 1 byte shorter.
12083 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12085 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12086 it is not needed after split.
12088 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12091 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12092 second label for nonlocal goto receivers. Don't output pic base labels
12093 unless we're producing PIC; mark that action unreachable().
12094 (ix86_save_reg): If the function contains a nonlocal label, save the
12096 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12097 * config/darwin.c (emitted_pic_label_num): New GTY.
12098 (update_pic_label_number_if_needed): New.
12099 (machopic_output_function_base_name): Adjust for nonlocal receiver
12101 (machopic_should_output_picbase_label): New.
12102 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12103 (nonlocal_goto_receiver): New insn and split.
12105 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12107 * config/aarch64/aarch64-builtins.c
12108 (aarch64_fold_builtin): Fold abs in all modes.
12109 * config/aarch64/aarch64-simd-builtins.def
12110 (abs): Enable for all modes.
12111 * config/aarch64/arm_neon.h
12112 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12113 (vabs_f64): Add missing intrinsic.
12115 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12117 * config/aarch64/arm_neon.h (vabs_s64): New function
12119 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12122 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12123 * config/avr/avr.md (adjust_len): Add `round'.
12124 * config/avr/avr-protos.h (avr_out_round): New prototype.
12125 (avr_out_plus): Add `out_label' argument.
12126 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12127 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12128 Handle the case where `insn' is just a pattern.
12129 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12130 (avr_out_round): New function.
12131 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12133 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12135 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12136 Add check for flag_stack_usage to handle -fstack-usage support
12138 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12140 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12141 interaction for new Power8 flags and VSX.
12143 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12145 PR middle-end/57698
12146 * tree-inline.c (expand_call_inline): Emit errors during
12147 early_inlining only if optimization is not turned on.
12149 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12153 * passes.c (init_optimization_passes): Move the construction of
12154 the pass hierarchy into a new passes.def file.
12156 * Makefile.in (passes.o): Add dependency on passes.def.
12158 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12160 * passes.c (init_optimization_passes): Introduce macros for
12161 constructing the tree of passes (INSERT_PASSES_AFTER,
12162 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12163 TERMINATE_PASS_LIST).
12165 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12166 Wei Mi <wmi@google.com>
12168 PR rtl-optimization/57878
12169 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12171 (reload_pseudo_compare_func): Check nregs first for reload
12174 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12176 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12178 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12180 * read-rtl.c (validate_const_int): Once an invalid character is
12181 seen, quit the loop.
12183 * gengtype.c (write_roots): Similarly once we find the "deletable"
12184 or "if_marked" option.
12186 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12188 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12189 "xtab" and "sat". Move value "clz" from here to ...
12190 (attriubte "type"): ... here.
12191 (satsi_<SAT:code>): Delete "insn" attribute.
12192 (satsi_<SAT:code>_shift): Likewise.
12193 (arm_zero_extendqisi2addsi): Likewise.
12194 (arm_extendqisi2addsi): Likewise.
12195 (clzsi2): Update for attribute changes.
12196 (rbitsi2): Likewise.
12197 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12199 (arm_usatsihi): Likewise.
12200 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12202 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12204 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12205 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12206 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12207 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12208 in alphabetical order.
12209 (attribute "core_cycles"): Update for attribute changes.
12210 (arm_addsi3): Likewise.
12211 (addsi3_compare0): Likewise.
12212 (addsi3_compare0_scratch): Likewise.
12213 (addsi3_compare_op1): Likewise.
12214 (addsi3_compare_op2): Likewise.
12215 (compare_addsi2_op0): Likewise.
12216 (compare_addsi2_op1): Likewise.
12217 (addsi3_carryin_shift_<optab>): Likewise.
12218 (subsi3_carryin_shift): Likewise.
12219 (rsbsi3_carryin_shift): Likewise.
12220 (arm_subsi3_insn): Likewise.
12221 (subsi3_compare0): Likewise.
12222 (subsi3_compare): Likewise.
12223 (arm_andsi3_insn): Likewise.
12224 (thumb1_andsi3_insn): Likewise.
12225 (andsi3_compare0): Likewise.
12226 (andsi3_compare0_scratch): Likewise.
12227 (zeroextractsi_compare0_scratch
12228 (andsi_not_shiftsi_si): Likewise.
12229 (iorsi3_insn): Likewise.
12230 (iorsi3_compare0): Likewise.
12231 (iorsi3_compare0_scratch): Likewise.
12232 (arm_xorsi3): Likewise.
12233 (thumb1_xorsi3_insn): Likewise.
12234 (xorsi3_compare0): Likewise.
12235 (xorsi3_compare0_scratch): Likewise.
12236 (satsi_<SAT:code>_shift): Likewise.
12238 (arm_shiftsi3): Likewise.
12239 (shiftsi3_compare0): Likewise.
12240 (not_shiftsi): Likewise.
12241 (not_shiftsi_compare0): Likewise.
12242 (not_shiftsi_compare0_scratch): Likewise.
12243 (arm_one_cmplsi2): Likewise.
12244 (thumb_one_complsi2): Likewise.
12245 (notsi_compare0): Likewise.
12246 (notsi_compare0_scratch): Likewise.
12247 (thumb1_zero_extendhisi2): Likewise.
12248 (arm_zero_extendhisi2): Likewise.
12249 (arm_zero_extendhisi2_v6): Likewise.
12250 (arm_zero_extendhisi2addsi): Likewise.
12251 (thumb1_zero_extendqisi2): Likewise.
12252 (thumb1_zero_extendqisi2_v6): Likewise.
12253 (arm_zero_extendqisi2): Likewise.
12254 (arm_zero_extendqisi2_v6): Likewise.
12255 (arm_zero_extendqisi2addsi): Likewise.
12256 (thumb1_extendhisi2): Likewise.
12257 (arm_extendhisi2): Likewise.
12258 (arm_extendhisi2_v6): Likewise.
12259 (arm_extendqisi): Likewise.
12260 (arm_extendqisi_v6): Likewise.
12261 (arm_extendqisi2addsi): Likewise.
12262 (thumb1_extendqisi2): Likewise.
12263 (thumb1_movdi_insn): Likewise.
12264 (arm_movsi_insn): Likewise.
12265 (movsi_compare0): Likewise.
12266 (movhi_insn_arch4): Likewise.
12267 (movhi_bytes): Likewise.
12268 (arm_movqi_insn): Likewise.
12269 (thumb1_movqi_insn): Likewise.
12270 (arm32_movhf): Likewise.
12271 (thumb1_movhf): Likewise.
12272 (arm_movsf_soft_insn): Likewise.
12273 (thumb1_movsf_insn): Likewise.
12274 (movdf_soft_insn): Likewise.
12275 (thumb_movdf_insn): Likewise.
12276 (arm_cmpsi_insn): Likewise.
12277 (cmpsi_shiftsi): Likewise.
12278 (cmpsi_shiftsi_swp): Likewise.
12279 (arm_cmpsi_negshiftsi_si): Likewise.
12280 (movsicc_insn): Likewise.
12281 (movsfcc_soft_insn): Likewise.
12282 (arith_shiftsi): Likewise.
12283 (arith_shiftsi_compare0
12284 (arith_shiftsi_compare0_scratch
12285 (sub_shiftsi): Likewise.
12286 (sub_shiftsi_compare0
12287 (sub_shiftsi_compare0_scratch
12288 (and_scc): Likewise.
12289 (cond_move): Likewise.
12290 (if_plus_move): Likewise.
12291 (if_move_plus): Likewise.
12292 (if_move_not): Likewise.
12293 (if_not_move): Likewise.
12294 (if_shift_move): Likewise.
12295 (if_move_shift): Likewise.
12296 (if_shift_shift): Likewise.
12297 (if_not_arith): Likewise.
12298 (if_arith_not): Likewise.
12299 (cond_move_not): Likewise.
12300 (thumb1_ashlsi3): Set type attribute.
12301 (thumb1_ashrsi3): Likewise.
12302 (thumb1_lshrsi3): Likewise.
12303 (thumb1_rotrsi3): Likewise.
12304 (shiftsi3_compare0_scratch): Likewise.
12305 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12306 (neon_mov<mode>): Likewise.
12307 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12309 (thumb2_movsi_insn): Likewise.
12310 (thumb2_cmpsi_neg_shiftsi): Likewise.
12311 (thumb2_extendqisi_v6): Likewise.
12312 (thumb2_zero_extendhisi2_v6): Likewise.
12313 (thumb2_zero_extendqisi2_v6): Likewise.
12314 (thumb2_shiftsi3_short): Likewise.
12315 (thumb2_addsi3_compare0_scratch): Likewise.
12316 (orsi_not_shiftsi_si): Likewise.
12317 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12318 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12320 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12321 (1020alu_shift_op): Likewise.
12322 (1020alu_shift_reg_op): Likewise.
12323 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12324 (alu_shift_op): Likewise.
12325 (alu_shift_reg_op): Likewise.
12326 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12327 (11_alu_shift_op): Likewise.
12328 (11_alu_shift_reg_op): Likewise.
12329 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12330 (9_alu_shift_reg_op): Likewise.
12331 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12333 (cortex_a15_alu_shift): Likewise.
12334 (cortex_a15_alu_shift_reg): Likewise.
12335 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12337 (cortex_a5_alu_shift): Likewise.
12338 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12340 (cortex_a53_alu_shift): Likewise.
12341 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12343 (cortex_a7_alu_reg): Likewise.
12344 (cortex_a7_alu_shift): Likewise.
12345 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12347 (cortex_a8_alu_shift): Likewise.
12348 (cortex_a8_alu_shift_reg): Likewise.
12349 (cortex_a8_mov): Likewise.
12350 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12351 (cortex_a9_dp_shift): Likewise.
12352 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12354 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12356 (cortex_r4_mov): Likewise.
12357 (cortex_r4_alu_shift): Likewise.
12358 (cortex_r4_alu_shift_reg): Likewise.
12359 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12360 (526_alu_shift_op): Likewise.
12361 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12362 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12363 (626te_alu_shift_op): Likewise.
12364 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12365 (726te_alu_op): Likewise.
12366 (726te_alu_shift_op): Likewise.
12367 (726te_alu_shift_reg_op): Likewise.
12368 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12369 (mp626_alu_shift_op): Likewise.
12370 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12371 (pj4_alu_e1_conds): Likewise.
12372 (pj4_alu): Likewise.
12373 (pj4_alu_conds): Likewise.
12374 (pj4_shift): Likewise.
12375 (pj4_shift_conds): Likewise.
12376 (pj4_alu_shift): Likewise.
12377 (pj4_alu_shift_conds): Likewise.
12378 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12380 (cortexa7_older_only): Likewise.
12381 (cortexa7_younger): Likewise.
12383 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12385 * ipa-pure-const.c (generate_summary): Rename to...
12386 (pure_const_generate_summary): ... this.
12388 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12390 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12392 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12395 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12398 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12400 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12401 enabled without -march=zEC12.
12402 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12405 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12407 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12408 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12409 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12410 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12411 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12412 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12413 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12415 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12416 and ISA_HAS_NMADD3_NMSUB3.
12417 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12418 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12419 (nmsub4<mode>, nmsub3<mode>): Likewise.
12420 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12422 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12424 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12425 TARGET_MIPS5400 checking.
12427 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12428 Peter Bergner <bergner@vnet.ibm.com>
12430 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12431 registers in the comment.
12432 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12433 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12434 rather than FIRST_PSEUDO_REGISTERS.
12436 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12438 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12439 enable extra ISA flags with TARGET_HTM.
12441 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12443 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12446 2013-07-15 Cong Hou <congh@google.com>
12448 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12449 in compare function for sorting.
12451 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12453 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12454 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12455 * config/rs6000/rs6000.opt: Add -mhtm option.
12456 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12457 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12458 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12459 __HTM__ if the HTM instructions are available.
12460 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12461 htm_spr_reg_operand): New define_predicates.
12462 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12463 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12465 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12466 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12467 HTM builtin functions.
12468 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12469 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12470 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12471 (rs6000_builtin_mask_calculate): Likewise.
12472 (rs6000_option_override_internal): Likewise.
12473 (bdesc_htm): Add new HTM builtin support.
12474 (htm_spr_num): New function.
12475 (htm_spr_regno): Likewise.
12476 (rs6000_htm_spr_icode): Likewise.
12477 (htm_expand_builtin): Likewise.
12478 (htm_init_builtins): Likewise.
12479 (rs6000_expand_builtin): Add support for HTM builtin functions.
12480 (rs6000_init_builtins): Likewise.
12481 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12483 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12484 (TARGET_HTM, MASK_HTM): Define macros.
12485 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12486 (FIXED_REGISTERS): Likewise.
12487 (CALL_USED_REGISTERS): Likewise.
12488 (CALL_REALLY_USED_REGISTERS): Likewise.
12489 (REG_ALLOC_ORDER): Likewise.
12490 (enum reg_class): Likewise.
12491 (REG_CLASS_NAMES): Likewise.
12492 (REG_CLASS_CONTENTS): Likewise.
12493 (REGISTER_NAMES): Likewise.
12494 (ADDITIONAL_REGISTER_NAMES): Likewise.
12495 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12496 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12497 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12498 * config/rs6000/htm.md: New file.
12499 * config/rs6000/htmintrin.h: New file.
12500 * config/rs6000/htmxlintrin.h: New file.
12502 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12504 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12505 Define SYMBOL_TINY_GOT, update comment.
12506 * config/aarch64/aarch64.c
12507 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12508 (aarch64_expand_mov_immediate): Likewise.
12509 (aarch64_print_operand): Likewise.
12510 (aarch64_classify_symbol): Likewise.
12511 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12512 (ldr_got_tiny): Define.
12514 2013-07-13 Tobias Grosser <tobias@grosser.es>
12516 PR tree-optimization/54094
12517 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12518 scheduling dimension for the parallelism check from the polyhedral
12519 information in the AST.
12520 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12523 2013-07-13 Jason Merrill <jason@redhat.com>
12525 * print-tree.c (debug_vec_tree): Use debug_raw.
12526 (debug_raw (vec<tree, va_gc> &)): New.
12527 (debug_raw (vec<tree, va_gc> *)): New.
12528 * tree.h: Declare them.
12530 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12532 * ifcvt.c (ifcvt_after_combine): New static variable.
12533 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12535 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12536 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12537 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12539 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12541 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12543 2013-07-12 Michael Matz <matz@suse.de>
12545 PR middle-end/55771
12546 * convert.c (convert_to_real): Reject non-float inner types.
12548 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12550 * config/aarch64/aarch64-protos.h
12551 (aarch64_simd_immediate_valid_for_move): Remove.
12552 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12553 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12555 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12557 2013-07-11 Steve Ellcey <sellcey@mips.com>
12559 * config/mips/mips.c (mips_conditional_register_usage): Do not
12560 use t[0-7] registers in MIPS16 mode when optimizing for size.
12562 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12564 * config/i386/i386.c (dispatch_function_versions): Fix array
12565 indexing of function_version_info to match actual_versions.
12567 2013-07-11 Teresa Johnson <tejohnson@google.com>
12569 * vec.h (struct va_gc): Move release out-of-line.
12570 (va_gc::release): Call ggc_free on released vec.
12572 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12574 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12575 Require GOT register as additional operand in UNSPEC.
12576 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12577 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12578 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12579 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12580 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12581 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12582 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12584 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12587 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12588 name seen by assembler/linker rather if available.
12590 2013-07-11 Andreas Schwab <schwab@suse.de>
12592 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12594 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12596 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12598 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12600 * doc/tm.texi.in: Move hook documentation to ....
12601 * target.def: ... here.
12603 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12604 text on @hook line.
12605 * doc/tm.texi: Regenerate.
12607 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12610 * doc/invoke.texi: Document -Wconditionally-supported.
12612 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12615 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12618 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12621 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12622 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12623 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12624 Remove duplicate devices.
12625 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12626 * config/avr/t-multilib: Regenerate.
12627 * config/avr/avr-tables.opt: Regenerate.
12628 * doc/avr-mmcu.texi: Regenerate.
12630 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12633 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12635 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12637 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12638 the cost of MULT when optimizing for size.
12640 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12642 * config/cr16/cr16-protos.h: Don't include target.h.
12644 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12646 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12647 adjust register size for TDmode and TFmode for VSX registers.
12649 2013-07-08 Kai Tietz <ktietz@redhat.com>
12652 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12653 hook_bool_const_tree_true.
12655 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12657 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12658 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12659 * config/s390/s390.md: Define FPR*_REGNUM constants.
12660 Fix FPR2_REGNUM constant (18 -> 17).
12661 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12662 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12663 ("trunc<BFP:mode><DFP_ALL:mode>2")
12664 ("trunc<DFP_ALL:mode><BFP:mode>2")
12665 ("*extend<BFP:mode><DFP_ALL:mode>2")
12666 ("*extend<DFP_ALL:mode><BFP:mode>2")
12667 ("extend<BFP:mode><DFP_ALL:mode>2")
12668 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12671 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12673 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12675 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12677 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12678 and cfun_fpr_bit_p to cfun_fpr_save_p.
12679 (s390_frame_area, s390_register_info, s390_frame_info)
12680 (s390_emit_prologue, s390_emit_epilogue)
12681 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12683 * config/s390/s390.h: Define *_REGNUM macros for floating point
12686 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12688 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12690 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12692 PR rtl-optimization/57786
12693 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12694 and break out of the loop when it is set to false.
12696 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12699 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12700 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12701 (const_int 63)) 0)).
12702 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12703 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12704 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12706 PR rtl-optimization/57829
12707 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12708 mask bits outside of mode are just sign-extension from mode to HWI.
12710 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12712 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12713 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12714 adjust_address instead of change_address to keep info about alignment.
12715 (emit_strmov): Remove.
12716 (emit_memmov): New function.
12717 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12718 (expand_movmem_epilogue): Likewise and return updated rtx for
12720 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12721 destination and source.
12722 (decide_alignment): Refactor, handle vector_loop.
12723 (ix86_expand_movmem): Likewise.
12724 (ix86_expand_setmem): Likewise.
12725 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12727 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12729 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12730 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12732 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12734 * config/i386/sse.md (sse_movlhps): Change alternative 3
12735 of operand 2 to "m".
12737 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12740 * config/i386/sse.md (iptr): New mode attribute.
12741 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12742 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12743 (<sse>_vmmul<mode>3): Ditto.
12744 (<sse>_vmdiv<mode>3): Ditto.
12745 (sse_vmrcpv4sf2): Ditto.
12746 (<sse>_vmsqrt<mode>2): Ditto.
12747 (sse_vmrsqrtv4sf2): Ditto.
12748 (<sse>_vm<code><mode>3): Ditto.
12749 (avx_vmcmp<mode>3): Ditto.
12750 (<sse>_vmmaskcmp<mode>3): Ditto.
12751 (<sse>_comi): Ditto.
12752 (<sse>_ucomi): Ditto.
12753 (*xop_vmfrcz_<mode>): Ditto.
12754 (*fmai_fmadd_<mode>): Ditto.
12755 (*fmai_fmsub_<mode>): Ditto.
12756 (*fmai_fnmadd_<mode>): Ditto.
12757 (*fmai_fnmsub_<mode>): Ditto.
12758 (*fma4i_vmfmadd_<mode>): Ditto.
12759 (*fma4i_vmfmsub_<mode>): Ditto.
12760 (*fma4i_vmfnmadd_<mode>): Ditto.
12761 (*fma4i_vmfnmsub_<mode>): Ditto.
12762 (*xop_vmfrcz_<mode>): Ditto.
12763 (sse_cvtps2pi): Ditto.
12764 (sse_cvttps2pi): Ditto.
12765 (sse_cvtss2si): Ditto.
12766 (sse_cvtss2si_2): Ditto.
12767 (sse_cvtss2siq_2): Ditto.
12768 (sse_cvttss2si): Ditto.
12769 (sse_cvttss2siq): Ditto.
12770 (sse_cvtsd2si): Ditto.
12771 (sse_cvtsd2si_2): Ditto.
12772 (sse_cvtsd2siq_2): Ditto.
12773 (sse_cvttsd2si): Ditto.
12774 (sse_cvttsd2siq): Ditto.
12775 (sse_cvtsd2ss): Ditto.
12776 (sse_cvtss2sd): Ditto.
12777 (avx2_pbroadcast<mode>): Ditto.
12778 (avx2_pbroadcast<mode>_1): Ditto.
12779 (*avx_vperm_broadcast_v4sf): Ditto.
12781 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12782 (sse_movlhps): Ditto.
12783 (sse_storehps): Ditto.
12784 (sse_loadhps): Ditto.
12785 (sse_storelps): Ditto.
12786 (sse_loadlps): Ditto.
12787 (*vec_concatv4sf): Ditto.
12788 (*vec_interleave_highv2df): Ditto.
12789 (*vec_interleave_lowv2df): Ditto.
12790 (*vec_extractv2df_1_sse): Ditto.
12791 (*vec_extractv2df_0_sse): Ditto.
12792 (sse2_storelpd): Ditto.
12793 (sse2_loadlpd): Ditto.
12794 (sse2_movsd): Ditto.
12795 (*vec_concatv4si): Ditto.
12796 (vec_concatv2di): Ditto.
12798 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12799 for Intel asm dialect.
12800 (mmx_punpcklwd): Ditto.
12801 (mmx_punpckldq): Ditto.
12803 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12804 for intel assembler dialect.
12806 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12809 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12810 for BUILT_IN_C{LZ,LRSB}*.
12811 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12812 * tree-vrp.c (extract_range_basic): Handle
12813 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12814 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12815 fall thru to code calling set_value*.
12816 * builtins.c (expand_builtin): Remove *IMAX cases.
12817 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12818 if width is bigger than 2*HWI.
12820 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12822 PR rtl-optimization/55342
12823 * lra-int.h (lra_subreg_reload_pseudos): New.
12824 * lra.c: Add undoing optional reloads to the block diagram.
12825 (lra_subreg_reload_pseudos): New.
12826 (lra_optional_reload_pseudos): Change comments.
12827 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12828 lra_optional_reload_pseudos after undo transformations.
12829 * lra-assigns.c (pseudo_prefix_title): New.
12830 (lra_setup_reg_renumber): Use it.
12831 (spill_for): Ditto. Check subreg reload pseudos too.
12832 (assign_by_spills): Consider subreg reload pseudos too.
12833 * lra-constraints.c (simplify_operand_subreg): Use
12834 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12835 (curr_insn_transform): Recognize and do optional reloads.
12836 (undo_optional_reloads): New.
12837 (lra_undo_inheritance): Call undo_optional_reloads.
12839 2013-07-05 Thomas Quinot <quinot@adacore.com>
12841 * tree-complex.c (expand_complex_operations_1): Fix typo.
12843 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12845 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12846 (tune_params): New member 'const vec_costs'.
12847 * config/aarch64/aarch64.c (generic_vector_cost): New.
12848 (generic_tunings): New member 'generic_vector_cost'.
12849 (aarch64_builtin_vectorization_cost): New.
12850 (aarch64_add_stmt_cost): New.
12851 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12852 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12854 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12857 * config/i386/predicates.md (vsib_address_operand): Disallow
12858 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12860 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12862 PR middle-end/55030
12863 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12864 expand_builtin_setjmp_receiver.
12865 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12866 with NULL for the label parameter.
12867 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12868 the frame-pointer. Adjust comments.
12869 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12870 only if LABEL is non-NULL.
12872 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12874 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12875 (ARM_ABI_AAPCS64): Ditto.
12877 (ARM_DEFAULT_ABI): Ditto.
12879 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12881 * config/aarch64/aarch64-builtins.c
12882 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12883 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12885 * config/aarch64/aarch64-simd.md
12886 (aarch64_ld1<VALL:mode>): New.
12887 (aarch64_st1<VALL:mode>): Likewise.
12888 * config/aarch64/arm_neon.h
12889 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12891 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12893 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12895 (ix86_option_override_internal):Turn on all -mavx target flags by
12896 default as they are dependent on AVX anyway.
12898 2013-07-02 Cary Coutant <ccoutant@google.com>
12900 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12901 deterministic hash.
12902 (loc_checksum_ordered): Likewise.
12903 (hash_loc_operands): Remove inline keyword.
12905 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12907 PR tree-optimization/57741
12908 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12909 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12910 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12911 Allow REAL_CST step_exprs if flag_associative_math.
12912 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12914 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12916 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12918 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12920 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12922 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12924 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12926 (iorsi3_insn): Likewise.
12927 (arm_xorsi3): Likewise.
12929 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12931 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12933 (attribute "type"): ... here, and prefix with "wmmx_".
12934 (attribute "core_cycles"): Update for attribute changes.
12935 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12936 (tbcstv4hi): Likewise.
12937 (tbcstv2si): Likewise.
12938 (iwmmxt_iordi3): Likewise.
12939 (iwmmxt_xordi3): Likewise.
12940 (iwmmxt_anddi3): Likewise.
12941 (iwmmxt_nanddi3): Likewise.
12942 (iwmmxt_arm_movdi): Likewise.
12943 (iwmmxt_movsi_insn): Likewise.
12944 (mov<mode>_internal): Likewise.
12945 (and<mode>3_iwmmxt): Likewise.
12946 (ior<mode>3_iwmmxt): Likewise.
12947 (xor<mode>3_iwmmxt): Likewise.
12948 (add<mode>3_iwmmxt): Likewise.
12949 (ssaddv8qi3): Likewise.
12950 (ssaddv4hi3): Likewise.
12951 (ssaddv2si3): Likewise.
12952 (usaddv8qi3): Likewise.
12953 (usaddv4hi3): Likewise.
12954 (usaddv2si3): Likewise.
12955 (sub<mode>3_iwmmxt): Likewise.
12956 (sssubv8qi3): Likewise.
12957 (sssubv4hi3): Likewise.
12958 (sssubv2si3): Likewise.
12959 (ussubv8qi3): Likewise.
12960 (ussubv4hi3): Likewise.
12961 (ussubv2si3): Likewise.
12962 (mulv4hi3_iwmmxt): Likewise.
12963 (smulv4hi3_highpart): Likewise.
12964 (umulv4hi3_highpart): Likewise.
12965 (iwmmxt_wmacs): Likewise.
12966 (iwmmxt_wmacsz): Likewise.
12967 (iwmmxt_wmacu): Likewise.
12968 (iwmmxt_wmacuz): Likewise.
12969 (iwmmxt_clrdi): Likewise.
12970 (iwmmxt_clrv8qi): Likewise.
12971 (iwmmxt_clr4hi): Likewise.
12972 (iwmmxt_clr2si): Likewise.
12973 (iwmmxt_uavgrndv8qi3): Likewise.
12974 (iwmmxt_uavgrndv4hi3): Likewise.
12975 (iwmmxt_uavgv8qi3): Likewise.
12976 (iwmmxt_uavgv4hi3): Likewise.
12977 (iwmmxt_tinsrb): Likewise.
12978 (iwmmxt_tinsrh): Likewise.
12979 (iwmmxt_tinsrw): Likewise.
12980 (iwmmxt_textrmub): Likewise.
12981 (iwmmxt_textrmsb): Likewise.
12982 (iwmmxt_textrmuh): Likewise.
12983 (iwmmxt_textrmsh): Likewise.
12984 (iwmmxt_textrmw): Likewise.
12985 (iwmxxt_wshufh): Likewise.
12986 (eqv8qi3): Likewise.
12987 (eqv4hi3): Likewise.
12988 (eqv2si3): Likewise.
12989 (gtuv8qi3): Likewise.
12990 (gtuv4hi3): Likewise.
12991 (gtuv2si3): Likewise.
12992 (gtv8qi3): Likewise.
12993 (gtv4hi3): Likewise.
12994 (gtv2si3): Likewise.
12995 (smax<mode>3_iwmmxt): Likewise.
12996 (umax<mode>3_iwmmxt): Likewise.
12997 (smin<mode>3_iwmmxt): Likewise.
12998 (umin<mode>3_iwmmxt): Likewise.
12999 (iwmmxt_wpackhss): Likewise.
13000 (iwmmxt_wpackwss): Likewise.
13001 (iwmmxt_wpackdss): Likewise.
13002 (iwmmxt_wpackhus): Likewise.
13003 (iwmmxt_wpackwus): Likewise.
13004 (iwmmxt_wpackdus): Likewise.
13005 (iwmmxt_wunpckihb): Likewise.
13006 (iwmmxt_wunpckihh): Likewise.
13007 (iwmmxt_wunpckihw): Likewise.
13008 (iwmmxt_wunpckilb): Likewise.
13009 (iwmmxt_wunpckilh): Likewise.
13010 (iwmmxt_wunpckilw): Likewise.
13011 (iwmmxt_wunpckehub): Likewise.
13012 (iwmmxt_wunpckehuh): Likewise.
13013 (iwmmxt_wunpckehuw): Likewise.
13014 (iwmmxt_wunpckehsb): Likewise.
13015 (iwmmxt_wunpckehsh): Likewise.
13016 (iwmmxt_wunpckehsw): Likewise.
13017 (iwmmxt_wunpckelub): Likewise.
13018 (iwmmxt_wunpckeluh): Likewise.
13019 (iwmmxt_wunpckeluw): Likewise.
13020 (iwmmxt_wunpckelsb): Likewise.
13021 (iwmmxt_wunpckelsh): Likewise.
13022 (iwmmxt_wunpckelsw): Likewise.
13023 (ror<mode>3): Likewise.
13024 (ashr<mode>3_iwmmxt): Likewise.
13025 (lshr<mode>3_iwmmxt): Likewise.
13026 (ashl<mode>3_iwmmxt): Likewise.
13027 (ror<mode>3_di): Likewise.
13028 (ashr<mode>3_di): Likewise.
13029 (lshr<mode>3_di): Likewise.
13030 (ashl<mode>3_di): Likewise.
13031 (iwmmxt_wmadds): Likewise.
13032 (iwmmxt_wmaddu): Likewise.
13033 (iwmmxt_tmia): Likewise.
13034 (iwmmxt_tmiaph): Likewise.
13035 (iwmmxt_tmiabb): Likewise.
13036 (iwmmxt_tmiatb): Likewise.
13037 (iwmmxt_tmiabt): Likewise.
13038 (iwmmxt_tmiatt): Likewise.
13039 (iwmmxt_tmovmskb): Likewise.
13040 (iwmmxt_tmovmskh): Likewise.
13041 (iwmmxt_tmovmskw): Likewise.
13042 (iwmmxt_waccb): Likewise.
13043 (iwmmxt_wacch): Likewise.
13044 (iwmmxt_waccw): Likewise.
13045 (iwmmxt_waligni): Likewise.
13046 (iwmmxt_walignr): Likewise.
13047 (iwmmxt_walignr0): Likewise.
13048 (iwmmxt_walignr1): Likewise.
13049 (iwmmxt_walignr2): Likewise.
13050 (iwmmxt_walignr3): Likewise.
13051 (iwmmxt_wsadb): Likewise.
13052 (iwmmxt_wsadh): Likewise.
13053 (iwmmxt_wsadbz): Likewise.
13054 (iwmmxt_wsadhz): Likewise.
13055 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13056 (iwmmxt_wabsdiffb): Likewise.
13057 (iwmmxt_wabsdiffh): Likewise.
13058 (iwmmxt_wabsdiffw): Likewise.
13059 (iwmmxt_waddsubhx): Likewise
13060 (iwmmxt_wsubaddhx): Likewise.
13061 (addc<mode>3): Likewise.
13062 (iwmmxt_avg4): Likewise.
13063 (iwmmxt_avg4r): Likewise.
13064 (iwmmxt_wmaddsx): Likewise.
13065 (iwmmxt_wmaddux): Likewise.
13066 (iwmmxt_wmaddsn): Likewise.
13067 (iwmmxt_wmaddun): Likewise.
13068 (iwmmxt_wmulwsm): Likewise.
13069 (iwmmxt_wmulwum): Likewise.
13070 (iwmmxt_wmulsmr): Likewise.
13071 (iwmmxt_wmulumr): Likewise.
13072 (iwmmxt_wmulwsmr): Likewise.
13073 (iwmmxt_wmulwumr): Likewise.
13074 (iwmmxt_wmulwl): Likewise.
13075 (iwmmxt_wqmulm): Likewise.
13076 (iwmmxt_wqmulwm): Likewise.
13077 (iwmmxt_wqmulmr): Likewise.
13078 (iwmmxt_wqmulwmr): Likewise.
13079 (iwmmxt_waddbhusm): Likewise.
13080 (iwmmxt_waddbhusl): Likewise.
13081 (iwmmxt_wqmiabb): Likewise.
13082 (iwmmxt_wqmiabt): Likewise.
13083 (iwmmxt_wqmiatb): Likewise.
13084 (iwmmxt_wqmiatt): Likewise.
13085 (iwmmxt_wqmiabbn): Likewise.
13086 (iwmmxt_wqmiabtn): Likewise.
13087 (iwmmxt_wqmiatbn): Likewise.
13088 (iwmmxt_wqmiattn): Likewise.
13089 (iwmmxt_wmiabb): Likewise.
13090 (iwmmxt_wmiabt): Likewise.
13091 (iwmmxt_wmiatb): Likewise.
13092 (iwmmxt_wmiatt): Likewise.
13093 (iwmmxt_wmiabbn): Likewise.
13094 (iwmmxt_wmiabtn): Likewise.
13095 (iwmmxt_wmiatbn): Likewise.
13096 (iwmmxt_wmiattn): Likewise.
13097 (iwmmxt_wmiawbb): Likewise.
13098 (iwmmxt_wmiawbt): Likewise.
13099 (iwmmxt_wmiawtb): Likewise.
13100 (iwmmxt_wmiawtt): Likewise.
13101 (iwmmxt_wmiawbbn): Likewise.
13102 (iwmmxt_wmiawbtn): Likewise.
13103 (iwmmxt_wmiawtbn): Likewise.
13104 (iwmmxt_wmiawttn): Likewise.
13105 (iwmmxt_wmerge): Likewise.
13106 (iwmmxt_tandc<mode>3): Likewise.
13107 (iwmmxt_torc<mode>3): Likewise.
13108 (iwmmxt_torvsc<mode>3): Likewise.
13109 (iwmmxt_textrc<mode>3): Likewise.
13110 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13111 (wmmxt_pack): Likewise.
13112 (wmmxt_mult_c1): Likewise.
13113 (wmmxt_mult_c2): Likewise.
13114 (wmmxt_alu_c1): Likewise.
13115 (wmmxt_alu_c2): Likewise.
13116 (wmmxt_alu_c3): Likewise.
13117 (wmmxt_transfer_c1): Likewise.
13118 (wmmxt_transfer_c2): Likewise.
13119 (wmmxt_transfer_c3): Likewise.
13120 (marvell_f_iwmmxt_wstr): Likewise.
13121 (marvell_f_iwmmxt_wldr): Likewise.
13123 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13125 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13127 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13130 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13131 * lra-constraints.c (need_for_split_p): Check call used hard regs
13132 living through calls.
13134 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13135 call used regs for call insn.
13137 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13140 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13141 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13143 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13145 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13147 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13149 * lra-constraints.c (need_for_split_p): Check call used hard regs
13150 living through calls.
13152 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13155 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13156 to tie with any other modes. Eliminate Altivec vector mode tests,
13157 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13158 code, to return 0 if testing MODE2 for a condition, if we've
13159 already tested MODE1 for the same condition.
13161 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13163 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13166 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13168 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13169 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13171 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13173 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13175 (aarch64_symbolic_constant_p): Remove.
13176 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13177 static. Fix line length and white space.
13178 (aarch64_symbolic_constant_p): Remove.
13179 * config/aarch64/predicates.md (aarch64_valid_symref):
13180 Use aarch64_classify_symbol_expression.
13182 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13184 * config/arm/constraints.md (Ts): New constraint.
13185 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13187 (compare_scc): Use "Ts" constraint for operand 0.
13188 (ior_scc_scc): Likewise.
13189 (and_scc_scc): Likewise.
13190 (and_scc_scc_nodom): Likewise.
13191 (ior_scc_scc_cmp): Likewise for operand 7.
13192 (and_scc_scc_cmp): Likewise.
13193 * config/arm/thumb2.md (thumb2_movsi_insn):
13194 Add alternatives for 16-bit encodings.
13195 (thumb2_movhi_insn): Likewise.
13196 (thumb2_movsicc_insn): Likewise.
13197 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13198 (thumb2_negscc): Use "Ts" constraint.
13199 Move mvn instruction outside cond_exec block.
13200 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13201 for 16-bit encodings.
13203 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13205 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13207 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13208 (mulsi3subsi): Likewise.
13209 (mulsidi3adddi): Likewise.
13210 (mulsidi3_v6): Likewise.
13211 (umulsidi3_v6): Likewise.
13212 (umulsidi3adddi_v6): Likewise.
13213 (smulsi3_highpart_v6): Likewise.
13214 (umulsi3_highpart_v6): Likewise.
13215 (mulhisi3tb): Likewise.
13216 (mulhisi3bt): Likewise.
13217 (mulhisi3tt): Likewise.
13218 (maddhisi4): Likewise.
13219 (maddhisi4tb): Likewise.
13220 (maddhisi4tt): Likewise.
13221 (maddhidi4): Likewise.
13222 (maddhidi4tb): Likewise.
13223 (maddhidi4tt): Likewise.
13224 (zeroextractsi_compare0_scratch): Likewise.
13225 (insv_zero): Likewise.
13226 (insv_t2): Likewise.
13227 (anddi_notzesidi_di): Likewise.
13228 (anddi_notsesidi_di): Likewise.
13229 (andsi_notsi_si): Likewise.
13230 (iordi_zesidi_di): Likewise.
13231 (xordi_zesidi_di): Likewise.
13232 (andsi_iorsi3_notsi): Likewise.
13233 (smax_0): Likewise.
13234 (smax_m1): Likewise.
13235 (smin_0): Likewise.
13236 (not_shiftsi): Likewise.
13237 (unaligned_loadsi): Likewise.
13238 (unaligned_loadhis): Likewise.
13239 (unaligned_loadhiu): Likewise.
13240 (unaligned_storesi): Likewise.
13241 (unaligned_storehi): Likewise.
13242 (extv_reg): Likewise.
13243 (extzv_t2): Likewise.
13244 (divsi3): Likewise.
13245 (udivsi3): Likewise.
13246 (arm_zero_extendhisi2addsi): Likewise.
13247 (arm_zero_extendqisi2addsi): Likewise.
13248 (compareqi_eq0): Likewise.
13249 (arm_extendhisi2_v6): Likewise.
13250 (arm_extendqisi2addsi): Likewise.
13251 (arm_movt): Likewise.
13252 (thumb2_ldrd): Likewise.
13253 (thumb2_ldrd_base): Likewise.
13254 (thumb2_ldrd_base_neg): Likewise.
13255 (thumb2_strd): Likewise.
13256 (thumb2_strd_base): Likewise.
13257 (thumb2_strd_base_neg): Likewise.
13258 (arm_negsi2): Add alternative for 16-bit encoding.
13259 (arm_one_cmplsi2): Likewise.
13261 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13263 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13264 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13265 (movdfcc): Likewise.
13266 * config/arm/vfp.md (*thumb2_movsf_vfp):
13267 Disable predication for arm_restrict_it.
13268 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13269 (*thumb2_movdfcc_vfp): Likewise.
13270 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13271 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13272 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13273 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13274 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13275 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13276 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13277 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13278 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13279 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13280 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13281 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13282 Disable predication for arm_restrict_it.
13284 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13286 * config/i386/bmiintrin.h (_bextr_u32): New.
13287 (_bextr_u64): Ditto.
13289 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13291 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13292 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13293 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13294 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13295 * config/mips/n32-elf.h: ...this new file.
13297 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13300 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13301 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13303 2013-06-27 Catherine Moore <clm@codesourcery.com>
13305 * config/mips/mips-tables.opt: Regenerate.
13306 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13307 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13308 * doc/invoke.texi: Add -m14kc.
13310 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13313 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13314 constraints of operand 1 and 2.
13317 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13318 to match RTL canonicalization. Swap predicates and
13319 constraints of operand 1 and 2.
13321 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13323 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13324 Process OP_INOUT regs for splitting too.
13326 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13328 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13329 decl before the loop, initialize to NULL.
13330 (vectorizable_load): Initialize ptr_incr to NULL.
13332 2013-06-27 Martin Jambor <mjambor@suse.cz>
13335 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13336 * ipa-ref.c (ipa_maybe_record_reference): New function.
13337 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13338 * ipa-cp.c (create_specialized_node): Record potential references from
13340 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13342 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13344 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13345 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13346 'mode' to force_reg.
13347 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13348 (aarch64_expand_mov_immediate): Likewise.
13350 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13352 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13353 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13355 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13357 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13358 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13359 (struct machine_function): Add tbegin_p.
13360 (s390_canonicalize_comparison): Fold CC mode compares to
13361 conditional jump if possible.
13362 (s390_emit_jump): Return the emitted jump.
13363 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13364 Handle CCRAWmode compares.
13365 (s390_option_override): Default to -mhtm if available.
13366 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13367 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13368 FPRs instead of df_regs_ever_live_p.
13369 (s390_optimize_nonescaping_tx): New function.
13370 (s390_init_frame_layout): Extend clobbered_regs array to cover
13372 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13373 (s390_expand_tbegin): New function.
13374 (enum s390_builtin): New enum definition.
13375 (code_for_builtin): New array definition.
13376 (s390_init_builtins): New function.
13377 (s390_expand_builtin): New function.
13378 (TARGET_INIT_BUILTINS): Define.
13379 (TARGET_EXPAND_BUILTIN): Define.
13380 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13381 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13382 (s390_alc_comparison): Likewise.
13383 * config/s390/s390-modes.def: Add CCRAWmode.
13384 * config/s390/s390.h (processor_flags): Add PF_TX.
13385 (TARGET_CPU_HTM): Define macro.
13386 (TARGET_HTM): Define macro.
13387 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13388 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13389 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13390 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13391 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13392 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13393 ("*cc_to_int"): Move up.
13394 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13395 constants other than 0.
13396 ("*ccraw_to_int"): New insn and splitter definition.
13397 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13398 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13399 ("tx_assist"): New expander.
13400 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13401 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13402 * config/s390/s390.opt: Add -mhtm option.
13403 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13404 * config/s390/htmxlintrin.h: New file.
13405 * config/s390/htmintrin.h: New file.
13406 * config/s390/s390intrin.h: New file.
13407 * doc/extend.texi: Document htm builtins.
13408 * config.gcc: Add the new header files to extra_headers.
13410 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13412 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13413 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13415 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13416 Pat Haugen <pthaugen@us.ibm.com>
13417 Peter Bergner <bergner@vnet.ibm.com>
13419 * config/rs6000/power8.md: New.
13420 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13421 setting for power8 entry.
13422 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13423 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13424 test for Power4/Power5 only.
13425 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13427 (force_new_group): Adjust comment.
13428 * config/rs6000/rs6000.md: Include power8.md.
13430 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13432 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13433 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13435 (tune_params): Update comment.
13436 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13437 (arm_max_conditional_execute): New function.
13438 (thumb2_final_prescan_insn): Use max_insn_skipped and
13439 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13441 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13443 PR tree-optimization/57705
13444 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13445 SSA_NAME step, provided that it is not defined inside the loop.
13446 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13447 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13449 2013-06-25 Martin Jambor <mjambor@suse.cz>
13451 PR middle-end/57670
13452 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13453 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13455 (ipa_note_param_call): Initialize member_ptr flag.
13456 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13457 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13458 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13459 (ipa_read_indirect_edge_info): Likewise.
13461 2013-06-25 Richard Biener <rguenther@suse.de>
13463 PR middle-end/56977
13464 * passes.c (init_optimization_passes): Move pass_fold_builtins
13465 and pass_dce earlier with -Og.
13467 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13469 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13470 <BIT_FIELD_REF>: Remove trailing TAB.
13471 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13474 2013-06-24 Martin Jambor <mjambor@suse.cz>
13476 PR tree-optimization/57358
13477 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13478 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13479 (ipa_analyze_params_uses): Generate pessimistic info when true.
13481 2013-06-24 Martin Jambor <mjambor@suse.cz>
13483 PR tree-optimization/57539
13484 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13485 global.inlined_to of the new node to it. All callers changed.
13486 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13487 inlining_into, pass it to cgraph_clone_node.
13488 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13489 ipa_free_edge_args_substructures.
13490 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13491 rdesc linked list. Do not assert rdesc edges have inlined caller.
13492 Assert we have found an rdesc in the rdesc list.
13494 2013-06-24 Richard Biener <rguenther@suse.de>
13496 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13497 (pointer_set_lookup): Declare.
13498 (class pointer_map): New template class implementing a
13499 generic pointer to T map.
13500 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13501 pointer_map<T>::contains, pointer_map<T>::insert,
13502 pointer_map<T>::traverse): New functions.
13503 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13504 (pointer_set_lookup): New function.
13505 (pointer_set_contains): Use pointer_set_lookup.
13506 (pointer_set_insert): Likewise.
13507 (insert_aux): Remove.
13508 (struct pointer_map_t): Embed a pointer_set_t.
13509 (pointer_map_create): Adjust.
13510 (pointer_map_destroy): Likewise.
13511 (pointer_map_contains): Likewise.
13512 (pointer_map_insert): Likewise.
13513 (pointer_map_traverse): Likewise.
13514 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13515 pointer_map<unsigned> instead of a pointer_map_t.
13516 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13517 (streamer_tree_cache_lookup): Likewise.
13518 (streamer_tree_cache_create): Likewise.
13519 (streamer_tree_cache_delete): Likewise.
13520 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13521 pointer_map<unsigned> instead of a pointer_map_t.
13522 (lto_init_tree_ref_encoder): Adjust.
13523 (lto_destroy_tree_ref_encoder): Likewise.
13524 * lto-section-out.c (lto_output_decl_index): Likewise.
13525 (lto_record_function_out_decl_state): Likewise.
13526 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13528 2013-06-24 Richard Biener <rguenther@suse.de>
13530 PR tree-optimization/57488
13531 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13533 2013-06-24 Alan Modra <amodra@gmail.com>
13535 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13536 (gen_easy_altivec_constant): Likewise.
13537 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13538 easy_vector_constant_msb): Likewise.
13540 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13543 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13544 add missing return true.
13546 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13549 * config/sh/predicates.md (general_extend_operand): Invoke
13550 general_movsrc_operand for memory operands.
13551 (general_movsrc_operand): Allow reg+reg addressing, do not use
13552 general_operand for memory operands.
13554 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13556 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13557 when current target options does not apply.
13558 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13559 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13560 * config/i386/bmiintrin.h: Pass appropriate target
13561 attributes to header.
13562 * config/i386/mmintrin.h: Ditto.
13563 * config/i386/nmmintrin.h: Ditto.
13564 * config/i386/avx2intrin.h: Ditto.
13565 * config/i386/fxsrintrin.h: Ditto.
13566 * config/i386/tbmintrin.h: Ditto.
13567 * config/i386/xsaveintrin.h: Ditto.
13568 * config/i386/f16cintrin.h: Ditto.
13569 * config/i386/xtestintrin.h: Ditto.
13570 * config/i386/xsaveoptintrin.h: Ditto.
13571 * config/i386/bmi2intrin.h: Ditto.
13572 * config/i386/lzcntintrin.h: Ditto.
13573 * config/i386/smmintrin.h: Ditto.
13574 * config/i386/wmmintrin.h: Ditto.
13575 * config/i386/x86intrin.h: Remove all header include guards.
13576 * config/i386/prfchwintrin.h: Ditto.
13577 * config/i386/pmmintrin.h: Ditto.
13578 * config/i386/tmmintrin.h: Ditto.
13579 * config/i386/xmmintrin.h: Ditto.
13580 * config/i386/popcntintrin.h: Ditto.
13581 * config/i386/rdseedintrin.h: Ditto.
13582 * config/i386/ammintrin.h: Ditto.
13583 * config/i386/emmintrin.h: Ditto.
13584 * config/i386/immintrin.h: Remove all header include guards.
13585 * config/i386/fma4intrin.h: Ditto.
13586 * config/i386/lwpintrin.h: Ditto.
13587 * config/i386/xopintrin.h: Ditto.
13588 * config/i386/ia32intrin.h: Ditto.
13589 * config/i386/avxintrin.h: Ditto.
13590 * config/i386/rtmintrin.h: Ditto.
13591 * config/i386/fmaintrin.h: Ditto.
13592 * config/i386/mm3dnow.h: Ditto.
13594 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13596 * common/config/i386/i386-common.c: Handle LZCNT.
13598 2013-06-22 Andi Kleen <ak@linux.intel.com>
13600 * doc/extend.texi: Use __atomic_store_n instead of
13601 __atomic_store in HLE example.
13603 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13605 * config/sh/sh.c: Remove <cstdlib> workaround.
13607 2013-06-21 Andi Kleen <ak@linux.intel.com>
13609 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13611 2013-06-21 Andi Kleen <ak@linux.intel.com>
13613 * doc/extend.texi: Document that __atomic_clear and
13614 __atomic_test_and_set should only be used with bool.
13616 2013-06-20 Jan Hubicka <jh@suse.cz>
13618 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13619 types_same_for_odr.
13620 * tree.c (decls_same_for_odr): New function.
13621 (same_for_edr): New function.
13622 (types_same_for_odr): New function.
13623 (get_binfo_at_offset): Use it.
13624 * tree.h (types_same_for_odr): Declare.
13626 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13627 Jason Merrill <jason@redhat.com>
13629 * system.h: Include <cstdlib> as well as <stdlib.h>.
13631 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13634 * config/i386/i386.c (construct_container): Report error if
13635 long double is used with disabled x87 float returns.
13637 2013-06-20 Jan Hubicka <jh@suse.cz>
13639 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13641 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13643 PR rtl-optimization/57425
13644 PR rtl-optimization/57569
13645 * alias.c (write_dependence_p): Remove parameters mem_mode and
13646 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13647 Changed all callers.
13648 (canon_anti_dependence): Get comments and semantics in sync.
13649 Add parameter mem_canonicalized. Changed all callers.
13650 * rtl.h (canon_anti_dependence): Update prototype.
13652 2013-06-20 Richard Biener <rguenther@suse.de>
13654 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13655 byte case, inline streamer_read_uchar and defer section
13658 2013-06-20 Richard Biener <rguenther@suse.de>
13660 PR tree-optimization/57584
13661 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13662 SSA names into the expanded expression that take part in
13663 abnormal coalescing.
13665 2013-06-19 Sharad Singhai <singhai@google.com>
13667 * gcov.c (print_usage): Handle new option.
13668 (process_args): Ditto.
13669 (get_gcov_intermediate_filename): New function.
13670 (output_intermediate_file): New function.
13671 (output_gcov_file): New function
13672 (generate_results): Handle new option.
13673 (release_function): Relase demangled name.
13674 (read_graph_file): Handle demangled name.
13675 (output_lines): Ditto.
13676 * doc/gcov.texi: Document gcov intermediate format.
13678 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13681 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13682 (lra_emit_add): Use the functions. Add comment about Y as an
13685 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13688 * collect2.c (collect_atexit): New.
13689 (collect_exit): Delete.
13690 (main): Register collect_atexit with atexit.
13691 (collect_wait): Change collect_exit to exit.
13693 * collect2.h (collect_exit): Delete.
13694 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13696 2013-06-19 Wei Mi <wmi@google.com>
13698 PR rtl-optimization/57518
13699 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13700 if regno is used in paradoxical subreg.
13701 (update_equiv_regs): Check pdx_subregs[regno] before
13702 set a reg to be equivalent with a mem.
13704 2013-06-19 Matthias Klose <doko@ubuntu.com>
13707 * file-find.h (find_a_file): Add a mode parameter.
13708 * file-find.c (find_a_file): Likewise.
13709 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13710 with X_OK for the executables.
13711 * collect2.c (main): Call find_a_file with X_OK.
13713 2013-06-19 Steve Ellcey <sellcey@mips.com>
13716 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13717 Use NEXT_INSN instead of next_real_insn.
13719 2013-06-19 Jan Hubicka <jh@suse.cz>
13721 * cgraph.h (const_value_known_p): Replace by ...
13722 (ctor_for_folding): .. this one.
13723 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13724 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13725 * expr.c (expand_expr_real_1): Likewise.
13726 (string_constant): Likewise.
13727 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13728 * ipa.c (process_references): Likewise.
13729 (symtab_remove_unreachable_nodes): Likewise.
13730 * ipa-inline-analysis.c (param_change_prob): Likewise.
13731 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13732 (get_base_constructor): Likwise.
13733 * varpool.c (varpool_remove_node): Likewise.
13734 (varpool_remove_initializer): LIkewise.
13735 (dump_varpool_node): LIkwise.
13736 (const_value_known_p): Rewrite to ...
13737 (ctor_for_folding): ... this one.
13739 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13742 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13743 PERSONALITY in $PATH derived prefixes.
13745 2013-06-19 Jeff Law <law@redhat.com>
13747 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13750 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13751 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13754 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13756 * config/arm/vfp.md: Move VFP instruction classification documentation
13758 * config/arm/arm.md: ... here. Update instruction classification
13761 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13763 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13764 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13765 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13767 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13770 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13771 with NEXT_INSN. Use tablejump_p to check for jump table data
13774 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13777 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13778 that now in C++ the value is correct per the C++ standards.
13780 2013-06-19 Richard Biener <rguenther@suse.de>
13782 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13783 for global context.
13785 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13788 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13791 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13792 with next_active_insn.
13794 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13796 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13797 functions are inlined during failures to flag an error.
13798 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13799 in early inline pass.
13801 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13803 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13806 2013-06-18 Julian Brown <julian@codesourcery.com>
13808 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13809 Permit virtual register pre-reload if !strict.
13810 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13812 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13814 * config/arm/neon.md (movmisalign<mode>): Use
13815 neon_perm_struct_or_reg_operand instead of
13816 neon_struct_or_register_operand.
13817 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13818 neon_permissive_struct_operand instead of neon_struct_operand.
13819 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13820 neon_vector_mem_operand.
13821 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13822 neon_vector_mem_operand.
13823 (neon_permissive_struct_operand): New.
13824 (neon_struct_or_register_operand): Rename to...
13825 (neon_perm_struct_or_reg_operand): This. Adjust call to
13826 neon_vector_mem_operand.
13828 2013-06-18 Richard Biener <rguenther@suse.de>
13830 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13831 * lto-streamer.h: Include pointer-set.h.
13832 (struct lto_decl_slot): Remove.
13833 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13834 Remove next_index entry.
13835 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13836 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13837 (lto_init_tree_ref_encoder): Adjust.
13838 (lto_destroy_tree_ref_encoder): Likewise.
13839 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13840 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13841 (lto_output_decl_index): Adjust.
13842 (lto_new_out_decl_state): Likewise.
13843 (lto_record_function_out_decl_state): Likewise.
13844 * lto-streamer-out.c (copy_function): Likewise.
13846 2013-06-18 Richard Biener <rguenther@suse.de>
13848 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13849 * cgraphunit.c: Include cfgloop.h.
13850 (init_lowered_empty_function): Initialize the loop tree.
13851 (assemble_thunk): Insert new BBs into loops.
13853 2013-06-18 Richard Biener <rguenther@suse.de>
13855 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13856 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13857 the map from cache entry to cache index optional.
13858 (streamer_tree_cache_replace_tree): Adjust accordingly.
13859 (streamer_tree_cache_append): Likewise.
13860 (streamer_tree_cache_delete): Likewise.
13861 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13862 streamer cache map from cache entry to cache index.
13863 * lto-streamer-out.c (create_output_block): Adjust.
13865 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13867 * config/arm/arm.md (attribute "insn"): Move multiplication and
13868 division attributes to...
13869 (attribute "type"): ... here. Remove mult.
13870 (attribute "mul32"): New attribute.
13871 (attribute "mul64"): Add umaal.
13872 (*arm_mulsi3): Update attributes.
13873 (*arm_mulsi3_v6): Likewise.
13874 (*thumb_mulsi3): Likewise.
13875 (*thumb_mulsi3_v6): Likewise.
13876 (*mulsi3_compare0): Likewise.
13877 (*mulsi3_compare0_v6): Likewise.
13878 (*mulsi_compare0_scratch): Likewise.
13879 (*mulsi_compare0_scratch_v6): Likewise.
13880 (*mulsi3addsi): Likewise.
13881 (*mulsi3addsi_v6): Likewise.
13882 (*mulsi3addsi_compare0): Likewise.
13883 (*mulsi3addsi_compare0_v6): Likewise.
13884 (*mulsi3addsi_compare0_scratch): Likewise.
13885 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13886 (*mulsi3subsi): Likewise.
13887 (*mulsidi3adddi): Likewise.
13888 (*mulsi3addsi_v6): Likewise.
13889 (*mulsidi3adddi_v6): Likewise.
13890 (*mulsidi3_nov6): Likewise.
13891 (*mulsidi3_v6): Likewise.
13892 (*umulsidi3_nov6): Likewise.
13893 (*umulsidi3_v6): Likewise.
13894 (*umulsidi3adddi): Likewise.
13895 (*umulsidi3adddi_v6): Likewise.
13896 (*smulsi3_highpart_nov6): Likewise.
13897 (*smulsi3_highpart_v6): Likewise.
13898 (*umulsi3_highpart_nov6): Likewise.
13899 (*umulsi3_highpart_v6): Likewise.
13900 (mulhisi3): Likewise.
13901 (*mulhisi3tb): Likewise.
13902 (*mulhisi3bt): Likewise.
13903 (*mulhisi3tt): Likewise.
13904 (maddhisi4): Likewise.
13905 (*maddhisi4tb): Likewise.
13906 (*maddhisi4tt): Likewise.
13907 (maddhidi4): Likewise.
13908 (*maddhidi4tb): Likewise.
13909 (*maddhidi4tt): Likewise.
13910 (divsi3): Likewise.
13911 (udivsi3): Likewise.
13912 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13913 (thumb2_mulsi_short_compare0): Likewise.
13914 (thumb2_mulsi_short_compare0_scratch): Likewise.
13915 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13916 (1020mult2): Likewise.
13917 (1020mult3): Likewise.
13918 (1020mult4): Likewise.
13919 (1020mult5): Likewise.
13920 (1020mult6): Likewise.
13921 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13923 (cortex_a15_mult64): Likewise.
13924 (cortex_a15_sdiv): Likewise.
13925 (cortex_a15_udiv): Likewise.
13926 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13932 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13933 (pj4_ir_div): Likewise.
13934 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13935 (11_mult2): Likewise.
13936 (11_mult3): Likewise.
13937 (11_mult4): Likewise.
13938 (11_mult5): Likewise.
13939 (11_mult6): Likewise.
13940 (11_mult7): Likewise.
13941 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13942 (cortex_a8_mla): Likewise.
13943 (cortex_a8_mull): Likewise.
13944 (cortex_a8_smulwy): Likewise.
13945 (cortex_a8_smlald): Likewise.
13946 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13947 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13948 (cortex_r4_mul_3): Likewise.
13949 (cortex_r4_mla_4): Likewise.
13950 (cortex_r4_mla_3): Likewise.
13951 (cortex_r4_smlald): Likewise.
13952 (cortex_r4_mull): Likewise.
13953 (cortex_r4_sdiv): Likewise.
13954 (cortex_r4_udiv): Likewise.
13955 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13956 (cortex_a7_idiv): Likewise.
13957 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13958 (9_mult2): Likewise.
13959 (9_mult3): Likewise.
13960 (9_mult4): Likewise.
13961 (9_mult5): Likewise.
13962 (9_mult6): Likewise.
13963 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13964 (cortex_a53_sdiv): Likewise.
13965 (cortex_a53_udiv): Likewise.
13966 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13967 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13968 (mp626_mult2): Likewise.
13969 (mp626_mult3): Likewise.
13970 (mp626_mult4): Likewise.
13971 * config/arm/fa526.md (526_mult1): Update attribute change.
13972 (526_mult2): Likewise.
13973 * config/arm/arm-generic.md (mult): Update attribute change.
13974 (mult_ldsched_strongarm): Likewise.
13975 (mult_ldsched): Likewise.
13976 (multi_cycle): Likewise.
13977 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13978 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13979 (606te_mult2): Likewise.
13980 (606te_mult3): Likewise.
13981 (606te_mult4): Likewise.
13982 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13983 (cortex_a9_mac16): Likewise.
13984 (cortex_a9_multiply): Likewise.
13985 (cortex_a9_mac): Likewise.
13986 (cortex_a9_multiply_long): Likewise.
13987 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13988 (626te_mult2): Likewise.
13989 (626te_mult3): Likewise.
13990 (626te_mult4): Likewise.
13992 2013-06-18 Richard Biener <rguenther@suse.de>
13995 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13997 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14000 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14001 with next_active_insn.
14003 2013-06-18 Alan Modra <amodra@gmail.com>
14005 * config/rs6000/rs6000.h (enum data_align): New.
14006 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14007 (DATA_ABI_ALIGNMENT): Define.
14008 (CONSTANT_ALIGNMENT): Correct comment.
14009 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14010 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14012 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14014 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14015 ATTRIBUTE_UNUSED marking.
14017 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14019 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14020 alternative and update.
14021 (aarch64_dup_lanedi): Delete.
14022 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14023 * config/aarch64/aarch64-simd-builtins.def: Update.
14025 2013-06-17 Richard Biener <rguenther@suse.de>
14027 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14028 (lto_input_scc): Declare.
14029 (lto_input_tree_1): Likewise.
14030 (struct lto_stats_d): Add num_tree_bodies_output and
14031 num_pickle_refs_output.
14032 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14033 (lto_read_tree_1): Split out from ...
14034 (lto_read_tree): ... this.
14035 (lto_input_scc): New function.
14036 (lto_input_tree_1): Split out from ...
14037 (lto_input_tree): ... this. Handle LTO_tree_scc.
14038 (lto_data_in_create): Create the streamer cache without hashes.
14039 * lto-streamer-out.c (create_output_block): Create the streamer
14040 cache with hashes when not doing WPA.
14041 (lto_write_tree_1): Split out from ...
14042 (lto_write_tree): ... this.
14043 (get_symbol_initial_value): New function.
14044 (lto_output_tree_1): Split out from ...
14045 (lto_output_tree): ... this. Write trees as series of SCCs
14046 using a DFS walk via DFS_write_tree.
14047 (struct sccs, struct scc_entry): New types.
14048 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14049 (DFS_write_tree_body): New function.
14050 (DFS_write_tree): Likewise.
14051 (hash_tree): Likewise.
14052 (scc_entry_compare): Likewise.
14053 (hash_scc): Likewise.
14054 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14055 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14056 TREE_CHAIN as regular reference.
14057 (streamer_read_integer_cst): Remove.
14058 (streamer_get_pickled_tree): Adjust.
14059 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14060 of DECL_EXTERNALs in BLOCK_VARS for now.
14061 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14063 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14064 Add hash value argument and record that if hashes are recorded
14066 (streamer_tree_cache_insert_1): Adjust.
14067 (streamer_tree_cache_insert): Likewise.
14068 (streamer_tree_cache_insert_at): Rename to ...
14069 (streamer_tree_cache_replace_tree): ... this and adjust.
14070 (streamer_tree_cache_append): Adjust.
14071 (record_common_node): Likewise.
14072 (streamer_tree_cache_create): Add argument whether to
14073 record hash values together with trees.
14074 (streamer_tree_cache_delete): Adjust.
14075 * tree-streamer.h (struct streamer_tree_cache_d): Add
14077 (streamer_read_integer_cst): Remove.
14078 (streamer_tree_cache_insert): Adjust.
14079 (streamer_tree_cache_append): Likewise.
14080 (streamer_tree_cache_insert_at): Rename to ...
14081 (streamer_tree_cache_replace_tree): ... this and adjust.
14082 (streamer_tree_cache_create): Add argument whether to record hashes.
14083 (streamer_tree_cache_get): Rename to ...
14084 (streamer_tree_cache_get_tree): ... this.
14085 (streamer_tree_cache_get_hash): New function.
14086 * tree.c (cache_integer_cst): New function.
14087 * tree.h (cache_integer_cst): Declare.
14088 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14089 * lto-symtab.c (lto_varpool_replace_node): Only release
14090 DECL_INITIAL of non-prevailing decls.
14091 * varpool.c (varpool_remove_initializer): Do not release
14092 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14094 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14096 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14097 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14098 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14099 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14100 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14101 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14102 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14103 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14104 instead of TARGET_64BIT.
14105 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14106 Require ISA_HAS_<D>DIV.
14108 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14110 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14111 (mips*-*-linux*): Move default with_llsc setting to where other
14113 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14115 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14116 Likewise. Remove default with_tune setting. Move default float
14117 setting to its own block. Handle with_llsc in the same block as above.
14119 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14121 PR rtl-optimization/57425
14122 PR rtl-optimization/57569
14123 * alias.c (write_dependence_p): Add new parameters mem_mode,
14124 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14125 Changed all callers.
14126 (canon_anti_dependence): New function.
14127 * cse.c (check_dependence): Use canon_anti_dependence.
14128 * cselib.c (cselib_invalidate_mem): Likewise.
14129 * rtl.h (canon_anti_dependence): Declare.
14131 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14133 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14134 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14135 ".set mips3" for 64-bit targets.
14137 2013-06-15 Dehao Chen <dehao@google.com>
14139 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14140 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14141 (gimple_check_call_args): Likewise.
14142 * value-prof.c (check_ic_target): Likewise.
14143 * ipa-inline.c (early_inliner): Likewise.
14144 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14145 * cgraph.c (cgraph_create_edge_1): Likewise.
14146 (cgraph_make_edge_direct): Likewise.
14148 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14151 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14152 rs6000_output_move_128bit to handle emitting quad memory
14153 operations. Set attribute length to 8 bytes.
14155 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14157 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14159 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14160 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14161 (aarch64_<su>mlsl<mode>): Likewise.
14163 2013-06-14 Mike Stump <mikestump@comcast.net>
14165 * Makefile.in (TARGET_H): Add insn-codes.h.
14167 2013-06-14 Alan Modra <amodra@gmail.com>
14169 PR middle-end/57134
14170 PR middle-end/57586
14171 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14172 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14173 bitfield expansion when EXPAND_MEMORY.
14174 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14176 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14178 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14179 test for clearing quad memory on 32-bit later.
14181 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14183 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14184 (fold_negate_expr): Likewise.
14185 (fold_real_zero_addition_p): Handle vectors.
14186 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14188 2013-06-14 Alan Modra <amodra@gmail.com>
14190 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14192 2013-06-13 Jan Hubicka <jh@suse.cz>
14194 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14195 Local comdats are not externally visible.
14196 * symtab.c (dump_symtab_base): Dump externally visible.
14197 (verify_symtab_base): Verify back links in the symtab hash.
14199 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14201 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14202 CONVERT_EXPR as equal nodes.
14204 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14206 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14208 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14210 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14211 Generalize to complex and vector.
14212 * tree.c (build_all_ones_cst): New function.
14213 * tree.h (build_all_ones_cst): Declare it.
14215 2013-06-13 Alan Modra <amodra@gmail.com>
14217 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14218 * config/rs6000/rs6000.md (signbittf2): New insn.
14219 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14220 (abstf2_internal, cmptf_internal2): Likewise.
14221 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14223 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14224 Pat Haugen <pthaugen@us.ibm.com>
14225 Peter Bergner <bergner@vnet.ibm.com>
14227 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14228 power8 byte, half-word, and quad-word atomic instructions.
14229 (emit_store_conditional): Likewise.
14230 (rs6000_expand_atomic_compare_and_swap): Likewise.
14231 (rs6000_expand_atomic_op): Likewise.
14233 * config/rs6000/sync.md (larx): Add new modes for power8.
14235 (AINT): New mode iterator to include TImode as well as normal
14236 integer modes on power8.
14237 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14238 that VSX registers are not considered. Use AINT mode iterator
14239 instead of INT1 to allow inclusion of quad word atomic operations
14241 (load_locked<mode>): Likewise.
14242 (store_conditional<mode>): Likewise.
14243 (atomic_compare_and_swap<mode>): Likewise.
14244 (atomic_exchange<mode>): Likewise.
14245 (atomic_nand<mode>): Likewise.
14246 (atomic_fetch_<fetchop_name><mode>): Likewise.
14247 (atomic_nand_fetch<mode>): Likewise.
14248 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14250 (ATOMIC): On power8, add QImode, HImode modes.
14251 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14252 modes that promote to SImode.
14253 (load_lockedti): Convert TImode arguments to PTImode, so that we
14254 get a guaranteed even/odd register pair.
14255 (load_lockedpti): Likewise.
14256 (store_conditionalti): Likewise.
14257 (store_conditionalpti): Likewise.
14259 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14260 atomic load/store instructions.
14263 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14265 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14267 (insn_count): New attribute, with most cases extracted from...
14268 (length): ...here. Redefine most cases in terms of insn_count.
14269 (single_insn): Delete.
14270 (can_delay): Use insn_count to check for single instructions.
14271 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14272 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14273 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14274 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14275 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14276 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14277 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14278 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14279 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14280 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14281 rather than "length".
14282 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14283 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14284 Use "insn_count" rather than "length".
14285 * config/mips/mips-dsp.md
14286 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14287 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14290 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14292 PR tree-optimization/57361
14293 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14295 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14297 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14299 (aarch64_simd_combine<mode>): New instruction expansion.
14300 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14301 function prototype.
14302 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14303 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14305 2013-06-12 Jan Hubicka <jh@suse.cz>
14307 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14308 decl has when in streaming stage.
14309 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14310 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14312 2013-06-12 Roland Stigge <stigge@antcom.de>
14315 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14317 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14319 PR tree-optimization/57537
14320 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14321 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14323 2013-06-12 Richard Biener <rguenther@suse.de>
14325 * data-streamer.h (streamer_write_char_stream): CSE
14326 obs->current_pointer.
14327 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14328 streamer_write_char_stream manually and optimize the resulting loop.
14329 (streamer_write_hwi_stream): Likewise.
14331 2013-06-12 Jan Hubicka <jh@suse.cz>
14333 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14334 * cgraph.h (varpool_create_empty_node): Declare.
14335 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14337 * symtab.c (symtab_unregister_node): Be lax about missin entries
14339 (symtab_get_node): Update comment.
14340 * varpool.c (varpool_create_empty_node): Break out from ...
14341 (varpool_node_for_decl): ... here.
14342 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14344 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14346 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14347 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14348 part. Use straight-line flow at the end.
14349 <COMPONENT_REF>: Remove superfluous else.
14350 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14352 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14355 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14356 target hook even for !TREE_PUBLIC decls. If no resolution info
14357 is available, return false for common and external decls.
14359 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14361 * config/rl78/constraints.md (U): New constraint.
14362 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14365 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14368 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14369 to allow returning address to AT_PLATFORM name.
14371 2013-06-11 Jan Hubicka <jh@suse.cz>
14373 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14374 * cgraph.h (symtab_node_base): Add weakref flag.
14375 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14376 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14377 (output_weakrefs): Use weakref flag.
14378 * fold-const.c (simple_operand_p): Handle WEAK.
14379 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14380 * ipa.c (varpool_externally_visible_p): Drop weakref.
14381 (function_and_variable_visibility): Update comment; fix weakref
14382 sanity checks; do not clear DECL_WEAK on them.
14383 * lto-cgraph.c (lto_output_node): update.
14384 (lto_output_varpool_node): Update.
14385 (input_overwrite_node): Update.
14386 (input_node): Update.
14387 (input_varpool_node): Update.
14388 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14389 (lto_symtab_merge_symbols): Add sanity check.
14390 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14391 * passes.c (rest_of_decl_compilation): Set static flag, too.
14392 * symtab.c (dump_symtab_base): Dump weakref.
14393 (verify_symtab_base): Sanity check weakrefs.
14394 (symtab_make_decl_local): Remove duplicated code.
14395 (symtab_alias_ultimate_target): Simplify.
14396 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14398 2013-06-11 Tom de Vries <tom@codesourcery.com>
14400 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14401 sequence_vect == NULL.
14403 2013-06-11 DJ Delorie <dj@redhat.com>
14405 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14406 (rl78_unwind_word_mode): New.
14408 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14410 * final.c (debug_prefix_maps): Make static.
14412 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14414 * function.c (initial_trampoline): Remove stray copy.
14416 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14418 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14420 2013-06-11 Martin Jambor <mjambor@suse.cz>
14422 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14423 within bounds at the beginning of the function.
14425 2013-06-11 Alan Modra <amodra@gmail.com>
14427 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14429 (get_named_section): Don't NULL !DECL_P decl.
14431 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14433 * doc/invoke.texi (core-avx2): Document.
14435 (atom): Updated with MOVBE.
14437 2013-06-11 Richard Biener <rguenther@suse.de>
14439 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14441 2013-06-11 Anton Blanchard <anton@samba.org>
14443 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14444 correct shift value in little-endian mode.
14446 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14449 * varasm.c (get_variable_align): Move #endif to the right place.
14451 2013-06-10 Cary Coutant <ccoutant@google.com>
14453 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14454 for hash so that hash table traversal order is deterministic.
14456 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14457 Pat Haugen <pthaugen@us.ibm.com>
14458 Peter Bergner <bergner@vnet.ibm.com>
14460 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14461 of vectors in GPRS if they are direct moves or quad word load or
14464 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14466 (direct_move_p): Likewise.
14467 (quad_load_store_p): Likewise.
14469 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14470 classes into bins based on the physical register type.
14471 (reg_class_to_reg_type): Likewise.
14472 (IS_STD_REG_TYPE): Likewise.
14473 (IS_FP_VECT_REG_TYPE): Likewise.
14474 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14475 use direct move instructions.
14476 (reload_gpr_vsx): Likewise.
14477 (reload_vsx_gpr): Likewise.
14478 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14479 information that is a simplification of register classes. Also
14480 precalculate direct move reload helpers.
14481 (direct_move_p): New function to return true if the operation can
14482 be done as a direct move instruciton.
14483 (quad_load_store_p): New function to return true if the operation
14484 is a quad memory operation.
14485 (rs6000_legitimize_address): If quad memory, only allow register
14486 indirect for TImode addresses.
14487 (rs6000_legitimate_address_p): Likewise.
14488 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14489 (rs6000_reload_register_type): Likewise.
14490 (register_to_reg_type): Return register type.
14491 (rs6000_secondary_reload_simple_move): New helper function for
14492 secondary reload and secondary memory needed to identify anything
14493 that is a simple move, and does not need reloading.
14494 (rs6000_secondary_reload_direct_move): New helper function for
14495 secondary reload to identify cases that can be done with several
14496 instructions via the direct move instructions.
14497 (rs6000_secondary_reload_move): New helper function for secondary
14498 reload to identify moves between register types that can be done.
14499 (rs6000_secondary_reload): Add support for quad memory operations
14500 and for direct move.
14501 (rs6000_secondary_memory_needed): Likewise.
14502 (rs6000_debug_secondary_memory_needed): Change argument names.
14503 (rs6000_output_move_128bit): New function to return the move to
14504 use for 128-bit moves, including knowing about the various
14505 limitations of quad memory operations.
14507 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14508 memory operations. call rs6000_output_move_128bit for the actual
14509 instruciton(s) to generate.
14510 (vsx_movti_64bit): Likewise.
14512 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14513 (UNSPEC_P8V_MTVSRWZ): Likewise.
14514 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14515 (UNSPEC_P8V_MTVSRD): Likewise.
14516 (UNSPEC_P8V_XXPERMDI): Likewise.
14517 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14518 (UNSPEC_FUSION_GPR): Likewise.
14519 (FMOVE128_GPR): New iterator for direct move.
14520 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14521 (f32_sv): Likewise.
14522 (f32_dm): Likewise.
14523 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14524 loads and direct move instructions.
14525 (zero_extendsidi2_lfiwzx): Likewise.
14526 (extendsidi2_lfiwax): Likewise.
14527 (extendsidi2_nocell): Likewise.
14528 (floatsi<mode>2_lfiwax): Likewise.
14529 (lfiwax): Likewise.
14530 (floatunssi<mode>2_lfiwzx): Likewise.
14531 (lfiwzx): Likewise.
14532 (fix_trunc<mode>_stfiwx): Likewise.
14533 (fixuns_trunc<mode>_stfiwx): Likewise.
14534 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14535 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14536 (parity<mode>2_cmpb): Set length/type attr.
14537 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14538 for 'mr.' to fast_compare.
14539 (bpermd_<mode>): Change type attr to popcnt.
14540 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14541 (p8_mtvsrwz_1): Likewise.
14542 (p8_mtvsrwz_2): Likewise.
14543 (reload_fpr_from_gpr<mode>): Likewise.
14544 (p8_mtvsrd_1): Likewise.
14545 (p8_mtvsrd_2): Likewise.
14546 (p8_xxpermdi_<mode>): Likewise.
14547 (reload_vsx_from_gpr<mode>): Likewise.
14548 (reload_vsx_from_gprsf): Likewise.
14549 (p8_mfvsrd_3_<mode>): LIkewise.
14550 (reload_gpr_from_vsx<mode>): Likewise.
14551 (reload_gpr_from_vsxsf): Likewise.
14552 (p8_mfvsrd_4_disf): Likewise.
14553 (multi-word GPR splits): Do not split direct moves or quad memory
14556 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14558 * tree-into-ssa.c (interesting_blocks): Make static.
14560 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14563 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14564 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14565 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14566 (get_variable_align): New function.
14567 (get_variable_section, emit_bss, emit_common,
14568 assemble_variable_contents, place_block_symbol): Use
14569 get_variable_align instead of DECL_ALIGN.
14570 (assemble_noswitch_variable): Add align argument, use it
14571 instead of DECL_ALIGN.
14572 (assemble_variable): Adjust caller. Use get_variable_align
14573 instead of DECL_ALIGN.
14574 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14576 (DATA_ABI_ALIGNMENT): Define.
14577 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14578 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14579 opt is false, only return the psABI mandated alignment increase.
14580 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14581 (DATA_ABI_ALIGNMENT): ... this.
14582 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14583 (DATA_ABI_ALIGNMENT): ... this.
14584 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14585 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14586 (DATA_ABI_ALIGNMENT): ... this.
14587 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14588 * doc/tm.texi: Regenerated.
14590 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14592 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14593 cmp_code to construct REG_EQUAL note.
14595 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14598 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14599 that operands[2] doesn't overlap with operands[0].
14601 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14602 Jan Hubicka <jh@suse.cz>
14604 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14605 hack to mark symbols as used.
14607 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14609 PR rtl-optimization/57559
14610 * lra-constraints.c (process_alt_operands): Don't discourage
14611 memory with known offset for offsetable memory constraint.
14612 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14614 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14616 * varasm.c (struct oc_local_state): Reorder fields.
14617 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14618 and adjust accordingly.
14619 (output_constructor): Reorder initialization code and adjust call to
14620 output_constructor_bitfield.
14622 2013-06-07 Jan Hubicka <jh@suse.cz>
14624 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14626 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14628 * tree-object-size.c (unknown): Make const.
14630 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14632 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14633 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14634 for last alternative in the cpu_facility attribute.
14636 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14639 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14640 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14641 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14642 * config/arm/constraints.md (Dg): New constraint.
14643 * config/arm/neon.md (xordi3_neon): Remove.
14644 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14645 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14647 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14649 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14650 Clean up alternatives.
14652 2013-06-07 Alan Modra <amodra@gmail.com>
14654 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14657 2013-06-07 Alan Modra <amodra@gmail.com>
14659 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14661 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14663 * config/arm/constraints.md (Df): New constraint.
14664 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14665 Correct length attribute for last two alternatives.
14667 2013-06-07 Alan Modra <amodra@gmail.com>
14669 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14670 override user -mfp-in-toc.
14671 (offsettable_ok_by_alignment): Consider just the current access
14672 rather than the whole object, unless BLKmode. Handle
14673 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14674 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14675 for -mcmodel=medium.
14676 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14677 override user -mfp-in-toc or -msum-in-toc. Default to
14678 -mno-fp-in-toc for -mcmodel=medium.
14680 2013-06-06 DJ Delorie <dj@redhat.com>
14682 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14683 TARGET_VALID_POINTER_MODE.
14685 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14686 Pat Haugen <pthaugen@us.ibm.com>
14687 Peter Bergner <bergner@vnet.ibm.com>
14689 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14690 Document new power8 builtins.
14692 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14693 condition code register, to allow 128-bit logical operations to be
14694 done in the VSX or GPR registers.
14695 (nor<mode>3): Use the canonical form for nor.
14696 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14697 vclz*, and vpopcnt* vector instructions.
14698 (nand<mode>3): Likewise.
14699 (orc<mode>3): Likewise.
14700 (clz<mode>2): LIkewise.
14701 (popcount<mode>2): Likewise.
14703 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14704 that only the GPRs are recognized.
14706 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14707 support for new power8 builtins.
14709 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14711 (xscvdpspn): Likewise.
14717 (vpopcnt): Likewise.
14718 (vpopcntb): Likewise.
14719 (vpopcnth): Likewise.
14720 (vpopcntw): Likewise.
14721 (vpopcntd): Likewise.
14723 (vmrgew): Likewise.
14724 (vmrgow): Likewise.
14726 (eqv_v16qi3): Likewise.
14727 (eqv_v8hi3): Likewise.
14728 (eqv_v4si3): Likewise.
14729 (eqv_v2di3): Likewise.
14730 (eqv_v4sf3): Likewise.
14731 (eqv_v2df3): Likewise.
14733 (nand_v16qi3): Likewise.
14734 (nand_v8hi3): Likewise.
14735 (nand_v4si3): Likewise.
14736 (nand_v2di3): Likewise.
14737 (nand_v4sf3): Likewise.
14738 (nand_v2df3): Likewise.
14740 (orc_v16qi3): Likewise.
14741 (orc_v8hi3): Likewise.
14742 (orc_v4si3): Likewise.
14743 (orc_v2di3): Likewise.
14744 (orc_v4sf3): Likewise.
14745 (orc_v2df3): Likewise.
14747 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14748 allow power8 quad mode in 64-bit.
14749 (rs6000_builtin_vectorized_function): Add support to vectorize
14750 ISA 2.07 count leading zeros, population count builtins.
14751 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14752 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14753 (builtin_function_type): Add vgbbd builtin function which takes an
14755 (altivec_expand_vec_perm_const): Add support for new power8 merge
14758 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14759 that does not include TImdoe for use with 32-bit.
14760 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14762 (UNSPEC_VSX_CVDPSPN): Likewise.
14763 (vsx_xscvdpspn): Likewise.
14764 (vsx_xscvspdpn): Likewise.
14765 (vsx_xscvdpspn_scalar): Likewise.
14766 (vsx_xscvspdpn_directmove): Likewise.
14767 (vsx_and<mode>3): Split logical operations into 32-bit and
14768 64-bit. Add support to do logical operations on TImode as well as
14769 VSX vector types. Allow logical operations to be done in either
14770 VSX registers or in general purpose registers in 64-bit mode. Add
14771 splitters if GPRs were used. For AND, add clobber of CCmode to
14772 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14774 (vsx_and<mode>3_32bit): Likewise.
14775 (vsx_and<mode>3_64bit): Likewise.
14776 (vsx_ior<mode>3): Likewise.
14777 (vsx_ior<mode>3_32bit): Likewise.
14778 (vsx_ior<mode>3_64bit): Likewise.
14779 (vsx_xor<mode>3): Likewise.
14780 (vsx_xor<mode>3_32bit): Likewise.
14781 (vsx_xor<mode>3_64bit): Likewise.
14782 (vsx_one_cmpl<mode>2): Likewise.
14783 (vsx_one_cmpl<mode>2_32bit): Likewise.
14784 (vsx_one_cmpl<mode>2_64bit): Likewise.
14785 (vsx_nor<mode>3): Likewise.
14786 (vsx_nor<mode>3_32bit): Likewise.
14787 (vsx_nor<mode>3_64bit): Likewise.
14788 (vsx_andc<mode>3): Likewise.
14789 (vsx_andc<mode>3_32bit): Likewise.
14790 (vsx_andc<mode>3_64bit): Likewise.
14791 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14792 and xxlorc instructions.
14793 (vsx_eqv<mode>3_64bit): Likewise.
14794 (vsx_nand<mode>3_32bit): Likewise.
14795 (vsx_nand<mode>3_64bit): Likewise.
14796 (vsx_orc<mode>3_32bit): Likewise.
14797 (vsx_orc<mode>3_64bit): Likewise.
14799 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14801 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14803 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14804 (p8_vmrgow): Likewise.
14805 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14806 GPRs to be split under VSX.
14807 (p8v_clz<mode>2): Add power8 count leading zero support.
14808 (p8v_popcount<mode>2): Add power8 population count support.
14809 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14812 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14815 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14817 (vec_nand): Likewise.
14818 (vec_vclz): Likewise.
14819 (vec_vclzb): Likewise.
14820 (vec_vclzd): Likewise.
14821 (vec_vclzh): Likewise.
14822 (vec_vclzw): Likewise.
14823 (vec_vgbbd): Likewise.
14824 (vec_vmrgew): Likewise.
14825 (vec_vmrgow): Likewise.
14826 (vec_vpopcnt): Likewise.
14827 (vec_vpopcntb): Likewise.
14828 (vec_vpopcntd): Likewise.
14829 (vec_vpopcnth): Likewise.
14830 (vec_vpopcntw): Likewise.
14832 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14834 PR rtl-optimization/57468
14835 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14838 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14840 PR rtl-optimization/57459
14841 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14842 type when setting live regs.
14844 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14846 * config/s390/s390.opt (mlra): New option.
14847 * config/s390/s390.c (s390_decompose_address): Check displacement
14848 for all registers for LRA.
14849 (s390_secondary_reload): Don't used secondary reloads for LRA.
14850 (s390_lra_p): New function.
14851 (TARGET_LRA_P): Define.
14852 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14853 of attribute cpu_facility to zarch for the last alternative.
14854 (*cmpmem_short): Ditto.
14856 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14858 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14859 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14860 (arm_expand_prologue): Likewise.
14862 2013-06-06 Teresa Johnson <tejohnson@google.com>
14865 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14866 as this is now done by redirect_edge_and_branch_force.
14867 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14868 barriers, and fix interaction with splitting.
14869 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14870 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14871 reflect changes made in the routine.
14872 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14873 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14874 since this is called in cfglayout mode, and replace partition fixup
14875 with assert as that is now done by force_nonfallthru_and_redirect.
14876 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14877 already be marked with region crossing note.
14878 (insert_section_boundary_note): Make non-static, gate on flag
14879 has_bb_partition, rewrite to also check for multiple partitions.
14880 (rest_of_handle_reorder_blocks): Remove call to
14881 insert_section_boundary_note, now done later during free_cfg.
14882 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14883 * bb-reorder.h (insert_section_boundary_note): Declare.
14884 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14885 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14886 invoke insert_section_boundary_note.
14887 (try_redirect_by_replacing_jump): Remove unnecessary
14888 check for region crossing note.
14889 (fixup_partition_crossing): New function.
14890 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14891 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14892 in non-cfglayout mode.
14893 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14894 remove old code that tried to do this. Emit barrier correctly
14895 when we are in cfglayout mode.
14896 (last_bb_in_partition): New function.
14897 (rtl_split_edge): Correctly fixup partition boundaries.
14898 (commit_one_edge_insertion): Remove old code that tried to
14899 fixup region crossing edge since this is now handled in
14900 split_block, and set up insertion point correctly since
14901 block may now end in a jump.
14902 (verify_hot_cold_block_grouping): Guard against checking when not in
14903 linearized RTL mode.
14904 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14906 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14907 rtl_verify_flow_info, so not called in cfglayout mode.
14908 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14909 (fixup_reorder_chain): Remove old code that attempted to fixup region
14910 crossing note as this is now handled in force_nonfallthru_and_redirect.
14911 (duplicate_insn_chain): Don't duplicate switch section notes.
14912 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14914 * basic-block.h (emit_barrier_after_bb): Declare.
14916 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14918 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14919 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14920 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14922 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14924 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14926 * config/arm/ldmstm.md: Regenerate.
14928 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14930 * config/arm/sync.md (atomic_loaddi_1):
14931 Disable predication for arm_restrict_it.
14932 (arm_load_exclusive<mode>): Likewise.
14933 (arm_load_exclusivesi): Likewise.
14934 (arm_load_exclusivedi): Likewise.
14935 (arm_load_acquire_exclusive<mode>): Likewise.
14936 (arm_load_acquire_exclusivesi): Likewise.
14937 (arm_load_acquire_exclusivedi): Likewise.
14938 (arm_store_exclusive<mode>): Likewise.
14939 (arm_store_exclusive<mode>): Likewise.
14940 (arm_store_release_exclusivedi): Likewise.
14941 (arm_store_release_exclusive<mode>): Likewise.
14943 2013-06-06 Richard Biener <rguenther@suse.de>
14945 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14947 (lto_tag_is_tree_code_p): Adjust.
14948 (lto_tag_is_gimple_code_p): Likewise.
14949 (lto_gimple_code_to_tag): Likewise.
14950 (lto_tag_to_gimple_code): Likewise.
14951 (lto_tree_code_to_tag): Likewise.
14952 (lto_tag_to_tree_code): Likewise.
14953 * data-streamer.h (streamer_write_hwi_in_range): Use
14954 uhwi streaming to stream the normalized range.
14955 (streamer_read_hwi_in_range): Likewise.
14957 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14959 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14960 (predicable_short_it): Likewise.
14961 (predicated): Likewise.
14962 (enabled): Handle above.
14963 (define_cond_exec): Set predicated attribute to yes.
14965 2013-06-05 Mike Stump <mikestump@comcast.net>
14967 * gdbinit.in (__FUNCTION__): Add.
14969 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14971 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14972 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14974 2013-06-05 Jan Hubicka <jh@suse.cz>
14976 * varasm.c (mark_decl_referenced): Revert the removal until targets
14979 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14981 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14982 instead of mark_decl_referenced.
14984 2013-06-05 Jan Hubicka <jh@suse.cz>
14986 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14987 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14988 and symtab_used_from_object_file_p.
14989 (cgraph_make_node_local_1): Clear forced_by_abi.
14990 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14991 * cgraph.h (symtab_node_base): Add forced_by_abi.
14992 (decide_is_variable_needed): Remove.
14993 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14994 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14995 (decide_is_symbol_needed): ... this one; handle symbols in general;
14996 always analyze virtuals; honnor forced_by_abi.
14997 (cgraph_finalize_function): Update.
14998 (varpool_finalize_decl): Update.
14999 (symbol_defined_and_needed): Remove.
15000 (analyze_functions): Update.
15001 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15002 output_refs, input_overwrite_node): Handle forced_by_abi.
15003 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15004 (address_taken_from_non_vtable_p): ... this one.
15005 (comdat_can_be_unshared_p_1): New function.
15006 (cgraph_comdat_can_be_unshared_p): Rename to ...
15007 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15008 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15009 (function_and_variable_visibility): Clear forced_by_abi as needed.
15010 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15011 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15012 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15013 * varpool.c (decide_is_variable_needed): Remove.
15015 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15017 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15018 (arm_option_override): Override arm_restrict_it where appropriate.
15019 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15020 * config/arm/arm.opt (mrestrict-it): New command-line option.
15021 * doc/invoke.texi: Document -mrestrict-it.
15023 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15025 * tsan.c (tsan_atomic_table): Make const.
15027 2013-06-05 Richard Biener <rguenther@suse.de>
15029 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15030 index associated with the tree we are supposed to replace.
15031 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15032 TREE_ASM_WRITTEN as zero for everything but SSA names.
15034 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15036 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15038 2013-06-04 Jan Hubicka <jh@suse.cz>
15040 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15041 (input_node, input_varpool_node): Handle correctly external same
15043 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15044 nodes at ltrans stage.
15046 2013-06-04 Jan Hubicka <jh@suse.cz>
15048 * ipa-inline.c (update_caller_keys): Fix availability test.
15049 (update_callee_keys): Likewise.
15050 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15051 to follow ELF standard.
15053 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15055 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15056 (mips64r5900el-*-elf*): New configurations.
15057 * config/mips/mips-cpus.def (r5900): New processor.
15058 * config/mips/mips-tables.opt: Regenerate.
15059 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15060 (mips_issue_rate): Handle PROCESSOR_R5900.
15061 (mips_reorg_process_insns): Force reorder mode for the R5900.
15062 * config/mips/mips.h (TARGET_MIPS5900): Define.
15063 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15065 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15067 * config/mips/mips.md (processor): Add r5900.
15068 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15070 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15072 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15073 into function to generate MOVI instruction.
15074 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15075 (aarch64_preferred_simd_mode): Turn into wrapper.
15076 (aarch64_output_scalar_simd_mov_immediate): New function.
15077 * config/aarch64/aarch64-protos.h: Add prototype for above.
15079 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15081 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15082 element_char member.
15083 (sizetochar): Return signed char.
15084 (aarch64_simd_valid_immediate): Remove elchar and other
15085 unnecessary variables.
15086 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15087 Calculate element_char as required.
15088 * config/aarch64/aarch64-protos.h: Update and move prototype
15089 for aarch64_output_simd_mov_immediate.
15090 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15093 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15095 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15096 information completed by aarch64_simd_valid_immediate.
15097 (aarch64_legitimate_constant_p): Update arguments.
15098 (aarch64_simd_valid_immediate): Work with struct rather than many
15100 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15101 (aarch64_simd_make_constant): Update arguments.
15102 (aarch64_output_simd_mov_immediate): Work with struct rather than
15103 many pointers. Output immediate directly rather than as operand.
15104 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15106 * config/aarch64/constraints.md (Dn): Update arguments.
15108 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15110 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15112 (aarch64_simd_immediate_valid_for_move): Remove.
15113 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15114 (aarch64_simd_make_constant): Update call.
15115 (aarch64_output_simd_mov_immediate): Update call.
15116 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15118 * config/aarch64/constraints.md (Dn): Update call.
15120 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15122 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15123 return type to bool for prototype.
15124 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15125 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15126 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15128 2013-06-04 Catherine Moore <clm@codesourcery.com>
15130 * config/mips/mips.opt (meva): New.
15131 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15132 (ASM_SPEC): Handle -meva.
15133 * doc/invoke.texi (meva): Document.
15135 2013-06-04 Alan Modra <amodra@gmail.com>
15137 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15140 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15142 * rtl.def: Add extra fourth optional field to define_cond_exec.
15143 * gensupport.c (process_one_cond_exec): Process attributes from
15145 * doc/md.texi: Document fourth field in define_cond_exec.
15147 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15149 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15150 out the processing order as in store_bit_field_1.
15152 2013-06-04 Jan Hubicka <jh@suse.cz>
15154 PR middle-end/57500
15155 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15156 non-VAR_DECL node if it does not exist yet.
15158 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15160 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15161 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15162 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15163 target_cpu_default setting.
15165 2013-06-03 Teresa Johnson <tejohnson@google.com>
15167 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15168 default to -fopt-info=optimized instead of all.
15169 * doc/invoke.texi: Ditto.
15170 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15171 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15172 (execute_vect_slp): Emit BB vectorization success under
15173 MSG_OPTIMIZED_LOCATIONS.
15174 * tree-vect-slp.c (vect_slp_transform_bb): Change
15175 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15176 * tree-vect-loop.c (vect_transform_loop): Ditto.
15178 2013-06-03 Jason Merrill <jason@redhat.com>
15181 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15182 Use TARGET_EXPR for C++.
15184 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15186 PR rtl-optimization/57268
15187 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15188 if DEBUG_INSN_P (insn).
15191 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15193 PR rtl-optimization/57268
15194 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15195 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15197 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15199 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15200 (ix86_avoid_lea_for_addr): Likewise.
15201 (exact_dependency_1): Likewise.
15202 (ix86_adjust_cost): Likewise.
15203 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15205 (do_reorder_for_imul): Fix typo, formatting and
15206 !reload_completed check removed.
15207 (ix86_sched_reorder): Fix typo and formatting.
15208 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15211 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15213 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15215 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15217 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15218 <STRING_CST>: Likewise.
15219 <VECTOR_CST>: Likewise.
15221 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15222 Mikael Morin <mikael@gcc.gnu.org>
15224 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15225 * config.in: Regenerated.
15226 * configure: Regenerated.
15228 2013-06-01 Jan Hubicka <jh@suse.cz>
15230 PR middle-end/57366
15231 * cgraphunit.c (compile): When weakref is not supported,
15232 set up transparent aliases before final output pass.
15233 * varasm.c (assemble_alias): Do not try to do it here.
15235 2013-06-01 Jan Hubicka <jh@suse.cz>
15237 PR middle-end/57467
15238 * passes.c (for_per_function): Skip unanalyzed functions.
15240 2013-06-01 Jan Hubicka <jh@suse.cz>
15242 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15243 (lto_symtab_merge_symbols_1): ... this one.
15244 (lto_symtab_merge_cgraph_nodes): Rename to ...
15245 (lto_symtab_merge_symbols): ... this one; simplify.
15246 * cgraph.c (same_body_aliases_done): Rename to ...
15247 (cpp_implicit_aliases_done): ... this one.
15248 (cgraph_create_function_alias): Update.
15249 (cgraph_same_body_alias): Update.
15250 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15251 (verify_edge_corresponds_to_fndecl): Simplify.
15252 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15253 (cgraph_node): Remove same_body_alias.
15254 (varpool_node): Remove alias_of and extra_name_alias.
15255 (same_body_aliases_done): Rename to ..
15256 (cpp_implicit_aliases_done): ... this one.
15257 (symtab_alias_ultimate_target): Add default parameter.
15258 (symtab_resolve_alias): New function.
15259 (fixup_same_cpp_alias_visibility): Declare.
15260 (cgraph_function_node): Add default parameter.
15261 (cgraph_node_asm_name): Likewise.
15262 (cgraph_function_or_thunk_node): Add default parameter; do
15263 not ICE when it is NULL.
15264 (varpool_variable_node): Likewise.
15265 * tree-emutls.c (create_emultls_var): Update.
15266 (ipa_lower_emutls): Update.
15267 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15268 (cgraph_reset_node): Reset alias info.
15269 (cgraph_finalize_function): Update.
15270 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15271 (analyze_function): Simplify.
15272 (cgraph_process_same_body_aliases): Simplify.
15273 (analyze_functions): Fixup same body aliases.
15274 (handle_alias_pairs): Simplify.
15275 (assemble_thunk): Update.
15276 (assemble_thunks_and_aliases): Update.
15277 (output_weakrefs): Rewrite.
15278 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15279 (lto_output_varpool_node): Likewise.
15280 (compute_ltrans_boundary): Remve assert.
15281 (get_alias_symbol): New functoin.
15282 (input_node): Rewrite alias handling.
15283 (input_varpool_node): Likewise.
15284 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15285 * ipa.c (process_references): Handle weakrefs correctly.
15286 (symtab_remove_unreachable_nodes): Likewise.
15287 * trans-mem.c (get_cg_data): Update.
15288 (ipa_tm_create_version_alias): Update.
15289 (ipa_tm_execute): Update.
15290 * symtab.c (dump_symtab_base): Dump aliases.
15291 (verify_symtab_base): Verify aliases.
15292 (symtab_node_availability): New function.
15293 (symtab_alias_ultimate_target): Simplify.
15294 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15295 handle all the fixup cases.
15296 (symtab_resolve_alias): New function.
15297 * passes.c (ipa_write_summaries): Handle weakrefs.
15298 * varpool.c (varpool_analyze_node): Simplify.
15299 (assemble_aliases): Update.
15300 (varpool_create_variable_alias): Simplify.
15301 (varpool_extra_name_alias): Simplify.
15302 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15303 (lto_symtab_merge_symbols): ... this one.
15305 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15308 PR rtl-optimization/57268
15309 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15310 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15312 2013-06-01 Tobias Burnus <burnus@net-b.de>
15314 Partially reverted:
15315 2013-05-31 Tobias Burnus <burnus@net-b.de>
15317 PR middle-end/57073
15318 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15321 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15323 PR rtl-optimization/57268
15324 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15325 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15327 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15329 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15330 unordered comparison operators when -fno-trapping-math is in effect
15332 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15333 and implement unordered comparison operators properly on the e500.
15335 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15337 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15338 for constant scalar integers.
15339 (simplify_relational_operation_1): Likewise.
15341 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15343 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15344 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15347 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15348 Igor Zamyatin <igor.zamyatin@intel.com>
15350 Silvermont (SLM) architecture performance tuning.
15351 * config/i386/i386.h (enum ix86_tune_indices): Add
15352 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15353 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15355 * config/i386/i386.c (initial_ix86_tune_features)
15356 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15357 (ix86_lea_outperforms): Handle Silvermont tuning.
15358 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15360 (ix86_use_lea_for_mov): Likewise.
15361 (ix86_avoid_lea_for_addr): Likewise.
15362 (ix86_lea_for_add_ok): Likewise.
15363 (exact_dependency_1): New function.
15364 (exact_store_load_dependency): Likewise.
15365 (ix86_adjust_cost): Handle Silvermont tuning.
15366 (do_reoder_for_imul): Likewise.
15367 (swap_top_of_ready_list): New function.
15368 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15370 * config/i386/i386.md (peepholes that split memory operand in fp
15373 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15375 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15376 Remove un-necessary braces.
15378 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15380 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15381 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15383 2013-05-31 Tobias Burnus <burnus@net-b.de>
15385 PR middle-end/57073
15386 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15388 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15391 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15392 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15393 * config/arm/neon.md (iordi3_neon): Remove.
15394 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15395 * config/arm/predicates.md (imm_for_neon_logic_operand):
15396 Move to earlier in the file.
15397 (neon_logic_op2): Likewise.
15398 (arm_iordi_operand_neon): New predicate.
15400 2013-05-31 Richard Biener <rguenther@suse.de>
15402 PR tree-optimization/57478
15403 PR tree-optimization/57453
15404 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15407 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15409 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15410 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15412 2013-05-30 Tobias Burnus <burnus@net-b.de>
15413 Thomas Koenig <tkoenig@gcc.gnu.org>
15415 PR middle-end/57073
15416 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15417 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15419 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15421 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15423 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15425 * target.def (register_usage_leveling_p): New hook.
15426 * targhooks.c (default_register_usage_leveling_p): New.
15427 * targhooks.h (default_register_usage_leveling_p): New prototype.
15428 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15429 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15430 * doc/tm.texi: Update.
15431 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15433 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15435 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15436 (*insv_reg<mode>): New define_insn.
15438 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15440 PR rtl-optimization/57439
15441 * postreload.c (move2add_valid_value_p): Check that we have
15442 a zero subreg_regno_offset when accessing the register in
15443 the requested mode.
15445 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15446 Igor Zamyatin <igor.zamyatin@intel.com>
15448 Silvermont (SLM) architecture pipeline model, tuning and
15450 * config.gcc: Add slm config options and target.
15452 * config/i386/slm.md: New.
15454 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15456 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15458 (ix86_target_macros_internal): Likewise.
15460 * config/i386/i386.c (slm_cost): New cost.
15461 (m_SLM): New macro flag.
15462 (initial_ix86_tune_features): Set m_SLM.
15463 (x86_accumulate_outgoing_args): Likewise.
15464 (x86_arch_always_fancy_math_387): Likewise.
15465 (processor_target_table): Add slm cost.
15466 (cpu_names): Add slm cpu name.
15467 (x86_option_override_internal): Set SLM ISA.
15468 (ix86_issue_rate): New case PROCESSOR_SLM.
15469 (ia32_multipass_dfa_lookahead): Likewise.
15470 (fold_builtin_cpu): Add slm.
15472 * config/i386/i386.h (TARGET_SLM): New target macro.
15473 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15474 (processor_type): Add PROCESSOR_SLM.
15476 * config/i386/i386.md (cpu): Add new value "slm".
15477 (slm.md): Include slm.md.
15479 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15480 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15482 * config/arm/arm-protos.h: Add and update function protos.
15483 * config/arm/arm.c (use_simple_return_p): New added.
15484 (thumb2_expand_return): Check simple_return flag.
15485 * config/arm/arm.md: Add simple_return and conditional simple_return.
15486 * config/arm/iterators.md: Add iterator for return and simple_return.
15488 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15490 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15491 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15492 (arm_emit_vfp_multi_reg_pop): Likewise.
15493 (thumb2_emit_ldrd_pop): Likewise.
15494 (arm_expand_epilogue): Add misc REG_CFA notes.
15495 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15497 2013-05-29 Lawrence Crowl <crowl@google.com>
15499 * config/arm/t-arm: Update for below.
15501 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15502 Change type to hash_table. Update dependent calls and types.
15504 * config/i386/t-cygming: Update for below.
15506 * config/i386/t-interix: Update for below.
15508 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15509 Change type to hash_table. Update dependent calls and types.
15510 (i386_find_on_wrapper_list::wrappers): Likewise.
15512 * config/ia64/t-ia64: Update for below.
15514 * config/ia64/ia64.c (bundle_state_table):
15515 Change type to hash_table. Update dependent calls and types.
15517 * config/mips/mips.c (mips_reorg_process_insns::htab):
15518 Change type to hash_table. Update dependent calls and types.
15520 * config/sol2.c (solaris_comdat_htab):
15521 Change type to hash_table. Update dependent calls and types.
15523 * config/t-sol2: Update for above.
15525 2013-05-29 Teresa Johnson <tejohnson@google.com>
15527 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15528 functions are not yet marked as defined.
15530 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15531 Pat Haugen <pthaugen@us.ibm.com>
15532 Peter Bergner <bergner@vnet.ibm.com>
15534 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15538 (vrotl<mode>3): Likewise.
15539 (vashl<mode>3): Likewise.
15540 (vlshr<mode>3): Likewise.
15541 (vashr<mode>3): Likewise.
15543 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15544 support for power8 V2DI builtins.
15546 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15547 power8 V2DI builtins.
15548 (vupkhsw): Likewise.
15549 (vupklsw): Likewise.
15550 (vaddudm): Likewise.
15551 (vminsd): Likewise.
15552 (vmaxsd): Likewise.
15553 (vminud): Likewise.
15554 (vmaxud): Likewise.
15555 (vpkudum): Likewise.
15556 (vpksdss): Likewise.
15557 (vpkudus): Likewise.
15558 (vpksdus): Likewise.
15563 (vsubudm): Likewise.
15564 (vcmpequd): Likewise.
15565 (vcmpgtsd): Likewise.
15566 (vcmpgtud): Likewise.
15567 (vcmpequd_p): Likewise.
15568 (vcmpgtsd_p): Likewise.
15569 (vcmpgtud_p): Likewise.
15570 (vupkhsw): Likewise.
15571 (vupklsw): Likewise.
15572 (vaddudm): Likewise.
15573 (vmaxsd): Likewise.
15574 (vmaxud): Likewise.
15575 (vminsd): Likewise.
15576 (vminud): Likewise.
15577 (vpksdss): Likewise.
15578 (vpksdus): Likewise.
15579 (vpkudum): Likewise.
15580 (vpkudus): Likewise.
15585 (vsubudm): Likewise.
15587 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15588 support for power8 V2DI instructions.
15590 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15591 power8 V2DI instructions. Combine pack and unpack insns to use an
15592 iterator for each mode. Check whether a particular mode supports
15593 Altivec instructions instead of just checking TARGET_ALTIVEC.
15594 (UNSPEC_VPKUWUM): Likewise.
15595 (UNSPEC_VPKSHSS): Likewise.
15596 (UNSPEC_VPKSWSS): Likewise.
15597 (UNSPEC_VPKUHUS): Likewise.
15598 (UNSPEC_VPKSHUS): Likewise.
15599 (UNSPEC_VPKUWUS): Likewise.
15600 (UNSPEC_VPKSWUS): Likewise.
15601 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15602 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15603 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15604 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15605 (UNSPEC_VUPKHSB): Likewise.
15606 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15607 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15608 (UNSPEC_VUPKHSH): Likewise.
15609 (UNSPEC_VUPKLSB): Likewise.
15610 (UNSPEC_VUPKLSH): Likewise.
15612 (VI_char): Likewise.
15613 (VI_scalar): Likewise.
15614 (VI_unit): Likewise.
15616 (VP_small): Likewise.
15617 (VP_small_lc): Likewise.
15618 (VU_char): Likewise.
15619 (add<mode>3): Likewise.
15620 (altivec_vaddcuw): Likewise.
15621 (altivec_vaddu<VI_char>s): Likewise.
15622 (altivec_vadds<VI_char>s): Likewise.
15623 (sub<mode>3): Likewise.
15624 (altivec_vsubcuw): Likewise.
15625 (altivec_vsubu<VI_char>s): Likewise.
15626 (altivec_vsubs<VI_char>s): Likewise.
15627 (altivec_vavgs<VI_char>): Likewise.
15628 (altivec_vcmpbfp): Likewise.
15629 (altivec_eq<mode>): Likewise.
15630 (altivec_gt<mode>): Likewise.
15631 (altivec_gtu<mode>): Likewise.
15632 (umax<mode>3): Likewise.
15633 (smax<mode>3): Likewise.
15634 (umin<mode>3): Likewise.
15635 (smin<mode>3): Likewise.
15636 (altivec_vpkuhum): Likewise.
15637 (altivec_vpkuwum): Likewise.
15638 (altivec_vpkshss): Likewise.
15639 (altivec_vpkswss): Likewise.
15640 (altivec_vpkuhus): Likewise.
15641 (altivec_vpkshus): Likewise.
15642 (altivec_vpkuwus): Likewise.
15643 (altivec_vpkswus): Likewise.
15644 (altivec_vpks<VI_char>ss): Likewise.
15645 (altivec_vpks<VI_char>us): Likewise.
15646 (altivec_vpku<VI_char>us): Likewise.
15647 (altivec_vpku<VI_char>um): Likewise.
15648 (altivec_vrl<VI_char>): Likewise.
15649 (altivec_vsl<VI_char>): Likewise.
15650 (altivec_vsr<VI_char>): Likewise.
15651 (altivec_vsra<VI_char>): Likewise.
15652 (altivec_vsldoi_<mode>): Likewise.
15653 (altivec_vupkhsb): Likewise.
15654 (altivec_vupkhs<VU_char>): Likewise.
15655 (altivec_vupkls<VU_char>): Likewise.
15656 (altivec_vupkhsh): Likewise.
15657 (altivec_vupklsb): Likewise.
15658 (altivec_vupklsh): Likewise.
15659 (altivec_vcmpequ<VI_char>_p): Likewise.
15660 (altivec_vcmpgts<VI_char>_p): Likewise.
15661 (altivec_vcmpgtu<VI_char>_p): Likewise.
15662 (abs<mode>2): Likewise.
15663 (vec_unpacks_hi_v16qi): Likewise.
15664 (vec_unpacks_hi_v8hi): Likewise.
15665 (vec_unpacks_lo_v16qi): Likewise.
15666 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15667 (vec_unpacks_lo_v8hi): Likewise.
15668 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15669 (vec_pack_trunc_v8h): Likewise.
15670 (vec_pack_trunc_v4si): Likewise.
15671 (vec_pack_trunc_<mode>): Likewise.
15673 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15675 (vec_vmaxsd): Likewise.
15676 (vec_vmaxud): Likewise.
15677 (vec_vminsd): Likewise.
15678 (vec_vminud): Likewise.
15679 (vec_vpksdss): Likewise.
15680 (vec_vpksdus): Likewise.
15681 (vec_vpkudum): Likewise.
15682 (vec_vpkudus): Likewise.
15683 (vec_vrld): Likewise.
15684 (vec_vsld): Likewise.
15685 (vec_vsrad): Likewise.
15686 (vec_vsrd): Likewise.
15687 (vec_vsubudm): Likewise.
15688 (vec_vupkhsw): Likewise.
15689 (vec_vupklsw): Likewise.
15691 2013-05-29 Jan Hubicka <jh@suse.cz>
15693 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15694 flags; reorder rest of fields in more consistent way.
15695 (varpool_node): Remove analyzed, finalized and alias.
15696 (cgraph_ndoe): Likewise.
15697 (symtab_alias_ultimate_target): New function.
15698 (cgraph_function_node): Move offline.
15699 (cgraph_reset_node): Declare.
15700 (cgraph_comdat_can_be_unshared_p): Remove.
15701 (varpool_remove_initializer): Declare.
15702 (varpool_first_defined_variable, varpool_next_defined_variable
15703 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15704 (cgraph_function_with_gimple_body_p): Update.
15705 (varpool_all_refs_explicit_p): Update.
15706 (symtab_alias_target): New function.
15707 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15708 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15709 (cgraph_function_or_thunk_node): Simplify using
15710 symtab_alias_ultimate_target.
15711 (varpool_variable_node): Likewise.
15712 * cgraph.c (cgraph_create_function_alias): Update.
15713 (cgraph_add_thunk): Update.
15714 (cgraph_remove_node): Update.
15715 (dump_cgraph_node): Do not dump removed flags.
15716 (cgraph_function_body_availability): Update.
15717 (cgraph_propagate_frequency): Update.
15718 (verify_cgraph_node): Check sanity of local flag.
15719 (cgraph_function_node): Move here from cgraph.h; revamp for
15720 cgraph_function_or_thunk_node.
15721 * lto-symtab.c (lto_varpool_replace_node): Update.
15722 (lto_symtab_resolve_can_prevail_p): Update.
15723 (lto_symtab_merge_cgraph_nodes): Update.
15724 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15725 propagate_constants_accross_call, devirtualization_time_bonus,
15726 ipcp_propagate_stage): Update.
15727 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15728 * ipa-inline-transform.c (clone_inlined_nodes,
15729 preserve_function_body_p): Update.
15730 * ipa-reference.c (propagate): Update.
15731 (write_node_summary_p): Update.
15732 * toplev.c (wrapup_global_declaration_2): Update.
15733 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15734 (analyze_function) ... this one.
15735 (cgraph_process_new_functions): Update.
15736 (cgraph_reset_node): Export.
15737 (cgraph_finalize_function): Update.
15738 (cgraph_add_new_function): Update.
15739 (process_function_and_variable_attributes): Update.
15740 (varpool_finalize_decl): Update.
15741 (symbol_finalized): Remove.
15742 (symbol_finalized_and_needed): Rename to ...
15743 (symbol_defined_and_needed): ... update.
15744 (cgraph_analyze_functions): Update.
15745 (handle_alias_pairs): Update.
15746 (mark_functions_to_output): Update.
15747 (assemble_thunk): Update.
15748 (output_in_order): Update.
15749 (output_weakrefs): Update.
15750 (finalize_compilation_unit): Update.
15751 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15752 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15753 input_node, input_varpool_node): Update.
15754 * dbxout.c (dbxout_expand_expr): Update.
15755 * cgraphclones.c (cgraph_clone_node): Update.
15756 (cgraph_copy_node_for_versioning): Update.
15757 (cgraph_materialize_clone): Update.
15758 (cgraph_materialize_all_clones): Update.
15759 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15760 propagate_pure_const, propagate_nothrow): Update.
15761 * lto-streamer-out.c (lto_output, write_symbol): Update.
15762 * ipa-utils.c (ipa_reverse_postorder): Update.
15763 * ipa-inline.c (can_inline_edge_p): Update.
15764 (update_caller_keys, ipa_inline): Update.
15765 * dwarf2out.c (reference_to_unused,
15766 premark_types_used_by_global_vars_helper): Update.
15767 * tree-eh.c (tree_could_trap_p): Update.
15768 * ipa-split.c (consider_split, execute_split_functions): Update.
15769 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15770 has_addr_references_p): Update; move ahead in file for better
15772 (process_references): Simplify.
15773 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15774 bodies are removed.
15775 (cgraph_comdat_can_be_unshared_p): Make static.
15776 (cgraph_externally_visible_p): Update.
15777 (varpool_externally_visible_p): Update.
15778 (function_and_variable_visibility): Update.
15779 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15780 ipa_tm_mark_force_output_node): Update.
15781 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15782 estimate_edge_devirt_benefit, inline_generate_summary,
15783 inline_write_summary): Update.
15784 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15785 * ipa-prop.c (ipa_compute_jump_functions): Update.
15786 (ipa_print_node_params, ipa_prop_read_section,
15787 ipa_update_after_lto_read, read_replacements_section): Update.
15788 * varasm.c (mark_decl_referenced): Update.
15789 (assemble_alias, dump_tm_clone_pairs): Update.
15790 * tree-inline.c (copy_bb): Update.
15791 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15793 * symtab.c (dump_symtab_base): Print new flags.
15794 (verify_symtab_base): Verify new flags.
15795 (symtab_alias_ultimate_target): New function.
15796 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15797 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15799 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15801 * i386.c (ix86_get_function_versions_dispatcher,
15802 ix86_generate_version_dispatcher_body): Update.
15803 (fold_builtin_cpu): Use varpool_add_new_variable.
15804 * varpool.c (varpool_remove_initializer): Break out from ...
15805 (varpool_remove_node): ... this one.
15806 (dump_varpool_node, varpool_node_for_asm,
15807 cgraph_variable_initializer_availability, varpool_analyze_node,
15808 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15809 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15812 2013-05-29 Jan Hubicka <jh@suse.cz>
15814 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15816 2013-05-29 Easwaran Raman <eraman@google.com>
15818 PR tree-optimization/57442
15819 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15820 when control exits the main loop.
15822 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15824 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15826 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15827 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15828 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15831 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15833 PR tree-optimization/57441
15834 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15835 Don't limit size of incr_vec to number of candidates.
15837 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15839 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15840 and mips16 directories.
15841 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15842 (MULTILIB_DIRNAMES): Ditto.
15843 (MULTILIB_EXCEPTIONS): Add new exceptions.
15844 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15845 (MULTILIB_DIRNAMES): Ditto.
15846 (MULTILIB_EXCEPTIONS): Add new exceptions.
15848 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15849 Marcus Shawcroft <marcus.shawcroft@arm.com>
15851 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15852 SYMBOL_TINY_ABSOLUTE.
15853 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15854 SYMBOL_TINY_ABSOLUTE.
15855 (aarch64_expand_mov_immediate): Likewise.
15856 (aarch64_classify_symbol): Likewise.
15857 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15858 Permit SYMBOL_TINY_ABSOLUTE.
15859 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15861 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15862 Marcus Shawcroft <marcus.shawcroft@arm.com>
15864 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15865 Refactor if/switch. Replace gcc_assert with if.
15867 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15869 * config/i386/i386.c (initial_ix86_tune_features): Enable
15870 FP Reassociation for AMD bdver1 and bdver2.
15872 2013-05-29 Martin Jambor <mjambor@suse.cz>
15874 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15875 and IMAGPART_EXPR do not occur within other handled_components.
15877 2013-05-29 Richard Biener <rguenther@suse.de>
15879 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15880 access on whether the use is in the BB we currently try to
15882 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15883 vectorize to vect_bb_slp_scalar_cost.
15885 2013-05-29 Richard Biener <rguenther@suse.de>
15887 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15888 computing scalar cost offsetted by stmts that are kept live
15890 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15891 for computation of scalar cost.
15893 2013-05-28 Steve Ellcey <sellcey@mips.com>
15895 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15897 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15899 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15900 array notation built-in reduction functions.
15901 * doc/passes.texi (Passes): Added documentation about changes done
15903 * doc/invoke.texi (C Dialect Options): Added documentation about
15904 the -fcilkplus flag.
15905 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15906 (BUILTINS_DEF): Depend on cilkplus.def.
15907 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15908 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15909 * cilkplus.def: New file.
15911 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15913 PR rtl-optimization/57439
15914 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15916 2013-05-28 Easwaran Raman <eraman@google.com>
15918 PR tree-optimization/57337
15919 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15920 (find_insert_point): Correctly identify the insertion point
15921 when two statements with the same UID is compared.
15923 2013-05-28 Richard Biener <rguenther@suse.de>
15925 PR tree-optimization/56787
15926 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15927 from the list of data references.
15928 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15930 (vect_analyze_loop_operations): Likewise.
15931 (vect_transform_loop): Remove clobbers.
15933 2013-05-28 Martin Jambor <mjambor@suse.cz>
15935 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15936 and REALPART_EXPRs have scalar type.
15938 2013-05-28 Richard Biener <rguenther@suse.de>
15940 PR tree-optimization/57411
15941 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15943 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15944 virtual operand propagation.
15946 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15948 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15949 destination register for bmasksi_vis.
15950 (vector_init_bshuffle): Likewise.
15951 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15953 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15955 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15956 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15957 mode if the instruction isn't available in the original mode.
15958 * config/sparc/sparc.opt (mfix-ut699): New option.
15959 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15960 (divdf3): Turn into expander.
15961 (divdf3_nofix): New insn.
15962 (divdf3_fix): Likewise.
15963 (divsf3): Disable if -mfix-ut699.
15964 (sqrtdf2): Turn into expander.
15965 (sqrtdf2_nofix): New insn.
15966 (sqrtdf2_fix): Likewise.
15967 (sqrtsf2): Disable if -mfix-ut699.
15969 2013-05-27 Richard Biener <rguenther@suse.de>
15971 PR middle-end/57412
15972 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15973 block for the new loop.
15975 2013-05-27 Richard Biener <rguenther@suse.de>
15977 PR tree-optimization/57343
15978 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15979 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15980 (number_of_iterations_cond): Do not build the folded tree.
15982 2013-05-27 Richard Biener <rguenther@suse.de>
15985 PR middle-end/57381
15986 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15987 OEP_CONSTANT_ADDRESS_OF retained.
15989 PR tree-optimization/57417
15990 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15991 for unchanged base.
15992 (set_ssa_val_to): Compare addresses using
15993 get_addr_base_and_unit_offset.
15995 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15997 PR rtl-optimization/56833
15998 * postreload.c (move2add_record_mode): New function.
15999 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16000 (move2add_use_add2_insn): Use move2add_record_sym_value.
16001 (move2add_use_add3_insn): Likewise.
16002 (reload_cse_move2add): Use move2add_valid_value_p and
16003 move2add_record_mode. Invalidate call-clobbered and REG_INC
16004 affected regs by setting reg_mode to VOIDmode.
16005 (move2add_note_store): Don't pretend the inside of a SUBREG is
16006 the actual destination. Invalidate single/leading registers by
16007 setting reg_mode to VOIDmode.
16008 Use move2add_record_sym_value, move2add_valid_value_p and
16009 move2add_record_mode.
16011 2013-05-27 Richard Biener <rguenther@suse.de>
16013 PR tree-optimization/57396
16014 * tree-affine.c (double_int_constant_multiple_p): Properly
16015 return false for val == 0 and div != 0.
16017 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16019 * config/mips/mips.h: Use #elif in preprocessor conditions.
16021 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16024 * config/mips/constraints.md (kl): New constraint.
16025 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16026 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16027 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16028 from LO for MIPS16.
16029 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16031 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16034 * config/mips/mips.c (mips_can_inline_p): New function.
16035 (TARGET_CAN_INLINE_P): Define.
16037 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16039 * sched-int.h (ds_t, dw_t): Make unsigned int.
16040 Fix documentation that describes how all the ds_t bits are used.
16041 Reserve the last bit for delayed-branch scheduling.
16042 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16043 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16044 (gen_dep_weak_1): Remove prototype.
16045 * sched-deps.c (get_dep_weak_1): Make static.
16046 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16047 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16048 * doc/tm.texi: Regenerate.
16049 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16051 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16054 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16056 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16057 Sandra Loosemore <sandra@codesourcery.com>
16059 * config.gcc (powerpc-*): Allow native for with-cpu.
16061 2013-05-24 Jeff Law <law@redhat.com>
16063 PR tree-optimization/57124
16064 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16065 conversion feeding a condition if the range has an overflow
16066 if -fstrict-overflow. Add warnings for when we do make the
16069 2013-05-24 Dehao Chen <dehao@google.com>
16071 * tree-cfg.c (locus_discrim_map): Fix the typo.
16072 (locus_discrim_hasher): Likewise.
16073 (locus_discrim_hasher::hash): Likewise.
16074 (locus_discrim_hasher::equal): Likewise.
16076 2013-05-24 Martin Jambor <mjambor@suse.cz>
16078 PR tree-optimization/57294
16079 * cgraph.h (ipa_record_stmt_references): Declare.
16080 * cgraphbuild.c (ipa_record_stmt_references): New function.
16081 (build_cgraph_edges): Use ipa_record_stmt_references.
16082 (rebuild_cgraph_edges): Likewise.
16083 (cgraph_rebuild_references): Likewise.
16084 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16085 associated with the old statement and build references from the
16086 newly built statements.
16087 * ipa-ref.c (ipa_remove_stmt_references): New function.
16088 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16090 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16092 * lra-constraints.c (emit_spill_move): Use smaller mode for
16094 (check_and_process_move): Consider mem-reg moves for secondary
16096 (curr_insn_transform): Don't lose insns emitted before for
16097 secondary memory moves.
16098 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16099 reg set up in the current insn.
16101 2013-05-24 Dehao Chen <dehao@google.com>
16103 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16105 (locus_descrim_hasher::equal): Likewise.
16106 (build_gimple_cfg): New discriminator assignment algorithm.
16107 (make_edges): Likewise.
16108 (next_discriminator_for_locus): Likewise.
16109 (same_line_p): Likewise.
16110 (assign_discriminators): Likewise.
16111 (make_cond_expr_edges): Likewise.
16112 (make_gimple_switch_edges): Likewise.
16113 (make_goto_expr_edges): Likewise.
16114 (make_gimple_asm_edges): Likewise.
16116 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16118 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16119 X format specifier to only display bottom 16 bits.
16120 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16121 immediate to match for operand 2, since it will be masked.
16123 2013-05-24 Richard Biener <rguenther@suse.de>
16125 PR tree-optimization/57287
16126 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16127 all SSA names that occur in abnormal PHIs.
16129 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16132 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16133 that index is not negative.
16135 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16137 PR rtl-optimization/55177
16138 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16139 (simplify_byte_swapping_operation): New.
16140 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16141 (simplify_relational_operation_1): Deal with BSWAP.
16143 2013-05-23 Richard Henderson <rth@redhat.com>
16146 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16147 (ix86_reorg): Call it.
16149 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16152 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16153 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16154 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16156 2013-05-23 Christian Bruel <christian.bruel@st.com>
16159 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16161 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16162 Marcus Shawcroft <marcus.shawcroft@arm.com>
16164 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16165 * config/aarch64/constraints.md (Usa): Remove.
16166 * doc/md.texi (AArch64 Usa): Remove.
16168 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16169 Marcus Shawcroft <marcus.shawcroft@arm.com>
16171 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16172 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16173 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16174 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16176 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16178 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16179 instruction (AdvSIMD).
16180 * config/aarch64/aarch64-builtins.c
16181 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16182 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16184 2013-05-23 Martin Jambor <mjambor@suse.cz>
16186 PR middle-end/57347
16187 * tree.h (contains_bitfld_component_ref_p): Declare.
16188 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16189 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16191 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16192 not access a bit-field. Assert all final offsets are byte-aligned.
16194 2013-05-23 Richard Biener <rguenther@suse.de>
16196 PR tree-optimization/57380
16197 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16198 least one invariant or re-used load.
16199 * passes.c (init_optimization_passes): Move pass_phiprop before
16202 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16204 * config/aarch64/aarch64-simd.md
16205 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16207 2013-05-23 Richard Biener <rguenther@suse.de>
16209 PR middle-end/57381
16210 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16211 OEP_CONSTANT_ADDRESS_OF retained.
16213 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16215 PR middle-end/57344
16216 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16217 don't lower unit. Handle unit not being always BITS_PER_WORD.
16219 2013-05-23 Richard Biener <rguenther@suse.de>
16221 PR rtl-optimization/57341
16222 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16223 instead of true_dependence.
16225 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16227 * bb-reorder.c (branch_threshold): Make const.
16228 (exec_threshold): Ditto.
16230 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16231 Pat Haugen <pthaugen@us.ibm.com>
16232 Peter Bergner <bergner@vnet.ibm.com>
16234 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16235 documentation for the power8 crypto builtins.
16237 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16239 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16240 macros for defining power8 builtin functions.
16241 (BU_P8V_AV_2): Likewise.
16242 (BU_P8V_AV_P): Likewise.
16243 (BU_P8V_VSX_1): Likewise.
16244 (BU_P8V_OVERLOAD_1): Likewise.
16245 (BU_P8V_OVERLOAD_2): Likewise.
16246 (BU_CRYPTO_1): Likewise.
16247 (BU_CRYPTO_2): Likewise.
16248 (BU_CRYPTO_3): Likewise.
16249 (BU_CRYPTO_OVERLOAD_1): Likewise.
16250 (BU_CRYPTO_OVERLOAD_2): Likewise.
16251 (XSCVSPDP): Fix typo, point to the correct instruction.
16252 (VCIPHER): Add power8 crypto builtins.
16253 (VCIPHERLAST): Likewise.
16254 (VNCIPHER): Likewise.
16255 (VNCIPHERLAST): Likewise.
16256 (VPMSUMB): Likewise.
16257 (VPMSUMH): Likewise.
16258 (VPMSUMW): Likewise.
16259 (VPERMXOR_V2DI): Likewise.
16260 (VPERMXOR_V4SI: Likewise.
16261 (VPERMXOR_V8HI: Likewise.
16262 (VPERMXOR_V16QI: Likewise.
16263 (VSHASIGMAW): Likewise.
16264 (VSHASIGMAD): Likewise.
16265 (VPMSUM): Likewise.
16266 (VPERMXOR): Likewise.
16267 (VSHASIGMA): Likewise.
16269 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16270 __CRYPTO__ if the crypto instructions are available.
16271 (altivec_overloaded_builtins): Add support for overloaded power8
16274 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16275 support for power8 crypto builtins.
16276 (builtin_function_type): Likewise.
16277 (altivec_init_builtins): Add support for builtins that take vector
16278 long long (V2DI) arguments.
16280 * config/rs6000/crypto.md: New file, define power8 crypto
16283 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16284 Pat Haugen <pthaugen@us.ibm.com>
16285 Peter Bergner <bergner@vnet.ibm.com>
16287 * doc/invoke.texi (Option Summary): Add power8 options.
16288 (RS/6000 and PowerPC Options): Likewise.
16290 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16291 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16292 wm, wn, wr documentation.
16294 * config/rs6000/constraints.md (wm): New constraint for VSX
16295 registers if direct move instructions are enabled.
16296 (wn): New constraint for no registers.
16297 (wq): New constraint for quad word even GPR registers.
16298 (wr): New constraint if 64-bit instructions are enabled.
16299 (wv): New constraint if power8 vector instructions are enabled.
16300 (wQ): New constraint for quad word memory locations.
16302 * config/rs6000/predicates.md (const_0_to_15_operand): New
16303 constraint for 0..15 for crypto instructions.
16304 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16305 as GPR and floating point registers.
16306 (int_reg_operand): New predicate to match only GPR registers.
16307 (base_reg_operand): New predicate to match base registers.
16308 (quad_int_reg_operand): New predicate to match even GPR registers
16309 for quad memory operations.
16310 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16311 operations in both GPR and VSX registers.
16312 (quad_memory_operand): New predicate for quad memory operations.
16313 (reg_or_indexed_operand): New predicate for direct move support.
16315 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16316 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16317 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16318 (POWERPC_MASKS): Add power8 options.
16319 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16322 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16323 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16325 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16326 (-mpower8-fusion): New power8 options.
16327 (-mpower8-fusion-sign): Likewise.
16328 (-mpower8-vector): Likewise.
16329 (-mcrypto): Likewise.
16330 (-mdirect-move): Likewise.
16331 (-mquad-memory): Likewise.
16333 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16334 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16336 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16337 (rs6000_debug_vector_unit): Add p8_vector.
16338 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16339 definitions. Also print fusion state.
16340 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16341 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16342 (rs6000_option_override_internal): Add support for power8.
16343 (rs6000_common_init_builtins): Add debugging for skipped builtins
16344 if -mdebug=builtin.
16345 (rs6000_adjust_cost): Add power8 support.
16346 (rs6000_issue_rate): Likewise.
16347 (insn_must_be_first_in_group): Likewise.
16348 (insn_must_be_last_in_group): Likewise.
16349 (force_new_group): Likewise.
16350 (rs6000_register_move_cost): Likewise.
16351 (rs6000_opt_masks): Likewise.
16353 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16354 power8 capable assembler, default to power7 options.
16355 (TARGET_DIRECT_MOVE): Likewise.
16356 (TARGET_CRYPTO): Likewise.
16357 (TARGET_P8_VECTOR): Likewise.
16358 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16359 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16360 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16361 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16362 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16363 (TARGET_XSCVDPSPN): Likewise.
16364 (TARGET_XSCVSPDPN): Likewsie.
16365 (TARGET_SYNC_HI_QI): Likewise.
16366 (TARGET_SYNC_TI): Likewise.
16367 (MASK_CRYPTO): Likewise.
16368 (MASK_DIRECT_MOVE): Likewise.
16369 (MASK_P8_FUSION): Likewise.
16370 (MASK_P8_VECTOR): Likewise.
16371 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16372 temporary used by some of the direct move instructions to get two FP
16373 temporary registers does not force creation of a stack frame.
16374 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16375 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16376 that any VSX registers are tieable, even if they are also an
16377 Altivec vector mode.
16378 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16379 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16380 (RS6000_BTM_CRYPTO): Likewise.
16381 (RS6000_BTM_COMMON): Likewise.
16383 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16384 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16385 (enum rs6000_vector): Add power8 vector support.
16387 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16391 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16392 (any_sibcall_could_use_r3): this and handle indirect calls.
16393 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16395 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16397 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16399 2013-05-22 Richard Biener <rguenther@suse.de>
16401 PR middle-end/57349
16402 * profile.c (branch_prob): Do not split blocks that are
16403 abnormally receiving from ECF_RETURNS_TWICE functions.
16405 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16407 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16408 address mode. Move pointer mode initialization to the same place.
16410 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16412 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16413 while it has any effect.
16415 2013-05-21 Easwaran Raman <eraman@google.com>
16417 PR tree-optimization/57322
16418 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16419 UID of the statement added to the BB to be 1.
16421 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16423 PR tree-optimization/57331
16424 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16425 of conversion from pointer type to integral type with integer.
16427 2013-05-21 Martin Jambor <mjambor@suse.cz>
16430 * ipa-prop.c (ipa_read_node_info): Process param_used and
16431 controlled_uses in the same order as when writing.
16433 2013-05-21 Magnus Granberg <baldrick@free.fr>
16436 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16438 2013-05-21 Richard Biener <rguenther@suse.de>
16440 PR tree-optimization/57318
16441 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16442 estimate stmts with side-effects as likely eliminated.
16444 2013-05-21 Richard Biener <rguenther@suse.de>
16446 PR tree-optimization/57330
16447 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16448 preserve the call stmts fntype.
16450 2013-05-21 Richard Biener <rguenther@suse.de>
16452 PR tree-optimization/57303
16453 * tree-ssa-sink.c (statement_sink_location): Improve killing
16454 stmt detection and properly handle self-assignments.
16456 2013-05-21 Christian Bruel <christian.bruel@st.com>
16458 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16459 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16460 registers. Set register size out of the PARALLEL loop.
16462 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16465 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16466 (*fmasf4, *fmasf4_media): New insns.
16468 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16470 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16471 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16472 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16473 (mips_idiv_insns): Update the comments to say that the returned
16474 instruction counts are in units of BASE_INSN_LENGTH.
16475 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16476 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16477 using 2 rather than 4 as the length of indirect MIPS16 and
16478 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16479 length of a NOP. Don't divide MIPS16 lengths by 2.
16480 (mips16_split_long_branches): Assume a branch is long if the
16481 length is greater than 4 rather than 8.
16482 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16483 rather than multiplying them by 2. Multiply instruction counts
16484 by BASE_INSN_LENGTH rather than 4.
16485 (*jump_mips16, tls_get_tp_mips16_<mode>)
16486 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16488 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16490 * config/mips/mips.md (extended_mips16): Remove branch case.
16491 (length): Remove duplicated extended_mips16 test.
16493 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16495 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16497 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16499 * recog.h (Recog_data): Rename to...
16500 (recog_data_d): ...this.
16501 (recog_data): Update accordingly.
16502 * recog.c (recog_data): Likewise.
16503 * reload.c (save_recog_data): Likewise.
16504 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16505 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16507 2013-05-17 Julian Brown <julian@codesourcery.com>
16509 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16510 found in a REG_EQUAL note, invalidate it.
16512 2013-05-17 Easwaran Raman <eraman@google.com>
16514 * tree-ssa-reassoc.c (find_insert_point): New function.
16515 (insert_stmt_after): Likewise.
16516 (get_def_stmt): Likewise.
16517 (ensure_ops_are_available): Likewise.
16518 (not_dominated_by): Likewise.
16519 (rewrite_expr_tree): Do not move statements beyond what is
16520 necessary. Remove call to swap_ops_for_binary_stmt...
16521 (reassociate_bb): ... and move it here.
16522 (build_and_add_sum): Assign UIDs for new statements.
16523 (linearize_expr): Likewise.
16524 (do_reassoc): Renumber gimple statement UIDs.
16526 2013-05-17 Jan Hubicka <jh@suse.cz>
16528 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16530 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16531 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16533 (output_weakrefs): Update.
16535 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16536 Martin Jambor <mjambor@suse.cz>
16538 PR middle-end/57276
16539 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16540 value that corresponds to the given aggval is found in values vector.
16542 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16544 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16545 sse, sse2, sse3, ssse3 and sse4a flags to options.
16547 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16549 * gengtype-state.c: (s_expr_writer): New class, to handle
16550 prettifying of output layout of s-expressions.
16551 (state_writer): New class, to write out gtype.state.
16552 (state_written_type_count): Move this variable into member data of
16554 (s_expr_writer::s_expr_writer): New code: constructor for new class
16555 (state_writer::state_writer(): ditto
16556 (s_expr_writer::write_new_line): New function
16557 (s_expr_writer::write_any_indent): ditto
16558 (s_expr_writer::begin_s_expr): ditto
16559 (s_expr_writer::end_s_expr): ditto
16560 (write_state_fileloc): convert to method of state_writer...
16561 (state_writer:: write_state_fileloc): ...and use methods of
16562 s_expr_writer to write indentation into the gtype.state output file
16563 to visually represent the hierarchical structure of the list
16565 (write_state_fields): ditto, renaming to...
16566 (state_writer::write_state_fields)
16567 (write_state_a_string): ditto, renaming to...
16568 (state_writer::write_state_a_string)
16569 (write_state_string_option): ditto, renaming to...
16570 (state_writer::write_state_string_option)
16571 (write_state_type_option): ditto, renaming to...
16572 (state_writer::write_state_type_option)
16573 (write_state_nested_option): ditto, renaming to...
16574 (state_writer::write_state_nested_option)
16575 (write_state_option): ditto, renaming to...
16576 (state_writer::write_state_option)
16577 (write_state_options): ditto, renaming to...
16578 (state_writer::write_state_options)
16579 (write_state_lang_bitmap): ditto, renaming to...
16580 (state_writer::write_state_lang_bitmap)
16581 (write_state_version): ditto, renaming to...
16582 (state_writer::write_state_version)
16583 (write_state_scalar_type): ditto, renaming to...
16584 (state_writer::write_state_scalar_type)
16585 (write_state_string_type): ditto, renaming to...
16586 (state_writer::write_state_string_type)
16587 (write_state_undefined_type): ditto, renaming to...
16588 (state_writer::write_state_undefined_type)
16589 (write_state_struct_union_type): ditto, renaming to...
16590 (state_writer::write_state_struct_union_type)
16591 (write_state_struct_type): ditto, renaming to...
16592 (state_writer::write_state_struct_type)
16593 (write_state_user_struct_type): ditto, renaming to...
16594 (state_writer::write_state_user_struct_type)
16595 (write_state_lang_struct_type): ditto, renaming to...
16596 (state_writer::write_state_lang_struct_type)
16597 (write_state_param_struct_type): ditto, renaming to...
16598 (state_writer::write_state_param_struct_type)
16599 (write_state_pointer_type): ditto, renaming to...
16600 (state_writer::write_state_pointer_type)
16601 (write_state_array_type): ditto, renaming to...
16602 (state_writer::write_state_array_type)
16603 (write_state_gc_used): ditto, renaming to...
16604 (state_writer::write_state_gc_used)
16605 (write_state_common_type_content): ditto, renaming to...
16606 (state_writer::write_state_common_type_content)
16607 (write_state_type): ditto, renaming to...
16608 (state_writer::write_state_type)
16609 (write_state_pair_list): ditto, renaming to...
16610 (state_writer::write_state_pair_list)
16611 (write_state_pair): ditto, renaming to...
16612 (state_writer::write_state_pair)
16613 (write_state_typedefs): ditto, renaming to...
16614 (state_writer::write_state_typedefs)
16615 (write_state_structures): ditto, renaming to...
16616 (state_writer::write_state_structures)
16617 (write_state_param_structs): ditto, renaming to...
16618 (state_writer::write_state_param_structs)
16619 (write_state_variables): ditto, renaming to...
16620 (state_writer::write_state_variables)
16621 (write_state_srcdir): ditto, renaming to...
16622 (state_writer::write_state_srcdir)
16623 (write_state_files_list): ditto, renaming to...
16624 (state_writer::write_state_files_list)
16625 (write_state_languages): ditto, renaming to...
16626 (state_writer::write_state_languages)
16627 (write_state): create a state_writer instance and use it when
16628 writing out the state file
16630 2013-05-17 Mike Stump <mikestump@comcast.net>
16632 PR rtl-optimization/57304
16633 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16634 accessing DF_REF_REAL_LOC.
16636 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16638 PR rtl-optimization/57281
16639 PR rtl-optimization/57300
16640 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16641 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16642 what the other splitter did if the registers are dead.
16644 2013-05-17 Richard Biener <rguenther@suse.de>
16646 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16649 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16651 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16654 2013-05-17 Marek Polacek <polacek@redhat.com>
16656 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16657 length when doing non-zero store of storing '\0' to '\0'.
16659 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16661 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16662 vect_external_def oprnd1 with loop_vinfo, try to emit
16663 optional cast, negation and and stmts on the loop preheader
16664 edge instead of into the pattern def seq.
16666 PR tree-optimization/57051
16667 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16668 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16670 2013-05-16 Nick Clifton <nickc@redhat.com>
16672 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16673 (rl78_is_naked_func): New function.
16674 (rl78_expand_prologue): Skip prologue generation for naked functions.
16675 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16676 * doc/extend.texi (naked): Add RL78 to the list of processors
16677 that supports this attribute.
16679 2013-05-16 Jeff Law <law@redhat.com>
16681 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16683 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16685 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16686 cache parameters using detect_caches_amd also for CYRIX,
16687 NSC and TM2 signatures.
16689 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16690 Dzianis Kahanovich <mahatma@eu.by>
16694 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16695 VIA/Centaur processors and determine their cache parameters
16696 using detect_caches_amd.
16698 2013-05-16 Teresa Johnson <tejohnson@google.com>
16700 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16701 (rtl_verify_edges): New function.
16702 (rtl_verify_bb_insns): Ditto.
16703 (rtl_verify_bb_pointers): Ditto.
16704 (rtl_verify_bb_insn_chain): Ditto.
16705 (rtl_verify_fallthru): Ditto.
16706 (rtl_verify_bb_layout): Ditto.
16707 (rtl_verify_flow_info_1): Outline checks into new functions.
16708 (rtl_verify_flow_info): Ditto.
16710 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16712 * cfghooks.c (copy_bbs): Add update_dominance argument.
16713 * cfghooks.h (copy_bbs): Update prototype.
16714 * tree-cfg.c (gimple_duplicate_sese_region):
16715 Add update_dominance argument.
16716 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16717 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16718 gimple_duplicate_sese_region call.
16719 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16720 Update copy_bbs call.
16721 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16722 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16724 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16726 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16727 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16728 vect_recog_rotate_pattern.
16729 (vect_recog_rotate_pattern): New function.
16731 2013-05-16 Jason Merrill <jason@redhat.com>
16733 * Makefile.in (LLINKER): New variable.
16734 (mostlyclean): Remove link mutex.
16735 * configure.ac: Handle --enable-link-mutex.
16736 * lock-and-run.sh: New script.
16738 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16741 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16744 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16746 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16748 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16750 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16751 * config/arm/arm.c (next_consecutive_mem): New function.
16752 (gen_movmem_ldrd_strd): Likewise.
16753 * config/arm/arm.md (movmemqi): Update condition and code.
16754 (unaligned_loaddi, unaligned_storedi): New patterns.
16756 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16758 * config.gcc: Obsolete *-*-solaris2.9*.
16759 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16761 2013-05-16 Richard Biener <rguenther@suse.de>
16763 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16764 earlier, after GRAPHITE transforms and IV canonicalization.
16766 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16768 * omp-low.c (extract_omp_for_data): For collapsed loops,
16769 if at least one of the loops is known at compile time to
16770 iterate zero times, set count to 0.
16771 (expand_omp_regimplify_p): New function.
16772 (expand_omp_for_generic): For collapsed loops, if at least
16773 one of the loops isn't known to iterate at least once,
16774 add runtime check with setting count to 0.
16775 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16776 For unsigned types if it isn't known at compile time that
16777 the loop will iterate at least once, add runtime check to bypass
16778 the whole loop if initial condition isn't true.
16780 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16782 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16784 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16786 PR middle-end/57286
16787 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16788 transformations to avoid an infinite loop.
16790 2013-05-16 Marek Polacek <polacek@redhat.com>
16792 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16794 2013-05-15 Leif Ekblad <leif@rdos.net>
16796 * config/i386/i386.c (ix86_decompose_address): Use
16797 DEFAULT_TLS_SEG_REG to access TLS segment register.
16798 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16799 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16800 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16802 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16805 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16806 sibling calls to functions that would normally be lazily bound,
16807 unless $gp is call-clobbered.
16809 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16811 * config/i386/i386.c (ix86_option_override_internal): Update
16812 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16813 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16814 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16815 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16817 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16819 2013-05-15 Andreas Schwab <schwab@suse.de>
16821 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16822 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16825 2013-05-15 Teresa Johnson <tejohnson@google.com>
16827 * loop-unroll.c (report_unroll_peel): Check decision before
16828 emitting unroll/peel message.
16830 2013-05-15 Teresa Johnson <tejohnson@google.com>
16832 * function.h (has_bb_partition): New rtl_data flag.
16833 (bb_reorder_complete): Ditto.
16834 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16835 instead of flag_reorder_blocks_and_partition.
16836 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16837 with some enhancements.
16838 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16839 * bb-reorder.c (connect_traces): Check for has_bb_partition
16840 instead of flag_reorder_blocks_and_partition.
16841 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16842 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16843 verify_hot_cold_block_grouping.
16844 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16846 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16849 * config/arm/predicates.md (call_insn_operand): New predicate.
16850 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16851 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16852 if insn is not a tail call.
16853 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16855 * config/arm/arm.h (enum reg_class): New caller save register class.
16856 (REG_CLASS_NAMES): Likewise.
16857 (REG_CLASS_CONTENTS): Likewise.
16858 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16861 2013-05-15 Richard Biener <rguenther@suse.de>
16863 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16864 of MSG_OPTIMIZED_LOCATIONS.
16865 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16866 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16868 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16869 of MSG_OPTIMIZED_LOCATIONS.
16870 (execute_vect_slp): Likewise.
16871 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16872 (vect_create_cond_for_alias_checks): Likewise.
16873 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16874 (vect_recog_widen_mult_pattern): Likewise.
16875 (vect_recog_widen_sum_pattern): Likewise.
16876 (vect_recog_over_widening_pattern): Likewise.
16877 (vect_recog_widen_shift_pattern): Likewise.
16878 (vect_recog_vector_vector_shift_pattern): Likewise.
16879 (vect_recog_divmod_pattern): Likewise.
16880 (vect_recog_mixed_size_cond_pattern): Likewise.
16881 (vect_recog_bool_pattern): Likewise.
16882 (vect_pattern_recog_1): Likewise.
16884 2013-05-15 Martin Jambor <mjambor@suse.cz>
16886 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16887 non-functions to builtin_unreachable.
16888 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16889 correct when new direct edges were discovered.
16891 2013-05-15 Martin Jambor <mjambor@suse.cz>
16893 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16894 header, print symbol order instead of node uid, print more information
16895 about indirect edge targets.
16896 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16898 (ipa_make_edge_direct_to_target): Likewise.
16899 (remove_described_reference): Likewise.
16900 (propagate_controlled_uses): Likewise.
16901 (ipa_print_node_params): Also print symbol order.
16902 (ipcp_transform_function): Print symbol order instead of node uids.
16903 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16904 (cgraph_get_create_real_symbol_node): Likewise.
16905 * ipa-cp.c (print_lattice): Likewise.
16906 (print_all_lattices): Likewise.
16907 (determine_versionability): Likewise.
16908 (initialize_node_lattices): Likewise.
16909 (estimate_local_effects): Likewise.
16910 (update_profiling_info): Likewise.
16911 (create_specialized_node): Likewise.
16912 (perhaps_add_new_callers): Likewise.
16913 (decide_about_value): Likewise.
16914 (decide_whether_version_node): Likewise.
16915 (identify_dead_nodes): Likewise.
16916 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16917 (dump_inline_summary): Likewise.
16918 (estimate_node_size_and_time): Likewise.
16919 (inline_analyze_function): Likewise.
16920 * ipa-inline.c (report_inline_failed_reason): Likewise.
16921 (want_early_inline_function_p): Likewise.
16922 (edge_badness): Likewise.
16923 (update_edge_key): Likewise.
16924 (inline_small_functions): Likewise. Add dumping of order to two other
16926 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16927 instead of node uids.
16928 (propagate_pure_const): Likewise.
16929 (propagate_pure_const): Likewise.
16930 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16931 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16932 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16934 * tree-pretty-print.c (dump_function_header): Likewise.
16935 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16936 Print symbol order instead of node uids.
16938 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16940 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16941 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16943 2013-05-15 Richard Biener <rguenther@suse.de>
16945 PR tree-optimization/57275
16946 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16947 return value for fail to do runtime alias checks for gather loads.
16949 2013-05-15 Jan Hubicka <jh@suse.cz>
16953 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16954 weakrefs are not external.
16955 (lto_symtab_merge_decls): Fix thinko when dealing with
16956 non-lto_symtab decls.
16957 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16958 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16959 * varpool.c (dump_varpool_node): Dump more flags.
16961 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16963 * config/i386/i386.c (processor_alias_table): Add instruction
16964 FSGSBASE for AMD bdver3 architecture.
16966 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16968 * tree.c (warn_deprecated_use): Print file:line using locus color.
16969 * diagnostic.c (diagnostic_report_current_module): Print file:line
16970 and file:line:column using locus color.
16972 2013-05-14 Mike Stump <mikestump@comcast.net>
16974 * gdbinit.in: Add __null.
16976 2013-05-14 Mike Stump <mikestump@comcast.net>
16978 * recog.h: Rename struct recog_data to Recog_data.
16979 * recog.c: Likewise.
16980 * reload.c (can_reload_into): Likewise.
16981 * config/picochip/picochip.c: Likewise.
16983 2013-05-14 Mike Stump <mikestump@comcast.net>
16985 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16987 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16989 * resource.h (struct resources): Remove unch_memory member.
16990 (CLEAR_RESOURCE): Don't clear unch_memory.
16991 * resource.c (mark_referenced_resources): Don't set it.
16992 (mark_set_resources): Likewise.
16993 (mark_target_live_regs): Don't clear it.
16994 (init_resource_info): Likewise.
16995 * reorg.c (resource_conflicts_p): Don't compare it.
16996 (redundant_insn): Don't set it.
16998 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17000 * emit-rtl.c (next_label): Remove unused function.
17001 (skip_consecutive_labels, link_cc0_insns): Move to ...
17002 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17003 only place where these functions are used, and make them static.
17005 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17007 * fold-const.c (fold_negate_expr): Handle vectors.
17008 (fold_truth_not_expr): Make it static.
17009 (fold_invert_truthvalue): New static function.
17010 (invert_truthvalue_loc): Handle vectors. Do not call
17011 fold_truth_not_expr directly.
17012 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17013 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17014 (fold_comparison): Handle vector constants.
17015 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17016 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17017 * tree.h (fold_truth_not_expr): Remove declaration.
17019 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17021 * config/aarch64/aarch64-simd.md
17022 (aarch64_vcond_internal<mode>): Rename to...
17023 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17024 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17025 float modes. Clarify all iterator modes.
17026 (vcond<mode><mode>): Use new name for vcond expanders.
17027 (vcond<v_cmp_result><mode>): Likewise.
17028 (vcondu<mode><mode>: Likewise.
17029 * config/aarch64/iterators.md (VDQF_COND): New.
17031 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17034 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17035 variable for the shift amount. Check that we shift by non-negative
17038 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17041 * config/arm/arm.h (EPILOGUE_USES): Only return true
17042 for LR_REGNUM after epilogue_completed.
17044 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17046 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17047 is error_mark_node.
17049 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17052 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17053 and Solaris 11+/x86 with gld.
17054 * configure: Regenerate.
17056 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17058 * expmed.c (expand_shift_1): Canonicalize rotates by
17059 constant bitsize / 2 to bitsize - 1.
17060 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17061 case ROTATERT>: Likewise.
17064 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17066 * config/i386/i386.md (rotateinv): New code attr.
17067 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17068 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17069 roll $31, %eax, etc.
17071 2013-05-14 Richard Biener <rguenther@suse.de>
17073 PR middle-end/57235
17074 * tree-eh.c (sink_clobbers): Give up for successors with
17075 multiple predecessors and no virtual uses.
17077 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17079 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17080 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17082 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17084 PR middle-end/57251
17085 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17086 the case when both op0 and op1 have VOIDmode.
17088 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17090 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17091 in multiply-accumulate mode.
17093 2013-05-13 Guozhi Wei <carrot@google.com>
17095 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17097 2013-05-13 Kai Tietz <ktietz@redhat.com>
17100 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17101 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17102 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17103 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17104 (ix86_expand_prologue): Likewise.
17105 (ix86_expand_split_stack_prologue): Likewise.
17106 (legitimate_pic_address_disp_p): Likewise.
17107 (legitimize_pic_address): Likewise.
17108 (legitimize_tls_address): Likewise.
17109 (legitimize_pe_coff_symbol): Likewise.
17110 (output_pic_addr_const): Likewise.
17111 (construct_plt_address): Likewise.
17112 (ix86_expand_call): Likewise.
17113 (x86_output_mi_thunk): Likewise.
17114 (x86_function_profiler): Likewise.
17116 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17118 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17119 similar switch cases.
17120 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17121 (aarch64_simd_mov_to_<mode>low): Delete.
17122 (aarch64_simd_mov_to_<mode>high): Delete.
17123 (move_lo_quad_<mode>): Add w<-r alternative.
17124 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17125 (aarch64_simd_mov_from_*): Update type attribute.
17126 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17129 2013-05-13 Jan Hubicka <jh@suse.cz>
17131 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17132 * config/i386/i386.c (ix86_compute_frame_layout,
17133 ix86_expand_epilogue, emit_i387_cw_initialization,
17134 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17135 ix86_local_alignment): Fix use of size/speed predicates.
17137 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17139 PR tree-optimization/45216
17140 PR tree-optimization/57157
17141 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17142 the (-Y) & (B - 1) variant if OP is |.
17143 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17144 return shifted. Use (-op1) & (prec - 1) as other_amount
17145 instead of prec - op1.
17147 2013-05-13 Martin Jambor <mjambor@suse.cz>
17149 PR middle-end/42371
17150 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17151 (ipa_constant_data): New type.
17152 (ipa_jump_func): Use ipa_constant_data to hold information about
17153 constant jump functions.
17154 (ipa_get_jf_constant): Adjust to jump function type changes.
17155 (ipa_get_jf_constant_rdesc): New function.
17156 (ipa_param_descriptor): New field controlled_uses.
17157 (ipa_get_controlled_uses): New function.
17158 (ipa_set_controlled_uses): Likewise.
17159 * ipa-ref.h (ipa_find_reference): Declare.
17160 * ipa-prop.c (ipa_cst_ref_desc): New type.
17161 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17163 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17164 New parameter cs. Adjust all callers.
17165 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17166 (remove_described_reference): New function.
17167 (jfunc_rdesc_usable): Likewise.
17168 (try_make_edge_direct_simple_call): Decrement controlled use count,
17169 attempt to remove reference if it hits zero.
17170 (combine_controlled_uses_counters): New function.
17171 (propagate_controlled_uses): Likewise.
17172 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17173 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17174 (ipa_print_node_params): Print described use counter.
17175 (ipa_write_jump_function): Adjust to jump function type changes.
17176 (ipa_read_jump_function): New parameter CS, pass it to
17177 ipa_set_jf_constant. Adjust caller.
17178 (ipa_write_node_info): Stream controlled use count
17179 (ipa_read_node_info): Likewise.
17180 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17182 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17183 count. Remove cloning-added reference if it reaches zero.
17184 * ipa-ref.c (ipa_find_reference): New function.
17186 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17188 * config/i386/i386.c (processor_target_table): Modified default
17189 alignment values for AMD BD and BT architectures.
17191 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17193 * tree-vect-generic.c (uniform_vector_p): Move ...
17194 * tree.c (uniform_vector_p): ... here.
17195 * tree.h (uniform_vector_p): Declare it.
17196 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17199 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17201 PR tree-optimization/57230
17202 * tree-ssa-strlen.c (handle_char_store): Record length for
17203 array store from STRING_CST.
17205 PR tree-optimization/57230
17206 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17209 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17211 * config/epiphany/epiphany.c (epiphany_init): Check size of
17212 NUM_MODES_FOR_MODE_SWITCHING.
17213 (epiphany_expand_prologue):
17214 Remove CONFIG_REGNUM initial value handling code.
17215 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17216 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17217 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17218 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17219 Don't return 1 for FP_MODE_NONE.
17220 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17221 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17222 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17223 * config/epiphany/epiphany.md (save_config): New pattern.
17225 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17227 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17229 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17231 * config/i386/i386.md (memory): Handle sseishft1.
17232 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17233 (*vec_extractv2di_1): Ditto.
17235 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17237 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17240 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17242 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17243 Add mthumb/march=armv7-a multilib.
17244 Add mthumb/march=armv7-r multilib.
17245 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17247 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17249 * config/v850/t-rtems: Add more multilibs.
17251 2013-05-10 Richard Biener <rguenther@suse.de>
17253 PR tree-optimization/57214
17254 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17255 not propagate from SSA names that occur in abnormal PHI nodes.
17257 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17259 * stor-layout.c (element_precision): New function.
17260 * machmode.h (element_precision): Declare it.
17261 * tree.c (build_minus_one_cst): New function.
17262 (element_precision): Likewise.
17263 * tree.h (build_minus_one_cst): Declare new function.
17264 (element_precision): Likewise.
17265 * fold-const.c (operand_equal_p): Use element_precision.
17266 (fold_binary_loc): Handle vector types.
17267 * convert.c (convert_to_integer): Use element_precision.
17268 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17271 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17273 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17274 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17275 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17276 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17277 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17278 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17279 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17280 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17281 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17282 (Uuw8): New constraints.
17283 (Usb4): Move into alphabetical order.
17284 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17285 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17286 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17287 previously unnamed patterns.
17288 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17289 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17290 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17291 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17292 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17293 of specific lengths.
17295 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17297 * config/i386/i386.md (rotateinv): New code attr.
17298 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17299 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17300 roll $31, %eax, etc.
17302 PR tree-optimization/45216
17303 PR tree-optimization/57157
17304 * tree-ssa-forwprop.c (simplify_rotate): New function.
17305 (ssa_forward_propagate_and_combine): Call it.
17307 2013-05-10 Richard Biener <rguenther@suse.de>
17309 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17310 disable peeling when we version for aliasing.
17311 (vector_alignment_reachable_p): Honor explicit user alignment.
17312 (vect_supportable_dr_alignment): Likewise.
17313 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17314 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17315 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17316 then peeling to arrange for the cost-model check to come first.
17318 2013-05-10 Alan Modra <amodra@gmail.com>
17320 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17321 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17322 * configure: Regenerate.
17324 2013-05-10 Alan Modra <amodra@gmail.com>
17327 * varasm.c (default_elf_select_section): Move !DECL_P check..
17328 (get_named_section): ..to here before calling get_section_name.
17330 (default_section_type_flags): Add DECL_P check.
17331 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17332 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17334 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17336 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17337 When using gen_stack_adjust_str with a register offset, add a
17338 REG_FRAME_RELATED_EXPR note.
17340 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17342 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17343 (*vec_extractv4si_zext_mem): Ditto.
17344 (*vec_extractv2di): Add 0->x and x->x alternatives.
17345 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17346 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17348 2013-05-09 Jason Merrill <jason@redhat.com>
17350 N3639 C++1y VLA support
17351 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17354 * tree.c (build_constructor_va): New.
17355 * tree.h: Declare it.
17357 2013-05-09 Martin Jambor <mjambor@suse.cz>
17360 * gimple-fold.c (canonicalize_constructor_val): Call
17361 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17363 2013-05-09 Jan Hubicka <jh@suse.cz>
17364 Richard Biener <rguenther@suse.de>
17367 * symtab.c (symtab_make_decl_local): Do not add private names.
17369 2013-05-09 Jan Hubicka <jh@suse.cz>
17372 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17373 (unlink_from_assembler_name_hash): Likewise.
17374 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17375 symtab_unregister_node, symtab_initialize_asm_name_hash,
17376 change_decl_assembler_name): Update.
17378 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17380 * config/aarch64/aarch64.md: New movtf split.
17381 (*movtf_aarch64): Update.
17382 (aarch64_movdi_tilow): Handle TF modes and rename to
17383 aarch64_movdi_<mode>low.
17384 (aarch64_movdi_tihigh): Handle TF modes and rename to
17385 aarch64_movdi_<mode>high
17386 (aarch64_movtihigh_di): Handle TF modes and rename to
17387 aarch64_mov<mode>high_di
17388 (aarch64_movtilow_di): Handle TF modes and rename to
17389 aarch64_mov<mode>low_di
17390 (aarch64_movtilow_tilow): Remove spurious whitespace.
17391 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17393 (aarch64_print_operand): Update.
17395 2013-05-09 Alan Modra <amodra@gmail.com>
17397 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17399 * configure: Regenerate.
17401 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17403 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17404 splitter preparation statements.
17405 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17406 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17409 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17411 * gimple-ssa-strength-reduction.c (count_candidates): Change
17412 return value to int.
17413 (analyze_candidates_and_replace): Change type of length to int.
17415 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17417 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17418 (*vec_extract<mode>): Use VI12_128 mode iterator.
17419 (*vec_extract<mode>_mem): Ditto.
17420 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17423 2013-05-08 Diego Novillo <dnovillo@google.com>
17428 2012-08-17 Diego Novillo <dnovillo@google.com>
17431 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17432 * config.in: Regenerate.
17433 * configure: Regenerate.
17434 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17436 2013-05-08 Jan Hubicka <jh@suse.cz>
17439 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17440 * cgraph.h (symtab_node_base): Add unique_name.
17441 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17442 input_overwrite_node, input_varpool_node): Stream unique_name.
17443 * cgraphclones.c (cgraph_create_virtual_clone,
17444 cgraph_function_versioning): Set unique_name.
17445 * ipa.c (function_and_variable_visibility): Set unique_name.
17447 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17449 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17450 (alloc_cand_and_find_basis): Restrict conditional candidate
17451 processing to CAND_MULTs.
17453 2013-05-08 Jan Hubicka <jh@suse.cz>
17456 lto-symtab.c (lto_symtab_symbol_p): New function.
17457 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17458 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17459 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17460 Skip static symbols.
17462 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17464 PR tree-optimization/57200
17465 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17466 Only call inform if the preceding warning_at returns true.
17468 2013-05-07 Han Shen <shenhan@google.com>
17470 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17471 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17472 * common.opt (fstack-protector-strong): New option.
17473 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17474 * doc/invoke.texi (Optimization Options): Document
17475 "-fstack-protector-strong".
17476 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17478 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17480 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17482 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17484 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17485 it is not lhs of the insn.
17486 (setup_reg_equiv): Remove insn having equiv memory which it is not
17488 * lra-constraints.c (process_address): Try to improve generation
17489 code for address base + disp.
17490 (lra_constraints): Make correct the code for checking insn setting
17491 up backward equivalence. Remove insn only if it is in the init
17493 * lra-eliminations.c (update_reg_eliminate): Change return value.
17494 (lra_eliminate): Use the result.
17496 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17498 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17499 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17500 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17501 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17502 register target operands.
17503 (*vec_extractv8hi_sse2): New pattern.
17504 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17505 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17506 (*vec_extract<mode>_mem): New insn and split pattern.
17508 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17510 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17511 (TARGET_ASAN_SHADOW_OFFSET): Define.
17512 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17513 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17515 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17517 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17518 (incr_vec_index): Return -1 if increment not found.
17519 (create_add_on_incoming_edge): Assert if increment not found.
17520 (record_increment): Limit number of increments recorded.
17521 (all_phi_incrs_profitable): Return false if an increment not found.
17522 (replace_profitable_candidates): Don't process increments that were
17524 (analyze_candidates_and_replace): Limit size of incr_vec.
17526 2013-05-07 Richard Biener <rguenther@suse.de>
17528 * calls.c (special_function_p): setjmp-like functions are leaf.
17529 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17530 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17532 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17534 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17535 (aarch64_simd_mov<mode>): New expander.
17536 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17537 (aarch64_simd_mov_to_<mode>high): Likewise.
17538 (aarch64_simd_mov_from_<mode>low): Likewise.
17539 (aarch64_simd_mov_from_<mode>high): Likewise.
17540 (aarch64_dup_lane<mode>): Update.
17541 (aarch64_dup_lanedi): New instruction pattern.
17542 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17543 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17545 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17547 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17548 (replace_mult_candidate): Remove unnecessary argument; remove
17549 unnecessary parameter from call to introduce_cast_before_cand.
17550 (replace_unconditional_candidate): Remove unnecessary parameter
17551 from call to replace_mult_candidate.
17552 (replace_conditional_candidate): Likewise.
17553 (insert_initializers): Use make_temp_ssa_name.
17554 (introduce_cast_before_cand): Remove unnecessary argument; use
17555 make_temp_ssa_name.
17556 (replace_one_candidate): Remove unnecessary argument; remove
17557 unnecessary parameter from calls to introduce_cast_before_cand.
17558 (replace_profitable_candidates): Remove unnecessary parameters
17559 from calls to replace_one_candidate.
17561 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17563 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17564 phi def as possibly hiding a basis for a CAND_ADD whose operands
17565 have been commuted in the analysis.
17566 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17568 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17570 * config/aarch64/aarch64.md
17571 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17572 shift value between 0-4.
17574 2013-05-07 Richard Biener <rguenther@suse.de>
17576 * double-int.h (rshift): New overload.
17577 * double-int.c (rshift): New function.
17578 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17579 (create_reference_ops_from_ref): Remove.
17580 (vn_reference_insert): Use shared ops for constructing the
17581 reference and copy it.
17583 2013-05-07 Richard Biener <rguenther@suse.de>
17585 PR middle-end/57190
17586 * tree-eh.c (sink_clobbers): Properly propagate
17587 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17589 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17591 PR tree-optimization/57149
17592 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17593 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17594 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17595 uninit_undefined_value_p instead of ssa_undefined_value_p.
17598 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17599 for modifier == EXPAND_INITIALIZER.
17601 2013-05-07 Anton Blanchard <anton@samba.org>
17603 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17604 for powerpc64 little endian.
17605 * configure: Regenerate.
17607 2013-05-06 Graham Stott <grahams@btinternet.com>
17609 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17610 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17611 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17612 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17614 2013-05-06 Graham Stott <grahams@btinternet.com>
17616 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17617 codes which allow non-lvalues.
17619 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17621 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17622 components are all 1s.
17623 (integer_minus_onep): New function.
17624 * tree.h (integer_minus_onep): Declare it.
17625 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17626 integer_minus_onep instead of integer_all_onesp.
17628 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17631 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17632 variations of these patterns.
17634 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17636 * config/i386/i386.md (isa): Add x64_sse4 member.
17637 (enabled): Handle x64_sse4.
17638 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17639 instruction for 64bit SSE4_1 targets. Update insn attributes.
17640 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17641 instruction for SSE4_1 targets. Update insn attributes.
17642 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17643 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17644 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17646 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17647 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17648 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17650 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17653 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17655 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17657 * final.c (do_assembler_dialects): Don't handle curly braces and
17658 vertical bar escaped by % as dialect delimiters.
17659 (output_asm_insn): Print curly braces and vertical bar if escaped
17660 by % and ASSEMBLER_DIALECT defined.
17661 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17662 * doc/tm.texi: Regenerated.
17664 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17666 * config/mips/mips.c: Include tree-pass.h.
17667 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17668 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17669 (pass_mips_machine_reorg2): New machine specific pass.
17670 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17671 (mips_option_override): Register the new pass.
17672 * rtl.h (cleanup_barriers): Remove prototype.
17673 (dbr_schedule): Likewise.
17674 * jump.c (cleanup_barriers): Make static.
17675 * reorg.c (dbr_schedule): Likewise.
17677 2013-05-06 Richard Biener <rguenther@suse.de>
17679 PR tree-optimization/57185
17680 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17681 SSA names properly.
17683 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17686 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17687 between "first_arg" and "insn", not "last" and "insn".
17689 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17691 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17692 (find_candidates_in_block): Re-enable slsr_process_phi.
17693 (create_phi_basis): Fix double counting of candidate adjustment.
17695 2013-05-06 Richard Biener <rguenther@suse.de>
17697 PR middle-end/57147
17698 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17699 the edge is also fallthru, preserve it and just clear the
17701 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17702 also complex, preserve that and just clear the fallthru flag.
17703 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17704 update virtual operands.
17706 2013-05-06 Alan Modra <amodra@gmail.com>
17708 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17709 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17710 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17711 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17712 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17713 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17715 2013-05-06 Alan Modra <amodra@gmail.com>
17717 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17718 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17719 (DEFAULT_ASM_ENDIAN): Define.
17720 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17721 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17722 Update -K PIC clause from sysv4.h.
17723 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17724 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17726 2013-05-06 Alan Modra <amodra@gmail.com>
17728 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17729 twice for little-endian.
17730 (ashrdi3_no_power, ashrdi3): Support little-endian.
17732 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17735 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17736 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17738 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17739 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17740 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17743 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17744 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17746 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17747 * config/pa/pa.opt: Make mbig-switch a no-op.
17748 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17749 (CASE_VECTOR_MODE): Always return SImode.
17750 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17751 for the !TARGET_BIG_SWITCH case.
17752 * config/pa/pa-linux.h: Likewise.
17753 * config/pa/pa-openbsd.h: Likewise.
17754 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17755 * config/pa/pa.md (short_jump): Remove define_insn.
17756 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17757 (casesi0): Remove define_insn.
17758 (type): Remove btable_branch.
17759 (pa_combine_type): Likewise.
17760 (in_nullified_branch_delay): Likewise.
17761 (in_call_delay): Likewise.
17762 (define_delay): Likewise.
17763 (define_insn_reservation "Z3"): Likewise.
17764 (define_insn_reservation "Z4"): Likewise.
17765 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17766 (pa_adjust_insn_length): Remove adjustment for btable branches.
17767 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17770 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17772 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17773 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17774 Add m->r,x alternatives.
17775 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17776 splitters using SWI48x mode iterator.
17777 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17778 TARGET_64BIT. Add m->x alternative.
17779 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17780 Add o->x alternative. Enable for TARGET_SSE.
17781 (sse_storeq): Remove expander.
17782 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17783 with memory input operand.
17784 (*vec_extractv2di_1 splitter): New.
17785 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17786 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17788 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17790 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17791 (INT_LOWPART): Delete.
17792 (extract_MB): Adjust.
17793 (extract_ME): Adjust.
17794 (print_operand): Adjust.
17796 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17798 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17799 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17800 (reg_or_logical_cint_operand, easy_fp_constant,
17801 logical_const_operand): Delete "CONST_DOUBLE" case.
17802 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17803 "HOST_BITS_PER_WIDE_INT == 64" test.
17804 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17805 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17806 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17807 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17809 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17810 CONST_DOUBLE DImode/VOIDmode case.
17811 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17812 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17813 CONST_DOUBLE VOIDmode case.
17814 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17815 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17816 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17817 Delete CONST_DOUBLE case.
17818 (splitters for mov FMOVE64 const_double): Delete
17819 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17820 "HOST_BITS_PER_WIDE_INT >= 64" test.
17821 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17823 (mov DI const_double): Delete.
17825 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17827 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17828 on op shows all bits zero in mode of a lowpart subreg, return zero.
17830 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17833 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17834 to save TFmode registers and DImode to save TImode registers for
17835 caller save operations.
17836 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17837 mark being partially clobbered since they only use the first
17840 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17841 and TDmode only use the upper 64-bits of each VSX register.
17843 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17845 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17846 (find_candidates_in_block): Disable slsr_process_phi.
17848 2013-05-03 Guozhi Wei <carrot@google.com>
17850 * coverage.c (coverage_obj_init): Move the construction of gcov
17852 (build_init_ctor): ... here.
17854 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17856 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17857 (slsr_cand_d): Redefine def_phi.
17858 (stride_status, phi_adjust_status, count_phis_status): New enums.
17859 (find_phi_def): New.
17860 (find_basis_for_base_expr): New.
17861 (find_basis_for_candidate): Handle hidden bases.
17862 (alloc_cand_and_find_basis): Handle phi candidates.
17863 (slsr_process_phi): New.
17864 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17865 (create_mul_imm_cand): Likewise.
17866 (create_add_ssa_cand): Exclude phi base candidates.
17867 (create_add_imm_cand): Likewise.
17868 (slsr_process_cast): Likewise.
17869 (slsr_process_copy): Likewise.
17870 (find_candidates_in_block): Handle phi candidates.
17871 (dump_candidate): Likewise.
17872 (unconditional_cands): Delete.
17873 (unconditional_cands_with_known_stride_p): Delete.
17874 (phi_dependent_cand_p): New.
17875 (cand_increment): Handle phi-dependent candidates.
17876 (replace_dependent): Delete.
17877 (replace_mult_candidate): New.
17878 (replace_unconditional_candidate): New.
17879 (incr_vec_index): Move to avoid forward reference.
17880 (create_add_on_incoming_edge): New.
17881 (create_phi_basis): New.
17882 (replace_dependents): Delete.
17883 (replace_conditional_candidate): New.
17884 (phi_add_costs): New.
17885 (replace_uncond_cands_and_profitable_phis): New.
17886 (record_increment): Handle phi adjustments.
17887 (record_phi_increments): New.
17888 (record_increments): Handle phi adjustments.
17889 (phi_incr_cost): New.
17890 (lowest_cost_path): Handle phis.
17891 (total_savings): Likewise.
17892 (analyze_increments): Likewise.
17893 (ncd_with_phi): New.
17894 (ncd_of_cand_and_phis): New.
17895 (nearest_common_dominator_for_cands): Handle phi increments.
17896 (all_phi_incrs_profitable): New.
17897 (replace_profitable_candidates): Handle phi-dependent candidates.
17898 (analyze_candidates_and_replace): Likewise.
17900 2013-05-03 Teresa Johnson <tejohnson@google.com>
17903 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17904 do not exceed REG_BR_PROB_BASE.
17906 2013-05-03 Jeff Law <law@redhat.com>
17908 PR tree-optimization/57144
17909 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17910 operand of the condition will bit into the new type when eliminating
17911 a cast feeding a condition.
17913 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17915 PR rtl-optimization/57130
17916 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17917 of COMPARE as in_code to the recursive call if needed.
17919 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17921 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17922 (enabled): Handle new members.
17923 * config/i386/sse.md (*vec_concatv2si): Merge from
17924 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17925 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17927 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17929 PR tree-optimization/57027
17930 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17931 for fnms opportunity, check we got the prerequisite kind
17932 of tree / gimple before using accessor functions.
17934 2013-05-03 Richard Biener <rguenther@suse.de>
17936 * double-int.h (lshift): New overload without precision
17937 and arith argument.
17938 (operator *=, operator +=, operator -=): Move ...
17939 * double-int.c (operator *=, operator +=, operator -=): ... here
17940 and implement more efficiently.
17941 (mul_double_with_sign): Remove.
17942 (lshift_double): Adjust to take unsinged shift argument, push
17943 dispatching code to callers.
17944 (mul_double_wide_with_sign): Add early out for callers that
17945 are not interested in high parts or overflow.
17946 (lshift): New function.
17947 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17948 dispatch code here.
17949 (lrotate, rrotate): Use logical shifts.
17950 * expr.c (get_inner_reference): Use lshift.
17951 * fixed-value.c (do_fixed_divide): Likewise.
17952 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17953 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17954 (indirect_refs_may_alias_p): Likewise.
17955 (stmt_kills_ref_p_1): Likewise.
17957 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17959 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17961 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17963 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17964 scalar form of FABD instruction.
17966 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17968 * lra-constraints.c (process_alt_operands): Add checking alt
17969 number to choose the best alternative.
17971 2013-05-02 Richard Biener <rguenther@suse.de>
17973 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17974 bitmap and its handling.
17975 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17977 2013-05-02 Richard Biener <rguenther@suse.de>
17979 PR middle-end/57140
17980 * tree-inline.c (copy_loops): Properly handle removed loops.
17981 (copy_cfg_body): Mark destination loops for fixup if source
17982 loops needed fixup.
17984 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17987 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17988 generating simple_return for naked functions.
17990 2013-05-02 Martin Jambor <mjambor@suse.cz>
17992 PR middle-end/56988
17993 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17994 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17996 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17997 ipa_agg_replacement_value structures.
17998 (known_aggs_to_agg_replacement_list): Likewise.
17999 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18000 (read_agg_replacement_chain): Likewise.
18001 (ipcp_transform_function): Also check that by_ref flags match.
18003 2013-05-02 Richard Biener <rguenther@suse.de>
18005 * graphds.h (struct graph): Add obstack member.
18006 * graphds.c (new_graph): Initialize obstack and allocate
18008 (add_edge): Allocate edge from the obstack.
18009 (free_graph): Free the obstack instead of all edges and vertices.
18011 2013-05-02 Teresa Johnson <tejohnson@google.com>
18013 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18015 * cfg.c (update_bb_profile_for_threading): Ditto.
18016 * tree-inline.c (copy_bb): Ditto.
18017 (copy_edges_for_bb): Ditto.
18018 (initialize_cfun): Ditto.
18019 (copy_cfg_body): Ditto.
18020 (expand_call_inline): Ditto.
18021 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18022 (estimate_node_size_and_time): Ditto.
18023 (inline_merge_summary): Ditto.
18024 * cgraphclones.c (cgraph_clone_edge): Ditto.
18025 (cgraph_clone_node): Ditto.
18026 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18027 (compute_trg_info): Ditto.
18029 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18031 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18032 S reg when fp attribute set.
18033 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18035 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18037 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18039 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18040 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18041 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18043 2013-05-02 Richard Biener <rguenther@suse.de>
18045 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18046 (struct instantiate_cache_entry): New type.
18047 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18048 (struct instantiate_cache_type): New type.
18049 (set_instantiated_value, get_instantiated_value): Remove.
18050 (get_instantiated_value_entry): New function.
18051 (instantiate_scev_name): Use the new cache and adjust.
18052 (instantiate_scev_poly): Adjust.
18053 (instantiate_scev_binary): Likewise.
18054 (instantiate_array_ref): Likewise.
18055 (instantiate_scev_convert): Likewise.
18056 (instantiate_scev_not): Likewise.
18057 (instantiate_scev_3): Likewise.
18058 (instantiate_scev_2): Likewise.
18059 (instantiate_scev_r): Likewise.
18060 (instantiate_scev): Likewise.
18061 (resolve_mixers): Likewise.
18063 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18066 * lra-constraints.c (best_small_class_operands_num): Remove.
18067 (process_alt_operands): Remove small_class_operands_num. Take
18068 small classes operands into losers and only if the operand is not
18069 matched. Modify debugging output.
18070 (curr_insn_transform): Remove best_small_class_operands_num.
18073 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18075 * config/aarch64/aarch64-builtins.c
18076 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18077 * config/aarch64/aarch64-simd-builtins.def
18078 (reduc_splus_): Add new modes.
18079 (reduc_uplus_): New.
18080 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18081 (reduc_uplus_v4sf): Likewise.
18082 (reduc_splus_v4sf): Likewise.
18083 (aarch64_addv<mode>): Likewise.
18084 (reduc_uplus_<mode>): Likewise.
18085 (reduc_splus_<mode>): Likewise.
18086 (aarch64_addvv2di): Likewise.
18087 (reduc_uplus_v2di): Likewise.
18088 (reduc_splus_v2di): Likewise.
18089 (aarch64_addvv2si): Likewise.
18090 (reduc_uplus_v2si): Likewise.
18091 (reduc_splus_v2si): Likewise.
18092 (reduc_<sur>plus_<mode>): New.
18093 (reduc_<sur>plus_v2di): Likewise.
18094 (reduc_<sur>plus_v2si): Likewise.
18095 (reduc_<sur>plus_v4sf): Likewise.
18096 (aarch64_addpv4sf): Likewise.
18097 * config/aarch64/arm_neon.h
18098 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18099 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18100 add UNSPEC_SADDV, UNSPEC_UADDV.
18102 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18104 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18106 * config/aarch64/arm_neon.h
18107 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18109 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18111 * config/aarch64/aarch64-builtins
18112 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18114 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18116 * config/aarch64/aarch64-simd-builtins.def
18117 (reduc_smax_): New.
18118 (reduc_smin_): Likewise.
18119 (reduc_umax_): Likewise.
18120 (reduc_umin_): Likewise.
18121 (reduc_smax_nan_): Likewise.
18122 (reduc_smin_nan_): Likewise.
18125 (smax): Update for V2SF, V4SF and V2DF modes.
18128 (smin_nan): Likewise.
18129 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18130 (<su><maxmin><mode>3): ...This, refactor.
18131 (s<maxmin><mode>3): New.
18132 (<maxmin_uns><mode>3): Likewise.
18133 (reduc_<maxmin_uns>_<mode>): Refactor.
18134 (reduc_<maxmin_uns>_v4sf): Likewise.
18135 (reduc_<maxmin_uns>_v2si): Likewise.
18136 (aarch64_<fmaxmin><mode>: Remove.
18137 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18139 (vmin<q>_f<32,64>): Likewise.
18140 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18142 (su): Add mappings for smax, smin, umax, umin.
18144 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18145 (FMAXMIN): Rename as...
18146 (FMAXMIN_UNS): ...This.
18148 (fmaxminv): Likewise.
18149 (fmaxmin): Likewise.
18151 (maxmin_uns_op): Likewise.
18153 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18155 * config/aarch64/arm_neon.h
18156 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18157 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18158 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18160 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18162 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18163 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18165 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18167 * config/aarch64/aarch64-simd.md
18168 (vcond<mode>_internal): Handle special cases for constant masks.
18169 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18170 (vcondu<mode><mode>): Likewise.
18171 (vcond<v_cmp_result><mode>): New.
18173 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18175 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18176 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18177 * config/aarch64/aarch64-simd-builtins.def
18178 (cmeq): Update to BUILTIN_VALLDI.
18183 * config/aarch64/arm_neon.h
18184 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18185 to builtins or C as appropriate.
18187 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18189 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18191 (cmhi): Rename to...
18193 * config/aarch64/aarch64-simd.md
18194 (simd_mode): Add SF.
18195 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18196 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18197 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18198 (cstore<mode>_neg): ...This.
18199 * config/aarch64/iterators.md
18201 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18202 (COMPARISONS): New.
18203 (UCOMPARISONS): Likewise.
18204 (optab): Add missing comparisons.
18210 (VCMP_S): Likewise.
18211 (VCMP_U): Likewise.
18212 (V_cmp_result): Add DF, SF modes.
18213 (v_cmp_result): Likewise.
18215 (vmtype): Likewise.
18216 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18218 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18220 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18221 define_insn to define_insn_and_split.
18222 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18223 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18224 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18225 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18226 (thumb2_negscc): Likewise.
18228 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18230 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18232 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18234 * config/arm/thumb2.md: Remove trailing whitespaces.
18236 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18238 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18239 Use gen_int_mode rather than GEN_INT.
18241 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18243 * value-prof.c (stream_in_histogram_value): Remove the strayed
18246 2013-04-30 Richard Biener <rguenther@suse.de>
18248 PR middle-end/57122
18249 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18251 2013-04-30 Richard Biener <rguenther@suse.de>
18253 PR middle-end/57107
18254 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18256 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18258 PR rtl-optimization/56957
18259 PR rtl-optimization/57105
18260 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18261 variable. Use just INSN_UID for determining whether an insn
18262 should be only disconnected from the insn stream.
18263 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18265 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18267 PR tree-optimization/57104
18268 * tsan.c (instrument_expr): Don't instrument accesses to
18269 DECL_HARD_REGISTER VAR_DECLs.
18271 2013-04-30 Richard Biener <rguenther@suse.de>
18273 * function.h (loops_for_fn): New inline function.
18274 (set_loops_for_fn): Likewise.
18275 * cfgloop.h (place_new_loop): Add struct function parameter.
18276 (get_loop): Likewise.
18277 (get_loops): Likewise.
18278 (number_of_loops): Likewise.
18279 (fel_next): Adjust.
18280 (fel_init): Likewise.
18281 * cfg.c (get_loop_copy): Adjust.
18282 * cfgloop.c (flow_loops_dump): Likewise.
18283 (record_loop_exits): Likewise.
18284 (verify_loop_structure): Likewise.
18285 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18286 (estimate_reg_pressure_cost): Likewise.
18287 (mark_loop_exit_edges): Likewise.
18288 * cfgloopmanip.c (place_new_loop): Likewise.
18289 (add_loop): Likewise.
18290 (duplicate_loop): Likewise.
18291 * graph.c (draw_cfg_nodes): Likewise.
18292 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18293 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18294 (extract_affine_chrec): Likewise.
18295 (build_scop_iteration_domain): Likewise.
18296 * graphite.c (graphite_initialize): Likewise.
18297 * ira-build.c (create_loop_tree_nodes): Likewise.
18298 (more_one_region_p): Likewise.
18299 (rebuild_regno_allocno_maps): Likewise.
18300 (mark_loops_for_removal): Likewise.
18301 (mark_all_loops_for_removal): Likewise.
18302 (remove_unnecessary_regions): Likewise.
18303 (ira_build): Likewise.
18304 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18305 * loop-init.c (fix_loop_structure): Likewise.
18306 (gate_rtl_move_loop_invariants): Likewise.
18307 (gate_rtl_unswitch): Likewise.
18308 (gate_rtl_unroll_and_peel_loops): Likewise.
18309 (rtl_doloop): Likewise.
18310 * lto-streamer-in.c (input_cfg): Likewise.
18311 * lto-streamer-out.c (output_cfg): Likewise.
18312 * modulo-sched.c (sms_schedule): Likewise.
18313 * predict.c (tree_estimate_probability): Likewise.
18314 (tree_estimate_probability_driver): Likewise.
18315 (estimate_loops): Likewise.
18316 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18317 (move_sese_region_to_fn): Likewise.
18318 (debug_loop_num): Likewise.
18319 * tree-chrec.c (chrec_evaluate): Likewise.
18320 (hide_evolution_in_other_loops_than_loop): Likewise.
18321 (chrec_component_in_loop_num): Likewise.
18322 (reset_evolution_in_loop): Likewise.
18323 (evolution_function_is_invariant_rec_p): Likewise.
18324 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18325 * tree-inline.c (copy_loops): Likewise.
18326 (copy_cfg_body): Likewise.
18327 (tree_function_versioning): Likewise.
18328 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18329 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18331 (add_to_evolution_1): Likewise.
18332 (scev_const_prop): Likewise.
18333 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18334 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18335 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18336 (tree_ssa_lim_initialize): Likewise.
18337 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18338 (verify_loop_closed_ssa): Likewise.
18339 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18340 (tree_ssa_loop_im): Likewise.
18341 (tree_ssa_loop_unswitch): Likewise.
18342 (tree_vectorize): Likewise.
18343 (check_data_deps): Likewise.
18344 (tree_ssa_loop_ivcanon): Likewise.
18345 (tree_ssa_loop_bounds): Likewise.
18346 (tree_complete_unroll): Likewise.
18347 (tree_complete_unroll_inner): Likewise.
18348 (tree_parallelize_loops): Likewise.
18349 (tree_ssa_loop_prefetch): Likewise.
18350 (tree_ssa_loop_ivopts): Likewise.
18351 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18352 * tree-vectorizer.c (vectorize_loops): Likewise.
18354 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18356 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18357 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18358 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18359 with EABI_LINK_SPEC.
18361 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18364 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18367 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18370 * lra-constraints.c (process_alt_operands): Discourage a bit more
18371 using memory for pseudos. Print cost dump for alternatives.
18372 Modify cost values for conflicts with early clobbers.
18373 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18375 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18378 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18380 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18382 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18383 from/to S register.
18384 (movdi_aarch64): Support LDR/STR from/to D register.
18386 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18388 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18389 or higher by default.
18391 2013-04-29 Richard Biener <rguenther@suse.de>
18393 PR middle-end/57075
18394 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18395 even if not adding abnormal edges for calls that can make
18398 2013-04-29 Richard Biener <rguenther@suse.de>
18400 PR middle-end/57103
18401 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18403 (move_stmt_r): Remove redundant checking.
18405 2013-04-29 Teresa Johnson <tejohnson@google.com>
18408 * basic-block.h (apply_scale): New function.
18409 (apply_probability): Use apply_scale.
18410 * gimple-streamer-in.c (input_bb): Ditto.
18411 * lto-streamer-in.c (input_cfg): Ditto.
18412 * lto-cgraph.c (merge_profile_summaries): Ditto.
18413 * tree-optimize.c (execute_fixup_cfg): Ditto.
18414 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18415 (copy_edges_for_bb): Ditto.
18416 (copy_cfg_body): Ditto.
18418 2013-04-29 Tom de Vries <tom@codesourcery.com>
18420 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18421 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18422 (tail_merge_optimize): Handle current_loops == NULL.
18424 2013-04-26 Jeff Law <law@redhat.com>
18426 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18427 (simplify_cond_using_ranges): Generalize code to simplify
18428 COND_EXPRs where one argument is a constant and the other
18429 is an SSA_NAME created by an integral type conversion.
18431 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18433 * config/arm/arm.md (store_minmaxsi): Use only when
18434 optimize_insn_for_size_p.
18436 2013-04-29 Christian Bruel <christian.bruel@st.com>
18439 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18441 2013-04-29 Richard Biener <rguenther@suse.de>
18443 PR middle-end/57089
18444 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18445 loop tree make sure to schedule a fixup for the child as well.
18446 (expand_omp_for_generic): Properly add loops.
18447 (expand_omp_for_static_nochunk): Likewise.
18448 (expand_omp_for_static_chunk): Likewise.
18449 (expand_omp_for): For the degenerate case fixup loops.
18450 (expand_omp_sections): Fix default bb placement in loops.
18451 (expand_omp_atomic_pipeline): Properly add loops.
18453 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18455 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18457 2013-04-29 Tom de Vries <tom@codesourcery.com>
18459 * tree-ssa-tail-merge.c: Update header comment.
18461 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18463 * config/aarch64/arm_neon.h
18464 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18465 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18466 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18467 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18468 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18469 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18470 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18471 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18473 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18475 * config/aarch64/aarch64-simd.md
18476 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18477 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18478 fix_trunc, fixuns_trunc.
18479 (ftrunc<VDQF:mode>2): New.
18480 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18481 (fix_trunc_optab): New.
18483 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18485 * config/aarch64/aarch64-builtins.c
18486 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18487 iceilf, lround, iroundf.
18489 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18492 * config/i386/i386.h (enum ix86_tune_indices)
18493 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18494 New, split from X86_TUNE_INTER_UNIT_MOVES.
18495 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18496 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18497 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18498 (TARGET_INTER_UNIT_MOVES): Remove.
18499 * config/i386/i386.c (initial_ix86_tune_features): Update.
18500 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18501 (ix86_expand_convert_uns_didf_sse): Use
18502 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18503 (ix86_expand_vector_init_one_nonzero): Ditto.
18504 (ix86_expand_vector_init_interleave): Ditto.
18505 (inline_secondary_memory_needed): Return true for moves from SSE class
18506 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18507 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18508 * config/i386/constraints.md (Yi, Ym): Depend on
18509 TARGET_INTER_UNIT_MOVES_TO_VEC.
18510 (Yj, Yn): New constraints.
18511 * config/i386/i386.md (*movdi_internal): Change constraints of
18512 operand 1 from Yi to Yj and from Ym to Yn.
18513 (*movsi_internal): Ditto.
18514 (*movdf_internal): Ditto.
18515 (*movsf_internal): Ditto.
18516 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18517 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18518 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18519 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18520 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18521 * config/i386/sse.md (movdi_to_sse): Ditto.
18522 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18523 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18524 TARGET_INTER_UNIT_MOVES.
18525 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18526 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18527 instead of TARGET_INTER_UNIT_MOVES.
18528 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18529 operand 1 from Yi to Yj and from Ym to Yn.
18531 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18533 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18534 (float_truncate_hi_): Likewise.
18535 (float_extend_lo_): Likewise.
18536 (float_truncate_lo_): Likewise.
18537 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18538 (aarch64_float_extend_lo_v2df): Likewise.
18539 (vec_unpacks_hi_v4sf): Likewise.
18540 (aarch64_float_truncate_lo_v2sf): Likewise.
18541 (aarch64_float_truncate_hi_v4sf): Likewise.
18542 (vec_pack_trunc_v2df): Likewise.
18543 (vec_pack_trunc_df): Likewise.
18545 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18547 * config/aarch64/aarch64-builtins.c
18548 (aarch64_fold_builtin): Fold float conversions.
18549 * config/aarch64/aarch64-simd-builtins.def
18550 (floatv2si, floatv4si, floatv2di): New.
18551 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18552 * config/aarch64/aarch64-simd.md
18553 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18554 * config/aarch64/iterators.md (FLOATUORS): New.
18555 (optab): Add float, floatuns.
18556 (su_optab): Likewise.
18558 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18560 * config/aarch64/aarch64-builtins.c
18561 (aarch64_builtin_vectorized_function): Use new names for
18563 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18564 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18565 (fcvtzu): Split as...
18566 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18567 (fcvtas): Split as...
18568 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18569 (fcvtau): Split as...
18570 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18571 (fcvtps): Split as...
18572 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18573 (fcvtpu): Split as...
18574 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18575 (fcvtms): Split as...
18576 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18577 (fcvtmu): Split as...
18578 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18579 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18580 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18581 (lfrintnusf, lfrintnudf): Likewise.
18582 * config/aarch64/aarch64-simd.md
18583 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18585 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18586 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18587 (fcvt_pattern): Likewise.
18589 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18591 * config/aarch64/aarch64-simd.md
18592 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18593 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18595 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18597 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18598 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18599 (vrnd<a,m,n,p>_f32): Implement using builtins.
18600 (vrnd<i,x><q>_f<32, 64>): New.
18602 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18604 * config/aarch64/aarch64-builtins.c
18605 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18606 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18607 (frintz): Rename to...
18609 (frintp): Rename to...
18611 (frintm): Rename to...
18613 (frinti): Rename to...
18614 (nearbyint): ...this.
18615 (frintx): Rename to...
18617 (frinta): Rename to...
18619 * config/aarch64/aarch64-simd.md
18620 (aarch64_frint<frint_suffix><mode>): Delete.
18621 (<frint_pattern><mode>2): Convert to insn.
18622 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18623 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18624 (frint_pattern): Likewise.
18625 (frint_suffix): Likewise.
18627 2013-04-29 Richard Biener <rguenther@suse.de>
18629 PR tree-optimization/57081
18630 * loop-init.c: Include tree-flow.h.
18631 (loop_optimizer_finalize): Free number of iteration estimates.
18632 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18634 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18636 PR tree-optimization/57083
18637 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18638 non-singleton shift count range, zero extend low_bound for uns case.
18640 * config/i386/predicates.md (general_vector_operand): New predicate.
18641 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18642 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18643 if they aren't nonimmediate operands. If their original values
18644 satisfy const_vector_equal_evenodd_p, don't shift them.
18645 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18646 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18647 if not nonimmediate_operand.
18648 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18649 instead of register_operand.
18650 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18652 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18654 * stor-layout.c (finalize_size_functions): Allocate a structure and
18655 reset cfun before dumping the functions.
18657 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18659 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18662 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18663 use xop_pmacsdqh if uns_p.
18664 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18665 the immediate rotate count.
18667 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18669 * rtl.h (struct rtx_def): Add comment for field jump.
18670 (LRA_SUBREG_P): New macro.
18671 * recog.c (register_operand): Check LRA_SUBREG_P.
18672 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18674 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18675 (lra_final_code_change): Skip subreg change for operators.
18676 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18677 if there are no operand changes.
18678 * lra-constraints.c (curr_insn_set): New.
18679 (match_reload): Set LRA_SUBREG_P.
18680 (emit_spill_move): Ditto.
18681 (check_and_process_move): Use curr_insn_set. Process only single
18682 set insns. Don't initialize sec_mem_p and change_p.
18683 (simplify_operand_subreg): Use LRA_SUBREG_P.
18684 (reg_in_class_p): New function.
18685 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18686 of #ifdef. Add code to remove cycling.
18687 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18688 non-null disp. Reload inner instead of disp when base and index
18689 are null. Try to put lo_sum into register.
18690 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18691 (check_and_process_move): Move code for move cost check to
18692 simple_move_p. Remove equiv_substitution.
18693 (simple_move_p): New function.
18694 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18695 curr_insn_set. Call check_and_process_move only for single set
18696 insns. Use the new function. Move call of check_and_process_move
18697 after operand equiv substitution and address process.
18699 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18702 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18703 with nonlocal goto receivers or returns twice calls, ignore
18704 unininitialized values from abnormal edges to nl goto receiver
18705 or returns twice call.
18707 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18709 PR tree-optimization/57051
18710 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18711 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18714 2013-04-26 Richard Biener <rguenther@suse.de>
18716 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18717 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18718 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18719 (fixup_loop_arrays_after_move): New function.
18720 (move_sese_region_to_fn): Properly outline the loop tree parts
18721 of the SESE region.
18723 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18725 * config/i386/i386.md (type, unit): Fix long lines.
18727 2013-04-26 Richard Biener <rguenther@suse.de>
18729 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18730 (lto-streamer-out.o): Likewise.
18731 * cfgloop.c (init_loops_structure): Export, add struct function
18732 argument and adjust.
18733 (flow_loops_find): Adjust.
18734 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18735 (init_loops_structure): Declare.
18736 * lto-streamer-in.c: Include cfgloop.h.
18737 (input_cfg): Input the loop tree.
18738 * lto-streamer-out.c: Include cfgloop.h.
18739 (output_cfg): Output the loop tree.
18740 (output_struct_function_base): Do not drop PROP_loops.
18742 2013-03-26 Richard Biener <rguenther@suse.de>
18744 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18745 (pass_build_cfg): Provide PROP_loops.
18746 (move_sese_region_to_fn): Remove loops that are outlined into fn
18748 * tree-inline.c: Include cfgloop.h.
18749 (initialize_cfun): Do not drop PROP_loops.
18750 (copy_loops): New function.
18751 (copy_cfg_body): Copy loop structure.
18752 (tree_function_versioning): Initialize destination loop tree.
18753 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18754 (pass_parallelize_loops): Do IL verification.
18755 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18756 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18757 the CFG make sure we fixup loops as well.
18758 * tree-ssa-tail-merge.c: Include cfgloop.h.
18759 (replace_block_by): When merging loop latches mark loops for fixup.
18760 * lto-streamer-out.c (output_struct_function_base): Drop
18761 PROP_loops for now.
18762 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18763 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18764 * ipa-split.c: Include cfgloop.h.
18765 (split_function): Add the new return block to the loop tree root.
18766 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18767 whether we have removed the forwarder block.
18768 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18769 * cfgloop.h (place_new_loop): Declare.
18770 * cfgloopmanip.c (place_new_loop): Export.
18771 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18772 (tree-switch-conversion.o): Likewise.
18773 (tree-complex.o): Likewise.
18774 (tree-inline.o): Likewise.
18775 (tree-ssa-tailmerge.o): Likewise.
18776 (ipa-split.o): Likewise.
18777 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18778 (tree-ssa-copy.o): Likewise.
18779 * tree-switch-conversion.c: Include cfgloop.h
18780 (process_switch): If we emit a bit-test cascade, schedule loops
18782 * tree-complex.c: Include cfgloop.h.
18783 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18784 * asan.c: Include cfgloop.h.
18785 (create_cond_insert_point): Properly add new basic-blocks to
18786 loops, schedule loop fixup.
18787 * cfgloop.c (verify_loop_structure): Check that looks are not
18789 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18791 (expand_omp_for_generic): Likewise.
18792 (expand_omp_sections): Likewise.
18793 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18794 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18795 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18796 is initialized, not when loops are present.
18797 * tree-parloops.c (parallelize_loops): Remove checking here.
18798 * passes.c (init_optimization_passes): Schedule a copy-propagation
18799 pass before complete unrolling of inner loops.
18801 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18803 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18804 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18805 (colorize_init): Add argument to _WIN32 version.
18806 * toplev.c: Include diagnostic-color.h.
18807 (process_options): Default to -fdiagnostics-color=auto if
18808 GCC_COLORS env var is in the environment.
18809 * common.opt (fdiagnostics-color=): Add Var and Init.
18810 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18811 env var is in the environment, the default is auto rather than never.
18813 * diagnostic.h (file_name_as_prefix): Add context argument.
18814 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18815 the string as locus.
18816 * langhooks.c (lhd_print_error_function): Adjust caller.
18818 2013-04-25 Lawrence Crowl <crowl@google.com>
18820 * var-tracking.c (shared_hash_def::htab):
18821 Change type to hash_table. Update dependent calls and types.
18823 2013-04-25 Lawrence Crowl <crowl@google.com>
18825 * Makefile.in: Update as needed below.
18827 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18828 Move declaration to after the type's method definitons.
18830 * attribs.c (htab_t scoped_attributes::attribute_hash):
18831 Change type to hash_table. Update dependent calls and types.
18833 * bitmap.c (htab_t bitmap_desc_hash):
18834 Change type to hash_table. Update dependent calls and types.
18836 * cselib.c (htab_t cselib_hash_table):
18837 Change type to hash_table. Update dependent calls and types.
18839 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18840 (hash_string_slot_node): Move implementation into lto-streamer.h
18841 struct string_slot_hasher.
18842 (eq_string_slot_node): Likewise.
18844 * data-streamer-out.c: Update output_block::string_hash_table
18845 dependent calls and types.
18847 * dwarf2cfi.c (htab_t trace_index):
18848 Change type to hash_table. Update dependent calls and types.
18850 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18851 Change type to hash_table. Update dependent calls and types.
18852 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18853 (htab_t optimize_external_refs::map): Likewise.
18854 (htab_t output_comp_unit::extern_map): Likewise.
18855 (htab_t output_comdat_type_unit::extern_map): Likewise.
18856 (htab_t output_macinfo::macinfo_htab): Likewise.
18857 (htab_t optimize_location_lists::htab): Likewise.
18858 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18860 * except.c (htab_t ehspec_hash_type):
18861 Change type to hash_table. Update dependent calls and types.
18862 (assign_filter_values::ttypes): Likewise.
18863 (assign_filter_values::ehspec): Likewise.
18864 (sjlj_assign_call_site_values::ar_hash): Likewise.
18865 (convert_to_eh_region_ranges::ar_hash): Likewise.
18867 * gcse.c (htab_t pre_ldst_table):
18868 Change type to hash_table. Update dependent calls and types.
18870 * ggc-common.c (htab_t saving_htab):
18871 Change type to hash_table. Update dependent calls and types.
18872 (htab_t loc_hash): Likewise.
18873 (htab_t ptr_hash): Likewise.
18874 (call_count): Rename ggc_call_count.
18875 (call_alloc): Rename ggc_call_alloc.
18876 (loc_descriptor): Rename make_loc_descriptor.
18877 (add_statistics): Rename ggc_add_statistics.
18879 * ggc-common.c (saving_htab):
18880 Change type to hash_table. Update dependent calls and types.
18882 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18883 (push_gimplify_context): Likewise.
18884 (pop_gimplify_context): Likewise.
18885 (struct gimple_temp_hash_elt): Added.
18886 (struct gimplify_hasher): Likewise.
18887 (struct gimplify_ctx.temp_htab):
18888 Change type to hash_table. Update dependent calls and types.
18890 * gimple-fold.c: Include gimplify-ctx.h.
18892 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18893 Change type to hash_table. Update dependent calls and types.
18894 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18895 avoid potential global name collision.
18897 * gimplify.c: Include gimplify-ctx.h.
18898 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18899 (htab_t gimplify_ctx::temp_htab):
18900 Update dependent calls and types for new type hash_table.
18901 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18902 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18904 * gimplify-ctx.h: New.
18905 (struct gimple_temp_hash_elt): Move from gimplify.c.
18906 (class gimplify_hasher): New.
18907 (struct gimplify_ctx): Move from gimple.h.
18908 (htab_t gimplify_ctx::temp_htab):
18909 Change type to hash_table. Update dependent calls and types.
18911 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18912 (htab_t ivs_params::newivs_index):
18913 Change type to hash_table. Update dependent calls and types.
18914 (htab_t ivs_params::params_index): Likewise.
18915 (htab_t print_generated_program::params_index): Likewise.
18916 (htab_t gloog::newivs_index): Likewise.
18917 (htab_t gloog::params_index): Likewise.
18919 * graphite.c: Include graphite-htab.h.
18920 4htab_t graphite_transform_loops::bb_pbb_mapping):
18921 Change type to hash_table. Update dependent calls and types.
18923 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18924 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18925 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18927 * graphite-dependences.c: Include graphite-htab.h.
18928 (loop_is_parallel_p): Change hash table type of parameter.
18930 * graphite-htab.h: New.
18931 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18932 (extern find_pbb_via_hash): Move from graphite-poly.h.
18933 (extern loop_is_parallel_p): Move from graphite-poly.h.
18934 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18936 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18937 (extern loop_is_parallel_p): Move to graphite-htab.h.
18938 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18940 * haifa-sched.c (htab_t delay_htab):
18941 Change type to hash_table. Update dependent calls and types.
18942 (htab_t delay_htab_i2): Likewise.
18944 * ira-color.c (htab_t allocno_hard_regs_htab):
18945 Change type to hash_table. Update dependent calls and types.
18947 * ira-costs.c (htab_t cost_classes_htab):
18948 Change type to hash_table. Update dependent calls and types.
18950 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18951 Change type to hash_table. Update dependent calls and types.
18953 * loop-iv.c (htab_t bivs):
18954 Change type to hash_table. Update dependent calls and types.
18956 * loop-unroll.c (htab_t opt_info::insns_to_split):
18957 Change type to hash_table. Update dependent calls and types.
18958 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18960 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18961 (struct string_slot_hasher): New.
18962 (htab_t output_block::string_hash_table):
18963 Change type to hash_table. Update dependent calls and types.
18965 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18966 (htab_t file_name_hash_table):
18967 Change type to hash_table. Update dependent calls and types.
18969 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18972 * lto-streamer.c (htab_t tree_htab):
18973 Change type to hash_table. Update dependent calls and types.
18975 * omp-low.c: Include gimplify-ctx.h.
18977 * passes.c (htab_t name_to_pass_map):
18978 Change type to hash_table. Update dependent calls and types.
18979 (pass_traverse): Rename to passes_pass_traverse.
18981 * plugin.c (htab_t event_tab):
18982 Change type to hash_table. Update dependent calls and types.
18984 * postreload-gcse.c (htab_t expr_table):
18985 Change type to hash_table. Update dependent calls and types.
18986 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18988 * sese.c (debug_rename_map_1): Make extern.
18989 (htab_t copy_bb_and_scalar_dependences::rename_map):
18990 Change type to hash_table. Update dependent calls and types.
18992 * sese.h (extern debug_rename_map): Move to .c file.
18994 * store-motion.c (htab_t store_motion_mems_table):
18995 Change type to hash_table. Update dependent calls and types.
18997 * trans-mem.c (htab_t tm_new_mem_hash):
18998 Change type to hash_table. Update dependent calls and types.
19000 * tree-browser.c (htab_t TB_up_ht):
19001 Change type to hash_table. Update dependent calls and types.
19003 * tree-cfg.c (htab_t discriminator_per_locus):
19004 Change type to hash_table. Update dependent calls and types.
19006 * tree-complex.c: Include tree-hasher.h
19007 (htab_t complex_variable_components):
19008 Change type to hash_table. Update dependent calls and types.
19010 * tree-eh.c (htab_t finally_tree):
19011 Change type to hash_table. Update dependent calls and types.
19013 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19014 struct int_tree_hasher.
19015 (extern int_tree_map_eq): Likewise.
19016 (uid_decl_map_hash): Removed.
19017 (extern decl_tree_map_eq): Likewise.
19019 * tree-hasher.h: New.
19020 (struct int_tree_hasher): New.
19021 (typedef int_tree_htab_type): New.
19023 * tree-inline.c: Include gimplify-ctx.h.
19025 * tree-mudflap.c: Include gimplify-ctx.h.
19027 * tree-parloops.c: Include tree-hasher.h.
19028 (htab_t eliminate_local_variables_stmt::decl_address):
19029 Change type to hash_table. Update dependent calls and types.
19030 (htab_t separate_decls_in_region::decl_copies): Likewise.
19032 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19033 Change type to hash_table. Update dependent calls and types.
19035 * tree-sra.c (candidates):
19036 Change type to hash_table. Update dependent calls and types.
19038 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19040 (int_tree_map_hash): Likewise.
19042 * tree-ssa-dom.c (htab_t avail_exprs):
19043 Change type to hash_table. Update dependent calls and types.
19045 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19046 Change type to hash_table. Update dependent calls and types.
19048 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19049 Change type to hash_table. Update dependent calls and types.
19051 * tree-ssa-phiopt.c (seen_ssa_names):
19052 Change type to hash_table. Update dependent calls and types.
19054 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19055 Change type to hash_table. Update dependent calls and types.
19057 * tree-ssa-uncprop.c (equiv):
19058 Change type to hash_table. Update dependent calls and types.
19060 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19062 PR rtl-optimization/57003
19063 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19064 call note_stores with kill_clobbered_value callback again after
19065 killing regs_invalidated_by_call.
19067 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19069 * config/aarch64/aarch64-simd.md
19070 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19071 (aarch64_simd_bsl<mode>): Likewise.
19072 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19074 2013-04-25 Marek Polacek <polacek@redhat.com>
19076 PR tree-optimization/57066
19077 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19079 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19081 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19083 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19085 * config/aarch64/aarch64-builtins.c
19086 (aarch64_fold_builtin): New.
19087 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19088 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19089 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19090 * config/aarch64/arm_neon.h
19091 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19093 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19094 Tejas Belagod <tejas.belagod@arm.com>
19096 * config/aarch64/aarch64-builtins.c
19097 (aarch64_gimple_fold_builtin): New.
19098 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19099 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19100 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19101 (addvv4sf): Update.
19102 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19104 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19106 * config/aarch64/aarch64.md
19107 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19109 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19111 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19112 (*ngcsi_uxtw): New pattern.
19114 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19115 Julian Brown <julian@codesourcery.com>
19117 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19118 (TB_DREG): Add T_V4HF.
19119 (v4hf_UP): New macro.
19120 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19121 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19122 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19124 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19125 (arm_vector_mode_supported_p): Handle V4HF.
19126 (arm_mangle_map): Handle V4HFmode.
19127 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19128 * config/arm/arm_neon_builtins.def: Add entries for
19129 vcvtv4hfv4sf, vcvtv4sfv4hf.
19130 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19131 (neon_vcvtv4hfv4sf): Likewise.
19132 * config/arm/neon-gen.ml: Handle half-precision floating point
19134 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19135 * config/arm/arm_neon.h: Regenerate.
19136 * config/arm/neon.ml (type elts): Add F16.
19137 (type vectype): Add T_float16x4, T_floatHF.
19138 (type vecmode): Add V4HF.
19139 (type features): Add Requires_FP_bit feature.
19140 (elt_width): Handle F16.
19141 (elt_class): Likewise.
19142 (elt_of_class_width): Likewise.
19143 (mode_of_elt): Refactor.
19144 (type_for_elt): Handle F16, fix error messages.
19145 (vectype_size): Handle T_float16x4.
19146 (vcvt_sh): New function.
19147 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19148 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19149 (string_of_mode): Handle V4HF.
19150 * doc/arm-neon-intrinsics.texi: Regenerate.
19152 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19154 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19155 format specifier in 'X' case.
19157 2013-04-25 Alan Modra <amodra@gmail.com>
19160 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19161 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19162 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19163 Repeat for many other rotate/shift and mask patterns using subregs.
19164 Name lshiftrt insns.
19165 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19166 on WORDS_BIG_ENDIAN.
19168 2013-04-25 Alan Modra <amodra@gmail.com>
19170 * config.gcc: Support little-endian powerpc-linux targets.
19171 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19172 (LINK_OS_LINUX_SPEC): Define.
19173 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19174 Preserve MASK_LITTLE_ENDIAN.
19175 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19176 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19177 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19178 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19179 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19180 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19181 Correct fp word order for little-endian. Don't shift toc entries
19182 smaller than a word for little-endian.
19183 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19184 (bswapdi2 splits): Correct low-part subreg for little-endian.
19185 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19186 low/high where such is correct only for be.
19187 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19188 little-endian for -mcall-aixdesc.
19190 2013-04-25 Alan Modra <amodra@gmail.com>
19192 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19193 replace_equiv_address_nv.
19195 2013-04-25 Alan Modra <amodra@gmail.com>
19197 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19199 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19202 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19203 * rtl.h (struct rtx_def): ...
19205 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19207 PR rtl-optimizations/57046
19208 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19209 for multi-reg splits.
19211 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19213 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19215 2013-04-24 Sterling Augustine <saugustine@google.com>
19217 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19218 (comp_dir_string, debug_str_dwo_section): New.
19219 (DEBUG_STR_DWO_SECTION): Rename to ...
19220 (DEBUG_DWO_STR_SECTION): ... this.
19221 (DEBUG_NORM_STR_SECTION): Delete.
19222 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19223 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19224 (find_AT_string): Move most logic to ...
19225 (find_AT_string_in_table): ... here. New.
19226 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19227 add_skeleton_AT_string. Delete logic.
19228 (output_skeleton_debug_sections): Remove call to
19229 add_top_level_skeleton_die_attrs.
19230 (add_comp_dir_attribute): Move logic to comp_dir_string.
19231 (dwarf2out_init): Initialize debug_str_dwo_section.
19232 (output_indirect_string): Call find_string_form.
19233 (output_indirect_strings): Rewrite.
19234 (prune_unused_types): Empty skeleton_debug_str_hash.
19235 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19236 (dwarf2out_finish): Call output_indirect_strings.
19238 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19240 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19242 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19244 * rtl.h (struct rtx_def): Add comment for field jump.
19245 (LRA_SUBREG_P): New macro.
19246 * recog.c (register_operand): Check LRA_SUBREG_P.
19247 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19249 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19250 (lra_final_code_change): Skip subreg change for operators.
19251 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19252 if there are no operand changes.
19253 * lra-constraints.c (curr_insn_set): New.
19254 (match_reload): Set LRA_SUBREG_P.
19255 (emit_spill_move): Ditto.
19256 (check_and_process_move): Use curr_insn_set. Process only single
19257 set insns. Don't initialize sec_mem_p and change_p.
19258 (simplify_operand_subreg): Use LRA_SUBREG_P.
19259 (reg_in_class_p): New function.
19260 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19261 of #ifdef. Add code to remove cycling.
19262 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19263 non-null disp. Reload inner instead of disp when base and index
19264 are null. Try to put lo_sum into register.
19265 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19266 (check_and_process_move): Move code for move cost check to
19267 simple_move_p. Remove equiv_substitution.
19268 (simple_move_p): New function.
19269 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19270 curr_insn_set. Call check_and_process_move only for single set
19271 insns. Use the new function. Move call of check_and_process_move
19272 after operand equiv substitution and address process.
19274 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19276 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19277 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19278 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19280 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19282 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19284 2013-04-24 Marek Polacek <polacek@redhat.com>
19286 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19287 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19288 (select_loops_exit_conditions): Likewise.
19289 (number_of_iterations_for_all_loops): Likewise.
19290 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19291 (scev_analysis): Likewise.
19293 2013-04-02 Catherine Moore <clm@codesourcery.com>
19294 Chao-ying Fu <fu@mips.com>
19296 * config/mips/micromips.md (jraddiusp): New pattern.
19297 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19298 instruction if possible.
19300 2013-04-24 Alan Modra <amodra@gmail.com>
19302 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19304 2013-04-24 Julian Brown <julian@codesourcery.com>
19305 Chung-Lin Tang <cltang@codesourcery.com>
19307 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19308 dependency behavior in enumeration type DIE generation. Add TODO note
19309 to comments about future DW_FORM_sdata/udata re-work of related code.
19311 2013-04-23 Lawrence Crowl <crowl@google.com>
19313 * Makefile.in: Update as needed below.
19315 * hash-table.h (class hash_table):
19316 Correct many methods with parameter types compare_type to the correct
19317 value_type. (Correct code was unlikely to notice the change.)
19318 (hash_table::elements_with_deleted) New.
19319 (class hashtable::iterator): New.
19320 (hashtable::begin()): New.
19321 (hashtable::end()): New.
19322 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19324 * statistics.c (statistics_hashes):
19325 Change type to hash_table. Update dependent calls and types.
19327 * tree-into-ssa.c (var_infos):
19328 Change type to hash_table. Update dependent calls and types.
19330 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19331 Change type to hash_table. Update dependent calls and types.
19333 * tree-ssa-loop-im.c (struct mem_ref.refs):
19334 Change type to hash_table. Update dependent calls and types.
19336 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19337 Change type to hash_table. Update dependent calls and types.
19339 * tree-ssa-sccvn.c (vn_tables_s::nary):
19340 Change type to hash_table. Update dependent calls and types.
19341 (vn_tables_s::phis): Likewise.
19342 (vn_tables_s::references): Likewise.
19344 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19345 (vn_reference_eq): Update parameter and return types.
19347 * tree-ssa-structalias.c (pointer_equiv_class_table):
19348 Change type to hash_table. Update dependent calls and types.
19349 (location_equiv_class_table): Likewise.
19351 * tree-vect-data-refs.c: Consequential changes for making
19352 peeling a hash_table.
19354 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19355 (destroy_loop_vec_info): Dependent hash_table update.
19357 * tree-vectorizer.h (peeling_htab):
19358 Change type to hash_table. Update dependent calls and types.
19360 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19362 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19363 to check the register content is equal or not.
19364 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19365 to assign register content record.
19366 * lra-eliminations.c (update_reg_eliminate): Use
19367 lra_update_reg_val_offset to update register content offset.
19368 * lra-int.h (struct lra_reg): Add offset member.
19369 (lra_reg_val_equal_p): New static inline function.
19370 (lra_update_reg_val_offset): New static inline function.
19371 (lra_assign_reg_val): New static inline function.
19372 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19373 to assign register content record.
19374 (initialize_lra_reg_info_element): Initial offset to zero.
19376 2013-04-23 Catherine Moore <clm@codesourcery.com>
19378 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19379 operands. Record compression.
19381 2013-04-23 Xinliang David Li <davidxl@google.com>
19383 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19385 2013-04-23 Richard Biener <rguenther@suse.de>
19387 PR middle-end/57036
19388 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19389 parameter, only add abnormal goto edges from the copied body
19390 if the call could perform abnormal gotos.
19391 (copy_cfg_body): Adjust.
19393 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19395 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19397 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19399 * coretypes.h (gimple_stmt_iterator): Add struct to make
19402 2013-04-23 Richard Biener <rguenther@suse.de>
19404 PR tree-optimization/57026
19405 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19406 from SSA names occuring in abnormal PHI nodes.
19408 2013-04-22 Andi Kleen <ak@linux.intel.com>
19410 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19412 2013-04-22 Andi Kleen <ak@linux.intel.com>
19414 * lto/lto.c (print_lto_report_1): Declare early.
19415 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19417 2013-04-22 Andi Kleen <ak@linux.intel.com>
19419 * common.opt (-flto-report-wpa): Add.
19420 * doc/invoke.texi (-flto-report-wpa): Add.
19421 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19424 2013-04-22 Xinliang David Li <davidxl@google.com>
19426 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19427 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19428 * Makefile.in: New dependency
19430 David Daney <ddaney.cavm@gmail.com>
19432 * configure.ac (gcc_cv_as_micromips_support): Use the
19433 --fatal-warnings option.
19434 * configure: Regenerate.
19436 2013-04-22 Marek Polacek <polacek@redhat.com>
19439 * tsan.c (instrument_expr): Don't instrument expression
19440 in case its size is zero.
19442 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19446 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19448 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19450 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19452 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19453 (gimple_stmt_iterator): New typedef.
19454 * gimple.h (gimple_stmt_iterator): Rename to...
19455 (gimple_stmt_iterator_d): ... This.
19456 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19457 trees be valid for GIMPLE and GENERIC.
19458 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19459 * gimple-fold.c (gimple_fold_call): Call target hook
19460 gimple_fold_builtin.
19461 * hooks.c (hook_bool_gsiptr_false): New.
19462 * hooks.h (hook_bool_gsiptr_false): New.
19463 * target.def (fold_stmt): New.
19464 * doc/tm.texi: Regenerate.
19466 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19469 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19470 a set sp if no stack realignment.
19472 2013-04-22 Nick Clifton <nickc@redhat.com>
19474 * config.gcc (tilegx-linux): Extend extra_objs rather than
19476 (tilepro-linux): Likewise.
19478 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19480 * config/aarch64/aarch64-builtins.c
19482 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19483 (VAR<1-12>): Add MAP parameter.
19484 (BUILTIN_*): Likewise.
19485 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19486 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19487 (aarch64_ushl_n<mode>): Likewise.
19488 (aarch64_sshr_n<mode>): Likewise.
19489 (aarch64_ushr_n<mode>): Likewise.
19490 (aarch64_<maxmin><mode>): Likewise.
19491 (aarch64_sqrt<mode>): Likewise.
19492 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19493 (vshr<q>_n_*): Likewise.
19495 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19497 * config/aarch64/aarch64-builtins.c
19498 (aarch64_simd_builtin_type_mode): Handle SF types.
19500 (BUILTIN_GPF): Define.
19501 (aarch64_init_simd_builtins): Handle SF types.
19502 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19503 (frecps): Likewise.
19504 (frecpx): Likewise.
19505 * config/aarch64/aarch64-simd.md
19506 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19507 (aarch64_frecpe<mode>): New.
19508 (aarch64_frecps<mode>): Likewise.
19509 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19510 (v8type): Add frecp<esx>.
19511 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19512 (aarch64_frecps<mode>): Likewise.
19513 * config/aarch64/iterators.md (FRECP): New.
19514 (frecp_suffix): Likewise.
19515 * config/aarch64/arm_neon.h
19516 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19518 2013-04-22 Christian Bruel <christian.bruel@st.com>
19521 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19522 (REG_CLASS_NAMES): Idem.
19523 (REG_CLASS_CONTENTS): Idem.
19524 (REGCLASS_HAS_FP_REG): Idem.
19525 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19526 (sh_conditional_register_usage): Idem.
19528 2013-04-21 Jeff Law <law@redhat.com>
19530 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19531 (ssa_forward_propagate_and_combine): Use it.
19533 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19535 * lra.c: Update the flow chart diagram.
19537 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19539 PR rtl-optimization/56847
19540 * lra-constraints.c (process_alt_operands): Discourage alternative
19541 with non-matche doffsettable memory constraint fro memory with
19544 2013-04-19 Richard Biener <rguenther@suse.de>
19546 PR tree-optimization/56982
19547 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19549 * gimplify.c (gimplify_call_expr): Notice special calls.
19550 (gimplify_modify_expr): Likewise.
19551 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19552 abnormal control flow receivers.
19553 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19554 in the same way as cfun->has_nonlocal_labels.
19555 (gimple_purge_dead_abnormal_call_edges): Likewise.
19556 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19557 receivers start a basic-block.
19559 2013-04-19 Richard Biener <rguenther@suse.de>
19561 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19563 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19564 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19565 (SLP_TREE_LOAD_PERMUTATION): Add.
19566 (vect_transform_slp_perm_load): Adjust prototype.
19567 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19568 (vect_free_slp_instance): Likewise.
19569 (vect_create_new_slp_node): Likewise.
19570 (vect_supported_slp_permutation_p): Remove.
19571 (vect_slp_rearrange_stmts): Adjust.
19572 (vect_supported_load_permutation_p): Likewise. Inline
19573 vect_supported_slp_permutation_p here.
19574 (vect_analyze_slp_instance): Compute load permutations per
19575 slp node instead of per instance.
19576 (vect_get_slp_defs): Adjust.
19577 (vect_transform_slp_perm_load): Likewise.
19578 (vect_schedule_slp_instance): Remove redundant code.
19579 (vect_schedule_slp): Remove hack for PR56270, add it ...
19580 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19581 CSE loads for SLP. Adjust.
19583 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19585 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19586 spelling in two comments.
19588 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19591 * config/arm/arm.c (load_multiple_sequence): Require SP
19592 as base register for loads if SP is in the register list.
19594 2013-04-19 Martin Jambor <mjambor@suse.cz>
19596 PR tree-optimization/56718
19597 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19598 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19599 and made public. Adjusted all callers.
19600 (ipa_intraprocedural_devirtualization): New function.
19601 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19602 (ipa_intraprocedural_devirtualization): Likewise.
19603 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19605 2013-04-19 Richard Biener <rguenther@suse.de>
19607 PR tree-optimization/57000
19608 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19610 2013-04-19 Terry Guo <terry.guo@arm.com>
19612 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19614 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19615 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19616 (cortex_m4_fmacs): Use new reservations.
19617 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19619 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19621 PR rtl-optimization/56999
19622 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19624 (lra_coalesce): Remove split_origin_bitmap and related code.
19625 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19626 ranges if necessary.
19628 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19630 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19632 (ix86_expand_call): Remove clobbered_registers array and use
19633 x86_64_ms_sysv_extra_clobbered_registers instead.
19634 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19636 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19638 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19639 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19640 (*call_value_rex64_ms_sysv): Ditto.
19642 2013-04-18 Cary Coutant <ccoutant@google.com>
19644 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19645 parent instead of die_mark.
19647 2013-04-18 Diego Novillo <dnovillo@google.com>
19649 * gimple.c (create_gimple_tmp): New.
19650 (get_expr_type): New.
19651 (build_assign): New.
19652 (build_type_cast): New.
19653 * gimple.h (enum ssa_mode): Define.
19654 (gimple_seq_set_location): New.
19655 * asan.c (build_check_stmt): Change some gimple_build_* calls
19656 to use build_assign and build_type_cast.
19658 2013-04-18 Richard Biener <rguenther@suse.de>
19660 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19661 handle negative step. Remove redundant checks.
19662 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19663 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19664 for negative step and grouped loads fail to vectorize.
19666 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19668 * emit-rtl.c (reset_insn_used_flags): New function.
19669 (reset_all_used_flags): Use it.
19670 (verify_insn_sharing): New function.
19671 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19673 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19675 PR tree-optimization/56984
19676 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19677 and (x >> M) >= N don't register any assertion if N << M is the
19680 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19682 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19683 do not delete_insn insns that have not yet been emitted, only
19684 unlink them with remove_insn.
19685 * df-scan.c (df_insn_delete): Revert r197492.
19687 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19689 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19690 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19692 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19694 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19695 define_insn_and_split.
19696 (and_scc,ior_scc,negscc): Likewise.
19697 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19699 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19701 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19702 can benefit from using a sequence of LDRD instructions in epilogue
19703 instead of a single LDM instruction.
19705 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19708 * doc/extend.texi: Fix typo.
19710 2013-04-17 Richard Biener <rguenther@suse.de>
19712 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19713 (vect_build_slp_tree): ... here.
19714 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19715 match. Remove special-casing of mismatched complex loads.
19716 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19717 re-try the match with swapped commutative operands.
19718 (vect_supported_load_permutation_p): Remove special-casing of
19719 mismatched complex loads.
19720 (vect_analyze_slp_instance): Adjust.
19722 2013-04-17 Richard Biener <rguenther@suse.de>
19724 PR rtl-optimization/56921
19725 * cfgloop.h (struct loop): Add simple_loop_desc member.
19726 (struct niter_desc): Mark with GTY(()).
19727 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19728 * loop-iv.c (get_simple_loop_desc): Likewise.
19729 (free_simple_loop_desc): Likewise.
19732 2013-04-16 Richard Biener <rguenther@suse.de>
19734 PR rtl-optimization/56921
19735 * loop-init.c (pass_rtl_move_loop_invariants): Add
19736 TODO_do_not_ggc_collect to todo_flags_finish.
19737 (pass_rtl_unswitch): Same.
19738 (pass_rtl_unroll_and_peel_loops): Same.
19739 (pass_rtl_doloop): Same.
19741 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19743 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19744 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19745 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19747 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19748 * tree-streamer.c (record_common_node): Adjust reference in comment.
19750 2013-04-17 Terry Guo <terry.guo@arm.com>
19752 * config/arm/cortex-m4.md: Add a new bypass.
19754 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19756 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19758 (*subs_<optab><mode>_multp2): New pattern.
19759 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19760 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19762 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19764 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19765 (*subs_mul_imm_<mode>): New pattern.
19767 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19770 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19771 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19772 (vsx_movti_32bit): Change j->wa to O->wa.
19774 2013-04-16 Richard Biener <rguenther@suse.de>
19776 PR rtl-optimization/56921
19777 * loop-init.c (pass_rtl_move_loop_invariants): Add
19778 TODO_do_not_ggc_collect to todo_flags_finish.
19779 (pass_rtl_unswitch): Same.
19780 (pass_rtl_unroll_and_peel_loops): Same.
19781 (pass_rtl_doloop): Same.
19783 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19785 * config/arm/arm.c (emit_multi_reg_push): New declaration
19786 for an existing function.
19787 (arm_emit_strd_push): New function.
19788 (arm_expand_prologue): Used here.
19789 (arm_emit_ldrd_pop): New function.
19790 (arm_expand_epilogue): Used here.
19791 (arm_get_frame_offsets): Update condition.
19792 (arm_emit_multi_reg_pop): Add a special case for load of a single
19793 register with writeback.
19795 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19797 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19800 2013-04-16 Richard Biener <rguenther@suse.de>
19802 PR tree-optimization/56756
19803 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19804 (first_mem_ref_loc): New.
19805 (execute_sm): Place the load temporarily before a previous
19806 access instead of in the latch edge to ensure its SSA dependencies
19807 are defined at points dominating the load.
19809 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19811 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19812 correct fix by moving header and footer insn to the footer of
19813 the merged basic block. Clear BB_END of the merged-away block.
19815 PR middle-end/43631
19816 * emit-rtl.c (make_note_raw): New function.
19817 (link_insn_into_chain): New static inline function.
19818 (add_insn): Use it.
19819 (add_insn_before, add_insn_after): Factor insn chain linking code...
19820 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19821 using link_insn_into_chain.
19822 (note_outside_basic_block_p): New helper function for emit_note_after
19823 and emit_note_before.
19824 (emit_note_after): Use nobb variant of add_insn_after if the note
19825 should not be contained in a basic block.
19826 (emit_note_before): Use nobb variant of add_insn_before if the note
19827 should not be contained in a basic block.
19828 (emit_note_copy): Use make_note_raw.
19829 (emit_note): Likewise.
19830 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19831 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19832 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19833 the moved barrier the tail of the basic block it follows.
19834 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19836 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19838 PR tree-optimization/56962
19839 * gimple-ssa-strength-reduction.c (record_increment): Only set
19840 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19841 either rhs1 or rhs2 is equal to c->base_expr.
19843 2013-04-15 Richard Biener <rguenther@suse.de>
19845 PR tree-optimization/56933
19846 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19848 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19849 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19850 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19851 dependence check ...
19852 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19854 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19855 GROUP_READ_WRITE_DEPENDENCE.
19857 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19859 * emit-rtl.c (reset_all_used_flags): New function.
19860 (verify_rtl_sharing): Call reset_all_used_flags before and after
19861 performing the checks.
19863 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19865 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19866 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19867 * config/arm/constraints.md (De): New constraint.
19868 * config/arm/neon.md (anddi3_neon): Delete.
19869 (neon_vand<mode>): Expand to standard anddi3 pattern.
19870 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19871 Move earlier in the file.
19872 (neon_inv_logic_op2): Likewise.
19873 (arm_anddi_operand_neon): New predicate.
19875 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19877 * configure.ac (gcc_cv_ld_as_needed): Set
19878 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19879 Use -z ignore, -z record on *-*-solaris2*.
19880 (HAVE_LD_AS_NEEDED): Update comment.
19881 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19882 * configure: Regenerate.
19883 * config.in: Regenerate.
19884 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19885 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19886 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19887 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19888 equivalents. Fix markup.
19889 * doc/tm.texi: Regenerate.
19891 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19893 * config/i386/i386.opt: New option mstack-protector-guard=.
19894 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19895 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19896 TARGET_SSP_TLS_GUARD.
19897 * config/i386/i386.c (ix86_option_override_internal): Set
19898 ix86_stack_protector_guard.
19899 * config/i386/i386.md (stack_protect_set): Enable for
19900 TARGET_SSP_TLS_GUARD only.
19901 (stack_protect_set_<mode>): Ditto.
19902 (stack_protect_test): Ditto.
19903 (stack_protect_test_<mode>): Ditto.
19904 * doc/invoke.texi (i386 Option): Document.
19906 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19909 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19910 (S_MODES): Set H_MODE bit.
19911 (SF_MODES): Set only S_MODE and SF_MODE bits.
19912 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19913 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19914 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19915 <MODE_FLOAT>: Likewise.
19917 2013-04-15 Joey Ye <joey.ye@arm.com>
19919 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19921 2013-04-15 Joey Ye <joey.ye@arm.com>
19923 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19925 (thumb_far_jump_used_p): Count instruction size and set
19928 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19930 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19931 * resource.c (mark_target_live_regs): Reformat conditional block.
19933 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19935 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19936 notes, they are emitted only just before final.
19937 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19939 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19941 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19942 * cfgrtl.c (delete_insn): Call it here instead.
19943 * lra-spills.c (lra_final_code_change): Use delete_insn.
19944 * haifa-sched.c (sched_remove_insn): Likewise.
19945 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19946 returning to the nop pool.
19947 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19948 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19950 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19952 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19953 * doc/tm.texi: Regenerated.
19955 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19957 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19960 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19962 * df-core.c (df_find_def): Compare register numbers.
19963 (df_find_use): Likewise.
19965 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19968 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19969 lra_in_progress for return.
19971 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19973 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19974 define_insn into define_insn_and_split and emit movsicc patterns.
19976 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19978 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19980 2013-04-12 Richard Biener <rguenther@suse.de>
19982 * tree-pass.h (TODO_do_not_ggc_collect): New.
19983 * passes.c (execute_one_ipa_transform_pass): Honor
19984 TODO_do_not_ggc_collect.
19985 (execute_one_pass): Likewise.
19988 2013-04-10 Richard Biener <rguenther@suse.de>
19990 * passes.c (init_optimization_passes): Remove reload pass.
19991 * ira.c (do_reload): Merge into ...
19993 (rest_of_handle_reload): Remove.
19994 (pass_reload): Likewise.
19995 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19996 of reload for vzeroupper pass placement.
19998 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20000 PR tree-optimization/56918
20001 PR tree-optimization/56920
20002 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20003 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20004 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20005 use wide_mul_with_sign method.
20007 2013-04-12 Richard Biener <rguenther@suse.de>
20009 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20010 not be considered a gimple constant.
20012 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20014 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20015 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20018 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20019 Jakub Jelinek <jakub@redhat.com>
20021 * opts.c: Include diagnostic-color.h.
20022 (common_handle_option): Handle OPT_fdiagnostics_color_.
20023 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20024 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20025 (diagnostic-color.o): New.
20026 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20027 (diagnostic_color_rule): New enum.
20028 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20029 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20030 the location string.
20031 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20032 either NULL, or color kind.
20033 * diagnostic-color.c: New file.
20034 * diagnostic-color.h: New file.
20035 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20037 * doc/invoke.texi (-fdiagnostics-color): Document.
20038 * pretty-print.h (pp_show_color): Define.
20039 (struct pretty_print_info): Add show_color field.
20040 * diagnostic.c: Include diagnostic-color.h.
20041 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20042 macros. Colorize error:, warning: etc. strings and also the location
20044 (diagnostic_show_locus): Colorize the caret line.
20045 * pretty-print.c: Include diagnostic-color.h.
20046 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20047 inside of %< %> quotes or quoted through q format modifier.
20049 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20051 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20053 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20055 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20056 code in CC_NZ mode.
20057 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20060 2013-04-11 Marek Polacek <polacek@redhat.com>
20062 PR tree-optimization/48184
20063 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20065 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20067 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20068 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20069 (skip_simple_arithmetic): Tidy up.
20070 * tree.h (skip_simple_constant_arithmetic): Declare.
20072 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20074 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20076 2013-04-11 Richard Biener <rguenther@suse.de>
20078 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20079 generate vector constants.
20081 2013-04-11 Richard Biener <rguenther@suse.de>
20083 PR tree-optimization/56878
20084 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20085 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20087 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20088 Prefer to align the DR with the most invariant base address.
20090 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20092 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20095 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20097 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20098 floating-point vector comparisons against 0.
20100 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20102 PR tree-optimization/56899
20103 * fold-const.c (extract_muldiv_1): Apply distributive law
20104 only if TYPE_OVERFLOW_WRAPS (ctype).
20106 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20109 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20110 loading has side effect.
20112 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20114 * config/sparc/sparc.c: Include tree-pass.h.
20115 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20116 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20117 head of file. Change return type. Split off gate function.
20118 (sparc_gate_work_around_errata): New function.
20119 (pass_work_around_errata): New pass definition.
20120 (insert_pass_work_around_errata) New pass insert definition to
20121 insert pass_work_around_errata just after delayed-branch scheduling.
20122 (sparc_option_override): Insert the pass.
20123 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20125 2013-04-10 David S. Miller <davem@davemloft.net>
20127 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20128 or -mcpu=hypersparc.
20130 * target.def (cstore_mode): New hook.
20131 * target.h: Include insn-codes.h
20132 * targhooks.c: Likewise.
20133 (default_cstore_mode): New function.
20134 * targhooks.h: Declare it.
20135 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20136 * doc/tm.texi: Rebuild.
20137 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20138 target hook, rather than inspecting the insn_data.
20139 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20140 (TARGET_CSTORE_MODE): Redefine.
20141 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20143 * config/sparc/predicates.md (cstore_result_operand): New special
20145 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20146 Use it for operand 0.
20147 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20148 (*snesi_special): Likewise.
20149 (*snesi_zero): Likewise.
20150 (*seqsi_zero): Likewise.
20151 (*sltu_insn): Likewise.
20152 (*sgeu_insn): Likewise.
20153 (*seqdi_special): Make operand 0 and comparison operation be of
20155 (*snedi_special): Likewise.
20156 (*snedi_special_vis3): Likewise.
20157 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20158 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20159 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20160 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20161 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20162 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20163 (*sltu_extend_sp64): Likewise.
20164 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20165 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20166 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20167 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20168 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20170 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20172 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20173 (aarch64_start_file): Use the new function.
20175 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20176 Jason Merrill <jason@redhat.com>
20178 * common.opt: Add -gdwarf.
20179 * opts.c (common_handle_option): Handle it.
20180 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20182 2013-04-10 Richard Biener <rguenther@suse.de>
20184 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20185 (execute_one_ipa_transform_pass): But unconditionally here.
20186 (execute_one_pass): And here.
20187 (init_optimization_passes): Remove reload pass.
20188 * tree-pass.h (TODO_ggc_collect): Remove.
20189 (pass_reload): Likewise.
20190 * ira.c (do_reload): Merge into ...
20192 (rest_of_handle_reload): Remove.
20193 (pass_reload): Likewise.
20194 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20195 of reload for vzeroupper pass placement.
20196 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20197 and todo_flags_finish of all passes.
20199 2013-04-10 Richard Biener <rguenther@suse.de>
20201 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20202 first_const_oprnd field, rename first_def_type to first_op_type.
20203 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20204 (vect_get_and_check_slp_defs): Always use the type of the
20205 operand. Allow mixed vect_external_def, vect_constant_def types.
20206 (vect_get_constant_vectors): Handle mixed vect_external_def,
20207 vect_constant_def types.
20209 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20211 PR tree-optimization/55524
20212 * tree-ssa-math-opts.c
20213 (convert_mult_to_fma): Don't use an fms construct
20214 when we don't have an fms operation, but fnma, and it looks
20215 likely that we'll be able to use the latter.
20217 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20219 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20221 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20222 inline fail caused by overwritable functions.
20224 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20226 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20227 unnecessary bits in the constant power of two case.
20229 2013-04-10 Richard Biener <rguenther@suse.de>
20231 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20232 broken code swapping operands.
20233 (vect_build_slp_tree): Do not compute load permutations here.
20234 (vect_analyze_slp_instance): Compute load permutations here,
20235 after building the SLP tree.
20237 2013-04-09 Christian Bruel <christian.bruel@st.com>
20239 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20240 of next/prev_real_insn.
20242 2013-04-09 Jan Hubicka <jh@suse.cz>
20244 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20245 Drop aliased parameter.
20246 (function_and_variable_visibility): Do not handle alias pairs.
20247 * cgraph.c (varpool_externally_visible_p): Update prototype.
20248 * varpool.c (varpool_add_new_variable): Update.
20250 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20252 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20254 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20256 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20258 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20260 2013-04-09 Marek Polacek <polacek@redhat.com>
20262 PR tree-optimization/48762
20263 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20265 2013-04-09 Richard Biener <rguenther@suse.de>
20267 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20269 (vect_build_slp_tree): Likewise.
20270 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20271 calculating the cost of a SLP instance.
20272 (vect_analyze_slp_instance): Use it from here, after building
20275 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20277 PR middle-end/56883
20278 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20279 expand_omp_for_static_chunk): Use simple_p = true in
20280 force_gimple_operand_gsi calls when assigning to addressable decls.
20282 2013-04-09 Jeff Law <law@redhat.com>
20284 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20285 when the boolean was created by converting a wider object which
20286 had a boolean range.
20288 2013-04-09 Richard Biener <rguenther@suse.de>
20290 * tree-vectorizer.h (slp_void_p): Remove.
20291 (slp_tree): Typedef before _slp_tree declaration.
20292 (struct _slp_tree): Use a vector of slp_tree as children.
20293 (vect_get_place_in_interleaving_chain): Remove.
20294 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20296 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20298 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20299 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20300 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20301 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20302 Use slp_node instead of slp_void_p and adjust.
20304 2013-04-09 Richard Biener <rguenther@suse.de>
20306 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20307 work that is not necessary.
20309 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20311 PR tree-optimization/56854
20312 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20313 forward into clobber stmts if it would change MEM_REF lhs into
20316 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20318 * tree.c (type_hash_lookup, type_hash_add): Make static.
20319 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20321 2013-04-09 Richard Biener <rguenther@suse.de>
20323 * tree.h (unsave_expr_now): Remove.
20324 * tree-inline.c (mark_local_for_remap_r): Remove.
20325 (unsave_expr_1): Likewise.
20326 (unsave_r): Likewise.
20327 (unsave_expr_now): Likewise.
20328 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20329 (propagate_tree_value): Likewise.
20331 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20333 * doc/rtl.texi (sequence): Rewrite documentation to match the
20334 current use of SEQUENCE rtl objects.
20335 * rtl.def (SEQUENCE): Likewise.
20337 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20338 Update documentation.
20339 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20340 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20342 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20344 2013-04-08 Teresa Johnson <tejohnson@google.com>
20346 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20347 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20349 (estimate_edge_size_and_time): Add comment to suggest using rounding
20351 (estimate_node_size_and_time): Ditto.
20352 (remap_edge_change_prob): Use helper rounding divide methods.
20353 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20354 (gimple_mod_pow2_value_transform): Ditto.
20355 (gimple_mod_subtract_transform): Ditto.
20356 (gimple_ic_transform): Ditto.
20357 (gimple_stringops_transform): Ditto.
20358 * stmt.c (conditional_probability): Ditto.
20359 (emit_case_dispatch_table): Ditto.
20360 * lto-cgraph.c (merge_profile_summaries): Ditto.
20361 * tree-optimize.c (execute_fixup_cfg): Ditto.
20362 * cfgcleanup.c (try_forward_edges): Ditto.
20363 * cfgloopmanip.c (scale_loop_profile): Ditto.
20365 (duplicate_loop_to_header_edge): Ditto.
20366 (lv_adjust_loop_entry_edge): Ditto.
20367 * tree-vect-loop.c (vect_transform_loop): Ditto.
20368 * profile.c (compute_branch_probabilities): Ditto.
20369 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20370 * lto-streamer-in.c (input_cfg): Ditto.
20371 * gimple-streamer-in.c (input_bb): Ditto.
20372 * ipa-cp.c (update_profiling_info): Ditto.
20373 (update_specialized_profile): Ditto.
20374 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20375 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20377 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20378 (compute_trg_info): Ditto.
20379 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20380 (purge_dead_edges): Ditto.
20381 * loop-unswitch.c (unswitch_loop): Ditto.
20382 * cgraphclones.c (cgraph_clone_edge): Ditto.
20383 (cgraph_clone_node): Ditto.
20384 * tree-inline.c (copy_bb): Ditto.
20385 (copy_edges_for_bb): Ditto.
20386 (initialize_cfun): Ditto.
20387 (copy_cfg_body): Ditto.
20388 (expand_call_inline): Ditto.
20390 2013-04-08 Kai Tietz <ktietz@redhat.com>
20392 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20393 TARGET_CYGWIN64 by TARGET_64BIT.
20395 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20397 * config/epiphany/epiphany.md (GPR_1): New constant.
20398 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20399 * config/epiphany/epiphany.c (gen_compare_reg):
20400 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20401 is already in place.
20402 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20403 Don't require being called during rtl expansion; If y operlaps r0,
20405 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20406 (epiphany_expand_epilogue): Likewise.
20408 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20409 Don't use CC_FPmode for ORDERED / UNORDERED.
20410 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20412 * config/epiphany/constraints.md (CnL): New constraint.
20413 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20414 * config/epiphany/predicates.md (add_operand): Allow 1024.
20416 * config/epiphany/epiphany.md (logical_op): New code iterator.
20417 (op_mnc): New code attribute.
20418 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20419 (mov_f+1, mov_f+2): New peephole2 patterns.
20421 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20422 (cstoresi4): Also allow re-use of zero result when doing a NE
20423 comparison to a non-zero operand.
20424 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20426 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20427 Use gen_addsi3_i / gen_subsi3_i.
20429 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20433 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20434 contain anything but clobbers, at most one __builtin_stack_restore,
20435 optionally debug stmts and final resx, and if it has at least one
20436 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20437 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20438 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20439 which isn't defaut definition, remove them.
20440 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20441 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20442 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20443 with MEM_REF LHS with SSA_NAME address.
20445 2013-04-08 Jeff Law <law@redhat.com>
20447 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20449 2013-04-08 Richard Biener <rguenther@suse.de>
20451 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20453 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20454 determined vector type.
20455 (vect_analyze_data_refs): Likewise.
20456 (vect_get_new_vect_var): Adjust.
20457 (vect_create_destination_var): Preserve SSA name versions.
20458 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20459 not dump anything here.
20461 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20463 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20464 Add member lr_slot_known.
20465 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20467 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20468 Remove code that sets lr_slot_offset according to what a previous
20469 version of epiphany_emit_save_restore used to do.
20470 (epiphany_emit_save_restore): When doing an lr save or restore,
20471 set/verify lr_slot_known and lr_slot_offset.
20473 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20476 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20479 2013-04-08 Richard Biener <rguenther@suse.de>
20481 * alias.c (find_base_term): Fix thinko in previous change.
20483 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20485 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20486 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20487 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20488 if possible to compute val.
20489 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20490 For QImode integers don't require anything about precision. Use
20491 const_with_all_bytes_same to find out if the constant doesn't have
20492 repeated bytes in it.
20494 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20496 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20499 2013-04-08 Marek Polacek <polacek@redhat.com>
20501 PR rtl-optimization/48182
20502 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20505 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20508 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20509 nuses, make sure we have a label.
20511 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20514 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20515 (rs6000_emit_swdiv_low_precision): Remove.
20516 (rs6000_emit_swdiv): Rewrite to handle between one and four
20517 iterations of Newton-Raphson generally; modify required number of
20518 iterations for some cases.
20519 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20521 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20523 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20524 set-but-unused variable.
20526 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20527 basic blocks of released function bodies garbage-collectable.
20529 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20530 (struct rtl_opt_pass): Add TODO_df_finish.
20532 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20534 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20536 * config/arm/constraints.md (q): New constraint.
20537 * config/arm/ldrdstrd.md: New file.
20538 * config/arm/arm.md (ldrdstrd.md) New include.
20539 (arm_movdi): Use "q" instead of "r" constraint
20540 for double-word memory access.
20541 (movdf_soft_insn): Likewise.
20542 * config/arm/vfp.md (movdi_vfp): Likewise.
20543 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20544 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20545 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20546 (mem_ok_for_ldrd_strd): Likewise.
20547 (output_move_double): Update assertion.
20549 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20551 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20553 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20555 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20556 define_insn_and_split.
20557 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20559 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20561 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20562 define_insn_and_split.
20563 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20564 (shiftsi3_compare): New pattern.
20565 (rrx): New pattern.
20566 * config/arm/unspecs.md (UNSPEC_RRX): New.
20568 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20570 * config/arm/arm.md (negdi_extendsidi): New pattern.
20571 (negdi_zero_extendsidi): Likewise.
20573 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20575 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20576 define_insn_and_split.
20577 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20578 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20582 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20583 define_insn_and_split.
20584 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20585 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20587 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20589 * config/arm/arm.md (subsi3_carryin): New pattern.
20590 (subsi3_carryin_const): Likewise.
20591 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20592 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20594 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20596 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20598 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20600 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20601 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20603 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20605 * config/arm/arm.c (arm_expand_builtin): Change fcode
20606 type to unsigned int.
20608 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20610 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20612 2013-04-04 Ian Lance Taylor <iant@google.com>
20614 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20617 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20619 PR middle-end/56729
20620 * df-scan.c (df_insn_delete): Disable failing assert.
20622 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20624 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20625 New function prototype.
20626 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20627 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20628 (arm_builtin_vectorized_function): New function.
20630 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20632 * config/arm/arm_neon_builtins.def: New file.
20633 * config/arm/arm.c (neon_builtin_data): Move contents to
20634 arm_neon_builtins.def.
20635 (enum arm_builtins): Include neon builtin definitions.
20636 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20637 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20639 2013-04-04 Marek Polacek <polacek@redhat.com>
20641 PR tree-optimization/48186
20642 * predict.c (maybe_hot_frequency_p): Return false if
20643 HOT_BB_FREQUENCY_FRACTION is 0.
20644 (cgraph_maybe_hot_edge_p): Likewise.
20646 2013-04-04 Richard Biener <rguenther@suse.de>
20648 PR tree-optimization/56826
20649 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20652 2013-04-04 Richard Biener <rguenther@suse.de>
20654 PR tree-optimization/56213
20655 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20656 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20658 2013-04-04 Richard Biener <rguenther@suse.de>
20660 PR tree-optimization/56837
20661 * tree-loop-distribution.c (classify_partition): For non-zero
20662 values require that the value has the same precision as its
20663 mode to be useful as memset value.
20665 2013-04-03 Nick Clifton <nickc@redhat.com>
20667 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20668 (fmssf4): Use fmsf.s on E3V5 architectures.
20669 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20670 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20672 2013-04-03 Jeff Law <law@redhat.com>
20674 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20675 (lra-eliminations.o): Likewise.
20677 2013-04-03 Teresa Johnson <tejohnson@google.com>
20679 * gcov-io.c (compute_working_sets): Moved most of body of old
20680 compute_working_sets here from profile.c.
20681 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20682 (gcov_working_set_t): Moved typedef here from basic-block.h
20683 (compute_working_set): Declare.
20684 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20685 (get_working_sets): Renamed from compute_working_set,
20686 replace most of body with call to new compute_working_sets.
20687 (get_exec_counts): Replace call to compute_working_sets
20688 to get_working_sets.
20689 * profile.h (get_working_sets): Renamed from compute_working_set.
20690 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20691 to get_working_sets.
20692 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20693 * gcov-dump.c (dump_working_sets): New function.
20695 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20697 * hwint.c (sext_hwi, zext_hwi): New functions.
20698 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20699 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20700 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20701 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20702 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20703 (sext_hwi, zext_hwi): New functions.
20705 2013-04-03 Jeff Law <law@redhat.com>
20707 PR tree-optimization/56799
20708 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20709 back test for widening conversion erroneously dropped in prior change.
20711 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20714 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20715 instead of next_real_insn.
20717 2013-04-03 Marek Polacek <polacek@redhat.com>
20720 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20722 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20725 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20727 (thumb1_output_casesi): Likewise.
20728 (thumb2_output_casesi): Likewise.
20730 2013-04-03 Richard Biener <rguenther@suse.de>
20732 PR tree-optimization/56817
20733 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20735 (tree_unroll_loops_completely_1): ... new function to manually
20736 walk the loop tree, properly defering outer loops of unrolled
20737 loops to later iterations.
20739 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20741 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20742 (vectorizable_load): Likewise.
20743 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20744 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20746 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20748 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20751 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20753 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20755 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20757 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20759 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20761 PR tree-optimization/56790
20762 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20765 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20767 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20769 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20770 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20773 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20776 * tree.h (force_folding_builtin_constant_p): New decl.
20777 * builtins.c (force_folding_builtin_constant_p): New variable.
20778 (fold_builtin_constant_p): Fold immediately also if
20779 force_folding_builtin_constant_p.
20781 2013-04-03 Richard Biener <rguenther@suse.de>
20783 PR tree-optimization/56812
20784 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20785 DRs of the same interleaving chain are independent.
20787 2013-04-02 Jason Merrill <jason@redhat.com>
20789 * gdbinit.in (pbb): Use debug fn.
20791 2013-04-02 Lawrence Crowl <crowl@google.com>
20793 * sese.h (struct ivtype_map_elt_s): Remove unused.
20794 (extern debug_ivtype_map): Remove unused.
20795 (extern eq_ivtype_map_elts): Remove unused.
20796 * sese.c (debug_ivtype_map): Removed unused.
20797 (debug_ivtype_map_1): Removed unused.
20798 (debug_ivtype_elt): Remove unused.
20799 (eq_ivtype_map_elts): Remove unused.
20802 2013-04-02 Kai Tietz <ktietz@redhat.com>
20805 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20806 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20807 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20809 (legitimize_pe_coff_symbol): Likewise.
20810 (is_imported_p): New helper-function.
20811 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20812 for Windows x64 targets.
20813 (ix86_expand_prologue): Optimize for pe-coff targets.
20814 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20815 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20816 medium/large code-model.
20817 (legitimize_pic_address): Likewise.
20818 (legitimize_tls_address): Likewise.
20819 (ix86_expand_call): Likewise.
20820 (x86_output_mi_thunk): Likewise.
20821 (get_dllimport_decl): Add new beimport argument.
20822 (construct_plt_address): Don't assert for x64 pe-coff targets.
20823 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20825 (SYMBOL_FLAG_STUBVAR): New macro.
20826 (SYMBOL_REF_STUBVAR_P): Likewise.
20827 * config/i386/winnt.c (stub_list): New structure.
20828 (stub_head): New local variable.
20829 (i386_pe_record_stub): New function.
20830 (i386_pe_file_end): Emit refptr-stubs.
20832 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20834 PR rtl-optimization/56745
20835 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20836 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20839 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20840 and both of them are MEM_REFs, just compare first argument for
20841 equality and attempt to deal even with differing offsets.
20844 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20845 of gimple_clobber_p to be MEM_REF.
20846 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20847 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20848 after gimplification.
20849 * asan.c (get_mem_ref_of_assignment): Don't instrument
20850 gimple_clobber_p stmts.
20851 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20852 gimple_clobber_p stmt if they have MEM_REF lhs and
20853 are dead because of another gimple_clobber_p stmt.
20854 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20855 gimple_clobber_p stmts like debug stmts.
20856 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20857 that refer to unused VAR_DECLs or uninitialized values.
20858 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20859 gimple_clobber_p stmts if they refer to removed parameters.
20860 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20863 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20865 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20866 using SWI48 mode attribute.
20868 2013-04-02 Wei Mi <wmi@google.com>
20870 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20871 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20872 *<rotate_insn><mode>3_mask in i386.md.
20874 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20876 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20878 2013-04-02 Richard Biener <rguenther@suse.de>
20880 PR tree-optimization/56778
20881 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20882 Runtime alias tests are not supported for gather loads.
20883 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20884 stmts referenced from SSA operands before updating SSA form.
20886 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20887 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20889 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20890 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20891 * config/arm/cortex-a53.md: New file.
20892 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20893 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20894 * config/arm/arm.c (arm_issue_rate): Likewise.
20895 * config/arm/arm-tune.md: Regenerate
20896 * config/arm/arm-tables.opt: Regenerate.
20897 * config/arm/arm-cores.def: Add cortex-a53.
20899 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20901 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20904 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20906 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20907 scalar load/store operations using B/H registers.
20908 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20910 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20912 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20914 * config/aarch64/aarch64.c
20915 (aarch64_simd_scalar_immediate_valid_for_move): New.
20916 * config/aarch64/aarch64-protos.h
20917 (aarch64_simd_scalar_immediate_valid_for_move): New.
20918 * config/aarch64/constraints.md (Dh, Dq): New.
20919 * config/aarch64/iterators.md (hq): New.
20921 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20923 * reorg.c (get_branch_condition): Deal with conditional returns.
20924 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20926 2013-04-01 Wei Mi <wmi@google.com>
20928 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20929 Truncate operand 2 using %b asm operand modifier.
20930 (*<shift_insn><mode>3_mask): Ditto.
20931 (*<rotate_insn><mode>3_mask): Ditto.
20933 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20935 PR middle-end/56798
20936 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20938 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20940 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20942 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20944 2013-03-30 Lawrence Crowl <crowl@google.com>
20946 * dse.c (clear_alias_sets): Remove never set.
20947 (disqualified_clear_alias_sets): Remove never set.
20948 (clear_alias_mode_pool): Remove never set.
20949 (dse_step0): Remove condition that is never true.
20950 (canon_address): Remove condition that is never true.
20951 (dse_step7): Remove condition that is never true.
20952 (rest_of_handle_dse): Remove condition that is never true.
20953 (rest_of_handle_dse::did_global): Remove never read from above.
20954 (dse_step2_spill): Remove never called from above.
20955 (dse_step5_spill): Remove never called from above.
20957 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20959 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20960 JUMP_TABLE_DATA changes.
20961 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20962 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20963 (Insns) <jump_table_data>: New entry.
20964 * doc/tm.texi: Regenerate.
20966 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20968 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20969 for table jump at the end of a basic block using tablejump_p.
20970 * targhooks.c (default_invalid_within_doloop): Likewise.
20971 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20972 target hook implementation that is identical to the default hook.
20973 (rs6000_invalid_within_doloop): Remove.
20975 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20976 unused variable from tablejump_p call.
20978 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20979 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20980 (INSN_DELETED_P): Likewise.
20981 (emit_jump_table_data): New prototype.
20982 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20983 after 4th as unused.
20984 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20985 * sched-vis.c (print_insn): Likewise.
20986 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20987 insn for compatibility with back ends that use next_active_insn to
20988 identify jump table data.
20989 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20990 (remove_insn): Likewise.
20991 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20993 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20994 (emit_jump_table_data): New function.
20996 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20997 basic block, a JUMP_TABLE_DATA never is.
20998 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20999 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21000 off from code handling real insns.
21001 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21002 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21003 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21004 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21005 is not a NONDEBUG_INSN_P.
21006 * ira-costs.c (scan_one_insn): Likewise.
21007 * jump.c (mark_all_labels): Likewise.
21008 (mark_jump_label_1): Likewise.
21009 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21010 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21012 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21013 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21014 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21015 JUMP_TABLE_DATA_P insns.
21016 (calculate_elim_costs_all_insns): Likewise.
21017 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21018 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21019 (delete_output_reload): Code style fixups.
21020 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21021 insn flags on this non-insn.
21022 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21023 as scheduling barriers, for pre-change compatibility.
21024 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21025 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21027 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21028 redundant JUMP_TABLE_DATA_P test.
21029 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21030 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21031 (frv_for_each_packet): Likewise.
21032 * config/i386/i386.c (min_insn_size): Likewise.
21033 (ix86_avoid_jump_mispredicts): Likewise.
21034 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21035 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21036 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21037 (mips16_insn_length): Robustify.
21038 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21039 (mips16_split_long_branches): Likewise.
21040 * config/pa/pa.c (pa_combine_instructions): Likewise.
21041 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21042 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21043 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21044 as contributing to pool range lengths.
21045 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21046 Remove redundant JUMP_TABLE_DATA_P test.
21047 (sh_loop_align): Likewise.
21048 (split_branches): Likewise.
21049 (sh_insn_length_adjustment): Likewise.
21050 * config/spu/spu.c (get_branch_target): Likewise.
21052 2013-03-29 Jan Hubicka <jh@suse.cz>
21054 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21055 gcov streaming; stream hot bb threshold to ltrans.
21056 * predict.c (get_hot_bb_threshold): Break out from ....
21057 (maybe_hot_count_p): ... here.
21058 (set_hot_bb_threshold): New function.
21059 * lto-section-in.c (lto_section_name): Add profile.
21060 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21061 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21062 and data-streamer.h
21063 (histogram_entry): New structure.
21064 (histogram, histogram_pool): New global vars.
21065 (histogram_hash): New structure.
21066 (histogram_hash::hash): New method.
21067 (histogram_hash::equal): Likewise.
21068 (account_time_size): New function.
21069 (cmp_counts): New function.
21070 (dump_histogram): New function.
21071 (ipa_profile_generate_summary): New function.
21072 (ipa_profile_write_summary): New function.
21073 (ipa_profile_read_summary): New function.
21074 (ipa_profile): Decide on threshold.
21075 (pass_ipa_profile): Add ipa_profile_write_summary and
21076 ipa_profile_read_summary.
21077 * Makefile.in (ipa.o): Update dependencies.
21078 * lto-streamer.h (LTO_section_ipa_profile): New section.
21080 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21082 * tree.h (VAR_P): New.
21084 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21087 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21089 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21091 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21092 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21093 (control_flow_insn_p): Likewise.
21094 * cfgrtl.c (duplicate_insn_chain): Likewise.
21095 * final.c (get_attr_length_1): Likewise.
21096 (shorten_branches): Likewise.
21097 (final_scan_insn): Likewise.
21098 * function.c (instantiate_virtual_regs): Likewise.
21099 * gcse.c (insert_insn_end_basic_block): Likewise.
21100 * ira-costs.c (scan_one_insn): Likewise.
21101 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21102 * lra.c (check_rtl): Likewise.
21103 * reload1.c (elimination_costs_in_insn): Likewise.
21104 * reorg.c (follow_jumps): Likewise.
21106 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21107 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21108 (thumb_far_jump_used_p): Likewise.
21109 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21110 (workaround_speculation): Likewise.
21111 (add_sched_insns_for_speculation): Likewise.
21112 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21113 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21114 (frv_for_each_packet): Likewise.
21115 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21116 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21117 (final_emit_insn_group_barriers): Likewise.
21118 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21119 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21120 (mips16_insn_length): Likewise.
21121 * config/pa/pa.c (pa_reorg): Likewise.
21122 (pa_combine_instructions): Likewise.
21123 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21124 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21125 (sh_reorg): Likewise.
21126 (split_branches): Likewise.
21127 * config/spu/spu.c (get_branch_target): Likewise.
21129 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21132 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21134 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21135 Fix declaration name.
21137 2013-03-28 Lawrence Crowl <crowl@google.com>
21139 * graphds.h (struct graph.indicies): Remove unused.
21140 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21141 (SCOP_ORIGINAL_PDDRS): Remove unused.
21142 * sese.h (extern insert_loop_close_phis): Removed unused.
21143 (extern insert_guard_phis): Removed unused.
21144 (extern ivtype_map_elt_info): Removed unused.
21145 (new_ivtype_map_elt): Removed unused.
21146 * sese.c (ivtype_map_elt_info): Removed unused.
21148 2013-03-28 Lawrence Crowl <crowl@google.com>
21150 * Makefile.in: Add several missing include dependences.
21152 (test-dump.o): New. This object is not added to any executable,
21153 but is present for ad-hoc testing.
21155 (debug (const bitmap_head_def &)): New.
21156 (debug (const bitmap_head_def *)): New.
21158 (extern debug (const bitmap_head_def &)): New.
21159 (extern debug (const bitmap_head_def *)): New.
21161 (debug (edge_def &)): New.
21162 (debug (edge_def *)): New.
21164 (debug (basic_block_def &)): New.
21165 (debug (basic_block_def *)): New.
21167 (dump_node (const_tree, int, FILE *)): Correct source file.
21169 (debug (die_struct &)): New.
21170 (debug (die_struct *)): New.
21172 (extern debug (die_struct &)): New.
21173 (extern debug (die_struct *)): New.
21174 * gimple-pretty-print.c
21175 (debug (gimple_statement_d &)): New.
21176 (debug (gimple_statement_d *)): New.
21177 * gimple-pretty-print.h
21178 (extern debug (gimple_statement_d &)): New.
21179 (extern debug (gimple_statement_d *)): New.
21181 (debug (ira_allocno_copy &)): New.
21182 (debug (ira_allocno_copy *)): New.
21183 (debug (ira_allocno &)): New.
21184 (debug (ira_allocno *)): New.
21186 (extern debug (ira_allocno_copy &)): New.
21187 (extern debug (ira_allocno_copy *)): New.
21188 (extern debug (ira_allocno &)): New.
21189 (extern debug (ira_allocno *)): New.
21191 (debug (live_range &)): New.
21192 (debug (live_range *)): New.
21194 (debug (lra_live_range &)): New.
21195 (debug (lra_live_range *)): New.
21197 (debug (lra_live_range &)): New.
21198 (debug (lra_live_range *)): New.
21200 (debug (omega_pb_d &)): New.
21201 (debug (omega_pb_d *)): New.
21203 (extern debug (omega_pb_d &)): New.
21204 (extern debug (omega_pb_d *)): New.
21206 (debug (const rtx_def &)): New.
21207 (debug (const rtx_def *)): New.
21209 (debug_tree (tree): Move within file.
21210 (debug_raw (const tree_node &)): New.
21211 (debug_raw (const tree_node *)): New.
21212 (dump_tree_via_hooks (const tree_node *, int)): New.
21213 (debug (const tree_node &)): New.
21214 (debug (const tree_node *)): New.
21215 (debug_verbose (const tree_node &)): New.
21216 (debug_verbose (const tree_node *)): New.
21217 (debug_head (const tree_node &)): New.
21218 (debug_head (const tree_node *)): New.
21219 (debug_body (const tree_node &)): New.
21220 (debug_body (const tree_node *)): New.
21221 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21222 (debug (vec<tree, va_gc> &)): New.
21223 (debug (vec<tree, va_gc> *)): New.
21225 (extern debug (const rtx_def &)): New.
21226 (extern debug (const rtx_def *)): New.
21228 (debug_raw (simple_bitmap_def &)): New.
21229 (debug_raw (simple_bitmap_def *)): New.
21230 (debug (simple_bitmap_def &)): New.
21231 (debug (simple_bitmap_def *)): New.
21233 (extern debug (simple_bitmap_def &)): New.
21234 (extern debug (simple_bitmap_def *)): New.
21235 (extern debug_raw (simple_bitmap_def &)): New.
21236 (extern debug_raw (simple_bitmap_def *)): New.
21238 (debug (vinsn_def &)): New.
21239 (debug (vinsn_def *)): New.
21240 (debug_verbose (vinsn_def &)): New.
21241 (debug_verbose (vinsn_def *)): New.
21242 (debug (expr_def &)): New.
21243 (debug (expr_def *)): New.
21244 (debug_verbose (expr_def &)): New.
21245 (debug_verbose (expr_def *)): New.
21246 (debug (vec<rtx> &)): New.
21247 (debug (vec<rtx> *)): New.
21249 (extern debug (vinsn_def &)): New.
21250 (extern debug (vinsn_def *)): New.
21251 (extern debug_verbose (vinsn_def &)): New.
21252 (extern debug_verbose (vinsn_def *)): New.
21253 (extern debug (expr_def &)): New.
21254 (extern debug (expr_def *)): New.
21255 (extern debug_verbose (expr_def &)): New.
21256 (extern debug_verbose (expr_def *)): New.
21257 (extern debug (vec<rtx> &)): New.
21258 (extern debug (vec<rtx> *)): New.
21260 (_list_iter_cond_expr): Make inline instead of static.
21262 (debug (sreal &)): New.
21263 (debug (sreal *)): New.
21265 (extern debug (sreal &)): New.
21266 (extern debug (sreal *)): New.
21268 (extern debug_raw (const tree_node &)): New.
21269 (extern debug_raw (const tree_node *)): New.
21270 (extern debug (const tree_node &)): New.
21271 (extern debug (const tree_node *)): New.
21272 (extern debug_verbose (const tree_node &)): New.
21273 (extern debug_verbose (const tree_node *)): New.
21274 (extern debug_head (const tree_node &)): New.
21275 (extern debug_head (const tree_node *)): New.
21276 (extern debug_body (const tree_node &)): New.
21277 (extern debug_body (const tree_node *)): New.
21278 (extern debug (vec<tree, va_gc> &)): New.
21279 (extern debug (vec<tree, va_gc> *)): New.
21281 (debug (struct loop &)): New.
21282 (debug (struct loop *)): New.
21283 (debug_verbose (struct loop &)): New.
21284 (debug_verbose (struct loop *)): New.
21285 * tree-dump.c: Add header dependence.
21287 (extern debug (struct loop &)): New.
21288 (extern debug (struct loop *)): New.
21289 (extern debug_verbose (struct loop &)): New.
21290 (extern debug_verbose (struct loop *)): New.
21292 (debug (data_reference &)): New.
21293 (debug (data_reference *)): New.
21294 (debug (vec<data_reference_p> &)): New.
21295 (debug (vec<data_reference_p> *)): New.
21296 (debug (vec<ddr_p> &)): New.
21297 (debug (vec<ddr_p> *)): New.
21299 (extern debug (data_reference &)): New.
21300 (extern debug (data_reference *)): New.
21301 (extern debug (vec<data_reference_p> &)): New.
21302 (extern debug (vec<data_reference_p> *)): New.
21303 (extern debug (vec<ddr_p> &)): New.
21304 (extern debug (vec<ddr_p> *)): New.
21306 (debug (pt_solution &)): New.
21307 (debug (pt_solution *)): New.
21309 (extern debug (pt_solution &)): New.
21310 (extern debug (pt_solution *)): New.
21312 (debug (_var_map &)): New.
21313 (debug (_var_map *)): New.
21314 (debug (tree_live_info_d &)): New.
21315 (debug (tree_live_info_d *)): New.
21317 (extern debug (_var_map &)): New.
21318 (extern debug (_var_map *)): New.
21319 (extern debug (tree_live_info_d &)): New.
21320 (extern debug (tree_live_info_d *)): New.
21322 2013-03-28 Jan Hubicka <jh@suse.cz>
21324 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21326 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21328 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21329 record only when desired or required.
21331 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21333 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21334 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21336 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21338 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21339 (*andsi3_compare0_uxtw): New pattern.
21340 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21341 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21343 2013-03-28 Jan Hubicka <jh@suse.cz>
21345 * data-streamer-in.c (streamer_read_gcov_count): New function.
21346 * gimple-streamer-out.c: Include value-prof.h.
21347 (output_gimple_stmt): Output histogram.
21348 (output_bb): Use streamer_write_gcov_count.
21349 * value-prof.c: Include data-streamer.h
21350 (dump_histogram_value): Add HIST_TYPE_MAX.
21351 (stream_out_histogram_value): New function.
21352 (stream_in_histogram_value): New function.
21353 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21354 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21355 * data-streamer-out.c (streamer_write_gcov_count): New function.
21356 (streamer_write_gcov_count_stream): New function.
21357 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21358 (lto_output_node): Likewise.
21359 (input_node, input_edge): Likewise.
21360 * lto-streamer-out.c (output_cfg): Update streaming.
21361 * lto-streamer-in.c (input_cfg): Likewise.
21362 * data-streamer.h (streamer_write_gcov_count,
21363 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21364 * gimple-streamer-in.c: Include value-prof.h
21365 (input_gimple_stmt): Input histograms.
21366 (input_bb): Update profile streaming.
21368 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21370 * genmodes.c (emit_max_int): New function.
21371 (emit_insn_modes_h): Added call to emit_max_function.
21372 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21374 * machmode.def: Fixed comment.
21376 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21378 * combine.c (try_combine): Removed useless assert.
21379 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21381 2013-03-28 Marek Polacek <polacek@redhat.com>
21382 Richard Biener <rguenther@suse.de>
21384 PR tree-optimization/56695
21385 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21386 build signed result of a vector comparison.
21387 * tree-cfg.c (verify_gimple_comparison): Check that a result
21388 of a vector comparison has signed type.
21390 2013-03-28 Richard Biener <rguenther@suse.de>
21392 PR tree-optimization/37021
21393 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21394 do not restrict gaps between groups.
21395 * tree-vect-stmts.c (vectorizable_load): Properly account for
21396 a gap between groups.
21398 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21400 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21401 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21404 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21406 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21407 * doc/extend.texi (Named Address Spaces): Ditto.
21408 (Variable Attributes): Ditto.
21410 2013-03-27 Kai Tietz <ktietz@redhat.com>
21412 * config.build: Add support for cygwin x64 target.
21413 * config.gcc: Likewise.
21414 * config.host: Likewise.
21415 * configure.ac: Likewise
21416 * configure: Regenerated.
21418 2013-03-27 Kai Tietz <ktietz@redhat.com>
21420 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21421 * config/i386/t-cygwin-w64: New file.
21422 * config/i386/cygwin-w64.h: New file.
21423 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21424 and add support for x64-cygwin target.
21425 (CPP_SPEC): Likewise.
21426 (CXX_WRAP_SPEC_LIST): Undefine before define.
21427 (LIBGCJ_SONAME): Use 15 as version.
21429 2013-03-27 Richard Biener <rguenther@suse.de>
21431 PR tree-optimization/56716
21432 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21433 dumping for ref nodes.
21435 2013-03-27 Martin Jambor <mjambor@suse.cz>
21437 PR tree-optimization/55334
21438 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21439 restricted pointers to arrays.
21441 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21443 * Makefile.in (.SUFFIXES): Add .cc.
21444 (.c.o): Apply same recipe for implicit rule .cc.o.
21446 2013-03-27 Richard Biener <rguenther@suse.de>
21448 PR tree-optimization/37021
21449 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21450 REALPART/IMAGPART_EXPRs around the supported refs.
21451 * tree-ssa-structalias.c (find_func_aliases): Assume that
21452 floating-point values are not used to transfer pointers.
21454 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21456 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21457 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21458 * doc/tm.texi: Regenerate.
21459 * targhooks.h (default_has_ifunc_p): New.
21460 * targhooks.c (default_has_ifunc_p): Ditto.
21461 * config/linux-protos.h: New file.
21462 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21463 hook for linux which disables support of indirect functions in android.
21464 * config/linux-android.c: New file.
21465 * config/t-linux-android.c: Ditto.
21466 * config.gcc: Added new object file linux-android.o.
21467 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21468 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21469 * varasm.c (do_assemble_alias): Likewise.
21470 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21471 doesn't support indirect functions.
21472 * configure: Regenerate.
21474 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21477 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21478 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21481 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21483 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21485 2013-03-27 Terry Guo <terry.guo@arm.com>
21487 * config/arm/arm-cores.def: Added core cortex-r7.
21488 * config/arm/arm-tune.md: Regenerated.
21489 * config/arm/arm-tables.opt: Regenerated.
21490 * doc/invoke.texi: Added entry for core cortex-r7.
21492 2013-03-27 Walter Lee <walt@tilera.com>
21494 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21495 double-decrement of next_scratch_regno.
21497 2013-03-27 Walter Lee <walt@tilera.com>
21499 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21501 (insn_v1mulus): Ditto.
21502 (insn_v2muls): Ditto.
21504 2013-03-27 Walter Lee <walt@tilera.com>
21506 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21507 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21509 2013-03-27 Walter Lee <walt@tilera.com>
21511 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21512 (*sibcall_value): Ditto.
21514 2013-03-27 Walter Lee <walt@tilera.com>
21516 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21517 (insn_mnz_v8qi): ... this ...
21518 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21520 (insn_v<n>mnz): Replaced by ...
21521 (insn_v1mnz): ... this ...
21522 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21524 (insn_mz_<mode>): Replaced by ...
21525 (insn_mz_v8qi): ... this ...
21526 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21528 (insn_v<n>mz): Replaced by ...
21529 (insn_v1mz): ... this ...
21530 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21533 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21535 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21537 2013-03-26 Roland McGrath <mcgrathr@google.com>
21539 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21540 than fprintf with a non-constant, non-format string.
21542 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21544 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21545 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21546 operand 0 predicate.
21547 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21548 attribute. Use general_x64nomem_operand as operand 1 predicate.
21549 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21550 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21551 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21552 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21553 (mov<mode>_insv_1): Remove expander. Merge insn with
21554 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21555 Use general_x64nomem_operand as operand 1 predicate.
21556 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21557 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21558 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21559 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21560 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21561 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21562 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21563 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21564 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21565 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21566 (general_x64nomem_operand): Ditto.
21568 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21570 * config/rtems.opt: Add -pthread option.
21572 2013-03-26 Richard Biener <rguenther@suse.de>
21574 * alias.c (find_base_term): Avoid redundant and not used recursion.
21575 (base_alias_check): Get the initial base term from the caller.
21576 (true_dependence_1): Compute and pass base terms to base_alias_check.
21577 (write_dependence_p): Likewise.
21578 (may_alias_p): Likewise.
21580 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21582 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21583 PC-relative load in SI modes and above only.
21585 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21587 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21588 * config/arm/iwmmxt.md (WCGR0): Update.
21589 (WCGR1, WCGR2, WCGR3): Likewise.
21591 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21593 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21594 Use x64 and nox64 isa attributes.
21596 2013-03-26 Richard Biener <rguenther@suse.de>
21598 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21599 alignment computations and rely on get_object_alignment_1
21600 for the !TYPE_P case.
21601 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21603 2013-03-26 Walter Lee <walt@tilera.com>
21605 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21606 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21608 2013-03-25 Jeff Law <law@redhat.com>
21610 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21611 check for INTEGRAL_TYPE_P that was missing due to checking in
21612 wrong version of prior patch.
21614 2013-03-25 Walter Lee <walt@tilera.com>
21616 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21617 TILEGX_INSN_SHUFFLEBYTES1.
21618 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21620 (tilegx_builtins): Ditto.
21621 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21623 2013-03-25 Walter Lee <walt@tilera.com>
21625 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21626 (floatunssisf2): New pattern.
21627 (floatsidf2): New pattern.
21628 (floatunssidf2): New pattern.
21630 2013-03-25 Walter Lee <walt@tilera.com>
21632 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21633 tests for constraint J, K, N, P.
21635 2013-03-25 Walter Lee <walt@tilera.com>
21637 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21638 Use indirect/pcrel encoding.
21639 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21642 2013-03-25 Steve Ellcey <sellcey@mips.com>
21644 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21645 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21646 * config/mips/mips.c (mips_option_override): Set IMADD default.
21647 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21648 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21649 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21650 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21651 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21653 2013-03-25 Jeff Law <law@redhat.com>
21655 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21656 slightly to avoid creating and folding useless trees. Simplify
21657 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21659 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21661 * config/i386/i386.md (*zero_extendsidi2): Merge with
21662 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21663 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21664 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21665 targets. Clarify comment.
21667 2013-03-25 Martin Jambor <mjambor@suse.cz>
21669 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21670 pass-through jump functions differently.
21671 (ipa_read_jump_function): Likewise. Also use setter functions to set
21674 2013-03-25 Martin Jambor <mjambor@suse.cz>
21676 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21677 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21679 (ipa_get_indirect_edge_target): New function.
21680 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21681 ipa_get_indirect_edge_target. Update all callers.
21682 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21683 ipa_get_indirect_edge_target_1 instead of calling
21684 ipa_get_indirect_edge_target.
21685 (create_specialized_node): Pass aggvlas to
21686 ipcp_discover_new_direct_edges.
21688 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21690 * config/arm/arm.md (f_sels, f_seld): New types.
21691 (*cmov<mode>): New pattern.
21692 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21695 2013-03-25 Kai Tietz <ktietz@redhat.com>
21697 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21698 POSIX-printf for mingw-hosted builds.
21700 2013-03-25 Richard Biener <rguenther@suse.de>
21702 PR middle-end/56694
21703 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21704 must-not-throw stmt location.
21706 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21708 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21709 Emit load-acquire versions when acq is true.
21710 (arm_emit_store_exclusive): Add rel parameter.
21711 Emit store-release versions when rel is true.
21712 (arm_split_compare_and_swap): Use acquire-release instructions
21714 of barriers when appropriate.
21715 (arm_split_atomic_op): Likewise.
21716 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21717 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21718 (VUNSPEC_SLX): Likewise.
21719 (VUNSPEC_LDA): Likewise.
21720 (VUNSPEC_STL): Likewise.
21721 * config/arm/sync.md (atomic_load<mode>): New pattern.
21722 (atomic_store<mode>): Likewise.
21723 (arm_load_acquire_exclusive<mode>): Likewise.
21724 (arm_load_acquire_exclusivesi): Likewise.
21725 (arm_load_acquire_exclusivedi): Likewise.
21726 (arm_store_release_exclusive<mode>): Likewise.
21728 2013-03-25 Catherine Moore <clm@codesourcery.com>
21730 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21731 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21732 * config/mip/predicates.md (lwsp_swsp_operand,
21733 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21734 sb16_operand, db4_operand, db7_operand, ib3_operand,
21735 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21736 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21737 andi16_operand): New predicates.
21738 * config/mips/mips.md (compression): New attribute.
21739 (enabled): New attribute.
21740 (length): Consider compression in computing length.
21741 (shift_compression): New code attribute.
21742 (*add<mode>3): New operands. Record compression.
21743 (sub<mode>3): Likewise.
21744 (one_cmpl<mode>2): Likewise.
21745 (*and<mode>3): Likewise.
21746 (*ior<mode>3): Likewise.
21747 (unnamed pattern for xor): Likewise.
21748 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21749 (*<optab><mode>3): Likewise.
21750 (*mov<mode>_internal: Likewise.
21751 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21752 (mips_unsigned_immediate_p): New.
21753 (umips_lwsp_swsp_address_p): New.
21754 (m16_based_address_p): New.
21755 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21756 (mips_unsigned_immediate_p): New prototype.
21757 (lwsp_swsp_address_p): New prototype.
21758 (m16_based_address_p): New prototype.
21759 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21760 (mips_signed_immediate_p): New function.
21761 (m16_based_address_p): New function.
21762 (lwsp_swsp_address_p): New function.
21763 (mips_print_operand_punctuation): Recognize short delay slot insns
21764 for microMIPS.add<mode>3"
21766 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21769 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21770 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21772 2013-03-25 Richard Biener <rguenther@suse.de>
21774 PR tree-optimization/56689
21775 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21778 2013-03-25 Richard Biener <rguenther@suse.de>
21780 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21782 (memory_references): Likewise.
21783 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21784 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21785 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21786 (gather_mem_refs_in_loops): Fold into ...
21787 (analyze_memory_references): ... this. Move initialization
21788 to tree_ssa_lim_initialize.
21789 (fill_always_executed_in): Rename to ...
21790 (fill_always_executed_in_1): ... this.
21791 (fill_always_executed_in): Move contains_call computation to
21792 this new function from ...
21793 (tree_ssa_lim_initialize): ... here.
21794 (tree_ssa_lim): Call fill_always_executed_in.
21796 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21798 * postreload.c (reload_combine): Fix code detecting returns.
21800 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21802 * function.c (emit_use_return_register_into_block): On cc0 targets,
21803 do not emit the sequence between cc0 setter and user.
21805 2013-03-25 Kai Tietz <ktietz@redhat.com>
21807 * config/i386/predicates.md (local_symbolic_operand): Interpret
21808 dll-imported symbols as none-local.
21810 2013-03-25 Richard Biener <rguenther@suse.de>
21812 * tree-ssa-loop-im.c (struct depend): Remove.
21813 (struct lim_aux_data): Make depends a vec of gimples.
21814 (free_lim_aux_data): Adjust.
21815 (add_dependency): Likewise.
21816 (set_level): Likewise.
21818 2013-03-25 Richard Biener <rguenther@suse.de>
21820 PR middle-end/56434
21821 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21822 the pointer returned by calls with ECF_MALLOC set.
21824 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21826 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21828 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21830 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21831 using MMXMODE mode iterator.
21832 (*move<mode>_internal): Merge with *movv2sf_internal and
21833 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21835 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21837 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21838 (record_last_mem_set_info): Likewise.
21840 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21841 of XNEWVEC followed by memset.
21842 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21844 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21846 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21847 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21848 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21849 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21850 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21851 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21852 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21853 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21854 BARRIER_P instead of GET_CODE.
21856 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21858 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21859 inaccuracy in the probing code.
21861 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21862 (ctrapdi4): Likewise.
21864 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21866 * calls.c (expand_call): Add missing guard to code handling return
21867 of non-BLKmode structures in MSB.
21868 * function.c (expand_function_end): Likewise.
21870 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21872 * combine.c (try_combine): Adjust comment. Do not add the set of
21873 insn #0 if the destination indirectly is set or dies in insn #2.
21874 Tidy up code to distribute a new note.
21876 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21878 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21879 also for alternatives 16 and 17.
21881 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21883 * config/i386/sse.md (*mov<mode>_internal): Merge with
21884 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21885 Emit insn template depending on type attribute. Use
21886 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21887 movd instead of movq mnemonic for interunit moves. Rewrite mode
21888 attribute calculation. Remove unit attribute calculation.
21889 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21890 Set prefix_data16 attribute for DImode ssemov types.
21891 Use Ym instead of y for SSE-MMX conversion alternatives.
21892 Reorder operand constraints.
21894 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21896 * df.h (df_insn_delete): Adjust prototype.
21897 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21898 and let it decide whether mark the basic block dirty.
21899 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21900 * df-scan.c (df_insn_info_delete): New helper function, split
21901 off from df_insn_delete.
21902 (df_scan_free_bb_info): Use it.
21903 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21905 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21906 that the insn is actually an insn and it has a non-NULL basic block.
21907 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21909 2013-03-22 Richard Biener <rguenther@suse.de>
21911 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21913 (mem_ref_alloc): Do not allocate them.
21914 (refs_independent_p): Do not query or maintain a cache.
21916 2013-03-22 Richard Biener <rguenther@suse.de>
21918 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21919 (gather_mem_refs_in_loops): Do not compute it.
21920 (analyze_memory_references): Do not allocate it.
21921 (tree_ssa_lim_finalize): Do not free it.
21922 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21924 2013-03-22 Richard Biener <rguenther@suse.de>
21926 * is-a.h (as_a): Use gcc_checking_assert.
21928 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21930 * config/aarch64/aarch64.c (aarch64_print_operand): New
21931 format specifier for printing a constant in hex.
21932 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21933 format specifier for printing second operand.
21935 2013-03-22 Richard Biener <rguenther@suse.de>
21937 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21939 (gather_mem_refs_in_loops): Perform store accumulation here.
21940 (create_vop_ref_mapping_loop): Remove.
21941 (create_vop_ref_mapping): Likewise.
21942 (analyze_memory_references): Initialize refs_stored_in_loop.
21943 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21944 (record_indep_loop): Remove.
21945 (record_dep_loop): New function.
21946 (ref_indep_loop_p_1): Adjust to only walk over references
21947 in the loop, not its subloops.
21948 (ref_indep_loop_p): Rename to ...
21949 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21950 maintaining a more fine-grained cache.
21951 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21952 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21954 2013-03-22 Richard Biener <rguenther@suse.de>
21956 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21957 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21958 aggregate mem_ref_loc.
21959 (free_mem_ref_locs): Inline into ...
21960 (memref_free): ... this and adjust.
21961 (mem_ref_alloc): Adjust.
21962 (mem_ref_locs_alloc): Remove.
21963 (record_mem_ref_loc): Adjust.
21964 (get_all_locs_in_loop): Rewrite into ...
21965 (for_all_locs_in_loop): ... this iterator.
21966 (rewrite_mem_ref_loc): New functor.
21967 (rewrite_mem_refs): Use for_all_locs_in_loop.
21968 (sm_set_flag_if_changed): New functor.
21969 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21970 (ref_always_accessed): New functor.
21971 (ref_always_accessed_p): Use for_all_locs_in_loop.
21973 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21975 * tree-pass.h (PROP_gimple_lvec): New.
21976 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21977 (init_optimization_passes): Move pass_lower_vector.
21978 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21980 (pass_lower_vector): Provide PROP_gimple_lvec.
21981 (pass_lower_vector_ssa): Likewise.
21982 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21984 2013-03-21 Mark Wielaard <mjw@redhat.com>
21986 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21988 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21990 * config/i386/i386.md (*movdi_internal): Disparage slightly
21991 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21992 conversion alternatives.
21994 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21996 PR middle-end/48087
21997 * diagnostic.def (DK_WERROR): New kind.
21998 * diagnostic.h (werrorcount): Define.
21999 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22000 promoted to DK_ERROR, increment DK_WERROR counter instead of
22002 * toplev.c (toplev_main): Call print_ignored_options even if
22003 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22004 even if just werrorcount is non-zero.
22007 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22009 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22010 (string_cst_pool_decl): New function.
22011 (optimize_one_addr_into_implicit_ptr): New function.
22012 (resolve_addr_in_expr): Optimize DWARF location expression
22013 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22014 which doesn't live in memory, but has DW_AT_location or
22015 DW_AT_const_value, or refers to a string literal, into
22016 DW_OP_GNU_implicit_pointer.
22017 (optimize_location_into_implicit_ptr): New function.
22018 (resolve_addr): If removing DW_AT_location of a variable because
22019 it was DW_OP_addr of address of the variable, but the variable doesn't
22020 live in memory, try to emit const value attribute for the initializer.
22022 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22024 * tree.h (VECTOR_TYPE_P): New macro.
22025 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22026 TYPE_MODE): Use it.
22027 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22028 VEC_COND_EXPR cannot be lvalues.
22029 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22031 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22033 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22034 Restrict the transformation to equal modes.
22036 2013-03-21 Richard Biener <rguenther@suse.de>
22038 PR tree-optimization/39326
22039 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22040 (MEM_ANALYZABLE): Adjust.
22041 (record_mem_ref_loc): Move bitmap ops ...
22042 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22043 unanalyzable refs, do not record locations for it.
22044 (analyze_memory_references): Allocate ref zero as shared
22046 (refs_independent_p): Do not test for unanalyzed mems here.
22047 (ref_indep_loop_p_1): Special-case disambiguation against
22048 the unanalyzed ref.
22049 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22051 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22053 * config/arm/arm-protos.h (tune_params): Add
22054 prefer_neon_for_64bits field.
22055 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22056 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22057 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22058 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22059 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22060 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22061 (arm_option_override): Handle -mneon-for-64bits new option.
22062 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22063 (prefer_neon_for_64bits): Declare new variable.
22064 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22065 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22066 (arch_enabled): Handle new arch types. Remove support for onlya8
22068 (one_cmpldi2): Use new arch names.
22069 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22070 * config/arm/arm.opt (mneon-for-64bits): Add option.
22071 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22072 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22073 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22075 * doc/invoke.texi (-mneon-for-64bits): Document.
22077 2013-03-21 Richard Biener <rguenther@suse.de>
22079 PR tree-optimization/39326
22080 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22081 (sort_bbs_in_loop_postorder_cmp): New function.
22082 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22084 2013-03-21 Richard Biener <rguenther@suse.de>
22086 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22087 (vect_insert_into_interleaving_chain): Likewise.
22088 (vect_drs_dependent_in_basic_block): Inline ...
22089 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22091 (vect_analyze_data_ref_dependence): ... here. Simplify.
22092 (vect_check_interleaving): Simplify.
22093 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22094 (vect_slp_analyze_data_ref_dependences): ... this new function.
22095 (dr_group_sort_cmp): New function.
22096 (vect_analyze_data_ref_accesses): Compute data-reference groups
22097 here instead of in vect_analyze_data_ref_dependence. Use
22098 a more efficient algorithm.
22099 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22100 vect_slp_analyze_data_ref_dependences. Call
22101 vect_analyze_data_ref_accesses earlier.
22102 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22103 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22104 (vect_slp_analyze_data_ref_dependences): New prototype.
22106 2013-03-21 Richard Biener <rguenther@suse.de>
22108 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22109 ref is stored in the loop.
22110 (find_refs_for_sm): Walk only over all stores.
22111 (store_motion_loop): Allocate from lim_bitmap_obstack.
22112 (store_motion): Likewise.
22114 2013-03-21 Richard Biener <rguenther@suse.de>
22116 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22117 Update virtual SSA form.
22119 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22121 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22122 * configure: Regenerate.
22123 * config.in: Regenerate.
22124 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22125 if !HAVE_LD_EH_FRAME_CIEV3.
22127 2013-03-21 Richard Biener <rguenther@suse.de>
22129 * tree-cfg.c (verify_expr_no_block): New function.
22130 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22131 nor DECL_VALUE_EXPR have locations with associated blocks.
22132 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22133 (clear_unused_block_pointer): Remove code dealing with
22134 blocks in DECL_DEBUG_EXPR locations.
22136 2013-03-21 Richard Biener <rguenther@suse.de>
22138 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22139 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22140 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22141 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22142 instead of DECL_DEBUG_EXPR_IS_FROM.
22143 * gimplify.c (gimplify_modify_expr): Likewise.
22144 * tree-cfg.c (verify_expr_location_1): Likewise.
22145 * tree-complex.c (create_one_component_var): Likewise.
22146 * tree-sra.c (create_access_replacement): Likewise.
22147 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22148 (clear_unused_block_pointer): Likewise.
22149 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22150 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22151 * var-tracking.c (var_debug_decl): Likewise.
22152 (track_expr_p): Likewise.
22153 * tree-inline.c (add_local_variables): Likewise. Set
22154 DECL_HAS_DEBUG_EXPR_P after copying it.
22155 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22156 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22158 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22161 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22162 * configure: Regenerate.
22163 * config.in: Regenerate.
22164 * config/i386/i386.md (*movdf_internal): Use
22165 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22166 movd instead of movq mnemonic for interunit moves.
22167 (*movdi_internal): Ditto.
22169 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22171 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22172 (abd<mode>_3): New pattern.
22173 (aba<mode>_3): New pattern.
22174 (fabd<mode>_3): New pattern.
22176 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22178 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22179 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22180 occurrence of REGISTER_PREFIX as its empty string.
22182 2013-03-20 Jeff Law <law@redhat.com>
22184 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22185 addititional equivalences for equality comparisons between an SSA_NAME
22186 and a constant where the SSA_NAME was set from a widening conversion.
22188 2013-03-20 Walter Lee <walt@tilera.com>
22190 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22192 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22194 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22195 depending on type attribute.
22196 (*movti_internal): Ditto.
22197 (*movtf_internal): Ditto.
22198 (*movxf_internal): Ditto.
22199 (*movdf_internal): Ditto.
22200 (*movsf_internal): Ditto.
22202 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22204 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22205 maybe_vex for sselog1 and ssemov types.
22206 (*movdi_internal): Reorder operand constraints.
22207 (*movsi_internal): Ditto. Set prefix attribute to
22208 maybe_vex for sselog1 and ssemov types.
22209 (*movtf_internal): Set prefix attribute to maybe_vex
22210 for sselog1 and ssemov types.
22211 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22212 DImode ssemov types. Reorder operand constraints.
22213 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22214 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22215 attribute for SImode ssemov types. Reorder operand constraints.
22217 2013-03-20 Martin Jambor <mjambor@suse.cz>
22219 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22220 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22222 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22224 * config/rs6000/predicates.md (indexed_address, update_address_mem
22225 update_indexed_address_mem): New predicates.
22226 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22227 attribute for load/store instructions.
22228 * config/rs6000/dfp.md (movsd_store): Likewise.
22229 (movsd_load): Likewise.
22230 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22231 (unnamed HI->DI extend define_insn): Likewise.
22232 (unnamed SI->DI extend define_insn): Likewise.
22233 (unnamed QI->SI extend define_insn): Likewise.
22234 (unnamed QI->HI extend define_insn): Likewise.
22235 (unnamed HI->SI extend define_insn): Likewise.
22236 (unnamed HI->SI extend define_insn): Likewise.
22237 (extendsfdf2_fpr): Likewise.
22238 (movsi_internal1): Likewise.
22239 (movsi_internal1_single): Likewise.
22240 (movhi_internal): Likewise.
22241 (movqi_internal): Likewise.
22242 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22243 attribute for load/store instructions.
22244 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22246 (mov<mode>_softfloat): Likewise.
22247 (mov<mode>_hardfloat32): Likewise.
22248 (mov<mode>_hardfloat64): Likewise.
22249 (mov<mode>_softfloat64): Likewise.
22250 (movdi_internal32): Likewise.
22251 (movdi_internal64): Likewise.
22252 (probe_stack_<mode>): Likewise.
22254 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22256 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22257 floating point, and decimal floating point to reload iterator.
22259 * config/rs6000/constraints.md (wl constraint): New constraints to
22260 return FLOAT_REGS if certain options are used to reduce the number
22261 of separate patterns that exist in the file.
22262 (wx constraint): Likewise.
22263 (wz constraint): Likewise.
22265 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22266 -mdebug=reg, print wg, wl, wx, and wz constraints.
22267 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22268 Initialize the reload functions for 64-bit binary/decimal floating
22270 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22271 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22272 create the buffer on the stack to overcome not having a 32-bit
22274 (rs6000_emit_move): Likewise.
22275 (rs6000_secondary_memory_needed_rtx): Likewise.
22276 (rs6000_alloc_sdmode_stack_slot): Likewise.
22277 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22278 via xxlxor, just like DFmode 0.0.
22280 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22281 define as 1 if we are running on a power7 or newer.
22282 (enum r6000_reg_class_enum): Add new constraints.
22284 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22285 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22286 with other moves by using conditional constraits (wg). Use LFIWZX
22287 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22288 (movsd splitter): Likewise.
22289 (movsd_hardfloat): Likewise.
22290 (movsd_softfloat): Likewise.
22292 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22293 binary and decimal floating point moves.
22294 (fmove_ok): New attributes to combine binary and decimal floating
22295 point moves, and to combine power6x (mfpgpr) moves along normal
22297 (real_value_to_target): Likewise.
22298 (f32_lr): Likewise.
22299 (f32_lm): Likewise.
22300 (f32_li): Likewise.
22301 (f32_sr): Likewise.
22302 (f32_sm): Likewise.
22303 (f32_si): Likewise.
22304 (movsf): Combine binary and decimal floating point moves. Combine
22305 power6x (mfpgpr) moves with other moves by using conditional
22306 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22307 (mov<mode> for SFmode/SDmode); Likewise.
22308 (SFmode/SDmode splitters): Likewise.
22309 (movsf_hardfloat): Likewise.
22310 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22311 (movsf_softfloat): Likewise.
22312 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22314 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22315 wx and wz constraints.
22317 * config/rs6000/constraints.md (wg constraint): New constraint to
22318 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22320 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22323 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22324 -mdebug=reg, print wg, wl, wx, and wz constraints.
22325 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22326 Initialize the reload functions for 64-bit binary/decimal floating
22328 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22329 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22330 create the buffer on the stack to overcome not having a 32-bit
22332 (rs6000_emit_move): Likewise.
22333 (rs6000_secondary_memory_needed_rtx): Likewise.
22334 (rs6000_alloc_sdmode_stack_slot): Likewise.
22335 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22336 via xxlxor, just like DFmode 0.0.
22338 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22339 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22340 with other moves by using conditional constraits (wg). Use LFIWZX
22341 and STFIWX for loading SDmode on power7.
22342 (movdd splitters): Likewise.
22343 (movdd_hardfloat32): Likewise.
22344 (movdd_softfloat32): Likewise.
22345 (movdd_hardfloat64_mfpgpr): Likewise.
22346 (movdd_hardfloat64): Likewise.
22347 (movdd_softfloat64): Likewise.
22349 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22350 64-bit binary and decimal floating point moves.
22351 (FMOVE64X): Likewise.
22352 (movdf): Combine 64-bit binary and decimal floating point moves.
22353 Combine power6x (mfpgpr) moves with other moves by using
22354 conditional constraits (wg).
22355 (mov<mode> for DFmode/DDmode): Likewise.
22356 (DFmode/DDmode splitters): Likewise.
22357 (movdf_hardfloat32): Likewise.
22358 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22359 (movdf_softfloat32): Likewise.
22360 (movdf_hardfloat64_mfpgpr): Likewise.
22361 (movdf_hardfloat64): Likewise.
22362 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22363 (movdf_softfloat64): Likewise.
22364 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22365 (reload_<mode>_load): Move to later in the file so they aren't in
22366 the middle of the floating point move insns.
22367 (reload_<mode>_store): Likewise.
22369 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22372 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22373 constraint if -mdebug=reg.
22374 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22375 Enable using dd reload support if needed.
22377 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22378 binary and decimal floating point moves in rs6000.md.
22379 (movtd_internal): Likewise.
22381 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22382 decimal floating point moves.
22384 (movtf_internal): Likewise.
22385 (mov<mode>_internal, TDmode/TFmode): Likewise.
22386 (movtf_softfloat): Likewise.
22387 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22389 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22390 movdi_internal64, using wg constraint for move direct operations.
22391 (movdi_internal64): Likewise.
22393 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22394 MODES_TIEABLE_P for selected modes. Print the numerical value of
22395 the various virtual registers. Use GPR/FPR first/last values,
22396 instead of hard coding the register numbers. Print which modes
22397 have reload functions registered.
22398 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22399 settings before/after setting cpu, target and subtarget settings.
22400 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22401 and for secondary reload failures in rs6000_secondary_reload_inner.
22402 (rs6000_secondary_reload_fail): Likewise.
22403 (rs6000_secondary_reload_inner): Likewise.
22405 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22406 macros for first/last GPR and FPR registers.
22407 (LAST_GPR_REGNO): Likewise.
22408 (FIRST_FPR_REGNO): Likewise.
22409 (LAST_FPR_REGNO): Likewise.
22411 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22412 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22413 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22414 (vcond<mode><mode>): Likewise.
22415 (vcondu<mode><mode>): Likewise.
22416 (vector_gtu<mode>): Likewise.
22417 (vector_gte<mode>): Likewise.
22418 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22419 to prevent the compiler from converting DImode operations to TImode.
22420 (ior<mode>3): Likewise.
22421 (and<mode>3): Likewise.
22422 (one_cmpl<mode>2): Likewise.
22423 (nor<mode>3): Likewise.
22424 (andc<mode>3): Likewise.
22426 * config/rs6000/constraints.md (wt constraint): New constraint
22427 that returns VSX_REGS if TImode is allowed in VSX registers.
22429 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22430 constant under VSX.
22432 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22433 similar to TImode, but it is restricted to being in the GPRs.
22435 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22436 TImode to occupy a single VSX register.
22438 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22439 -mvsx-timode for power7/power8.
22440 (power7 cpu): Likewise.
22441 (power8 cpu): Likewise.
22443 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22444 sure that TFmode/TDmode take up two registers if they are ever
22445 allowed in the upper VSX registers.
22446 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22448 (rs6000_init_hard_regno_mode_ok): Likewise.
22449 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22450 constraint. Print if LRA is turned on.
22451 (rs6000_option_override_internal): Give an error if -mvsx-timode
22452 and VSX is not enabled.
22453 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22454 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22455 to reg+offset addressing. Use PTImode when checking offset
22456 addresses for validity.
22457 (reg_offset_addressing_ok_p): Likewise.
22458 (rs6000_legitimate_offset_address_p): Likewise.
22459 (rs6000_legitimize_address): Likewise.
22460 (rs6000_legitimize_reload_address): Likewise.
22461 (rs6000_legitimate_address_p): Likewise.
22462 (rs6000_eliminate_indexed_memrefs): Likewise.
22463 (rs6000_emit_move): Likewise.
22464 (rs6000_secondary_reload): Likewise.
22465 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22466 reloads to fpr registers to continue to use reg+offset addressing,
22467 but 64-bit reloads to altivec registers need reg+reg addressing.
22468 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22469 it. Treat LO_SUM like a PLUS operation.
22470 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22471 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22472 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22473 registers to share a register with a smaller sized type, since VSX
22474 puts scalars in the upper 64-bits.
22475 (print_operand): Add support for PTImode.
22476 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22477 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22478 registers, but don't have arithmetic support.
22479 (rs6000_memory_move_cost): Add test for VSX.
22480 (rs6000_opt_masks): Add -mvsx-timode.
22482 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22485 (VSr): Use wt constraint for TImode.
22486 (VSv): Drop TImode support.
22487 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22488 (vsx_movti_64bit): Likewise.
22489 (vsx_movti_32bit): Likewise.
22490 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22491 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22492 one '?' on the appropriate output constraint. Do not allow TImode
22493 logical operations on 32-bit systems.
22494 (vsx_ior<mode>3): Likewise.
22495 (vsx_xor<mode>3): Likewise.
22496 (vsx_one_cmpl<mode>2): Likewise.
22497 (vsx_nor<mode>3): Likewise.
22498 (vsx_andc<mode>3): Likewise.
22499 (vsx_concat_<mode>): Likewise.
22500 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22502 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22503 OPTION_MASK_VSX_TIMODE.
22504 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22505 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22507 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22508 (TI2 iterator): New iterator for TImode, PTImode.
22509 (wd mode attribute): Add values for vector types.
22510 (movti_string): Replace TI move operations with operations for TImode
22511 and PTImode. Add support for TImode being allowed in VSX registers.
22512 (mov<mode>_string, TImode/PTImode): Likewise.
22513 (movti_ppc64): Likewise.
22514 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22515 (TI mode splitters): Likewise.
22517 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22520 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22522 PR tree-optimization/56355
22523 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22524 Also handle integers with undefined overflow.
22526 2013-03-20 Catherine Moore <clm@codesourcery.com>
22527 Maciej W. Rozycki <macro@codesourcery.com>
22528 Tom de Vries <tom@codesourcery.com>
22529 Nathan Sidwell <nathan@codesourcery.com>
22530 Iain Sandoe <iain@codesourcery.com>
22531 Nathan Froyd <froydnj@codesourcery.com>
22532 Chao-ying Fu <fu@mips.com>
22534 * doc/extend.texi: (micromips, nomicromips, nocompression):
22535 Document new function attributes.
22536 * doc/invoke.texi (minterlink-compressed, mmicromips,
22537 m14k, m14ke, m14kec): Document new options.
22538 (minterlink-mips16): Update documentation.
22539 * doc/md.texi (ZC, ZD): Document new constraints.
22540 * configure.ac (gcc_cv_as_micromips): Check if linker
22541 supports the .set micromips directive.
22542 * configure: Regenerate.
22543 * config.in: Regenerate.
22544 * config/mips/mips-tables.opt: Regenerate.
22545 * config/mips/micromips.md: New file.
22546 * constraints.md (ZC, ZD): New constraints.
22547 * config/mips/predicates.md (movep_src_register): New predicate.
22548 (movep_src_operand): New predicate.
22549 (non_volatile_mem_operand): New predicate.
22550 * config/mips/mips.md (multimem): New type.
22551 (length): Differentiate between 17-bit and 18-bit branch offsets.
22552 (MOVEP1, MOVEP2): New mode iterator.
22553 (mov_<load>l): Use ZC constraint.
22554 (mov_<load>r): Likewise.
22555 (mov_<store>l): Likewise.
22556 (mov_<store>r): Likewise.
22557 (*branch_equality<mode>_inverted): Add microMIPS support.
22558 (*branch_equality<mode>): Likewise.
22559 (*jump_absolute): Likewise.
22560 (indirect_jump_<mode>): Likewise.
22561 (tablejump_<mode>): Likewise.
22562 (<optab>_internal): Likewise.
22563 (sibcall_internal): Likewise.
22564 (sibcall_value_internal): Likewise.
22565 (prefetch): Use constraint ZD.
22566 * config/mips/mips.opt (minterlink-compressed): New option.
22567 (minterlink-mips16): Now an alias for minterlink-compressed.
22568 (mmicromips): New option.
22569 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22570 (compare_and_swap_12): Likewise.
22571 (sync_add<mode>): Likewise.
22572 (sync_<optab>_12): Likewise.
22573 (sync_old_<optab>_12): Likewise.
22574 (sync_new_<optab>_12): Likewise.
22575 (sync_nand_12): Likewise.
22576 (sync_old_nand_12): Likewise.
22577 (sync_new_nand_12): Likewise.
22578 (sync_sub<mode>): Likewise.
22579 (sync_old_add<mode>): Likewise.
22580 (sync_old_sub<mode>): Likewise.
22581 (sync_new_add<mode>): Likewise.
22582 (sync_new_sub<mode>): Likewise.
22583 (sync_<optab><mode>): Likewise.
22584 (sync_old_<optab><mode>): Likewise.
22585 (sync_new_<optab><mode>): Likewise.
22586 (sync_nand<mode>): Likewise.
22587 (sync_old_nand<mode>): Likewise.
22588 (sync_new_nand<mode>): Likewise.
22589 (sync_lock_test_and_set<mode>): Likewise.
22590 (test_and_set_12): Likewise.
22591 (atomic_compare_and_swap<mode>): Likewise.
22592 (atomic_exchange<mode>_llsc): Likewise.
22593 (atomic_fetch_add<mode>_llsc): Likewise.
22594 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22595 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22596 (umips_save_restore_pattern_p): Likewise.
22597 (umips_load_store_pair_p): Likewise.
22598 (umips_output_load_store_pair): Likewise.
22599 (umips_movep_target_p): Likewise.
22600 (umips_12bit_offset_address_p): Likewise.
22601 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22602 (mips_base_mips16): Rename this...
22603 (mips_base_compression_flags): ...to this. Update all uses.
22604 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22605 (mips_mips16_decl_p): Delete.
22606 (mips_nomips16_decl_p): Delete.
22607 (mips_get_compress_on_flags): New function.
22608 (mips_get_compress_off_flags): New function.
22609 (mips_get_compress_mode): New function.
22610 (mips_get_compress_on_name): New function.
22611 (mips_get_compress_off_name): New function.
22612 (mips_insert_attributes): Support multiple compression types.
22613 (mips_merge_decl_attributes): Likewise.
22614 (umips_12bit_offset_address_p): New function.
22615 (mips_start_function_definition): Emit .set micromips directive.
22616 (mips_call_may_need_jalx_p): New function.
22617 (mips_function_ok_for_sibcall): Add microMIPS support.
22618 (mips_print_operand_punctuation): Support short delay slots and
22620 (umips_swm_mask, umips_swm_encoding): New.
22621 (umips_build_save_restore): New function.
22622 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22623 (was_mips16_p): Remove.
22624 (old_compression_mode): New.
22625 (mips_set_compression_mode): New function.
22626 (mips_set_current_function): Add microMIPS support.
22627 (mips_option_override): Likewise.
22628 (umips_save_restore_pattern_p): New function.
22629 (umips_output_save_restore): New function.
22630 (umips_load_store_pair_p_1): New function.
22631 (umips_load_store_pair_p): New function.
22632 (umips_output_load_store_pair_1): New function.
22633 (umips_output_load_store_pair): New function.
22634 (umips_movep_target_p) New function.
22635 (mips_prepare_pch_save): Add microMIPS support.
22636 * config/mips/mips.h (TARGET_COMPRESSION): New.
22637 (TARGET_CPU_CPP_BUILTINS): Update macro
22638 to use new compression flags and to support microMIPS.
22639 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22640 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22641 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22642 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22643 (ASM_SPEC): Support mmicromips and mno-micromips.
22644 (M16STORE_REG_P): New macro.
22645 (MIPS_CALL): Support TARGET_MICROMIPS.
22646 (MICROMIPS_J): New macro.
22647 (mips_base_mips16): Rename this...
22648 (mips_base_compression_flags): ...to this.
22649 (UMIPS_12BIT_OFFSET_P): New macro.
22650 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22651 (MULTILIB_DIRNAMES): Likewise.
22652 2013-03-20 Richard Biener <rguenther@suse.de>
22654 PR tree-optimization/56661
22655 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22656 the result does not have to be distinct.
22658 2013-03-20 Richard Biener <rguenther@suse.de>
22660 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22663 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22664 Steven Bosscher <steven@gcc.gnu.org>
22666 PR rtl-optimization/56605
22667 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22669 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22672 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22673 that require movd instead of movq.
22675 2013-03-20 Richard Biener <rguenther@suse.de>
22677 * tree-ssa-structalias.c (struct variable_info): Add pointer
22678 to the first field of an aggregate with sub-vars. Make
22679 this and the pointer to the next subfield its ID.
22680 (vi_next): New function.
22681 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22682 storedanything_id, integer_id): Increment by one.
22683 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22684 get_call_clobber_vi): Adjust.
22685 (solution_set_expand): Simplify and speedup.
22686 (solution_set_add): Inline into ...
22687 (set_union_with_increment): ... this. Adjust accordingly.
22688 (do_sd_constraint): Likewise.
22689 (do_ds_constraint): Likewise.
22690 (do_complex_constraint): Simplify.
22691 (build_pred_graph): Adjust.
22692 (solve_graph): Likewise. Simplify and speedup.
22693 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22694 get_constraint_for_component_ref, get_constraint_for_1,
22695 first_vi_for_offset, first_or_preceding_vi_for_offset,
22696 create_function_info_for, create_variable_info_for_1,
22697 create_variable_info_for, intra_create_variable_infos): Adjust.
22698 (init_base_vars): Push NULL for ID zero.
22699 (compute_points_to_sets): Adjust.
22701 2013-03-20 Richard Biener <rguenther@suse.de>
22703 * cfgloop.c (verify_loop_structure): Streamline and avoid
22704 ICEing on corrupt loop tree.
22705 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22708 2013-03-20 Richard Biener <rguenther@suse.de>
22710 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22711 check whether an SSA update is needed.
22713 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22715 * config/mips/constraints.md (T): Rename to...
22719 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22720 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22722 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22724 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22725 (*subsi3_carryin_uxtw): Likewise.
22727 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22729 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22730 (*rorsi3_insn_uxtw): Likewise.
22732 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22734 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22735 (*extrsi5_insn_uxtw): Likewise.
22737 2013-03-19 Richard Biener <rguenther@suse.de>
22739 PR tree-optimization/56273
22740 * passes.c (init_optimization_passes): Move second VRP after DOM.
22742 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22744 * config/i386/i386.md (*movti_internal): Merge from
22745 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22746 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22747 nox64 isa attributes.
22749 2013-03-18 Richard Biener <rguenther@suse.de>
22751 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22753 (merge_node_constraints): Likewise.
22754 (build_succ_graph): Likewise.
22755 (valid_graph_edge): Inline into single caller.
22756 (unify_nodes): Likewise. Use bitmap_set_bit return value
22758 (scc_visit): Fix formatting and variable use.
22759 (do_sd_constraint): Use gcc_checking_assert.
22760 (do_ds_constraint): Likewise.
22761 (do_complex_constraint): Likewise.
22762 (condense_visit): Likewise. Cleanup.
22763 (dump_pred_graph): New function.
22764 (perform_var_substitution): Dump the pred-graph before
22765 variable substitution.
22766 (find_equivalent_node): Use gcc_checking_assert.
22767 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22769 2013-03-18 Richard Biener <rguenther@suse.de>
22771 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22772 Remove cond_expr_stmt_list argument and do not gimplify the
22774 (vect_loop_versioning): Adjust.
22775 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22776 Cleanup to use less temporaries.
22777 (vect_create_data_ref_ptr): Cleanup.
22779 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22781 PR tree-optimization/56635
22782 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22783 require types_compatible_p types.
22785 2013-03-18 Nick Clifton <nickc@redhat.com>
22787 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22788 spurious backslash.
22790 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22791 Add missing line to comment describing function.
22793 2013-03-18 Richard Biener <rguenther@suse.de>
22795 PR tree-optimization/56210
22796 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22797 Handle string / character search functions.
22798 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22800 2013-03-18 Richard Biener <rguenther@suse.de>
22802 PR middle-end/56483
22803 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22804 and implement properly.
22805 * gimple.h (gimple_cond_single_var_p): Remove.
22807 2013-03-18 Richard Biener <rguenther@suse.de>
22809 * tree-data-ref.h (find_data_references_in_loop): Declare.
22810 * tree-data-ref.c (get_references_in_stmt): Use a stack
22811 vector pre-allocated in the callers.
22812 (find_data_references_in_stmt): Adjust.
22813 (graphite_find_data_references_in_stmt): Likewise.
22814 (create_rdg_vertices): Likewise.
22815 (find_data_references_in_loop): Export.
22816 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22817 Compute dependences here...
22818 (vect_analyze_data_refs): ...not here. When we encounter
22819 a non-vectorizable data reference in basic-block vectorization
22820 truncate the data reference vector. Do not bother to
22821 fixup data-dependence information for gather loads.
22822 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22823 of data references, as reported.
22825 2013-03-18 Richard Biener <rguenther@suse.de>
22827 PR tree-optimization/3713
22828 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22829 has_constants and expr.
22830 (stmt_has_constants): Properly valueize SSA names when deciding
22831 whether the stmt has constants.
22833 2013-03-18 Richard Biener <rguenther@suse.de>
22835 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22836 whole function when there is nothing to do.
22837 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22838 * tree-vectorizer.c (vectorize_loops): Update virtual and
22839 loop-closed SSA once.
22840 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22842 2013-03-18 Richard Biener <rguenther@suse.de>
22844 PR middle-end/56113
22845 * domwalk.c (bb_postorder): New global static.
22846 (cmp_bb_postorder): New function.
22847 (walk_dominator_tree): Replace scheme imposing an order for
22848 visiting dominator sons by one sorting them at the time they
22849 are pushed on the stack.
22851 2013-03-18 Richard Biener <rguenther@suse.de>
22853 PR tree-optimization/39326
22854 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22855 (struct mem_ref): Replace mem member with ao_ref typed member.
22856 (MEM_ANALYZABLE): Adjust.
22857 (memref_eq): Likewise.
22858 (mem_ref_alloc): Likewise.
22859 (gather_mem_refs_stmt): Likewise.
22860 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22861 (execute_sm_if_changed_flag_set): Adjust.
22862 (execute_sm): Likewise.
22863 (ref_always_accessed_p): Likewise.
22864 (refs_independent_p): Likewise.
22865 (can_sm_ref_p): Likewise.
22867 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22870 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22871 return 1 even for !unsignedp.
22873 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22875 * config/i386/i386.md (isa): Add x64 and nox64.
22876 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22877 (*pushtf): Enable *roF alternative for x64 isa only.
22878 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22879 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22880 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22881 (*movtf_internal): Merge from *movtf_internal_rex64 and
22882 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22883 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22884 nox64 isa attributes.
22885 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22886 nox64 isa attributes.
22887 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22889 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22891 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22893 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22896 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22897 class names. Remove trailing comma after "ALL_REGS".
22899 2013-03-16 Jan Hubicka <jh@suse.cz>
22901 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22902 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22903 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22904 of cgraph_get_create_node.
22905 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22907 2013-03-16 Jason Merrill <jason@redhat.com>
22910 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22911 with DW_AT_default_value.
22913 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22915 * BASE-VER: Set to 4.9.0.
22917 2013-03-14 Andi Kleen <ak@linux.intel.com>
22920 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22921 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22922 Document _x* TSX intrinsics.
22924 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22925 David Holsgrove <david.holsgrove@xilinx.com>
22927 * configure.ac: Add MicroBlaze TLS support detection.
22928 * configure: Regenerate.
22929 * config/microblaze/microblaze-protos.h
22930 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22931 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22932 * config/microblaze/microblaze.c (microblaze_address_type): Add
22933 ADDRESS_TLS and tls_reloc address types.
22934 (microblaze_address_info): Add tls_reloc.
22935 (TARGET_HAVE_TLS): Define.
22936 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22937 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22938 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22939 load_tls_operand, microblaze_call_tls_get_addr,
22940 microblaze_legitimize_tls_address): New functions.
22941 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22942 (get_base_reg): Use microblaze_tls_symbol_p.
22943 (microblaze_classify_address): Handle TLS.
22944 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22945 label_mentioned_p and microblaze_tls_referenced_p.
22946 (microblaze_legitimize_address): Handle TLS.
22947 (microblaze_address_insns): Handle ADDRESS_TLS.
22948 (pic_address_needs_scratch): Handle TLS.
22949 (print_operand_address): Handle TLS.
22950 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22951 (microblaze_expand_move): Handle TLS.
22952 (microblaze_legitimate_constant_p): Check
22953 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22954 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22955 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22956 (PIC_OFFSET_TABLE_REGNUM): Set.
22957 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22958 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22959 (addsi3, movsi_internal2, movdf_internal): Update constraints
22960 * config/microblaze/predicates.md (arith_plus_operand): Define
22961 (move_operand): Redefine as move_src_operand,
22962 check microblaze_tls_referenced_p.
22964 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22966 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22967 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22969 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22971 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22974 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22976 PR tree-optimization/53265
22977 * common.opt (Waggressive-loop-optimizations): New option.
22978 * tree-ssa-loop-niter.c: Include tree-pass.h.
22979 (do_warn_aggressive_loop_optimizations): New function.
22980 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22981 if number_of_latch_executions returned constant.
22982 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22983 early. If number_of_latch_executions returned constant, set
22984 nb_iterations_upper_bound back to it.
22985 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22987 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22988 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22990 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22991 (MULTILIB_OSDIRNAMES): Set.
22992 * genmultilib: If defaultosdirname doesn't start with :: , set
22993 defaultosdirname2 instead, clear it and emit two . multilib_raw
22994 entries instead of just one.
22996 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22998 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22999 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23000 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23001 (SUBTARGET_OVERRIDE_OPTIONS): New.
23003 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23006 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23007 (musermode): Convert to Var(TARGET_USERMODE).
23008 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23009 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23010 * config/sh/sh.c (sh_option_override): Use
23011 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23012 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23014 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23016 (udivsi3_i4_single, divsi3_i4_single): Use
23017 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23019 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23021 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23024 2013-03-13 Richard Biener <rguenther@suse.de>
23026 PR tree-optimization/56608
23027 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23028 calls when vectorizing basic-blocks.
23030 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23033 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23036 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23038 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23040 2013-03-11 Jan Hubicka <jh@suse.cz>
23043 * lto-streamer-out.c (output_symbol_p): Skip references from
23044 constructors of external variables.
23046 2013-03-11 Jan Hubicka <jh@suse.cz>
23048 PR middle-end/56571
23049 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23051 * emit-rtl.c (verify_rtx_sharing): Likewise.
23052 (copy_insn_1): Likewise.
23053 * rtl.c (copy_rtx): Likewise.
23055 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23058 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23059 output_operand_lossage message.
23061 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23064 * arm.c (shift_op): Validate RTL pattern on the fly.
23065 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23068 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23071 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23072 new call patterns that clobber %fr12.
23073 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23074 split and postreload patterns.
23075 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23076 registers %fr12 and %fr12R as call used.
23078 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23080 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23081 (canon_address, record_store, replace_read, check_mem_read_rtx,
23082 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23083 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23084 rest_of_handle_dse): Likewise.
23086 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23088 PR middle-end/56524
23089 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23091 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23092 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23093 (save_optabs_if_changed): Replace with...
23094 (init_tree_optimization_optabs): ...this.
23095 * optabs.c (save_optabs_if_changed): Rename to...
23096 (init_tree_optimization_optabs): ...this. Take the optimization node
23097 as argument. Do nothing if the base optabs are already correct.
23098 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23099 to recompute optabs.
23100 * function.h (function): Remove optabs field.
23101 * function.c (invoke_set_current_function_hook): Call
23102 init_tree_optimization_optabs. Use the result to initialize
23105 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23107 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23108 if GTMA_HAS_NO_INSTRUMENTATION.
23109 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23110 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23111 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23112 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23113 GTMA_HAS_NO_INSTRUMENTATION.
23115 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23117 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23120 2013-03-08 Marek Polacek <polacek@redhat.com>
23121 Jakub Jelinek <jakub@redhat.com>
23123 PR tree-optimization/56478
23124 * predict.c (is_comparison_with_loop_invariant_p): Change the
23125 type of loop_step to tree.
23126 (predict_loops): Adjust.
23127 (predict_iv_comparison): Perform the computations on double_ints.
23129 2013-03-08 Richard Biener <rguenther@suse.de>
23131 PR tree-optimization/56570
23132 * tree-cfg.c (verify_expr_location_1): Verify locations for
23134 * tree-sra.c (create_access_replacement): Strip locations
23135 from DECL_DEBUG_EXPRs.
23137 2013-03-08 Richard Biener <rguenther@suse.de>
23139 * tree-inline.c (expand_call_inline): Do not associate
23140 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23141 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23143 2013-03-08 Richard Biener <rguenther@suse.de>
23145 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23146 or block changes with -Og. Fix for location / block encoding
23147 changes and PHI arguments with locations.
23149 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23151 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23153 (struct output_info): Likewise.
23154 (register_overhead): Remove bad gcc_assert.
23155 (bitmap_find_bit): If there is only a single bitmap element, do not
23156 count a miss as a search.
23157 (print_statistics): Update for counter type changes.
23158 (dump_bitmap_statistics): Likewise. Print headers such that they
23159 are properly lined up with the printed counters.
23161 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23163 PR tree-optimization/56559
23164 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23165 check that it has only a single use.
23167 2013-03-07 Richard Biener <rguenther@suse.de>
23169 * doc/invoke.texi (fwhole-program): Discourage use in combination
23172 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23174 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23176 PR tree-optimization/56539
23177 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23178 instead of GSI_CONTINUE_LINKING as last argument to
23179 force_gimple_operand_gsi. Adjust function comment.
23181 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23184 PR middle-end/56548
23185 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23186 promoted mode, convert the result back to the original mode.
23188 2013-03-06 Richard Biener <rguenther@suse.de>
23190 PR middle-end/56294
23191 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23192 (insert_updated_phi_nodes_compare_uids): New function.
23193 (update_ssa): Sort symbols_to_rename after UID before
23194 traversing it to insert PHI nodes.
23196 2013-03-06 Richard Biener <rguenther@suse.de>
23198 PR middle-end/50494
23199 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23200 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23203 2013-02-13 Richard Biener <rguenther@suse.de>
23206 * varasm.c (output_constant_def_1): Get the decl representing
23207 the constant as argument.
23208 (output_constant_def): Wrap output_constant_def_1.
23209 (make_decl_rtl): Use output_constant_def_1 with the decl
23210 representing the constant.
23211 (build_constant_desc): Optionally re-use a decl already
23212 representing the constant.
23213 (tree_output_constant_def): Adjust.
23215 2013-03-06 Joey Ye <joey.ye@arm.com>
23218 * gcc.c (convert_white_space): New function.
23219 (main): Handles white space in function name.
23221 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23224 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23225 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23226 to SH_DIV_CALL_TABLE for TARGET_SH2.
23227 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23229 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23230 call-table options.
23232 2013-03-05 Sterling Augustine <saugustine@google.com>
23233 Cary Coutant <ccoutant@google.com>
23236 * dwarf2out.c (resolve_addr): Don't call
23237 remove_loc_list_addr_table_entries a second time for the same
23240 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23243 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23244 (avoid_complex_debug_insns): New function.
23245 (expand_debug_locations): Call it.
23247 PR rtl-optimization/56484
23248 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23249 lifetimes of hard registers on small register class machines.
23251 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23253 * config/microblaze/microblaze-protos.h: Rename
23254 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23255 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23257 (microblaze_fast_interrupt_function_p): New function.
23258 (microblaze_is_interrupt_handler): Rename to
23259 microblaze_is_interrupt_variant and add fast_interrupt check.
23260 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23261 (save_restore_insns): Likewise.
23262 (compute_frame_size): Likewise.
23263 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23264 (microblaze_globalize_label): Likewise.
23265 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23266 * config/microblaze/microblaze.md: Use wrapper
23267 microblaze_is_interrupt_variant.
23269 2013-03-05 Kai Tietz <ktietz@redhat.com>
23271 * sdbout.c (sdbout_one_type): Switch to current function's section
23272 supporting cold/hot.
23274 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23276 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23279 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23281 PR middle-end/56461
23282 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23283 if VALGRIND_GET_VBITS is defined, temporarily make object
23284 memory all defined, and restore previous valgrind addressability
23285 and definability afterwards. Free this_object at the end.
23287 PR middle-end/56461
23288 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23289 right before calling lra_create_live_ranges, also call it
23290 when clearing live_p. Only call lra_clear_live_ranges
23291 at the end if live_p.
23293 PR middle-end/56461
23294 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23296 2013-03-05 Richard Biener <rguenther@suse.de>
23298 PR tree-optimization/56521
23299 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23302 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23305 * except.h (remove_unreachable_eh_regions): New prototype.
23306 * except.c (remove_eh_handler_splicer): New function, split out
23307 of remove_eh_handler.
23308 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23309 warning about running it on many EH regions one at a time.
23310 (remove_unreachable_eh_regions_worker): New function, walk the
23311 EH tree in depth-first order and remove non-marked regions.
23312 (remove_unreachable_eh_regions): New function.
23313 * tree-eh.c (mark_reachable_handlers): New function, split out
23314 from remove_unreachable_handlers.
23315 (remove_unreachable_handlers): Use mark_reachable_handlers and
23316 remove_unreachable_eh_regions.
23317 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23318 and remove_unreachable_eh_regions.
23320 2013-03-05 Richard Biener <rguenther@suse.de>
23322 PR middle-end/56525
23323 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23324 not freeing them until the end.
23326 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23328 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23330 2013-03-05 Richard Biener <rguenther@suse.de>
23332 PR tree-optimization/56270
23333 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23334 of loads after scheduling an SLP instance.
23336 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23338 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23340 (check_gcc_parallelize): Run guality.exp as a separate job from
23341 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23342 struct-layout-1.exp with stackalign.exp.
23344 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23346 PR middle-end/56461
23347 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23348 load_index sbitmap even if some bit in it isn't set.
23350 PR middle-end/56461
23351 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23352 (discover_iteration_bound_by_body_walk): Change queues to
23353 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23354 spelling in comment. Call safe_push on queues[bound_index] directly.
23355 Release queues[queue_index] in every iteration unconditionally.
23356 Release bounds vector.
23358 PR middle-end/56461
23359 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23360 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23361 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23364 2013-03-05 Richard Biener <rguenther@suse.de>
23367 * tree-inline.c (remap_blocks_to_null): New function.
23368 (expand_call_inline): When expanding a call stmt without
23369 an associated block inline remap all callee blocks to NULL.
23371 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23373 PR rtl-optimization/56494
23374 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23375 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23376 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23378 PR middle-end/56461
23379 * sel-sched-ir.c (free_sched_pools): Release
23380 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23381 if succs_info_pool.max_top isn't -1.
23384 * opts.c (opts_obstack, opts_concat): Moved to...
23385 * opts-common.c (opts_obstack, opts_concat): ... here.
23387 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23389 PR middle-end/56461
23390 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23392 2013-03-04 Martin Jambor <mjambor@suse.cz>
23394 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23395 all appropriate places.
23397 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23399 PR tree-optimization/56424
23400 * ipa-split.c (split_function): Do not set the RSO flag if result is
23401 not by reference and its type is a register type.
23403 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23405 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23406 (microblaze_legitimate_pic_operand): Likewise
23407 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23408 new function microblaze_legitimate_pic_operand
23409 * config/microblaze/microblaze-protos.h
23410 (microblaze_legitimate_pic_operand): Declare.
23412 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23414 * config/microblaze/predicates.md (call_insn_simple_operand):
23415 New predicate for supported rtx code types.
23416 * config/microblaze/microblaze.md (call_internal1): Use
23417 call_insn_simple_operand predicate.
23419 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23421 PR middle-end/56461
23422 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23423 partitions.ordered_remove.
23425 PR middle-end/56461
23426 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23427 vec_oprnds0.create (1) for modifier == NONE.
23429 PR middle-end/56461
23430 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23431 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23432 vec_oprnds1 right before pushing anything to it for
23435 PR middle-end/56461
23436 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23437 set nbbs to 0 instead of having separate code path.
23438 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23439 instead of false as last argument if returning NULL.
23441 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23443 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23444 the attribute is now called "target" instead of "option".
23445 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23446 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23447 attribute/pragma name for TARGET_OPTION_VALID_P and
23448 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23449 * doc/tm.texi: Regenerated.
23451 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23453 * config/microblaze/microblaze.c:
23454 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23455 * config/microblaze/microblaze.h: Add -mxl-reorder to
23457 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23458 instructions emitted if TARGET_REORDER.
23459 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23460 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23463 2013-03-01 Xinliang David Li <davidxl@google.com>
23465 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23468 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23470 PR middle-end/56461
23471 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23472 vector even when returning true. Fix up function comment formatting.
23474 PR middle-end/56461
23475 * ira-build.c (ira_loop_nodes_count): New variable.
23476 (create_loop_tree_nodes): Initialize it.
23477 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23479 PR middle-end/56461
23480 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23481 method on dr_chain and result_chain.
23482 * tree-vect-stmts.c (vectorizable_store): Only call
23483 result_chain.create if j == 0.
23485 PR middle-end/56461
23486 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23487 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23488 before overwriting it.
23490 2013-03-01 Tobias Burnus <burnus@net-b.de>
23492 * doc/extended.texi (C Extensions): Change order in @menu
23494 (Other MIPS Built-in Functions): Move last MIPS entry before
23495 "picoChip Built-in Functions".
23496 (SH Built-in Functions): Move after RX Built-in Functions.
23497 * doc/gcc.texi (Introduction): Change order in @menu
23499 * doc/md.texi (Constraints): Ditto.
23500 * gty.texi (Type Information): Ditto.
23501 (User-provided marking routines for template types): Make
23503 * doc/invoke.texi (AArch64 Options): Move before
23504 "Adapteva Epiphany Options".
23506 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23507 Jakub Jelinek <jakub@redhat.com>
23510 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23511 no_address_safety_analysis attribute.
23512 * doc/extend.texi (no_address_safety_attribute): Rename to
23513 no_sanitize_address attribute, mention no_address_safety_analysis
23514 attribute as deprecated alias.
23516 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23518 PR middle-end/56461
23519 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23520 type to vec<vec<tree> > *.
23521 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23522 to be vec<tree> instead of vec<tree> *, set vec_defs
23523 to vNULL and call vec_defs.create (number_of_vects), adjust other
23525 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23526 vectorizable_condition): Adjust vect_get_slp_defs callers.
23528 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23530 * config/aarch64/aarch64.c
23531 (aarch64_float_const_representable): Remove unused variable.
23533 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23535 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23537 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23539 * config/aarch64/aarch64-builtins.c
23540 (aarch64_init_simd_builtins): Make static.
23542 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23544 * config/aarch64/aarch64.c
23545 (aarch64_simd_make_constant): Make static.
23547 2013-02-28 Martin Jambor <mjambor@suse.cz>
23549 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23550 with no initialization to the RHS of debug statements.
23552 2013-02-28 Martin Jambor <mjambor@suse.cz>
23554 PR tree-optimization/56294
23555 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23557 (get_access_replacement): Do not call create_access_replacement.
23558 Assert a replacement exists.
23559 (get_repl_default_def_ssa_name): Create the replacement declaration
23562 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23564 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23565 final_end_function.
23567 2013-02-28 Marek Polacek <polacek@redhat.com>
23569 PR rtl-optimization/56466
23570 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23571 if we're changing a loop.
23572 (peel_loops_completely): Likewise.
23574 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23577 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23579 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23582 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23583 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23584 INTX_FTYPE_FX, FX_FTYPE_INTX.
23585 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23587 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23589 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23590 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23591 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23592 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23593 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23594 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23595 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23596 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23597 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23598 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23599 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23600 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23601 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23602 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23603 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23604 (avrxmega6): Increase max flash segments from 5 to 6.
23605 * config/avr/t-multilib: Regenerate.
23606 * config/avr/avr-tables.opt: Regenerate.
23607 * doc/avr-mmcu.texi: Regenerate.
23609 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23611 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23612 (avr_device_to_arch): Rename to avr_device_to_ld.
23613 (avr_device_to_as): New prototype.
23614 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23615 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23616 * config/avr/driver-avr.c (avr_device_to_as): New.
23617 (avr_device_to_arch): Rename to avr_device_to_ld.
23619 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23621 PR middle-end/56461
23622 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23623 method on dr_chain and result_chain.
23625 PR middle-end/56461
23626 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23627 pointer_set_destroy on not_executed_last_iteration.
23629 PR middle-end/56461
23630 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23632 PR middle-end/56461
23633 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23634 FOR_EACH_DEFINED_FUNCTION when freeing state.
23636 PR middle-end/56461
23637 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23639 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23642 PR middle-end/56461
23643 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23644 known_aggs[i].items and release known_aggs vector.
23646 PR middle-end/56461
23647 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23649 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23651 * config/microblaze/microblaze.c (microblaze_emit_compare):
23652 Use xor for EQ/NE comparisions.
23653 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23654 (cbranchsf4): Adjust operator to comparison_operator.
23656 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23658 PR middle-end/56461
23659 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23661 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23662 vec_safe_push, always update *slot.
23663 (redirect_edge_var_map_clear): Use vec_free.
23664 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23665 (free_var_map_entry): Use vec_free.
23666 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23667 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23669 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23671 PR middle-end/45472
23672 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23673 when the may_trap_p bit of the exprs being merged differs.
23674 Reorder tests for speculativeness in the logical and operator.
23676 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23678 * incpath.c (add_standard_paths): Use reconcat instead of concat
23679 where appropriate and avoid leaking memory.
23681 * opts.h: Include obstack.h.
23682 (opts_concat): New prototype.
23683 (opts_obstack): New declaration.
23684 * opts.c (opts_concat): New function.
23685 (opts_obstack): New variable.
23686 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23687 (finish_options): Use opts_concat instead of concat
23688 and XOBNEWVEC instead of XNEWVEC.
23689 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23690 generate_option): Likewise.
23691 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23692 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23695 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23696 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23698 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23700 PR middle-end/56461
23701 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23703 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23705 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23706 (arm_block_move_unaligned_straight): Likewise.
23707 (arm_adjust_block_mem): Likewise.
23709 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23712 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23713 temp, cond and label.
23714 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23717 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23718 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23721 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23722 of prologue/epilogue functions.
23725 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23726 Remove unused variable cfa_offset.
23727 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23730 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23731 type promotion to unsigned.
23734 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23735 for HOST_WIDE_INT of 32 bit / same size as int.
23736 (arm_block_move_unaligned_straight): Likewise.
23737 (arm_adjust_block_mem): Likewise.
23740 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23743 2013-02-26 Marek Polacek <polacek@redhat.com>
23745 PR tree-optimization/56426
23746 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23748 2013-02-26 Richard Biener <rguenther@suse.de>
23751 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23752 unused variable loops.
23754 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23756 PR tree-optimization/56448
23757 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23758 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23759 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23760 later operands of the references, or even first operand for
23761 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23763 PR tree-optimization/56443
23764 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23765 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23766 to type_for_mode langhook.
23768 2013-02-25 Matt Turner <mattst88@gmail.com>
23770 * doc/invoke.texi: Document r4700.
23772 2013-02-25 Richard Biener <rguenther@suse.de>
23774 PR tree-optimization/56175
23775 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23777 (simplify_bitwise_binary): ... here. Also guard the conversion
23778 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23780 2013-02-25 Catherine Moore <clm@codesourcery.com>
23783 2013-02-24 Catherine Moore <clm@codesourcery.com>
23784 Maciej W. Rozycki <macro@codesourcery.com>
23785 Tom de Vries <tom@codesourcery.com>
23786 Nathan Sidwell <nathan@codesourcery.com>
23787 Iain Sandoe <iain@codesourcery.com>
23788 Nathan Froyd <froydnj@codesourcery.com>
23789 Chao-ying Fu <fu@mips.com>
23791 * doc/extend.texi: (micromips, nomicromips, nocompression):
23792 Document new function attributes.
23793 * doc/invoke.texi (minterlink-compressed, mmicromips,
23794 m14k, m14ke, m14kec): Document new options.
23795 (minterlink-mips16): Update documentation.
23796 * doc/md.texi (ZC, ZD): Document new constraints.
23797 * configure.ac (gcc_cv_as_micromips): Check if linker
23798 supports the .set micromips directive.
23799 * configure: Regenerate.
23800 * config.in: Regenerate.
23801 * config/mips/mips-tables.opt: Regenerate.
23802 * config/mips/micromips.md: New file.
23803 * constraints.md (ZC, AD): New constraints.
23804 * config/mips/predicates.md (movep_src_register): New predicate.
23805 (movep_src_operand): New predicate.
23806 (non_volatile_mem_operand): New predicate.
23807 * config/mips/mips.md (multimem): New type.
23808 (length): Differentiate between 17-bit and 18-bit branch offsets.
23809 (MOVEP1, MOVEP2): New mode iterator.
23810 (mov_<load>l): Use ZC constraint.
23811 (mov_<load>r): Likewise.
23812 (mov_<store>l): Likewise.
23813 (mov_<store>r): Likewise.
23814 (*branch_equality<mode>_inverted): Add microMIPS support.
23815 (*branch_equality<mode>): Likewise.
23816 (*jump_absolute): Likewise.
23817 (indirect_jump_<mode>): Likewise.
23818 (tablejump_<mode>): Likewise.
23819 (<optab>_internal): Likewise.
23820 (sibcall_internal): Likewise.
23821 (sibcall_value_internal): Likewise.
23822 (prefetch): Use constraint ZD.
23823 * config/mips/mips.opt (minterlink-compressed): New option.
23824 (minterlink-mips16): Now an alias for minterlink-compressed.
23825 (mmicromips): New option.
23826 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23827 (compare_and_swap_12): Likewise.
23828 (sync_add<mode>): Likewise.
23829 (sync_<optab>_12): Likewise.
23830 (sync_old_<optab>_12): Likewise.
23831 (sync_new_<optab>_12): Likewise.
23832 (sync_nand_12): Likewise.
23833 (sync_old_nand_12): Likewise.
23834 (sync_new_nand_12): Likewise.
23835 (sync_sub<mode>): Likewise.
23836 (sync_old_add<mode>): Likewise.
23837 (sync_old_sub<mode>): Likewise.
23838 (sync_new_add<mode>): Likewise.
23839 (sync_new_sub<mode>): Likewise.
23840 (sync_<optab><mode>): Likewise.
23841 (sync_old_<optab><mode>): Likewise.
23842 (sync_new_<optab><mode>): Likewise.
23843 (sync_nand<mode>): Likewise.
23844 (sync_old_nand<mode>): Likewise.
23845 (sync_new_nand<mode>): Likewise.
23846 (sync_lock_test_and_set<mode>): Likewise.
23847 (test_and_set_12): Likewise.
23848 (atomic_compare_and_swap<mode>): Likewise.
23849 (atomic_exchange<mode>_llsc): Likewise.
23850 (atomic_fetch_add<mode>_llsc): Likewise.
23851 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23852 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23853 (umips_save_restore_pattern_p): Likewise.
23854 (umips_load_store_pair_p): Likewise.
23855 (umips_output_load_store_pair): Likewise.
23856 (umips_movep_target_p): Likewise.
23857 (umips_12bit_offset_address_p): Likewise.
23858 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23859 (mips_base_mips16): Rename this...
23860 (mips_base_compression_flags): ...to this. Update all uses.
23861 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23862 (mips_mips16_decl_p): Delete.
23863 (mips_nomips16_decl_p): Delete.
23864 (mips_get_compress_on_flags): New function.
23865 (mips_get_compress_off_flags): New function.
23866 (mips_get_compress_mode): New function.
23867 (mips_get_compress_on_name): New function.
23868 (mips_get_compress_off_name): New function.
23869 (mips_insert_attributes): Support multiple compression types.
23870 (mips_merge_decl_attributes): Likewise.
23871 (umips_12bit_offset_address_p): New function.
23872 (mips_start_function_definition): Emit .set micromips directive.
23873 (mips_call_may_need_jalx_p): New function.
23874 (mips_function_ok_for_sibcall): Add microMIPS support.
23875 (mips_print_operand_punctuation): Support short delay slots and
23877 (umips_swm_mask, umips_swm_encoding): New.
23878 (umips_build_save_restore): New function.
23879 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23880 (was_mips16_p): Remove.
23881 (old_compression_mode): New.
23882 (mips_set_compression_mode): New function.
23883 (mips_set_current_function): Add microMIPS support.
23884 (mips_option_override): Likewise.
23885 (umips_save_restore_pattern_p): New function.
23886 (umips_output_save_restore): New function.
23887 (umips_load_store_pair_p_1): New function.
23888 (umips_load_store_pair_p): New function.
23889 (umips_output_load_store_pair_1): New function.
23890 (umips_output_load_store_pair): New function.
23891 (umips_movep_target_p) New function.
23892 (mips_prepare_pch_save): Add microMIPS support.
23893 * config/mips/mips.h (TARGET_COMPRESSION): New.
23894 (TARGET_CPU_CPP_BUILTINS): Update macro
23895 to use new compression flags and to support microMIPS.
23896 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23897 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23898 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23899 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23900 (ASM_SPEC): Support mmicromips and mno-micromips.
23901 (M16STORE_REG_P): New macro.
23902 (MIPS_CALL): Support TARGET_MICROMIPS.
23903 (MICROMIPS_J): New macro.
23904 (mips_base_mips16): Rename this...
23905 (mips_base_compression_flags): ...to this.
23906 (UMIPS_12BIT_OFFSET_P): New macro.
23907 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23908 (MULTILIB_DIRNAMES): Likewise.
23910 2013-02-25 Tom de Vries <tom@codesourcery.com>
23912 PR rtl-optimization/56131
23913 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23914 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23915 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23917 2013-02-25 Tobias Burnus <burnus@net-b.de>
23919 * doc/invoke.texi (-fsanitize=): Move from optimization
23920 to debugging options.
23922 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23924 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23926 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23927 Alexander Monakov <amonakov@ispras.ru>
23929 PR middle-end/56077
23930 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23931 flush pending lists also on non-jumps. Adjust comment.
23933 2013-02-24 Catherine Moore <clm@codesourcery.com>
23934 Maciej W. Rozycki <macro@codesourcery.com>
23935 Tom de Vries <tom@codesourcery.com>
23936 Nathan Sidwell <nathan@codesourcery.com>
23937 Iain Sandoe <iain@codesourcery.com>
23938 Nathan Froyd <froydnj@codesourcery.com>
23939 Chao-ying Fu <fu@mips.com>
23941 * doc/extend.texi: (micromips, nomicromips, nocompression):
23942 Document new function attributes.
23943 * doc/invoke.texi (minterlink-compressed, mmicromips,
23944 m14k, m14ke, m14kec): Document new options.
23945 (minterlink-mips16): Update documentation.
23946 * doc/md.texi (ZC, ZD): Document new constraints.
23947 * configure.ac (gcc_cv_as_micromips): Check if linker
23948 supports the .set micromips directive.
23949 * configure: Regenerate.
23950 * config.in: Regenerate.
23951 * config/mips/mips-tables.opt: Regenerate.
23952 * config/mips/micromips.md: New file.
23953 * constraints.md (ZC, AD): New constraints.
23954 * config/mips/predicates.md (movep_src_register): New predicate.
23955 (movep_src_operand): New predicate.
23956 (non_volatile_mem_operand): New predicate.
23957 * config/mips/mips.md (multimem): New type.
23958 (length): Differentiate between 17-bit and 18-bit branch offsets.
23959 (MOVEP1, MOVEP2): New mode iterator.
23960 (mov_<load>l): Use ZC constraint.
23961 (mov_<load>r): Likewise.
23962 (mov_<store>l): Likewise.
23963 (mov_<store>r): Likewise.
23964 (*branch_equality<mode>_inverted): Add microMIPS support.
23965 (*branch_equality<mode>): Likewise.
23966 (*jump_absolute): Likewise.
23967 (indirect_jump_<mode>): Likewise.
23968 (tablejump_<mode>): Likewise.
23969 (<optab>_internal): Likewise.
23970 (sibcall_internal): Likewise.
23971 (sibcall_value_internal): Likewise.
23972 (prefetch): Use constraint ZD.
23973 * config/mips/mips.opt (minterlink-compressed): New option.
23974 (minterlink-mips16): Now an alias for minterlink-compressed.
23975 (mmicromips): New option.
23976 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23977 (compare_and_swap_12): Likewise.
23978 (sync_add<mode>): Likewise.
23979 (sync_<optab>_12): Likewise.
23980 (sync_old_<optab>_12): Likewise.
23981 (sync_new_<optab>_12): Likewise.
23982 (sync_nand_12): Likewise.
23983 (sync_old_nand_12): Likewise.
23984 (sync_new_nand_12): Likewise.
23985 (sync_sub<mode>): Likewise.
23986 (sync_old_add<mode>): Likewise.
23987 (sync_old_sub<mode>): Likewise.
23988 (sync_new_add<mode>): Likewise.
23989 (sync_new_sub<mode>): Likewise.
23990 (sync_<optab><mode>): Likewise.
23991 (sync_old_<optab><mode>): Likewise.
23992 (sync_new_<optab><mode>): Likewise.
23993 (sync_nand<mode>): Likewise.
23994 (sync_old_nand<mode>): Likewise.
23995 (sync_new_nand<mode>): Likewise.
23996 (sync_lock_test_and_set<mode>): Likewise.
23997 (test_and_set_12): Likewise.
23998 (atomic_compare_and_swap<mode>): Likewise.
23999 (atomic_exchange<mode>_llsc): Likewise.
24000 (atomic_fetch_add<mode>_llsc): Likewise.
24001 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24002 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24003 (umips_save_restore_pattern_p): Likewise.
24004 (umips_load_store_pair_p): Likewise.
24005 (umips_output_load_store_pair): Likewise.
24006 (umips_movep_target_p): Likewise.
24007 (umips_12bit_offset_address_p): Likewise.
24008 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24009 (mips_base_mips16): Rename this...
24010 (mips_base_compression_flags): ...to this. Update all uses.
24011 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24012 (mips_mips16_decl_p): Delete.
24013 (mips_nomips16_decl_p): Delete.
24014 (mips_get_compress_on_flags): New function.
24015 (mips_get_compress_off_flags): New function.
24016 (mips_get_compress_mode): New function.
24017 (mips_get_compress_on_name): New function.
24018 (mips_get_compress_off_name): New function.
24019 (mips_insert_attributes): Support multiple compression types.
24020 (mips_merge_decl_attributes): Likewise.
24021 (umips_12bit_offset_address_p): New function.
24022 (mips_start_function_definition): Emit .set micromips directive.
24023 (mips_call_may_need_jalx_p): New function.
24024 (mips_function_ok_for_sibcall): Add microMIPS support.
24025 (mips_print_operand_punctuation): Support short delay slots and
24027 (umips_swm_mask, umips_swm_encoding): New.
24028 (umips_build_save_restore): New function.
24029 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24030 (was_mips16_p): Remove.
24031 (old_compression_mode): New.
24032 (mips_set_compression_mode): New function.
24033 (mips_set_current_function): Add microMIPS support.
24034 (mips_option_override): Likewise.
24035 (umips_save_restore_pattern_p): New function.
24036 (umips_output_save_restore): New function.
24037 (umips_load_store_pair_p_1): New function.
24038 (umips_load_store_pair_p): New function.
24039 (umips_output_load_store_pair_1): New function.
24040 (umips_output_load_store_pair): New function.
24041 (umips_movep_target_p) New function.
24042 (mips_prepare_pch_save): Add microMIPS support.
24043 * config/mips/mips.h (TARGET_COMPRESSION): New.
24044 (TARGET_CPU_CPP_BUILTINS): Update macro
24045 to use new compression flags and to support microMIPS.
24046 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24047 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24048 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24049 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24050 (ASM_SPEC): Support mmicromips and mno-micromips.
24051 (M16STORE_REG_P): New macro.
24052 (MIPS_CALL): Support TARGET_MICROMIPS.
24053 (MICROMIPS_J): New macro.
24054 (mips_base_mips16): Rename this...
24055 (mips_base_compression_flags): ...to this.
24056 (UMIPS_12BIT_OFFSET_P): New macro.
24057 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24058 (MULTILIB_DIRNAMES): Likewise.
24060 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24063 * target-globals.c (save_target_globals): For init_reg_sets and
24064 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24066 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24068 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24069 * config/aarch64/t-aarch64
24070 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24072 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24074 PR inline-asm/56148
24075 * lra-constraints.c (process_alt_operands): Reload operand
24076 conflicting with earlier clobber only if no more other conflicting
24079 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24082 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24083 if not linking a shared library.
24085 2013-02-22 Seth LaForge <sethml@google.com>
24087 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24089 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24091 * config/arm/arm.md (split for extendsidi): Update condition.
24092 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24093 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24094 (qhs_zextenddi_cstr): Likewise.
24096 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24098 PR middle-end/56420
24099 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24100 avoid signed wrapping.
24101 (expand_mult): Handle properly multiplication by
24102 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24103 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24104 in the compiler if coeff is HOST_WIDE_INT_MIN.
24105 (expand_divmod): Don't make ext_op1 static, change it's type to
24106 uhwi. Avoid undefined behavior in -INTVAL (op1).
24108 PR rtl-optimization/50339
24109 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24111 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24112 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24113 into splitting_ashiftrt field.
24114 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24116 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24119 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24121 PR middle-end/56108
24122 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24123 are sure to go irrevocable.
24125 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24127 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24128 scalars are valid operands.
24130 2013-02-21 Martin Jambor <mjambor@suse.cz>
24132 PR tree-optimization/56310
24133 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24134 only matching indices and non-negative final offsets.
24135 (intersect_aggregates_with_edge): Pass src_idx to
24136 agg_replacements_to_vector. Pass src_idx insstead of index to
24137 intersect_with_agg_replacements.
24139 2013-02-21 Martin Jambor <mjambor@suse.cz>
24141 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24142 instead of hard-wired defaults.
24144 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24146 * doc/invoke.texi (MIPS Options): Update documentation of the
24147 floating-point multiply-accumulate instruction restrictions.
24149 2013-02-21 Kostya Serebryany <kcc@google.com>
24151 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24152 asan_shadow_offset on x86_64 linux.
24154 2013-02-21 Richard Biener <rguenther@suse.de>
24156 PR tree-optimization/56415
24158 2013-02-11 Richard Biener <rguenther@suse.de>
24160 PR tree-optimization/56273
24161 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24164 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24167 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24170 PR inline-asm/56405
24171 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24172 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24174 2013-02-20 Jan Hubicka <jh@suse.cz>
24176 PR tree-optimization/56265
24177 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24178 when target is referenced for first time.
24180 2013-02-20 Richard Biener <rguenther@suse.de>
24182 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24183 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24184 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24185 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24186 not return anything.
24187 (rename_ssa_copies): Do not remove unused locals.
24188 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24189 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24190 * passes.c (execute_function_todo): Do not schedule unused locals
24191 removal if cleanup_tree_cfg did something.
24192 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24193 about the number of removed locals.
24195 2013-02-20 Richard Biener <rguenther@suse.de>
24197 PR tree-optimization/56398
24198 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24200 2013-02-20 Martin Jambor <mjambor@suse.cz>
24202 PR tree-optimization/55334
24203 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24204 restricted pointers to arrays.
24206 2013-02-20 Richard Biener <rguenther@suse.de>
24207 Jakub Jelinek <jakub@redhat.com>
24209 PR tree-optimization/56396
24210 * tree-ssa-ccp.c (n_const_val): New static variable.
24211 (get_value): Return NULL for SSA names we don't have a lattice
24213 (ccp_initialize): Initialize n_const_val.
24214 * tree-ssa-copy.c (n_copy_of): New static variable.
24215 (init_copy_prop): Initialize n_copy_of.
24216 (get_value): Return NULL_TREE for SSA names we don't have a
24219 2013-02-20 Martin Jambor <mjambor@suse.cz>
24221 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24223 2013-02-20 Richard Biener <rguenther@suse.de>
24225 * genpreds.c (write_lookup_constraint): Do not compare first
24226 letter of the constraint again.
24228 2013-02-20 Richard Biener <rguenther@suse.de>
24230 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24232 (get_use_iv_cost): Terminate hashtable walk when coming across
24235 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24237 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24238 reassociation for avx2 targets.
24240 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24242 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24243 Add version check for v8.10.a to enable microblaze_has_clz
24244 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24245 version and TARGET_PATTERN_COMPARE check
24246 * config/microblaze/microblaze.md: New clzsi2 instruction
24248 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24250 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24251 function before branching.
24253 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24255 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24257 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24259 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24261 PR middle-end/55889
24262 * sel-sched.c: Include ira.h.
24263 (implicit_clobber_conflict_p): New function.
24264 (moveup_expr): Use it.
24265 * Makefile.in (sel-sched.o): Depend on ira.h.
24267 2013-02-19 Richard Biener <rguenther@suse.de>
24269 PR tree-optimization/56384
24270 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24271 (vn_hash_type): Split out from ...
24272 (vn_hash_constant_with_type): ... here.
24273 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24274 (vn_phi_eq): Compare types from vn_phi_s structure.
24275 (vn_phi_lookup): Populate vn_phi_s type.
24276 (vn_phi_insert): Likewise.
24278 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24280 PR tree-optimization/56350
24281 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24282 if haven't found reduction or nested cycle operand, rather than
24283 asserting we must find it.
24285 PR tree-optimization/56381
24286 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24289 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24290 Jakub Jelinek <jakub@redhat.com>
24293 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24294 (swap_optab_enable): Same.
24295 (init_all_optabs): Use argument instead of global.
24296 * tree.h (struct tree_optimization_option): New field target_optabs.
24297 * expr.h (init_all_optabs): Add argument to prototype.
24298 (TREE_OPTIMIZATION_OPTABS): New.
24299 (save_optabs_if_changed): Protoize.
24300 * optabs.h: Declare this_fn_optabs.
24301 * optabs.c (save_optabs_if_changed): New.
24302 Declare this_fn_optabs.
24303 (init_optabs): Add argument to init_all_optabs() call.
24304 * function.c (invoke_set_current_function_hook): Handle per
24306 * function.h (struct function): New field optabs.
24307 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24308 optimization_current_node has changed.
24309 * target-globals.h (save_target_globals_default_opts): Protoize.
24310 * target-globals.c (save_target_globals_default_opts): New.
24312 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24315 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24316 registers %fr12 and %fr12R as call used.
24319 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24320 and HImode, require all displacements to be an integer multiple of
24322 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24323 only allow QImode and HImode when reload is in progress and strict is
24324 true. Likewise for symbolic addresses. Use base14_operand to check
24325 displacements in REG+BASE addresses.
24327 2013-02-18 Richard Biener <rguenther@suse.de>
24329 PR tree-optimization/56366
24330 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24331 handle sign-conversion of outer-loop initial induction value.
24333 2013-02-18 Richard Biener <rguenther@suse.de>
24335 PR middle-end/56349
24336 * cfghooks.c (merge_blocks): If we merge a latch into another
24337 block adjust references to it.
24338 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24339 (verify_loop_structure): Verify that a recorded latch is in fact
24342 2013-02-18 Richard Biener <rguenther@suse.de>
24344 PR tree-optimization/56321
24345 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24346 order SSA name release and virtual operand unlinking.
24348 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24350 * config/microblaze/microblaze.md (save_stack_block): Define.
24351 (restore_stack_block): Likewise.
24353 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24355 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24356 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24357 * config/microblaze/microblaze.c (microblaze_option_override):
24358 Bail out early for PIC modes when target does not support PIC.
24360 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24362 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24363 Replace with a microblaze version.
24364 (microblaze_trampoline_init): Adapt for microblaze.
24365 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24368 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24369 Dodji Seketeli <dodji@redhat.com>
24372 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24373 (instrument_mem_region_access): Do not forget to always put
24374 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24375 0) statement, even for cases where either 'base' or 'base + len'
24376 are not instrumented -- because they have been previously
24377 instrumented. Simplify the logic by putting all the statements
24378 instrument 'base + len' inside a sequence, and then insert that
24379 sequence right before the current insertion point. Then, to
24380 instrument 'base + len', just get an iterator on that statement.
24381 And do not forget to update the pointer to iterator the function
24382 received as argument.
24384 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24386 PR rtl-optimization/56348
24387 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24389 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24391 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24392 (clean_graph_dump_file): Pass base to start_graph_dump.
24394 2013-02-14 Richard Henderson <rth@redhat.com>
24397 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24399 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24401 * collect2-aix.h: Define F_LOADONLY.
24403 2013-02-14 Richard Biener <rguenther@suse.de>
24406 * varasm.c (output_constant_def_1): Get the decl representing
24407 the constant as argument.
24408 (output_constant_def): Wrap output_constant_def_1.
24409 (make_decl_rtl): Use output_constant_def_1 with the decl
24410 representing the constant.
24411 (build_constant_desc): Optionally re-use a decl already
24412 representing the constant.
24413 (tree_output_constant_def): Adjust.
24415 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24418 * asan.c (instrument_builtin_call): Really put the length of the
24419 second source argument into src1_len.
24421 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24423 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24424 argument. If it is false, don't create edge from then_bb to
24426 (insert_if_then_before_iter): Pass true to it.
24427 (build_check_stmt): Pass false to it.
24428 (transform_statements): Flush hash table only on extended basic
24429 block boundaries, rather than at the beginning of every bb.
24430 Don't flush hash table on nonfreeing_call_p calls.
24431 * tree-flow.h (nonfreeing_call_p): New prototype.
24432 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24434 2013-02-13 David S. Miller <davem@davemloft.net>
24436 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24438 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24441 * ira.c (max_regno_before_ira): Move from ...
24443 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24446 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24448 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24450 2013-02-13 Richard Biener <rguenther@suse.de>
24453 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24454 globals in MEM_REFs.
24456 2013-02-13 Richard Biener <rguenther@suse.de>
24458 * loop-init.c (loop_optimizer_init): Clear loop state when
24459 re-initializing preserved loops.
24460 * loop-unswitch.c (unswitch_single_loop): Return whether
24461 we unswitched the loop. Do not verify loop state here.
24462 (unswitch_loops): When we unswitched a loop discover new loops.
24464 2013-02-13 Kostya Serebryany <kcc@google.com>
24466 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24468 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24470 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24472 Avoid instrumenting duplicated memory access in the same basic block
24473 * Makefile.in (asan.o): Add new dependency on hash-table.h
24474 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24475 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24476 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24477 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24478 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24479 (get_mem_ref_of_assignment): New functions.
24480 (get_mem_refs_of_builtin_call): Extract from
24481 instrument_builtin_call and tweak a little bit to make it fit with
24483 (instrument_builtin_call): Use the new
24484 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24485 of is_gimple_builtin_call.
24486 (instrument_derefs, instrument_mem_region_access): Insert the
24487 instrumented memory reference into the hash table.
24488 (maybe_instrument_assignment): Renamed instrument_assignment into
24489 this, and change it to advance the iterator when instrumentation
24490 actually happened and return true in that case. This makes it
24491 homogeneous with maybe_instrument_assignment, and thus give a
24492 chance to callers to be more 'regular'.
24493 (transform_statements): Clear the memory reference hash table
24494 whenever we enter a new BB, when we cross a function call, or when
24495 we are done transforming statements. Use
24496 maybe_instrument_assignment instead of instrumentation. No more
24497 need to special case maybe_instrument_assignment and advance the
24498 iterator after calling it; it's now handled just like
24499 maybe_instrument_call. Update comment.
24501 2013-02-13 Richard Biener <rguenther@suse.de>
24503 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24504 Fix loop discovery code.
24506 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24508 PR inline-asm/56148
24509 * lra-constraints.c (process_alt_operands): Match early clobber
24510 operand with itself. Check conflicts with earlyclobber only if
24511 the operand is not reloaded. Prefer to reload conflicting operand
24512 if earlyclobber and matching operands are the same.
24514 2013-02-12 Richard Biener <rguenther@suse.de>
24517 * lto-streamer-out.c (write_symbol): Do not output symbols
24518 for hard register variables.
24520 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24523 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24524 (umulsidi3_insn, mulsidi3_insn): New insns.
24526 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24528 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24529 (struct tune_params): Add vec_costs field.
24530 * config/arm/arm.c (arm_builtin_vectorization_cost)
24531 (arm_add_stmt_cost): New functions.
24532 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24533 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24534 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24535 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24536 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24537 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24538 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24540 2013-02-12 Richard Biener <rguenther@suse.de>
24543 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24544 decls again if possible.
24546 2013-02-12 Richard Biener <rguenther@suse.de>
24548 PR middle-end/56288
24549 * tree-ssa.c (verify_ssa_name): Fix check, move
24550 SSA_NAME_IN_FREE_LIST check up.
24552 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24553 Steven Bosscher <steven@gcc.gnu.org>
24555 PR rtl-optimization/56151
24556 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24557 equal to op0 or op1, and last_insn pattern is CODE operation
24558 with MEM dest and one of the operands matches that MEM.
24560 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24562 * doc/extend.texi: Document Function Multiversioning and "default"
24563 parameter string to target attribute.
24564 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24565 target attribute parameter is "default".
24566 (ix86_compare_version_priority): Remove checks for target attribute.
24567 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24568 Remove check for target attribute equal to NULL. Add assert.
24569 (ix86_generate_version_dispatcher_body): Change error to sorry.
24571 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24572 Jack Howarth <howarth@bromo.med.uc.edu>
24573 Patrick Marlier <patrick.marlier@gmail.com>
24576 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24577 define ENDFILE_SPEC as TM_DESTRUCTOR.
24578 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24580 2013-02-11 Alexander Potapenko <glider@google.com>
24581 Jack Howarth <howarth@bromo.med.uc.edu>
24582 Jakub Jelinek <jakub@redhat.com>
24585 * config/darwin.c (cdtor_record): Rename ctor_record.
24586 (sort_cdtor_records): Rename sort_ctor_records.
24587 (finalize_dtors): New routine to sort destructors by
24588 priority before use in assemble_integer.
24589 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24591 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24593 PR rtl-optimization/56275
24594 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24595 offset is non-negative and less than cmode size before
24596 calling simplify_subreg.
24598 2013-02-11 Richard Biener <rguenther@suse.de>
24600 PR tree-optimization/56264
24601 * cfgloop.h (fix_loop_structure): Adjust prototype.
24602 * loop-init.c (fix_loop_structure): Return the number of
24603 newly discovered loops.
24604 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24605 are discovered, do a full loop-closed SSA rewrite.
24607 2013-02-11 Richard Biener <rguenther@suse.de>
24609 PR tree-optimization/56273
24610 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24612 (check_array_ref): Fix missing newline in dumps.
24613 (search_for_addr_array): Likewise.
24615 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24617 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24619 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24622 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24624 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24626 PR rtl-optimization/56246
24627 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24629 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24630 constraints are satisfied.
24632 2013-02-08 Jeff Law <law@redhat.com>
24635 * emit-rtl.c (reg_is_parm_p): New function.
24636 * regs.h (reg_is_parm_p): New prototype.
24637 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24638 callee-clobbered registers.
24640 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24643 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24644 If there is no implicit builtin declaration, just return NULL.
24646 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24648 * config/i386/sse.md (FMAMODEM): New mode iterator.
24649 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24650 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24652 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24654 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24655 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24656 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24658 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24660 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24661 (microblaze*-*-elf): Likewise.
24662 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24664 * config/microblaze/microblaze-c.c: Add builtin defines for
24665 _LITTLE_ENDIAN and _BIG_ENDIAN.
24666 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24667 add to TARGET_DEFAULT flags.
24668 Expand ASM_SPEC and LINK_SPEC.
24669 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24670 * config/microblaze/microblaze.md: Update extendsidi2 and
24671 movdi_internal instructions to use low-order / high-order reg
24673 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24674 options and inversemask / mask of LITTLE_ENDIAN.
24675 * config/microblaze/t-microblaze: Expand multilib options to
24676 include mlittle-endian (le) and update exceptions patterns.
24678 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24680 PR rtl-optimization/56195
24681 * lra-constraints.c (get_reload_reg): Don't reuse regs
24682 if they have smaller mode than requested, if they have
24683 wider mode than requested, try to return a SUBREG.
24685 PR tree-optimization/56250
24686 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24687 if type is unsigned and code isn't MULT_EXPR.
24689 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24691 PR tree-optimization/56064
24692 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24693 bits according to mode.
24694 * fixed-value.h (fixed_from_double_int)
24695 (const_fixed_from_double_int): Adjust comments.
24697 2013-02-08 Richard Biener <rguenther@suse.de>
24700 * lto-streamer.h (struct data_in): Remove current_file, current_line
24701 and current_col members.
24702 * lto-streamer-out.c (lto_output_location): Stream changed bits
24703 en-block for efficiency.
24704 * lto-streamer-in.c (clear_line_info): Remove.
24705 (lto_input_location): Cache current file, line and column
24706 globally via local statics. Read changed bits en-block.
24707 (input_function): Do not call clear_line_info.
24708 (lto_read_body): Likewise.
24709 (lto_input_toplevel_asms): Likewise.
24711 2013-02-08 Michael Matz <matz@suse.de>
24713 PR tree-optimization/52448
24714 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24715 (nt_call_phase): New static.
24716 (add_or_mark_expr): Only mark accesses with newer phase than any
24718 (nonfreeing_call_p): New.
24719 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24720 (nt_fini_block): Keep blocks marked as visited.
24721 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24723 2013-02-08 Richard Biener <rguenther@suse.de>
24725 * ira.c (ira): Free broken dominator information.
24727 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24729 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24731 2013-02-08 Marek Polacek <polacek@redhat.com>
24733 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24735 2013-02-08 Richard Biener <rguenther@suse.de>
24737 PR middle-end/56181
24738 * cfgloop.h (flow_loops_find): Adjust.
24739 (bb_loop_header_p): Declare.
24740 * cfgloop.c (bb_loop_header_p): New function split out from ...
24741 (flow_loops_find): ... here. Adjust function signature,
24742 support incremental loop structure update.
24743 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24744 * cfgloopmanip.c (fix_loop_structure): Move ...
24745 * loop-init.c (fix_loop_structure): ... here.
24746 (apply_loop_flags): Split out from ...
24747 (loop_optimizer_init): ... here.
24748 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24749 in incremental mode, only remove dead loops here.
24751 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24754 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24755 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24756 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24757 (*round<mode>3.libgcc): New insns for fixed-modes.
24758 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24759 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24760 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24761 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24762 implementations. Define to __builtin_avr_absFX,
24763 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24764 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24765 __builtin_avr_countlsFX, respectively.
24766 * config/avr/avr-c.c (target.h): Include it.
24767 (enum avr_builtin_id): New enum.
24768 (avr_resolve_overloaded_builtin): New static function.
24769 (avr_register_target_pragmas): Use it to set
24770 targetm.resolve_overloaded_builtin.
24771 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24772 tree nodes used by DEF_BUILTIN.
24773 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24774 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24775 <AVR_BUILTIN_xxBITS>: Same.
24777 2013-02-08 Richard Biener <rguenther@suse.de>
24779 * cfgloop.c (verify_loop_structure): Properly handle
24780 a loop exiting to another loop header.
24781 * ira-int.h (ira_loops): Remove.
24782 * ira.c (ira_loops): Remove.
24783 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24784 (do_reload): Use loop_optimizer_finalize.
24785 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24786 number_of_loops to access the loop tree.
24787 (more_one_region_p): Likewise.
24788 (finish_loop_tree_nodes): Likewise.
24789 (rebuild_regno_allocno_maps): Likewise.
24790 (mark_loops_for_removal): Likewise.
24791 (mark_all_loops_for_removal): Likewise.
24792 (remove_unnecessary_regions): Likewise.
24793 (ira_build): Likewise.
24794 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24796 2013-02-08 Richard Biener <rguenther@suse.de>
24798 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24799 * ipa-pure-const.c (analyze_function): Avoid calling
24800 mark_irreducible_loops twice.
24801 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24803 2013-02-07 David S. Miller <davem@davemloft.net>
24805 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24807 * var-tracking.c (vt_add_function_parameter): Test the presence of
24808 HAVE_window_save properly and do not remap argument registers when
24809 we have a leaf function.
24811 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24814 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24816 * config/i386/i386.c (ix86_print_operand): Ditto.
24818 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24820 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24822 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24824 PR rtl-optimization/56225
24825 * lra-constraints.c (process_alt_operands): Check that reload hard
24826 reg can hold value for strict_low_part.
24828 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24831 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24832 dwarf2out_end_function.
24833 (in_first_function_p, maybe_at_text_label_p,
24834 first_loclabel_num_not_at_text_label): New variables.
24835 (dwarf2out_var_location): In the first function find out
24836 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24837 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24840 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24842 PR rtl-optimization/56178
24843 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24844 SUBREG of a register. Tidy up related block of code.
24845 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24846 note if the source is a register or a SUBREG of a register.
24848 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24851 * config/rs6000/rs6000.md (ptrm): New mode attr.
24852 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24853 call_value_indirect_aix<pttrsize>,
24854 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24857 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24859 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24860 if -bnortl. Convert to strcmp and strncmp.
24862 2013-02-07 Alan Modra <amodra@gmail.com>
24865 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24866 addresses won't wrap when offsetting.
24867 (rs6000_secondary_reload): Provide secondary reloads needed for
24868 wrapping LO_SUM addresses.
24870 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24872 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24873 MACH, just __MACH__.
24875 2013-02-06 Richard Biener <rguenther@suse.de>
24877 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24878 instead of calling fix_loop_structure.
24880 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24882 PR middle-end/56217
24883 * omp-low.c (use_pointer_for_field): Return false if
24884 lower_send_shared_vars doesn't generate any copy-out code.
24886 2013-02-06 Tom de Vries <tom@codesourcery.com>
24888 PR rtl-optimization/56131
24889 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24890 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24891 of the label is NULL. Add comment.
24893 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24895 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24898 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24899 (STATIC_LIBTSAN_LIBS): Likewise.
24900 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24901 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24902 is defined, don't add anything else beyond that.
24903 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24904 (LINK_COMMAND_SPEC): Use them.
24906 PR tree-optimization/56205
24907 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24908 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24909 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24911 2013-02-05 Richard Biener <rguenther@suse.de>
24913 PR tree-optimization/53342
24914 PR tree-optimization/53185
24915 * tree-vectorizer.h (vect_check_strided_load): Remove.
24916 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24917 not disallow peeling for vectorized strided loads.
24918 (vect_check_strided_load): Make static and simplify.
24919 (vect_analyze_data_refs): Adjust.
24920 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24921 correctly when vectorizing strided loads.
24923 2013-02-05 Richard Biener <rguenther@suse.de>
24925 * doc/install.texi: Refer to ISL, not PPL.
24927 2013-02-05 Jan Hubicka <jh@suse.cz>
24929 PR tree-optimization/55789
24930 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24932 2013-02-05 Jan Hubicka <jh@suse.cz>
24934 PR tree-optimization/55789
24935 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24936 the dead call anyway.
24938 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24941 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24943 2013-02-04 Alexander Potapenko <glider@google.com>
24944 Jack Howarth <howarth@bromo.med.uc.edu>
24945 Jakub Jelinek <jakub@redhat.com>
24948 * config/darwin.c (sort_ctor_records): Stabilized qsort
24949 on constructor priority by using original position.
24950 (finalize_ctors): New routine to sort constructors by
24951 priority before use in assemble_integer.
24952 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24954 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24957 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24958 about visibility on artificial decls.
24959 * config/sol2.c (solaris_assemble_visibility): Likewise.
24961 2013-02-04 Kai Tietz <ktietz@redhat.com>
24964 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24965 argument and improve checking of return-argument types for 16-byte
24967 (ix86_function_value_1): Add additional valtype argument on call
24968 of function_value_64.
24969 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24970 handling infunction_value_64 function.
24972 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24974 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24976 2013-02-04 Richard Biener <rguenther@suse.de>
24978 PR tree-optimization/56188
24979 * tree-ssa-structalias.c (label_visit): Consider case with
24980 initially non-empty points-to set.
24981 (perform_var_substitution): Dump node mapping and clean up.
24983 2013-02-04 Richard Guenther <rguenther@suse.de>
24986 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24987 node prevail as last resort.
24988 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24989 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24991 2013-02-04 Richard Biener <rguenther@suse.de>
24993 PR tree-optimization/56113
24994 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24996 (equiv_class_lookup_or_add): ... this.
24997 (label_visit): Adjust and fix error in previous patch.
24998 (perform_var_substitution): Adjust.
25000 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25002 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25003 * config/sh/sh4-300.md: Likewise.
25004 * config/sh/sh4a.md: Likewise.
25005 * config/sh/constraints.md: Likewise.
25006 * config/sh/sh.md: Likewise.
25007 * config/sh/netbsd-elf.h: Likewise.
25008 * config/sh/predicates.md: Likewise.
25009 * config/sh/sh-protos.h: Likewise.
25010 * config/sh/ushmedia.h: Likewise.
25011 * config/sh/linux.h: Likewise.
25012 * config/sh/sh.c: Likewise.
25013 * config/sh/superh.h: Likewise.
25014 * config/sh/elf.h: Likewise.
25015 * config/sh/sh4.md: Likewise.
25016 * config/sh/sh.h: Likewise.
25018 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25020 * config/pa/constraints.md: Adjust unused letters. Change "T"
25021 constraint to match_test floating_point_store_memory_operand().
25022 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25023 (base14_operand): New.
25024 (floating_point_store_memory_operand): New.
25025 (integer_store_memory_operand): Revise to use base14_operand and
25026 reg_plus_base_memory_operand.
25027 (move_dest_operand): Allow symbolic_memory_operands.
25028 (symbolic_memory_operand): Check for LO_SOM.
25029 (symbolic_operand): Change default case to break.
25030 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25031 CONST_DOUBLE values to be reloaded by putting them into memory when
25032 the destination is a floating point register.
25033 (movdf): Remove code to handle CONST_DOUBLE.
25035 (reload_indf_r1): New.
25036 (reload_insf_r1): New.
25037 Consistently use "Q" and "T" constraints with integer and floating
25038 point move instructions, respectively.
25039 (movdi): Remove FAIL.
25040 Change predicate for source operand unamed DImode move from
25041 general_operand to move_src_operand.
25042 (umulsidi3): Change predicate for destination operand to
25044 Likewise for similar unamed patterns.
25045 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25046 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25047 (hppa_legitimize_address): Simplify mask calculation.
25048 (pa_emit_move_sequence): Revised handling of secondary reloads from
25049 REG+D addresses for floating point loads and stores. Directly handle
25050 loading CONST0_RTX (mode) to a floating point register.
25051 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25052 to floating point registers. Don't restrict secondary reloads to
25053 floating point registers to integer modes. Revise some comments and
25055 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25056 (pa_legitimate_address_p): New.
25057 (pa_legitimize_reload_address): New.
25058 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25059 (STRICT_REG_OK_FOR_BASE_P): New.
25060 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25061 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25063 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25064 Andrew Dixie <andrewd@gentrack.com>
25066 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25069 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25071 * expmed.c (extract_bit_field_1): Pass the full width of the
25072 structure to get_best_reg_extraction_insn.
25074 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25077 * configure.ac (use_cxa_atexit): Add AIX.
25078 * configure: Regenerate.
25080 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25082 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25085 * final.c (need_profile_function): New variable.
25086 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25087 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25088 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25089 notes, targetm.asm_out.function_prologue doesn't emit anything,
25090 HAVE_prologue and profiler should be emitted before prologue,
25091 set need_profile_function instead of emitting it.
25092 (final_scan_insn): If need_profile_function, emit
25093 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25094 NOTE_INSN_FUNCTION_BEG note.
25096 2013-02-01 Richard Henderson <rth@redhat.com>
25098 * config/rs6000/rs6000.md (smulditi3): New.
25101 * config/alpha/alpha.md (umulditi3): New.
25103 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25105 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25106 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25108 2013-02-01 Richard Biener <rguenther@suse.de>
25110 PR tree-optimization/56113
25111 * tree-ssa-structalias.c (label_visit): Reduce work for
25112 single-predecessor nodes.
25114 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25116 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25117 range isn't testing for zero.
25119 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25121 PR middle-end/56113
25122 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25124 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25125 Nick Clifton <nickc@redhat.com>
25127 * config/v850/constraints.md (Q): Define as a memory constraint.
25128 * config/v850/predicates.md (label_ref_operand): New predicate.
25129 (e3v5_shift_operand): New predicate.
25130 (ior_operator): New predicate.
25131 * config/v850/t-v850: Add e3v5 multilib.
25132 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25133 (v850_gen_movdi): Prototype.
25134 * config/v850/v850.c: Add support for e3v5 architecture.
25135 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25137 (construct_save_jarl): Add e3v5 long JARL support.
25138 (v850_adjust_insn_length): New function. Adjust length of call
25139 insns when using e3v5 instructions.
25140 (v850_gen_movdi): New function: Generate instructions to move a
25142 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25143 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25144 (TARGET_USE_FPU): Enable for e3v5.
25145 (CONST_OK_FOR_W): New macro.
25146 (ADJUST_INSN_LENGTH): Define.
25147 * config/v850/v850.md (UNSPEC_LOOP): Define.
25148 (attr cpu): Add v850e3v5.
25149 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25150 (movdi): New pattern.
25151 (movdi_internal): New pattern.
25152 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25153 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25154 (cstoresf4): Likewise.
25155 (cstoredf4): Likewise.
25156 (insv): New pattern.
25157 (rotlso3_a): New pattern.
25158 (rotlsi3_b): New pattern
25159 (rotlsi3_v850e3v5): New pattern.
25160 (doloop_begin): New pattern.
25161 (fix_loop_counter): New pattern.
25162 (doloop_end): New pattern.
25163 (branch_normal): Add e3v5 long branch support.
25164 (branch_invert): Likewise.
25165 (branch_z_normal): Likewise.
25166 (branch_z_invert): Likewise.
25167 (branch_nz_normal): Likewise.
25168 (branch_nz_invert): Likewise.
25169 (call_internal_short): Add e3v5 register-indirect JARL support.
25170 (call_internal_long): Likewise.
25171 (call_value_internal_short): Likewise.
25172 (call_value_internal_long): Likewise.
25173 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25174 (mloop): New option.
25175 * config.gcc: Add support for configuring v840e3v5 target.
25176 * doc/invoke.texi: Document new v850 specific command line options.
25178 2013-01-31 Paul Koning <ni1d@arrl.net>
25182 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25183 children if parent is a class.
25184 (prune_unused_types_prune): Don't add DW_AT_declaration.
25186 2013-01-31 Richard Biener <rguenther@suse.de>
25188 PR tree-optimization/56157
25189 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25190 match up operand with SLP child.
25192 2013-01-31 Jason Merrill <jason@redhat.com>
25195 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25196 parameters the first time.
25197 (gen_scheduled_generic_parms_dies): Check completeness here.
25199 2013-01-31 Richard Biener <rguenther@suse.de>
25201 PR middle-end/53073
25202 * common.opt (faggressive-loop-optimizations): New flag,
25203 enabled by default.
25204 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25205 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25206 infer_loop_bounds_from_undefined by it.
25208 2013-01-31 Richard Biener <rguenther@suse.de>
25210 PR tree-optimization/56150
25211 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25212 visit virtual operands.
25213 (find_uses_to_rename_bb): Likewise.
25215 2013-01-31 Richard Biener <rguenther@suse.de>
25217 PR tree-optimization/56150
25218 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25219 mixed store non-store stmts.
25221 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25224 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25225 LIBASAN_EARLY_SPEC is defined.
25226 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25227 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25229 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25232 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25233 invalid args instead of ICEing on it.
25234 (ix86_valid_target_attribute_tree): Return error_mark_node if
25235 ix86_valid_target_attribute_inner_p failed.
25236 (ix86_valid_target_attribute_p): Return false only if
25237 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25238 target("default") attribute.
25239 (sorted_attr_string): Change argument from const char * to tree,
25240 merge in all target attribute arguments rather than just one.
25241 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25242 instead of free. Avoid using strcat.
25243 (ix86_mangle_function_version_assembler_name): Mangle
25244 target("default") as if no target attribute is present. Adjust
25245 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25246 instead of xmalloc and XDELETEVEC instead of free.
25247 (ix86_function_versions): Don't return true if one of the decls
25248 doesn't have target attribute. If they don't and one of the decls
25249 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25250 sorted_attr_string caller. Use XDELETEVEC instead of free.
25251 (ix86_supports_function_versions): Remove.
25252 (make_name): Fix up formatting.
25253 (make_dispatcher_decl): Remove resolver_name and its initialization.
25254 Avoid leaking memory.
25255 (is_function_default_version): Return true if there is
25256 target("default") attribute rather than no target attribute at all.
25257 (make_resolver_func): Avoid leaking memory.
25258 (ix86_generate_version_dispatcher_body): Likewise.
25259 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25260 * target.def (supports_function_versions): Remove.
25261 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25262 * doc/tm.texi: Regenerated.
25264 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25266 PR rtl-optimization/56144
25267 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25268 for values with side effects.
25270 2013-01-30 Richard Biener <rguenther@suse.de>
25272 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25273 (sparseset_pop): Likewise.
25274 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25275 to be able to use quick_push in the worker loop.
25277 2013-01-30 Marek Polacek <polacek@redhat.com>
25279 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25281 2013-01-30 Richard Biener <rguenther@suse.de>
25284 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25286 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25288 PR tree-optimization/56064
25289 * fixed-value.c (fixed_from_double_int): New function.
25290 * fixed-value.h (fixed_from_double_int): New prototype.
25291 (const_fixed_from_double_int): New static inline function.
25292 * fold-const.c (native_interpret_fixed): New static function.
25293 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25294 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25295 (native_encode_fixed): New static function.
25296 (native_encode_expr) <FIXED_CST>: Use it.
25297 (native_interpret_int): Move double_int worker code to...
25298 * double-int.c (double_int::from_buffer): ...this new static method.
25299 * double-int.h (double_int::from_buffer): Prototype it.
25301 2013-01-30 Richard Biener <rguenther@suse.de>
25303 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25304 New pointer-map and obstack.
25305 (init_alias_vars): Allocate pointer-map and obstack.
25306 (delete_points_to_sets): Free them.
25307 (find_what_var_points_to): Cache result.
25308 (find_what_p_points_to): Adjust for changed interface of
25309 find_what_var_points_to.
25310 (compute_points_to_sets): Likewise.
25311 (ipa_pta_execute): Likewise.
25313 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25315 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25316 * configure: Regenerate.
25317 * config.in: Regenerate.
25318 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25319 #nobits/#progbits if supported.
25321 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25324 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25325 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25326 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25328 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25330 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25331 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25333 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25335 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25336 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25338 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25340 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25342 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25343 * config/arm/cortex-a7.md: New bypasses using
25344 arm_mac_accumulator_is_result.
25346 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25348 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25349 (cortex_a7_neon_mla): Likewise.
25350 (cortex_a7_fpfmad): New reservation.
25351 (cortex_a7_fpmacs): Use ffmas and update required units.
25352 (cortex_a7_fpmuld): Update required units and latency.
25353 (cortex_a7_fpmacd): Likewise.
25354 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25355 (cortex_a7_neon). Likewise.
25356 (bypass) Update participating units.
25358 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25360 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25361 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25363 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25364 (vfp_fmul): Use ffmad.
25365 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25366 (cortex_r4_fmacd): Use ffmad.
25367 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25368 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25369 (cortex_a9_fmacd): Use ffmad.
25370 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25371 (cortex_a8_vfp_macd): Use ffmad.
25372 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25373 (cortex_a5_fpmacd): Use ffmad.
25374 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25375 (cortex_a15_vfp_macd): Use ffmad.
25376 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25378 2013-01-29 Jason Merrill <jason@redhat.com>
25381 * varasm.c (default_assemble_visibility): Don't warn about
25382 visibility on artificial decls.
25384 2013-01-29 Richard Biener <rguenther@suse.de>
25386 PR tree-optimization/56113
25387 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25389 (label_visit): Free duplicate bitmaps and record the leader instead.
25390 (perform_var_substitution): Adjust.
25392 2013-01-29 Richard Biener <rguenther@suse.de>
25394 PR tree-optimization/55270
25395 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25396 the CFG, schedule loops for fixup.
25398 2013-01-29 Nick Clifton <nickc@redhat.com>
25400 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25403 2013-01-28 Leif Ekblad <leif@rdos.net>
25405 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25406 * config/i386/i386.h (TARGET_RDOS): New macro.
25407 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25408 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25409 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25410 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25411 DEFAULT_LARGE_SECTION_THRESHOLD.
25412 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25413 * config/i386/rdos.h: New file.
25414 * config/i386/rdos64.h: New file.
25416 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25419 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25422 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25424 PR rtl-optimization/56117
25425 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25426 call cselib_lookup_from_insn on the MEM before calling
25427 add_insn_mem_dependence.
25429 2013-01-28 Richard Biener <rguenther@suse.de>
25431 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25432 to a stmt that didn't have one.
25433 (copy_phis_for_bb): Likewise for PHI arguments.
25434 (copy_debug_stmt): Likewise for debug stmts.
25436 2013-01-28 Richard Biener <rguenther@suse.de>
25438 PR tree-optimization/56034
25439 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25440 (partition_builtin_p): Adjust.
25441 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25442 it is the last partition.
25443 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25444 up the vertex for the definition.
25445 (classify_partition): Classify whether a partition is a
25446 PKIND_REDUCTION, thus has uses outside of the loop.
25447 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25448 Merge all PKIND_REDUCTION partitions into the last partition.
25449 (tree_loop_distribution): Seed partitions from reductions as well.
25451 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25453 PR tree-optimization/56125
25454 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25455 pow(x,c) into sqrt(x) * powi(x, n/2) or
25456 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25457 optimizing for size.
25458 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25459 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25462 PR tree-optimization/56094
25463 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25464 to UNKNOWN_LOCATION while gimplifying expr.
25466 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25469 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25470 operand 0 in movabs insn template for -masm=intel asm alternative.
25471 (*movabs<mode>_2): Ditto for operand 1.
25473 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25476 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25479 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25481 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25484 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25486 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25487 Undef to avoid warning.
25489 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25491 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25492 * configure: Regenerate.
25494 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25496 PR tree-optimization/56098
25497 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25498 for stmts with volatile ops.
25499 (cond_store_replacement): Don't optimize if assign has volatile ops.
25500 (cond_if_else_store_replacement_1): Don't optimize if either
25501 then_assign or else_assign have volatile ops.
25502 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25505 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25507 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25509 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25511 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25512 missing ':' in asm example.
25514 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25516 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25517 entries into lane and laneq entries.
25518 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25519 Remove AdvSIMD scalar modes.
25520 (aarch64_sq<r>dmulh_laneq<mode>): New.
25521 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25523 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25524 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25525 * config/aarch64/iterators.md (VCOND): New.
25528 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25531 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25532 Add NULL LIBNAME argument to existing definitions.
25533 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25534 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25535 * config/avr/avr.c (DEF_BUILTIN): Same.
25536 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25537 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25538 implementation is available (DECL_ASSEMBLER_NAME is set).
25539 (avr_fold_absfx): New static function.
25540 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25541 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25542 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25543 AVR_BUILTIN_ABSLLK.
25544 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25545 (abshk, absk, abslk, absllk): Provide as static inline functions.
25547 2013-01-25 Marek Polacek <polacek@redhat.com>
25549 PR tree-optimization/56035
25550 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25552 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25554 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25555 (*movtf_internal_rex64): Add (!o,C) alternative
25556 (*movxf_internal_rex64): Ditto.
25557 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25559 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25561 * doc/invoke.texi: fix typo.
25562 * doc/objc.texi: fix typo.
25564 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25566 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25567 for the first two alternatives.
25569 2013-01-24 Diego Novillo <dnovillo@google.com>
25571 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25572 (ggc-zone.o): Remove.
25573 * configure.ac: Remove option --with-gc.
25574 * configure: Re-generate.
25575 * doc/install.texi: Remove documentation for --with-gc.
25576 * gengtype.c (write_enum_defn): Remove. Update all users.
25577 (write_Types_process_field): Remove generation of gt_e_* argument.
25578 (output_type_enum): Remove. Update all users.
25579 (write_enum_defn): Remove. Update all users.
25580 (enum alloc_zone): Remove. Update all users.
25581 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25582 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25583 Update all callers.
25584 (struct ptr_data): Remove field TYPE. Update all users.
25585 (gt_pch_note_object): Remove argument TYPE. Update all users.
25586 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25588 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25589 (struct alloc_zone): Remove.
25590 (ggc_internal_alloc_zone_stat): Remove.
25591 (ggc_internal_cleared_alloc_zone_stat): Remove.
25592 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25593 (ggc_pch_count_object): Remove last argument. Update all users.
25594 (ggc_pch_alloc_object): Remove last argument. Update all users.
25595 (struct alloc_zone): Remove.
25596 * ggc-zone.c: Remove.
25597 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25598 (struct alloc_zone): Remove.
25599 (ggc_alloc_typed_stat): Remove.
25600 (ggc_alloc_typed): Remove.
25601 (ggc_splay_alloc): Remove first argument.
25602 (rtl_zone): Remove. Update all users.
25603 (tree_zone): Remove. Update all users.
25604 (tree_id_zone): Remove. Update all users.
25605 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25606 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25607 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25608 * tree-ssanames.c: Remove references to zone allocator in comments.
25610 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25612 * config/avr/avr.c (avr_out_fract): Make register numbers that
25613 might be outside of source operand signed.
25615 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25617 * config/i386/constraints.md (Yf): New constraint.
25618 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25619 of f constraint to conditionaly disable x87 register preferences.
25620 (*movdf_internal): Ditto.
25621 (*movsf_internal): Ditto.
25623 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25625 PR inline-asm/55934
25626 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25627 that have operands with impossible constraints.
25628 Add a FIXME for a speed-up opportunity.
25629 * lra-constraints.c (process_alt_operands): Verify that a class
25630 selected from constraints on asms is valid for the operand mode.
25631 (curr_insn_transform): Remove incorrect comment.
25633 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25635 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25636 TOC operand is a valid symbol ref in the constant pool.
25638 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25640 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25642 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25645 * config/avr/stdfix.h: New file.
25646 * t-avr (stdfix-gcc.h): New rule to build it.
25647 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25649 2013-01-23 Kostya Serebryany <kcc@google.com>
25651 * config/darwin.h: remove dependency on
25652 CoreFoundation (asan on Mac OS).
25654 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25657 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25658 instead of cmpdi_operand for first comparison operand.
25659 Don't assert that comparison operands aren't both constants.
25661 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25663 * doc/install.texi (Downloading the Source): Update references to
25664 downloading separate components.
25666 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25668 * doc/extend.texi (__int128): Improve grammar.
25670 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25673 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25674 alternative to (o,r).
25675 (*movdi_internal_rex64): Remove (!o,n) alternative.
25676 (DImode immediate->memory splitter): Remove.
25677 (DImode immediate->memory peephole2): Remove.
25678 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25679 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25680 alternative to (!o,*r).
25681 (*movtf_internal_sse): New pattern.
25682 (*movxf_internal_rex64): New pattern.
25683 (*movxf_internal): Disable for TARGET_64BIT.
25684 (*movdf_internal_rex64): Remove (!o,F) alternative.
25686 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25688 PR middle-end/56074
25689 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25690 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25691 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25692 stmt locations where LOCATION_LOCUS of the stmt location is
25693 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25696 * config/i386/i386.md (UNSPEC_STOS): New.
25697 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25698 *strsetqi_1): Add UNSPEC_STOS.
25700 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25703 * doc/invoke.texi: Remove left over -Wsynth example.
25705 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25707 PR tree-optimization/56051
25708 * fold-const.c (fold_binary_loc): Don't fold
25709 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25710 a narrowing conversion, or widening conversion from signed
25713 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25715 PR rtl-optimization/56023
25716 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25717 dependent on debug instruction.
25719 2013-01-21 Martin Jambor <mjambor@suse.cz>
25721 PR middle-end/56022
25722 * function.c (allocate_struct_function): Call
25723 invoke_set_current_function_hook earlier.
25725 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25727 * reload1.c (init_reload): Only initialize reload_obstack
25728 during the first call.
25730 2013-01-21 Marek Polacek <polacek@redhat.com>
25732 * cfgloop.c (verify_loop_structure): Fix up grammar.
25734 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25736 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25737 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25739 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25742 * config/arm/marvell-pj4.md: Update copyright year.
25743 Fix up use of alu to alu_reg and simple_alu_imm.
25745 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25747 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25749 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25752 * lra-constraints.c (curr_insn_transform): Don't reuse original
25753 insn for secondary memory move when memory mode should be different.
25755 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25757 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25758 atomic_storedi_1): New patterns.
25760 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25762 btver2 pipeline descriptions.
25763 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25765 * config/i386/i386.md (btver2_decode): New type attributes.
25766 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25768 * config/i386/btver2.md: New file describing btver2 pipelines.
25770 2013-01-19 Andrew Pinski <apinski@cavium.com>
25772 PR tree-optimization/52631
25773 * tree-ssa-sccvn (visit_use): Before looking up the original
25774 statement, try looking up the simplified expression.
25776 2013-01-19 Anthony Green <green@moxielogic.com>
25778 * config/moxie/moxie.c (moxie_expand_prologue): Set
25779 current_function_static_stack_size.
25781 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25783 PR tree-optimization/56029
25784 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25785 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25787 2013-01-18 Sharad Singhai <singhai@google.com>
25789 PR tree-optimization/55995
25790 * dumpfile.c (dump_loc): Print location only if available.
25791 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25793 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25796 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25797 secondary memory move.
25798 (inherit_reload_reg): Use rclass instead of cl for
25799 check_secondary_memory_needed_p.
25801 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25803 PR middle-end/56015
25804 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25805 the case where writing real complex part of target modifies op1.
25807 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25809 * config/aarch64/aarch64-simd.md
25810 (aarch64_vcond_internal<mode>): Handle unordered cases.
25811 * config/aarch64/iterators.md (v_cmp_result): New.
25813 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25814 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25816 * config/arm/marvell-pj4.md: New file.
25817 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25818 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25819 (generic_vfp): Likewise.
25820 * config/arm/arm-cores.def: Add marvell-pj4.
25821 * config/arm/arm-tune.md: Regenerate.
25822 * config/arm/arm-tables.opt: Regenerate.
25823 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25824 * doc/invoke.texi: Document marvell-pj4.
25826 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25828 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25830 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25835 * var-tracking.c (negative_power_of_two_p): New.
25836 (global_get_addr_cache, local_get_addr_cache): New.
25837 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25838 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25840 (vt_stack_offset_p): Remove.
25841 (vt_canon_true_dep): Always canonicalize loc's address.
25842 (clobber_overlapping_mems): Make sure we have a MEM.
25843 (local_get_addr_clear_given_value): New.
25844 (val_reset): Clear local cached entries.
25845 (compute_bb_dataflow): Create and release the local cache.
25846 Disable duplicate MEMs clobbering.
25847 (emit_notes_in_bb): Clobber MEMs likewise.
25848 (vt_emit_notes): Create and release the local cache.
25849 (vt_initialize, vt_finalize): Create and release the global
25850 cache, respectively.
25851 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25853 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25855 PR libmudflap/53359
25856 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25857 not found in the symtab.
25859 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25862 PR rtl-optimization/55547
25863 PR rtl-optimization/53827
25866 * alias.c (offset_overlap_p): New, factored out of...
25867 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25868 the conservative special case for symbolic constants. Don't
25869 adjust zero sizes on alignment.
25871 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25873 PR rtl-optimization/52573
25874 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25875 REG_UNUSED for the same register.
25877 2013-01-17 Richard Biener <rguenther@suse.de>
25878 Marek Polacek <polacek@redhat.com>
25880 PR rtl-optimization/55833
25881 * loop-unswitch.c (unswitch_loops): Move loop verification...
25882 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25883 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25884 Set it to true when we're removing a loop from hierarchy tree in
25885 an irreducible region.
25886 (fix_bb_placements): Adjust caller.
25887 (fix_loop_placements): Likewise.
25889 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25891 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25892 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25893 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25894 Remove ID. Adjust comments.
25895 * config/avr/avr-c.c (avr_builtin_name): Remove.
25896 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25897 * config/avr/avr.c (avr_tolower): New static function.
25898 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25899 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25900 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25903 2013-01-17 Jan Hubicka <jh@suse.cz>
25905 PR tree-optimization/55273
25906 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25908 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25911 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25912 store through atomic_store<mode>_1.
25913 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25915 2013-01-17 Martin Jambor <mjambor@suse.cz>
25917 PR tree-optimizations/55264
25918 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25919 for virtual methods.
25920 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25921 virtual methods before inlining is over.
25922 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25924 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25927 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25929 PR rtl-optimization/56005
25930 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25931 pending reads for prefetch.
25933 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25935 * config/aarch64/aarch64.md
25936 (*cstoresi_neg_uxtw): New pattern.
25937 (*cmovsi_insn_uxtw): New pattern.
25938 (*<optab>si3_uxtw): New pattern.
25939 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25940 (*<optab>si3_insn_uxtw): New pattern.
25941 (*bswapsi2_uxtw): New pattern.
25943 2013-01-16 Richard Biener <rguenther@suse.de>
25945 * tree-inline.c (tree_function_versioning): Remove set but
25946 never used variable.
25948 2013-01-16 Richard Biener <rguenther@suse.de>
25950 PR tree-optimization/55964
25951 * tree-flow.h (rename_variables_in_loop): Remove.
25952 (rename_variables_in_bb): Likewise.
25953 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25954 (copy_loop_before): Adjust and delete update-ssa status.
25955 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25956 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25957 (rename_variables_in_loop): Remove.
25958 (slpeel_update_phis_for_duplicate_loop): Likewise.
25959 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25960 use available cfg machinery instead of duplicating it.
25961 Update PHI nodes and perform poor-mans SSA update here.
25962 (slpeel_tree_peel_loop_to_edge): Adjust.
25964 2013-01-16 Richard Biener <rguenther@suse.de>
25966 PR tree-optimization/54767
25967 PR tree-optimization/53465
25968 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25969 (vrp_visit_phi_node): For PHI arguments coming via backedges
25970 drop all symbolical range information.
25971 (execute_vrp): Compute backedges.
25973 2013-01-16 Richard Biener <rguenther@suse.de>
25975 * doc/install.texi: Update CLooG and ISL requirements to
25978 2013-01-16 Christian Bruel <christian.bruel@st.com>
25981 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25982 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25983 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25985 2013-01-16 DJ Delorie <dj@redhat.com>
25987 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25988 (UNSPECV_SP_SWITCH_E): New.
25989 (sp_switch_1): Change to an unspec.
25990 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25993 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25995 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25996 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25997 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25998 (expand_mem_thread_fence): Ditto.
25999 (expand_mem_signal_fence): Ditto.
26000 (expand_atomic_load): Ditto.
26001 (expand_atomic_store): Ditto.
26003 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26005 PR rtl-optimization/55547
26006 PR rtl-optimization/53827
26009 * alias.c (memrefs_conflict_p): Set sizes to negative after
26012 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26015 * function.c (thread_prologue_and_epilogue_insns): Always
26016 add crtl->drap_reg to set_up_by_prologue.set, even if
26017 stack_realign_drap is false.
26019 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26021 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26022 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26023 *call): Fix indention.
26025 2013-01-15 Tom de Vries <tom@codesourcery.com>
26028 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26031 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26033 PR rtl-optimization/55153
26034 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26036 2013-01-15 Martin Jambor <mjambor@suse.cz>
26038 PR tree-optimization/55920
26039 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26040 accesses as grp_to_be_debug_replaced.
26042 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26044 PR tree-optimization/55920
26045 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26046 there is non-useless type conversion needed from debug rhs to lhs,
26047 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26049 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26050 Mikael Pettersson <mikpe@it.uu.se>
26053 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26055 (ASM_OUTPUT_CASE_LABEL): Remove.
26056 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26057 * final.c (shorten_branches): Update alignment of labels before
26058 jump tables if CASE_VECTOR_SHORTEN_MODE.
26060 2013-01-15 Richard Biener <rguenther@suse.de>
26063 * system.h: Do not include gmp.h for building host tools.
26065 2013-01-15 Richard Biener <rguenther@suse.de>
26067 PR middle-end/55882
26068 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26069 account for bitpos when computing alignment.
26071 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26073 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26074 (ix86_target_macros_internal): Likewise.
26076 * config/i386/i386.c (m_CORE2I7): Removed.
26077 (m_CORE_HASWELL): New macro.
26078 (m_CORE_ALL): Likewise.
26079 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26080 (initial_ix86_arch_features): Likewise.
26081 (processor_target_table): Initializations for Core avx2.
26082 (cpu_names): New names "core-avx2".
26083 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26084 PROCESSOR_CORE_HASWELL.
26085 (ix86_issue_rate): New case.
26086 (ia32_multipass_dfa_lookahead): Likewise.
26087 (ix86_sched_init_global): Likewise.
26089 * config/i386/i386.h (TARGET_HASWELL): New macro.
26090 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26091 (processor_type): New PROCESSOR_HASWELL.
26093 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26095 PR tree-optimization/55955
26096 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26097 *SHIFT_EXPR and *ROTATE_EXPR codes.
26099 PR tree-optimization/48766
26100 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26101 -ftrapv disable -fwrapv.
26103 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26106 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26107 etc. to 1 and not to __flash.
26108 Use LL suffix for __INT24_MAX__ with -mint8.
26109 Use ULL suffix for __UINT24_MAX__ with -mint8.
26111 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26113 * config/avr/avr-arch.h
26114 (struct base_arch_s): Use typedef avr_arch_t instead.
26115 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26116 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26117 * config/avr/avr.c: Same.
26118 * config/avr/avr-devices.c: Same.
26119 * config/avr/driver-avr.c: Same.
26120 * config/avr/gen-avr-mmcu-texi.c: Same.
26121 * config/avr/avr-mcus.def: Adjust comment.
26123 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26125 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26126 * config/aarch64/iterators.md (VALLDI): New.
26128 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26129 Andi Kleen <ak@linux.intel.com>
26132 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26133 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26136 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26138 * config/avr/avr-stdint.h: Remove trailing blanks.
26139 * config/avr/avr-log.h: Same.
26140 * config/avr/avr-arch.h: Same.
26141 * config/avr/avr-devices.c: Same.
26142 * config/avr/avr-dimode.md: Same.
26143 * config/avr/predicates.md: Same.
26144 * config/avr/avr-c.c: Same. And fix typo.
26146 * config/avr/avr-protos.h: Same. And:
26147 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26148 (init_cumulative_args): Rename to avr_init_cumulative_args.
26149 (expand_prologue): Rename to avr_expand_prologue.
26150 (expand_epilogue): Rename to avr_expand_epilogue.
26151 (adjust_insn_length): Rename to avr_adjust_insn_length.
26152 (notice_update_cc): Rename to avr_notice_update_cc.
26153 (final_prescan_insn): Rename to avr_final_prescan_insn.
26154 * config/avr/avr.c: Same.
26155 * config/avr/avr.h: Same.
26156 * config/avr/avr.md: Remove trailing blanks.
26157 (prologue): Use avr_expand_prologue.
26158 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26160 2013-01-14 Richard Biener <rguenther@suse.de>
26162 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26163 verify_location, collect_subblocks): New functions.
26164 (verify_gimple_in_cfg): Verify that locations only reference
26165 BLOCKs in the functions BLOCK tree.
26167 2013-01-14 Richard Biener <rguenther@suse.de>
26169 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26171 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26173 (insert_out_of_ssa_copy_on_edge): Likewise.
26174 (rewrite_close_phi_out_of_ssa): Likewise.
26175 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26177 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26178 propagated constants.
26179 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26182 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26184 * config/avr/avr-modes.def: Add GPL copyright notice.
26186 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26188 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26189 MEMMODEL_MASK to determine memory model.
26190 (atomic_store<mode>): Ditto from operands[2].
26191 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26193 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26196 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26197 (fold_gimple_assign): Don't call unshare_expr here.
26198 (fold_ctor_reference): Call unshare_expr.
26200 2013-01-13 Terry Guo <terry.guo@arm.com>
26202 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26203 * doc/fragments.texi: Document MULTILIB_REUSE.
26204 * gcc.c (multilib_reuse): New internal spec.
26205 (set_multilib_dir): Also search multilib from multilib_reuse.
26206 * genmultilib (tmpmultilib3): Refactor code.
26207 (tmpmultilib4): Ditto.
26208 (multilib_reuse): New multilib argument.
26210 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26212 * Makefile.in: Update copyright.
26214 2013-01-12 Tom de Vries <tom@codesourcery.com>
26216 PR middle-end/55890
26217 * calls.c (expand_call): Check if arg_nr is valid.
26219 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26221 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26222 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26223 documentation. Add missing '__' in front of
26224 __builtin_ia32_packssdw256.
26226 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26229 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26230 NO_REGS for larl operands.
26231 (s390_reload_larl_operand): Use s390_load_address instead of
26234 2013-01-11 Richard Biener <rguenther@suse.de>
26236 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26237 (verify_node_sharing): ... here.
26238 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26240 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26242 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26243 Substitute TREECHECKING.
26244 * configure: Regenerate.
26245 * Makefile.in (TREECHECKING): New.
26247 2013-01-11 Richard Guenther <rguenther@suse.de>
26249 PR tree-optimization/44061
26250 * tree-vrp.c (extract_range_basic): Compute zero as
26251 value-range for __builtin_constant_p of function parameters.
26253 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26255 Update copyright years.
26257 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26259 PR rtl-optimization/55672
26260 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26261 const to be eliminable.
26263 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26265 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26266 * configure: Regenerate.
26268 2013-01-10 Richard Biener <rguenther@suse.de>
26270 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26272 2013-01-10 Richard Biener <rguenther@suse.de>
26275 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26276 locations for virtual PHI arguments.
26277 (rewrite_update_phi_arguments): Likewise.
26279 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26281 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26284 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26286 PR tree-optimization/55921
26287 * tree-complex.c (expand_complex_asm): New function.
26288 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26290 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26293 * config/s390/s390.c (s390_symref_operand_p)
26294 (s390_loadrelative_operand_p): Merge the two functions.
26295 (s390_check_qrst_address, print_operand_address): Add parameters
26296 to s390_loadrelative_operand_p invokation.
26297 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26298 (s390_reload_larl_operand, s390_secondary_reload): Use
26299 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26300 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26302 2013-01-09 Mike Stump <mikestump@comcast.net>
26304 * dse.c (record_store): Remove unnecessary assert.
26306 2013-01-09 Jan Hubicka <jh@suse.cz>
26308 PR tree-optimization/55569
26309 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26310 * cfgloop.h (scale_loop_profile): Likewise.
26312 2013-01-09 Jan Hubicka <jh@suse.cz>
26315 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26317 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26319 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26321 PR middle-end/55114
26322 * expr.h (maybe_emit_group_store): Declare.
26323 * expr.c (maybe_emit_group_store): New function.
26324 * builtins.c (expand_builtin_int_roundingfn): Call it.
26325 (expand_builtin_int_roundingfn_2): Likewise.
26327 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26329 PR rtl-optimization/55829
26330 * lra-constraints.c (match_reload): Add code for absent output.
26331 (curr_insn_transform): Add code for reloads of matched inputs
26334 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26336 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26337 attribute of movddup insn to DF.
26338 (*vec_interleave_lowv2df): Ditto.
26339 (vec_dupv2df): Ditto.
26341 2013-01-09 Jan Hubicka <jh@suse.cz>
26343 PR tree-optimiation/55875
26344 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26345 EVERY_ITERATION parameter.
26346 (number_of_iterations_exit): Check if exit is executed every iteration.
26347 (idx_infer_loop_bounds): Similarly here.
26348 (n_of_executions_at_most): Simplify
26349 to only test for cases where statement is dominated by the
26350 particular bound; handle correctly the "postdominance" test.
26351 (scev_probably_wraps_p): Use max loop iterations info
26352 as a global bound first.
26354 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26355 Nick Clifton <nickc@redhat.com>
26357 * config/v850/v850.md (cbranchsf4): New pattern.
26358 (cstoresf4): New pattern.
26359 (cbranchdf4): New pattern.
26360 (cstoredf4): New pattern.
26361 (movsicc): Disallow floating point comparisons.
26362 (cmpsf_le_insn): Fix order of operators.
26363 (cmpsf_lt_insn): Likewise.
26364 (cmpsf_eq_insn): Likewise.
26365 (cmpdf_le_insn): Likewise.
26366 (cmpdf_lt_insn): Likewise.
26367 (cmpdf_eq_insn): Likewise.
26368 (cmpsf_ge_insn): Use LE comparison.
26369 (cmpdf_ge_insn): Likewise.
26370 (cmpsf_gt_insn): Use LT comparison.
26371 (cmpdf_gt_insn): Likewise.
26372 (cmpsf_ne_insn): Delete pattern.
26373 (cmpdf_ne_insn): Delete pattern.
26374 * config/v850/v850.c (v850_gen_float_compare): Use
26375 gen_cmpdf_eq_insn for NE comparison.
26376 (v850_float_z_comparison_operator)
26377 (v850_float_nz_comparison_operator): Move from here ...
26378 * config/v850/predicates.md: ... to here. Move GT and GE
26379 comparisons into v850_float_z_comparison_operator.
26380 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26382 (v850_float_nz_comparison_operator): Likewise.
26384 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26386 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26387 with calls to gen_insvsi/gen_insvdi.
26389 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26391 * config/i386/i386.c (initial_ix86_tune_features): Set up
26392 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26394 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26395 Jakub Jelinek <jakub@redhat.com>
26397 PR tree-optimization/48189
26398 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26399 If nitercst is 0, don't predict the exit edge.
26401 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26403 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26404 in asm_fprintf with reg_names.
26405 (aarch64_print_operand_address): Likewise.
26406 (aarch64_return_addr): Likewise.
26407 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26409 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26411 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26412 (INT_U6_BITS): Likewise.
26413 * config/pa/predicates.md (uint6_operand): New predicate.
26414 (shift5_operand, shift6_operand): Likewise.
26415 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26417 (lshrdi3): Use shift6_operand.
26418 (shrpsi4, shrpdi4): New insn patterns.
26419 (extzv): Delete expander.
26420 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26421 predicates in unamed zero extract patterns. Tighten common constraint.
26422 (extv): Delete expander.
26423 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26424 predicates in unamed sign extract patterns. Tighten common constraint.
26425 (insv): Delete expander.
26426 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26427 predicates in unamed insert patterns. Tighten common constraint.
26428 Change uint32_operand predicate to uint6_operand predicate in unamed
26429 DImode pattern to insert constant values of type 1...1xxxx.
26431 2013-01-04 Jan Hubicka <jh@suse.cz>
26433 PR tree-optimization/55823
26434 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26437 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26438 Uros Bizjak <ubizjak@gmail.com>
26440 PR rtl-optimization/55845
26441 * df-problems.c (can_move_insns_across): Stop scanning at
26442 volatile_insn_p source instruction or give up if
26443 across_from .. across_to range contains any volatile_insn_p
26446 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26448 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26449 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26451 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26452 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26454 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26457 * asan.c (asan_clear_shadow): New function.
26458 (asan_emit_stack_protection): Use it.
26460 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26462 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26463 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26464 with tab instead of space.
26466 2013-01-08 Nick Clifton <nickc@redhat.com>
26468 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26469 register bank 0 at the start of an interrupt handler.
26470 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26473 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26475 * config/aarch64/aarch64-simd.md
26476 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26477 (aarch64_simd_bsl): Likewise.
26478 (aarch64_vcond_internal<mode>): Likewise.
26479 (vcond<mode><mode>): Likewise.
26480 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26481 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26483 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26485 * config/aarch64/aarch64-builtins.c
26486 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26488 2013-01-08 Martin Jambor <mjambor@suse.cz>
26491 * tree-sra.c (analyze_access_subtree): Return true also after
26492 potentially creating a debug-only replacement.
26494 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26496 PR middle-end/55890
26497 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26499 PR tree-optimization/54120
26500 * tree-vrp.c (range_fits_type_p): Don't allow
26501 src_precision < precision from signed vr to unsigned_p
26502 if vr->min or vr->max is negative.
26503 (simplify_float_conversion_using_ranges): Test can_float_p
26504 against CODE_FOR_nothing.
26506 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26507 Richard Biener <rguenther@suse.de>
26509 PR middle-end/55851
26510 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26511 types instead of just INTEGER_TYPE types.
26513 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26515 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26518 2013-01-07 Steve Ellcey <sellcey@mips.com>
26521 * config/mips/mips.opt: Change mad to mmad to match documentation.
26523 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26526 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26527 .progmemx.data now.
26529 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26532 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26533 (avr_addrspace_t): Add .section_name field.
26534 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26536 (avr_addrspace): Same. Initialize .section_name. Remove last
26537 NULL entry. Put __memx into .progmemx.data.
26538 (progmem_section_prefix): Remove.
26539 (avr_asm_init_sections): No need to initialize progmem_section.
26540 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26541 section name prefix.
26542 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26543 retrieve the progmem section.
26544 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26545 boundary to run over avr_addrspace[].
26546 (avr_register_target_pragmas): Ditto.
26548 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26550 * varasm.c (output_constant_def_contents): For asan_protect_global
26551 protected strings, adjust DECL_ALIGN if needed, before testing for
26553 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26554 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26556 (output_object_block): For asan protected decls, emit asan padding
26557 after their contents.
26558 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26559 (asan_finish_file): Test it here instead.
26561 2013-01-07 Nick Clifton <nickc@redhat.com>
26562 Matthias Klose <doko@debian.org>
26563 Doug Kwan <dougkwan@google.com>
26564 H.J. Lu <hongjiu.lu@intel.com>
26567 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26569 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26571 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26573 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26575 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26577 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26580 * doc/install.texi (Cross-Compiler-Specific Options): Document
26583 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26585 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26586 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26587 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26588 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26589 vqmovun_high_s64): Fix source operand number and update copyright.
26591 2013-01-07 Richard Biener <rguenther@suse.de>
26593 PR middle-end/55890
26594 * gimple.h (gimple_call_builtin_p): New overload.
26595 * gimple.c (validate_call): New function.
26596 (gimple_call_builtin_p): Likewise.
26597 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26598 Use gimple_call_builtin_p.
26599 (find_func_clobbers): Likewise.
26600 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26601 (strlen_optimize_stmt): Likewise.
26603 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26605 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26606 (vld1q_dup_*): Likewise.
26607 (vld1_*): Likewise.
26608 (vld1q_*): Likewise.
26609 (vld1_lane_*): Likewise.
26610 (vld1q_lane_*): Likewise.
26612 2013-01-07 Richard Biener <rguenther@suse.de>
26614 * lto-streamer.h (LTO_minor_version): Bump to 2.
26616 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26618 * config/aarch64/aarch64-protos.h
26619 (aarch64_const_double_zero_rtx_p): Rename to...
26620 (aarch64_float_const_zero_rtx_p): ...this.
26621 (aarch64_float_const_representable_p): New.
26622 (aarch64_output_simd_mov_immediate): Likewise.
26623 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26624 move immediate case.
26625 * config/aarch64/aarch64.c
26626 (aarch64_const_double_zero_rtx_p): Rename to...
26627 (aarch64_float_const_zero_rtx_p): ...this.
26628 (aarch64_print_operand): Allow printing of new constants.
26629 (aarch64_valid_floating_const): New.
26630 (aarch64_legitimate_constant_p): Check for valid floating-point
26632 (aarch64_simd_valid_immediate): Likewise.
26633 (aarch64_vect_float_const_representable_p): New.
26634 (aarch64_float_const_representable_p): Likewise.
26635 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26636 (aarch64_output_simd_mov_immediate): New.
26637 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26638 (*movdf_aarch64): Likewise.
26639 * config/aarch64/constraints.md (Ufc): New.
26640 (Y): call aarch64_float_const_zero_rtx.
26641 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26643 2013-01-07 Richard Biener <rguenther@suse.de>
26645 PR tree-optimization/55888
26646 PR tree-optimization/55862
26647 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26648 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26649 not if it is contained therein.
26651 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26653 * config/avr/t-avr: Typo.
26655 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26658 * config/avr/t-avr: Don't automatically rebuild
26659 $(srcdir)/config/avr/t-multilib
26660 $(srcdir)/config/avr/avr-tables.opt
26661 $(srcdir)/doc/avr-mmcu.texi
26662 (avr-mcus): New phony target to build them on request.
26663 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26664 * avr/avr-mcus.def: Adjust comments.
26666 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26668 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26670 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26672 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26674 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26676 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26678 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26680 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26681 to generate profiling.
26682 * config/rs6000/aix64.h (LIB_SPEC): Same.
26684 2013-01-04 Andrew Pinski <apinski@cavium.com>
26686 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26688 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26690 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26692 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26694 (ix86_expand_move): Ditto.
26695 (ix86_zero_extend_to_Pmode): Ditto.
26696 (ix86_expand_call): Ditto.
26697 (ix86_expand_special_args_builtin): Ditto.
26698 (ix86_expand_builtin): Ditto.
26700 2013-01-04 Richard Biener <rguenther@suse.de>
26702 PR tree-optimization/55862
26703 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26704 translating them through PHI nodes.
26706 2013-01-04 Martin Jambor <mjambor@suse.cz>
26708 PR tree-optimization/55755
26709 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26710 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26712 2013-01-04 Marek Polacek <polacek@redhat.com>
26714 PR middle-end/55859
26715 * opts.c (default_options_optimization): Clarify error message.
26717 2013-01-04 Richard Biener <rguenther@suse.de>
26719 PR middle-end/55863
26720 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26723 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26726 * config/pa/pa.md (movsi): Revert previous change.
26727 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26730 2013-01-03 Richard Henderson <rth@redhat.com>
26732 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26733 after eliminating TLS symbols.
26735 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26738 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26739 * graphite-poly.c (debug_gmp_value): Likewise.
26741 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26744 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26745 selected code model, define __code_mode_small__, __code_model_medium__,
26746 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26747 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26748 xchg temporary register with %k. Declare temporary register as
26750 [__x86_64__]: For medium and large code models, preserve %rbx register.
26752 2013-01-03 Richard Biener <rguenther@suse.de>
26754 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26755 (dump_subscript): Adjust.
26756 (finalize_ddr_dependent): Do not dump redundant info.
26757 (analyze_siv_subscript): Adjust.
26758 (subscript_dependence_tester): Likewise.
26759 (compute_affine_dependence): Likewise.
26761 2013-01-03 Richard Biener <rguenther@suse.de>
26764 2013-01-03 Richard Biener <rguenther@suse.de>
26766 PR tree-optimization/55857
26767 * tree-vect-stmts.c (vectorizable_load): Do not setup
26768 re-alignment for invariant loads.
26770 2013-01-02 Richard Biener <rguenther@suse.de>
26772 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26773 invariant load do not generate a vector load from the scalar location.
26775 2013-01-03 Richard Biener <rguenther@suse.de>
26777 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26778 for not vectorizing.
26779 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26780 not build INDIRECT_REFs, call get_name once only.
26781 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26782 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26784 2013-01-03 Richard Biener <rguenther@suse.de>
26786 PR tree-optimization/55857
26787 * tree-vect-stmts.c (vectorizable_load): Do not setup
26788 re-alignment for invariant loads.
26790 2013-01-03 Richard Biener <rguenther@suse.de>
26793 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26794 prefer a built-in decl.
26796 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26798 * gcc.c (process_command): Update copyright notice dates.
26799 * gcov.c (print_version): Likewise.
26800 * gcov-dump.c (print_version): Likewise.
26802 PR rtl-optimization/55838
26803 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26804 iv0.step, iv1.step and step.
26806 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26807 Marc Glisse <marc.glisse@inria.fr>
26809 PR tree-optimization/55832
26810 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26811 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26812 integer_{one,zero}_node.
26814 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26817 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26818 * var-tracking.c (reverse_op): Don't add reverse ops to
26819 VALUEs that have already
26820 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26822 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26824 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26826 2013-01-02 Teresa Johnson <tejohnson@google.com>
26828 * dumpfile.c (dump_loc): Print filename with location.
26829 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26830 new location_t parameter to emit complete unroll message with
26831 new dump framework.
26832 (canonicalize_loop_induction_variables): Compute loops location
26833 and pass to try_unroll_loop_completely.
26834 * loop-unroll.c (report_unroll_peel): New function.
26835 (peel_loops_completely): Use new dump format with location
26836 for main dumpfile message, and invoke report_unroll_peel on success.
26837 (decide_unrolling_and_peeling): Ditto.
26838 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26839 by report_unroll_peel.
26840 (decide_peel_completely): Ditto.
26841 (decide_unroll_constant_iterations): Ditto.
26842 (decide_unroll_runtime_iterations): Ditto.
26843 (decide_peel_simple): Ditto.
26844 (decide_unroll_stupid): Ditto.
26845 * cfgloop.c (get_loop_location): New function.
26846 * cfgloop.h (get_loop_location): Declare.
26848 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26850 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26853 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26855 PR middle-end/55198
26856 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26857 BLKmode objects when EXPAND_MEMORY is specified.
26859 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26861 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26863 (fold_builtin_cpu): Do not share cpu model decls across statements.
26865 2013-01-02 Jason Merrill <jason@redhat.com>
26868 * tree.c (build_array_type_1): Revert earlier change.
26870 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26872 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26874 * config/aarch64/aarch64-tune.md: Re-generate.
26876 2013-01-02 Richard Biener <rguenther@suse.de>
26878 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26879 invariant load do not generate a vector load from the scalar location.
26881 2013-01-02 Richard Biener <rguenther@suse.de>
26884 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26885 * configure: Regenerate.
26887 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26889 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26890 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26891 (expand_builtin_int_roundingfn_2): Keep the original target around
26892 for the fallback case.
26894 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26896 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26897 to be clear for sign changes.
26899 2013-01-01 Jan Hubicka <jh@suse.cz>
26901 * ipa-inline-analysis.c: Fix formatting.
26903 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26905 PR tree-optimization/55831
26906 * tree-vect-loop.c (get_initial_def_for_induction): Use
26907 gsi_after_labels instead of gsi_start_bb.
26909 Copyright (C) 2013 Free Software Foundation, Inc.
26911 Copying and distribution of this file, with or without modification,
26912 are permitted in any medium without royalty provided the copyright
26913 notice and this notice are preserved.