1 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3 * opth-gen.awk: Define target_flags_explicit.
5 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
7 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
8 fields to the reg_addr array that describes the valid addressing
9 mode for any register, general purpose registers, floating point
10 registers, and Altivec registers.
11 (FIRST_RELOAD_REG_CLASS): Likewise.
12 (LAST_RELOAD_REG_CLASS): Likewise.
13 (struct reload_reg_map_type): Likewise.
14 (reload_reg_map_type): Likewise.
15 (RELOAD_REG_VALID): Likewise.
16 (RELOAD_REG_MULTIPLE): Likewise.
17 (RELOAD_REG_INDEXED): Likewise.
18 (RELOAD_REG_OFFSET): Likewise.
19 (RELOAD_REG_PRE_INCDEC): Likewise.
20 (RELOAD_REG_PRE_MODIFY): Likewise.
22 (mode_supports_pre_incdec_p): New helper functions to say whether
23 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
24 (mode_supports_pre_modify_p): Likewise.
25 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
26 print the valid address mode bits for each mode.
27 (rs6000_debug_print_mode): Likewise.
28 (rs6000_debug_reg_global): Likewise.
29 (rs6000_setup_reg_addr_masks): New function to set up the address
30 mask bits for each type.
31 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
32 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
33 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
34 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
35 PRE_MODIFY are supported.
36 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
37 registers, instead of {src,dest}_av_p.
38 (rs6000_print_options_internal): Tweak the debug output slightly.
40 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
42 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
45 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
47 * tree-flow.h (struct omp_region): Move to omp-low.c.
48 Remove omp_ prototypes and variables.
49 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
50 (copy_var_decl): Relocate prototype from tree-flow.h.
51 * gimple.c (copy_var_decl): Relocate from omp-low.c.
52 * tree.h: Move prototype to omp-low.h.
53 * omp-low.h: New File. Relocate prototypes here.
54 * omp-low.c (struct omp_region): Make local here.
55 (root_omp_region): Make static.
56 (copy_var_decl) Move to gimple.c.
57 (new_omp_region): Make static.
58 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
59 * tree-cfg.c: Include omp-low.h.
60 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
61 * gimplify.c: Include omp-low.h.
62 * tree-parloops.c: Likewise.
64 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
66 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
69 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
71 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
72 operands are in memory, prefer to force non-matched operand 1 to
75 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
78 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
79 restrict TImode addresses to single indirect registers if both
80 -mquad-memory and -mvsx-timode are used.
81 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
82 we should emit load/store quad. Remove using %y for quad memory
85 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
86 constraints to allow load/store quad on machines where TImode is
87 not allowed in VSX registers. Use 'n' instead of 'F' constraint
88 for TImode to load integer constants.
90 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
92 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
94 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
96 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
97 handling of STACK_REG.
99 2013-10-17 Richard Biener <rguenther@suse.de>
101 PR tree-optimization/58143
102 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
104 (rewrite_to_defined_overflow): Likewise.
105 (move_computations_dom_walker::before_dom): Rewrite stmts
106 with undefined signed overflow that are not always executed
107 into unsigned arithmetic.
109 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
112 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
113 explicit isa flag to be an options variable, instead of using
114 global_options_set. Remove define from rs6000.h.
115 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
117 * config/rs6000/rs6000.c (rs6000_option_override_internal):
118 Initialize rs6000_isa_flags_explicit.
119 (rs6000_function_specific_save): Add gcc_options* parameter, so
120 that the powerpc builds after the 2013-10-15 changes.
121 (rs6000_function_specific_restore): Likewise.
123 2013-10-16 DJ Delorie <dj@redhat.com>
125 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
126 op is a REG before checking REGNO.
127 (rl78_alloc_physical_registers): Verify pattern is a SET before
130 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
132 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
134 (vec_unpacks_lo_v4sf): Likewise.
135 (vec_unpacks_float_hi_v4si): Likewise.
136 (vec_unpacks_float_lo_v4si): Likewise.
137 (vec_unpacku_float_hi_v4si): Likewise.
138 (vec_unpacku_float_lo_v4si): Likewise.
140 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
142 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
143 (vsx_concat_v2sf): Likewise.
145 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
147 * config/aarch64/aarch64.md
148 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
150 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
152 PR tree-optimization/58697
153 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
154 estimated_loop_iterations_int.
155 (max_stmt_executions_int): Call get_max_loop_iterations_int.
156 (get_max_loop_iterations_int): New. HWINT version of
157 get_max_loop_iterations.
158 * cfgloop.h: Add prototypes.
159 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
160 * loop-unroll.c (decide_peel_once_rolling): Call
161 get_estimated_loop_iterations_int.
162 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
163 * tree-ssa-loop-niter.h: Tweak prototypes.
165 2013-10-16 David Malcolm <dmalcolm@redhat.com>
167 * gengtype-parse.c (struct_field_seq): Ignore access-control
168 keywords ("public:" etc).
170 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
172 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
173 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
175 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
177 * config/arm/arm.opt (mlra): New option.
178 * config/arm/arm.c (arm_lra_p): New function.
179 (TARGET_LRA_P): Define.
181 2013-10-16 Paulo Matos <pmatos@broadcom.com>
183 * tree-core.h (tree_code_name): Remove.
184 * tree.h (get_tree_code_name): New prototype.
185 * tree.c (tree_code_name): Make static.
186 (get_tree_code_name): New function.
187 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
188 tree_class_check_failed, tree_range_check_failed,
189 tree_not_class_check_failed, omp_clause_check_failed,
190 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
191 wrapper get_tree_code_name instead of calling tree_code_name directly.
192 * tree-vrp.c (dump_asserts_for): Likewise.
193 * tree-dump.c (dequeue_and_dump): Likewise.
194 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
195 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
196 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
197 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
198 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
199 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
200 dump_gimple_omp_for): Likewise.
201 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
202 * tree-ssa-pre.c (print_pre_expr): Likewise.
203 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
204 * print-tree.c (print_node_brief, print_node): Likewise.
205 * gimple.c (gimple_check_failed): Likewise.
206 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
207 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
208 * config/mep/mep.c (mep_validate_vliw): Likewise.
209 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
210 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
212 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
214 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
217 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
219 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
220 (MULTILIB_MATCHES): Add multilib for -march=v8.
222 2013-10-15 Sriraman Tallam <tmsriram@google.com>
225 * optc-save-gen.awk: Add extra parameter to the save and restore
227 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
228 * tree.c (build_optimization_node): New parameter. Add extra parameter
229 to call to cl_optimization_save.
230 (build_target_option_node): New parameter. Add extra parameter
231 to call to cl_target_option_save.
232 * tree.h (build_optimization_node): New parameter.
233 (build_target_option_node): New parameter.
234 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
235 build_optimization_node and build_target_option_node.
236 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
237 (handle_pragma_push_options): Ditto.
238 * toplev.c (process_options): Ditto.
239 * opts.c (init_options_struct): Check for opts_set non-null.
240 * target.def (target_option.save): New parameter.
241 (target_option.restore): New parameter.
243 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
244 (ix86_pragma_target_parse): Ditto.
245 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
247 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
248 to build_optimization_node and build_target_option_node.
249 (rs6000_valid_attribute_p): Ditto.
250 (rs6000_pragma_target_parse): Ditto.
251 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
253 * config/i386/i386.h:
254 TARGET_64BIT_P: New Macro
255 TARGET_MMX_P: New Macro.
256 TARGET_3DNOW_P: New Macro.
257 TARGET_3DNOW_A_P: New Macro.
258 TARGET_SSE_P: New Macro.
259 TARGET_SSE2_P: New Macro.
260 TARGET_SSE3_P: New Macro.
261 TARGET_SSSE3_P: New Macro.
262 TARGET_SSE4_1_P: New Macro.
263 TARGET_SSE4_2_P: New Macro.
264 TARGET_AVX_P: New Macro.
265 TARGET_AVX2_P: New Macro.
266 TARGET_AVX512F_P: New Macro.
267 TARGET_AVX512PF_P: New Macro.
268 TARGET_AVX512ER_P: New Macro.
269 TARGET_AVX512CD_P: New Macro.
270 TARGET_FMA_P: New Macro.
271 TARGET_SSE4A_P: New Macro.
272 TARGET_FMA4_P: New Macro.
273 TARGET_XOP_P: New Macro.
274 TARGET_LWP_P: New Macro.
275 TARGET_ABM_P: New Macro.
276 TARGET_BMI_P: New Macro.
277 TARGET_BMI2_P: New Macro.
278 TARGET_LZCNT_P: New Macro.
279 TARGET_TBM_P: New Macro.
280 TARGET_POPCNT_P: New Macro.
281 TARGET_SAHF_P: New Macro.
282 TARGET_MOVBE_P: New Macro.
283 TARGET_CRC32_P: New Macro.
284 TARGET_AES_P: New Macro.
285 TARGET_PCLMUL_P: New Macro.
286 TARGET_CMPXCHG16B_P: New Macro.
287 TARGET_FSGSBASE_P: New Macro.
288 TARGET_RDRND_P: New Macro.
289 TARGET_F16C_P: New Macro.
290 TARGET_RTM_P: New Macro.
291 TARGET_HLE_P: New Macro.
292 TARGET_RDSEED_P: New Macro.
293 TARGET_PRFCHW_P: New Macro.
294 TARGET_ADX_P: New Macro.
295 TARGET_FXSR_P: New Macro.
296 TARGET_XSAVE_P: New Macro.
297 TARGET_XSAVEOPT_P: New Macro.
298 TARGET_LP64_P: New Macro.
299 TARGET_X32_P: New Macro.
300 TARGET_FPMATH_DEFAULT_P: New Macro.
301 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
302 * config/i386/i386.c (ix86_option_override_internal): New parameters.
304 Change ix86_tune_string to access opts->x_ix86_tune_string.
305 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
306 Change ix86_arch_string to access opts->x_ix86_arch_string.
307 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
308 Change ix86_pmode to access opts->x_ix86_pmode.
309 Change ix86_abi to access opts->x_ix86_abi.
310 Change ix86_cmodel to access opts->x_ix86_cmodel.
311 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
312 Change ix86_isa_flags_explicit to access
313 opts->x_ix86_isa_flags_explicit.
314 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
315 Change ix86_regparm to access opts->x_ix86_regparm.
316 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
317 Change ix86_preferred_stack_boundary_arg to access
318 opts->x_ix86_preferred_stack_boundary_arg.
319 Change ix86_force_align_arg_pointer to access
320 opts->x_ix86_force_align_arg_pointer.
321 Change ix86_incoming_stack_boundar_arg to access
322 opts->x_ix86_incoming_stack_boundar_arg.
323 Change ix86_fpmath to access opts->x_ix86_fpmath.
324 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
325 Change ix86_recip_name to access opts->x_ix86_recip_name.
326 Change ix86_stack_protector_guard to access
327 opts->x_ix86_stack_protector_guard.
328 Change ix86_tune_memcpy_strategy to access
329 opts->x_ix86_tune_memcpy_strategy.
330 Change ix86_tune_memset_strategy to access
331 opts->x_ix86_tune_memset_strategy.
332 Change global_options to access opts.
333 Change global_options_set to access opts_set.
334 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
335 Change TARGET_MMX to TARGET_MMX_P (opts->...).
336 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
337 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
338 Change TARGET_SSE to TARGET_SSE_P (opts->...).
339 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
340 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
341 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
342 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
343 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
344 Change TARGET_AVX to TARGET_AVX_P (opts->...).
345 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
346 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
347 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
348 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
349 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
350 Change TARGET_FMA to TARGET_FMA_P (opts->...).
351 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
352 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
353 Change TARGET_XOP to TARGET_XOP_P (opts->...).
354 Change TARGET_LWP to TARGET_LWP_P (opts->...).
355 Change TARGET_ABM to TARGET_ABM_P (opts->...).
356 Change TARGET_BMI to TARGET_BMI_P (opts->...).
357 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
358 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
359 Change TARGET_TBM to TARGET_TBM_P (opts->...).
360 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
361 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
362 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
363 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
364 Change TARGET_AES to TARGET_AES_P (opts->...).
365 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
366 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
367 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
368 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
369 Change TARGET_F16C to TARGET_F16C_P (opts->...).
370 Change TARGET_RTM to TARGET_RTM_P (opts->...).
371 Change TARGET_HLE to TARGET_HLE_P (opts->...).
372 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
373 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
374 Change TARGET_ADX to TARGET_ADX_P (opts->...).
375 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
376 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
377 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
378 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
379 Change TARGET_X32 to TARGET_X32_P (opts->...).
380 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
381 Change TARGET_FLOAT_RETURNS_IN_80387 to
382 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
383 (ix86_function_specific_save): New parameter. Use opts-> fields
384 to replace global fields.
385 (ix86_function_specific_restore): Ditto.
386 (ix86_valid_target_attribute_inner_p): New parameters.
388 Fix call to ix86_handle_option and set_option.
389 (ix86_valid_target_attribute_tree): New parameters.
390 Change global_options to access opts.
391 Change global_options_set to access opts_set.
392 Fix call to ix86_valid_target_attribute_inner_p.
393 Change ix86_tune_string to access opts->x_ix86_tune_string.
394 Change ix86_arch_string to access opts->x_ix86_arch_string.
395 Change ix86_fpmath to access opts->x_ix86_fpmath
396 Fix call to ix86_option_override_internal.
397 Fix call to ix86_add_new_builtins.
398 Fix calls to build_optimization_node and build_target_option_node.
399 (ix86_valid_target_attribute_p): Remove access to global_options.
400 Use new gcc_options structure func_options.
401 Fix call to ix86_valid_target_attribute_tree.
402 Fix call to build_optimization_node.
403 (get_builtin_code_for_version): Fix call to
404 ix86_valid_target_attribute_tree.
406 2013-10-15 David Malcolm <dmalcolm@redhat.com>
408 * Makefile.in (PICFLAG): New.
409 (enable_host_shared): New.
410 (INTERNAL_CFLAGS): Use PICFLAG.
411 (LIBIBERTY): Use pic build of libiberty.a if configured with
412 --enable-host-shared.
413 * configure.ac: Add --enable-host-shared, setting up new
415 * configure: Regenerate.
416 * doc/install.texi (--enable-shared): Add note contrasting it with ...
417 (--enable-host-shared): New option.
419 2013-10-15 Richard Biener <rguenther@suse.de>
421 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
422 for built-in functions.
424 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
426 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
427 (optimize_range_tests_1): New function,
428 extracted from optimize_range_tests.
429 (optimize_range_tests_xor): Similarly.
430 (optimize_range_tests_diff): New function.
431 (optimize_range_tests): Use optimize_range_tests_1.
433 2013-10-15 Cong Hou <congh@google.com>
435 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
436 requirement of the reduction pattern so that one operand of the
437 reduction operation can come from outside of the loop.
439 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
441 * config/arm/neon-schedgen.ml: Remove.
442 * config/arm/cortex-a9-neon.md: Remove comment regarding
445 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
447 * config/arm/types: Remove old neon types.
449 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
451 * config/arm/cortex-a7.md
452 (cortex_a7_neon_type): New.
453 (cortex_a7_neon_mul): Update for new types.
454 (cortex_a7_neon_mla): Likewise.
455 (cortex_a7_neon): Likewise.
457 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
459 * config/arm/cortex-a15-neon.md
460 (cortex_a15_neon_type): New,
462 (cortex_a15_neon_int_1): Remove.
463 (cortex_a15_neon_int_2): Likewise.
464 (cortex_a15_neon_int_3): Likewise.
465 (cortex_a15_neon_int_4): Likewise.
466 (cortex_a15_neon_int_5): Likewise.
467 (cortex_a15_neon_vqneg_vqabs): Likewise.
468 (cortex_a15_neon_vmov): Likewise.
469 (cortex_a15_neon_vaba): Likewise.
470 (cortex_a15_neon_vaba_qqq): Likewise.
471 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
472 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
473 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
475 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
476 (cortex_a15_neon_mla_qqq_8_16): Likewise.
477 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
478 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
479 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
480 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
481 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
482 (cortex_a15_neon_shift_1): Likewise.
483 (cortex_a15_neon_shift_2): Likewise.
484 (cortex_a15_neon_shift_3): Likewise.
485 (cortex_a15_neon_vshl_ddd): Likewise.
486 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
487 (cortex_a15_neon_vsra_vrsra): Likewise.
488 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
489 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
490 (cortex_a15_neon_bp_3cycle): Likewise.
491 (cortex_a15_neon_ldm_2): Likewise.
492 (cortex_a15_neon_stm_2): Likewise.
493 (cortex_a15_neon_mcr): Likewise.
494 (cortex_a15_neon_mrc): Likewise.
495 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
496 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
497 (cortex_a15_neon_fp_vmul_ddd): Likewise.
498 (cortex_a15_neon_fp_vmul_qqd): Likewise.
499 (cortex_a15_neon_fp_vmla_ddd): Likewise.
500 (cortex_a15_neon_fp_vmla_qqq): Likewise.
501 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
502 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
503 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
504 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
505 (cortex_a15_neon_bp_simple): Likewise.
506 (cortex_a15_neon_bp_2cycle): Likewise.
507 (cortex_a15_neon_bp_3cycle): Likewise.
508 (cortex_a15_neon_vld1_1_2_regs): Likewise.
509 (cortex_a15_neon_vld1_3_4_regs): Likewise.
510 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
511 (cortex_a15_neon_vld2_4_regs): Likewise.
512 (cortex_a15_neon_vld3_vld4): Likewise.
513 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
514 (cortex_a15_neon_vst1_3_4_regs): Likewise.
515 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
516 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
517 (cortex_a15_neon_vst3_vst4): Rename to...
518 (cortex_a15_neon_vst4): This, update for new attributes.
519 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
520 (cortex_a15_neon_vld3_vld4_lane): Likewise.
521 (cortex_a15_neon_vst1_vst2_lane): Likewise.
522 (cortex_a15_neon_vst3_vst4_lane): Likewise.
523 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
524 (cortex_a15_neon_ldm_2): Likewise.
525 (cortex_a15_neon_stm_2): Likewise.
526 (cortex_a15_neon_mcr): Likewise.
527 (cortex_a15_neon_mcr_2_mcrr): Likewise.
528 (cortex_a15_neon_mrc): Likewise.
529 (cortex_a15_neon_mrrc): Likewise.
531 (cortex_a15_neon_abd): New.
532 (cortex_a15_neon_abd_q): Likewise.
533 (cortex_a15_neon_aba): Likewise.
534 (cortex_a15_neon_aba_q): Likewise.
535 (cortex_a15_neon_acc): Likewise.
536 (cortex_a15_neon_acc_q): Likewise.
537 (cortex_a15_neon_arith_basic): Likewise.
538 (cortex_a15_neon_arith_complex): Likewise.
539 (cortex_a15_neon_multiply): Likewise.
540 (cortex_a15_neon_multiply_q): Likewise.
541 (cortex_a15_neon_mla): Likewise.
542 (cortex_a15_neon_mla_q): Likewise.
543 (cortex_a15_neon_sat_mla_long): Likewise.
544 (cortex_a15_neon_shift_acc): Likewise.
545 (cortex_a15_neon_shift_imm_basic): Likewise.
546 (cortex_a15_neon_shift_imm_complex): Likewise.
547 (cortex_a15_neon_shift_reg_basic): Likewise.
548 (cortex_a15_neon_shift_reg_basic_q): Likewise.
549 (cortex_a15_neon_shift_reg_complex): Likewise.
550 (cortex_a15_neon_shift_reg_complex_q): Likewise.
551 (cortex_a15_neon_fp_negabs): Likewise
552 (cortex_a15_neon_fp_arith): Likewise
553 (cortex_a15_neon_fp_arith_q): Likewise
554 (cortex_a15_neon_fp_cvt_int): Likewise
555 (cortex_a15_neon_fp_cvt_int_q): Likewise
556 (cortex_a15_neon_fp_cvt_16): Likewise
557 (cortex_a15_neon_fp_mul): Likewise
558 (cortex_a15_neon_fp_mul_q): Likewise
559 (cortex_a15_neon_fp_mla): Likewise
560 (cortex_a15_neon_fp_mla_q): Likewise
561 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
562 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
563 (cortex_a15_neon_bitops): Likewise.
564 (cortex_a15_neon_bitops_q): Likewise.
565 (cortex_a15_neon_from_gp): Likewise.
566 (cortex_a15_neon_from_gp_q): Likewise.
567 (cortex_a15_neon_tbl3_tbl4): Likewise.
568 (cortex_a15_neon_zip_q): Likewise.
569 (cortex_a15_neon_to_gp): Likewise.
570 (cortex_a15_neon_load_a): Likewise.
571 (cortex_a15_neon_load_b): Likewise.
572 (cortex_a15_neon_load_c): Likewise.
573 (cortex_a15_neon_load_d): Likewise.
574 (cortex_a15_neon_load_e): Likewise.
575 (cortex_a15_neon_load_f): Likewise.
576 (cortex_a15_neon_store_a): Likewise.
577 (cortex_a15_neon_store_b): Likewise.
578 (cortex_a15_neon_store_c): Likewise.
579 (cortex_a15_neon_store_d): Likewise.
580 (cortex_a15_neon_store_e): Likewise.
581 (cortex_a15_neon_store_f): Likewise.
582 (cortex_a15_neon_store_g): Likewise.
583 (cortex_a15_neon_store_h): Likewise.
584 (cortex_a15_vfp_to_from_gp): Likewise.
586 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
588 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
590 (cortex_a9_neon_vshl_ddd): Remove.
591 (cortex_a9_neon_vst3_vst4): Likewise.
592 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
594 (cortex_a9_neon_bit_ops_q): New.
596 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
597 (cortex_a9_neon_int_2): Likewise.
598 (cortex_a9_neon_int_3): Likewise.
599 (cortex_a9_neon_int_4): Likewise.
600 (cortex_a9_neon_int_5): Likewise.
601 (cortex_a9_neon_vqneg_vqabs): Likewise.
602 (cortex_a9_neon_vmov): Likewise.
603 (cortex_a9_neon_vaba): Likewise.
604 (cortex_a9_neon_vaba_qqq): Likewise.
605 (cortex_a9_neon_shift_1): Likewise.
606 (cortex_a9_neon_shift_2): Likewise.
607 (cortex_a9_neon_shift_3): Likewise.
608 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
609 (cortex_a9_neon_vsra_vrsra): Likewise.
610 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
611 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
612 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
614 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
615 (cortex_a9_neon_mla_qqq_8_16): Likewise.
616 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
618 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
619 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
620 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
621 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
622 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
623 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
624 (cortex_a9_neon_fp_vsum): Likewise.
625 (cortex_a9_neon_fp_vmul_ddd): Likewise.
626 (cortex_a9_neon_fp_vmul_qqd): Likewise.
627 (cortex_a9_neon_fp_vmla_ddd): Likewise.
628 (cortex_a9_neon_fp_vmla_qqq): Likewise.
629 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
630 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
631 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
632 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
633 (cortex_a9_neon_bp_simple): Likewise.
634 (cortex_a9_neon_bp_2cycle): Likewise.
635 (cortex_a9_neon_bp_3cycle): Likewise.
636 (cortex_a9_neon_ldr): Likewise.
637 (cortex_a9_neon_str): Likewise.
638 (cortex_a9_neon_vld1_1_2_regs): Likewise.
639 (cortex_a9_neon_vld1_3_4_regs): Likewise.
640 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
641 (cortex_a9_neon_vld2_4_regs): Likewise.
642 (cortex_a9_neon_vld3_vld4): Likewise.
643 (cortex_a9_neon_vld1_vld2_lane): Likewise.
644 (cortex_a9_neon_vld3_vld4_lane): Likewise.
645 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
646 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
647 (cortex_a9_neon_vst1_3_4_regs): Likewise.
648 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
649 (cortex_a9_neon_vst1_vst2_lane): Likewise.
650 (cortex_a9_neon_vst3_vst4_lane): Likewise.
651 (cortex_a9_neon_mcr): Likewise.
652 (cortex_a9_neon_mcr_2_mcrr): Likewise.
653 (cortex_a9_neon_mrc): Likewise.
654 (cortex_a9_neon_mrrc): Likewise.
656 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
658 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
660 (cortex_a8_neon_vshl_ddd): Remove.
661 (cortex_a8_neon_vst3_vst4): Likewise.
662 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
664 (cortex_a8_neon_bit_ops_q): New.
666 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
667 (cortex_a8_neon_int_2): Likewise..
668 (cortex_a8_neon_int_3): Likewise.
669 (cortex_a8_neon_int_5): Likewise.
670 (cortex_a8_neon_vqneg_vqabs): Likewise.
671 (cortex_a8_neon_int_4): Likewise.
672 (cortex_a8_neon_vaba): Likewise.
673 (cortex_a8_neon_vaba_qqq): Likewise.
674 (cortex_a8_neon_shift_1): Likewise.
675 (cortex_a8_neon_shift_2): Likewise.
676 (cortex_a8_neon_shift_3): Likewise.
677 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
678 (cortex_a8_neon_vsra_vrsra): Likewise.
679 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
680 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
681 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
683 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
684 (cortex_a8_neon_mla_qqq_8_16): Likewise.
685 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
687 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
688 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
689 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
690 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
691 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
692 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
693 (cortex_a8_neon_fp_vsum): Likewise.
694 (cortex_a8_neon_fp_vmul_ddd): Likewise.
695 (cortex_a8_neon_fp_vmul_qqd): Likewise.
696 (cortex_a8_neon_fp_vmla_ddd): Likewise.
697 (cortex_a8_neon_fp_vmla_qqq): Likewise.
698 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
699 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
700 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
701 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
702 (cortex_a8_neon_bp_simple): Likewise.
703 (cortex_a8_neon_bp_2cycle): Likewise.
704 (cortex_a8_neon_bp_3cycle): Likewise.
705 (cortex_a8_neon_ldr): Likewise.
706 (cortex_a8_neon_str): Likewise.
707 (cortex_a8_neon_vld1_1_2_regs): Likewise.
708 (cortex_a8_neon_vld1_3_4_regs): Likewise.
709 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
710 (cortex_a8_neon_vld2_4_regs): Likewise.
711 (cortex_a8_neon_vld3_vld4): Likewise.
712 (cortex_a8_neon_vld1_vld2_lane): Likewise.
713 (cortex_a8_neon_vld3_vld4_lane): Likewise.
714 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
715 (cortex_a8_neon_vst1_3_4_regs): Likewise.
716 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
717 (cortex_a8_neon_vst1_vst2_lane): Likewise.
718 (cortex_a8_neon_vst3_vst4_lane): Likewise.
719 (cortex_a8_neon_mcr): Likewise.
720 (cortex_a8_neon_mcr_2_mcrr): Likewise.
721 (cortex_a8_neon_mrc): Likewise.
722 (cortex_a8_neon_mrrc): Likewise.
724 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
726 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
728 * config/aarch64/aarch64-simd.md (neon_type): Remove.
729 (aarch64_simd_dup<mode>): Add "type" attribute.
730 (aarch64_dup_lane<mode>): Likewise.
731 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
732 (*aarch64_simd_mov<mode>): Likewise.
733 (aarch64_simd_mov_from_<mode>low): Likewise.
734 (aarch64_simd_mov_from_<mode>high): Likewise.
735 (orn<mode>3): Likewise.
736 (bic<mode>3): Likewise.
737 (add<mode>3): Likewise.
738 (sub<mode>3): Likewise.
739 (mul<mode>3): Likewise.
740 (*aarch64_mul3_elt<mode>): Likewise.
741 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
742 (*aarch64_mul3_elt_to_128df): Likewise.
743 (*aarch64_mul3_elt_to_64v2df): Likewise.
744 (neg<mode>2): Likewise.
745 (abs<mode>2): Likewise.
746 (abd<mode>_3): Likewise.
747 (aba<mode>_3): Likewise.
748 (fabd<mode>_3): Likewise.
749 (*fabd_scalar<mode>3): Likewise.
750 (and<mode>3): Likewise.
751 (ior<mode>3): Likewise.
752 (xor<mode>3): Likewise.
753 (one_cmpl<mode>2): Likewise.
754 (aarch64_simd_vec_set<mode>): Likewise.
755 (aarch64_simd_lshr<mode>): Likewise.
756 (aarch64_simd_ashr<mode>): Likewise.
757 (aarch64_simd_imm_shl<mode>): Likewise.
758 (aarch64_simd_reg_sshl<mode): Likewise.
759 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
760 (aarch64_simd_reg_shl<mode>_signed): Likewise.
761 (aarch64_simd_vec_setv2di): Likewise.
762 (aarch64_simd_vec_set<mode>): Likewise.
763 (aarch64_mla<mode>): Likewise.
764 (*aarch64_mla_elt<mode>): Likewise.
765 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
766 (aarch64_mls<mode>): Likewise.
767 (*aarch64_mls_elt<mode>): Likewise.
768 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
769 (<su><maxmin><mode>3): Likewise.
770 (move_lo_quad_<mode>): Likewise.
771 (aarch64_simd_move_hi_quad_<mode>): Likewise.
772 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
773 (vec_pack_trunc_<mode>): Likewise.
774 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
775 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
776 (*aarch64_<su>mlal_lo<mode>): Likewise.
777 (*aarch64_<su>mlal_hi<mode>): Likewise.
778 (*aarch64_<su>mlsl_lo<mode>): Likewise.
779 (*aarch64_<su>mlsl_hi<mode>): Likewise.
780 (*aarch64_<su>mlal<mode>): Likewise.
781 (*aarch64_<su>mlsl<mode>): Likewise.
782 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
783 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
784 (add<mode>3): Likewise.
785 (sub<mode>3): Likewise.
786 (mul<mode>3): Likewise.
787 (div<mode>3): Likewise.
788 (neg<mode>2): Likewise.
789 (abs<mode>2): Likewise.
790 (fma<mode>4): Likewise.
791 (*aarch64_fma4_elt<mode>): Likewise.
792 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
793 (*aarch64_fma4_elt_to_128df): Likewise.
794 (*aarch64_fma4_elt_to_64v2df): Likewise.
795 (fnma<mode>4): Likewise.
796 (*aarch64_fnma4_elt<mode>): Likewise.
797 (*aarch64_fnma4_elt_<vswap_width_name><mode>
798 (*aarch64_fnma4_elt_to_128df): Likewise.
799 (*aarch64_fnma4_elt_to_64v2df): Likewise.
800 (<frint_pattern><mode>2): Likewise.
801 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
802 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
803 (vec_unpacks_lo_v4sf): Likewise.
804 (aarch64_float_extend_lo_v2df): Likewise.
805 (vec_unpacks_hi_v4sf): Likewise.
806 (aarch64_float_truncate_lo_v2sf): Likewise.
807 (aarch64_float_truncate_hi_v4sf): Likewise.
808 (aarch64_vmls<mode>): Likewise.
809 (<su><maxmin><mode>3): Likewise.
810 (<maxmin_uns><mode>3): Likewise.
811 (reduc_<sur>plus_<mode>): Likewise.
812 (reduc_<sur>plus_v2di): Likewise.
813 (reduc_<sur>plus_v2si): Likewise.
814 (reduc_<sur>plus_<mode>): Likewise.
815 (aarch64_addpv4sf): Likewise.
816 (clz<mode>2): Likewise.
817 (reduc_<maxmin_uns>_<mode>): Likewise.
818 (reduc_<maxmin_uns>_v2di): Likewise.
819 (reduc_<maxmin_uns>_v2si): Likewise.
820 (reduc_<maxmin_uns>_<mode>): Likewise.
821 (reduc_<maxmin_uns>_v4sf): Likewise.
822 (aarch64_simd_bsl<mode>_internal): Likewise.
823 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
824 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
825 (aarch64_get_lane<mode>): Likewise.
826 (*aarch64_combinez<mode>): Likewise.
827 (aarch64_combine<mode>): Likewise.
828 (aarch64_simd_combine<mode>): Likewise.
829 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
830 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
831 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
832 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
833 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
834 (aarch64_<sur>h<addsub><mode>): Likewise.
835 (aarch64_<sur><addsub>hn<mode>): Likewise.
836 (aarch64_<sur><addsub>hn2<mode>): Likewise.
837 (aarch64_pmul<mode>): Likewise.
838 (aarch64_<su_optab><optab><mode>): Likewise.
839 (aarch64_<sur>qadd<mode>): Likewise.
840 (aarch64_sqmovun<mode>): Likewise.
841 (aarch64_<sur>qmovn<mode>): Likewise.
842 (aarch64_s<optab><mode>): Likewise.
843 (aarch64_sq<r>dmulh<mode>): Likewise.
844 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
845 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
846 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
847 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
848 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
849 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
850 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
851 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
852 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
853 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
854 (aarch64_sqdmull<mode>): Likewise.
855 (aarch64_sqdmull_lane<mode>_internal): Likewise.
856 (aarch64_sqdmull_n<mode>): Likewise.
857 (aarch64_sqdmull2<mode>_internal): Likewise.
858 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
859 (aarch64_sqdmull2_n<mode>_internal): Likewise.
860 (aarch64_<sur>shl<mode>): Likewise.
861 (aarch64_<sur>q<r>shl<mode>
862 (aarch64_<sur>shll_n<mode>): Likewise.
863 (aarch64_<sur>shll2_n<mode>): Likewise.
864 (aarch64_<sur>shr_n<mode>): Likewise.
865 (aarch64_<sur>sra_n<mode>): Likewise.
866 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
867 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
868 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
869 (aarch64_cm<optab><mode>): Likewise.
870 (aarch64_cm<optab>di): Likewise.
871 (aarch64_cm<optab><mode>): Likewise.
872 (aarch64_cm<optab>di): Likewise.
873 (aarch64_cmtst<mode>): Likewise.
874 (aarch64_cmtstdi): Likewise.
875 (aarch64_cm<optab><mode>): Likewise.
876 (*aarch64_fac<optab><mode>): Likewise.
877 (aarch64_addp<mode>): Likewise.
878 (aarch64_addpdi): Likewise.
879 (sqrt<mode>2): Likewise.
880 (vec_load_lanesoi<mode>): Likewise.
881 (vec_store_lanesoi<mode>): Likewise.
882 (vec_load_lanesci<mode>): Likewise.
883 (vec_store_lanesci<mode>): Likewise.
884 (vec_load_lanesxi<mode>): Likewise.
885 (vec_store_lanesxi<mode>): Likewise.
886 (*aarch64_mov<mode>): Likewise.
887 (aarch64_ld2<mode>_dreg): Likewise.
888 (aarch64_ld2<mode>_dreg): Likewise.
889 (aarch64_ld3<mode>_dreg): Likewise.
890 (aarch64_ld3<mode>_dreg): Likewise.
891 (aarch64_ld4<mode>_dreg): Likewise.
892 (aarch64_ld4<mode>_dreg): Likewise.
893 (aarch64_tbl1<mode>): Likewise.
894 (aarch64_tbl2v16qi): Likewise.
895 (aarch64_combinev16qi): Likewise.
896 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
897 (aarch64_st2<mode>_dreg): Likewise.
898 (aarch64_st2<mode>_dreg): Likewise.
899 (aarch64_st3<mode>_dreg): Likewise.
900 (aarch64_st3<mode>_dreg): Likewise.
901 (aarch64_st4<mode>_dreg): Likewise.
902 (aarch64_st4<mode>_dreg): Likewise.
903 (*aarch64_simd_ld1r<mode>): Likewise.
904 (aarch64_frecpe<mode>): Likewise.
905 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
906 (aarch64_frecps<mode>): Likewise.
908 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
910 * config/arm/iterators.md (V_elem_ch): New.
912 (VQH_type): Likewise.
913 * config/arm/arm.md (is_neon_type): New.
914 (conds): Use is_neon_type.
915 (anddi3_insn): Update type attribute.
916 (xordi3_insn): Likewise.
917 (one_cmpldi2): Likewise.
918 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
919 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
920 (*movmisalign<mode>_neon_store): Likewise.
921 (*movmisalign<mode>_neon_load): Likewise.
922 (vec_set<mode>_internal): Likewise.
923 (vec_set<mode>_internal): Likewise.
924 (vec_setv2di_internal): Likewise.
925 (vec_extract<mode>): Likewise.
926 (vec_extract<mode>): Likewise.
927 (vec_extractv2di): Likewise.
928 (*add<mode>3_neon): Likewise.
929 (adddi3_neon): Likewise.
930 (*sub<mode>3_neon): Likewise.
931 (subdi3_neon): Likewise.
932 (fma<VCVTF:mode>4): Likewise.
933 (fma<VCVTF:mode>4_intrinsic): Likewise.
934 (*fmsub<VCVTF:mode>4): Likewise.
935 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
936 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
937 (ior<mode>3): Likewise.
938 (and<mode>3): Likewise.
939 (orn<mode>3_neon): Likewise.
940 (orndi3_neon): Likewise.
941 (bic<mode>3_neon): Likewise.
942 (bicdi3_neon): Likewise.
943 (xor<mode>3): Likewise.
944 (one_cmpl<mode>2): Likewise.
945 (abs<mode>2): Likewise.
946 (neg<mode>2): Likewise.
947 (negdi2_neon): Likewise.
948 (*umin<mode>3_neon): Likewise.
949 (*umax<mode>3_neon): Likewise.
950 (*smin<mode>3_neon): Likewise.
951 (*smax<mode>3_neon): Likewise.
952 (vashl<mode>3): Likewise.
953 (vashr<mode>3_imm): Likewise.
954 (vlshr<mode>3_imm): Likewise.
955 (ashl<mode>3_signed): Likewise.
956 (ashl<mode>3_unsigned): Likewise.
957 (neon_load_count): Likewise.
958 (ashldi3_neon_noclobber): Likewise.
959 (ashldi3_neon): Likewise.
960 (signed_shift_di3_neon): Likewise.
961 (unsigned_shift_di3_neon): Likewise.
962 (ashrdi3_neon_imm_noclobber): Likewise.
963 (lshrdi3_neon_imm_noclobber): Likewise.
964 (<shift>di3_neon): Likewise.
965 (widen_ssum<mode>3): Likewise.
966 (widen_usum<mode>3): Likewise.
967 (quad_halves_<code>v4si): Likewise.
968 (quad_halves_<code>v4sf): Likewise.
969 (quad_halves_<code>v8hi): Likewise.
970 (quad_halves_<code>v16qi): Likewise.
971 (reduc_splus_v2di): Likewise.
972 (neon_vpadd_internal<mode>): Likewise.
973 (neon_vpsmin<mode>): Likewise.
974 (neon_vpsmax<mode>): Likewise.
975 (neon_vpumin<mode>): Likewise.
976 (neon_vpumax<mode>): Likewise.
977 (*ss_add<mode>_neon): Likewise.
978 (*us_add<mode>_neon): Likewise.
979 (*ss_sub<mode>_neon): Likewise.
980 (*us_sub<mode>_neon): Likewise.
981 (neon_vadd<mode>_unspec): Likewise.
982 (neon_vaddl<mode>): Likewise.
983 (neon_vaddw<mode>): Likewise.
984 (neon_vhadd<mode>): Likewise.
985 (neon_vqadd<mode>): Likewise.
986 (neon_vaddhn<mode>): Likewise.
987 (neon_vmul<mode>): Likewise.
988 (neon_vfms<VCVTF:mode>): Likewise.
989 (neon_vmlal<mode>): Likewise.
990 (neon_vmls<mode>): Likewise.
991 (neon_vmlsl<mode>): Likewise.
992 (neon_vqdmulh<mode>): Likewise.
993 (neon_vqdmlal<mode>): Likewise.
994 (neon_vqdmlsl<mode>): Likewise.
995 (neon_vmull<mode>): Likewise.
996 (neon_vqdmull<mode>): Likewise.
997 (neon_vsub<mode>_unspec): Likewise.
998 (neon_vsubl<mode>): Likewise.
999 (neon_vsubw<mode>): Likewise.
1000 (neon_vqsub<mode>): Likewise.
1001 (neon_vhsub<mode>): Likewise.
1002 (neon_vsubhn<mode>): Likewise.
1003 (neon_vceq<mode>): Likewise.
1004 (neon_vcge<mode>): Likewise.
1005 (neon_vcgeu<mode>): Likewise.
1006 (neon_vcgt<mode>): Likewise.
1007 (neon_vcgtu<mode>): Likewise.
1008 (neon_vcle<mode>): Likewise.
1009 (neon_vclt<mode>): Likewise.
1010 (neon_vcage<mode>): Likewise.
1011 (neon_vcagt<mode>): Likewise.
1012 (neon_vtst<mode>): Likewise.
1013 (neon_vabd<mode>): Likewise.
1014 (neon_vabdl<mode>): Likewise.
1015 (neon_vaba<mode>): Likewise.
1016 (neon_vabal<mode>): Likewise.
1017 (neon_vmax<mode>): Likewise.
1018 (neon_vmin<mode>): Likewise.
1019 (neon_vpaddl<mode>): Likewise.
1020 (neon_vpadal<mode>): Likewise.
1021 (neon_vpmax<mode>): Likewise.
1022 (neon_vpmin<mode>): Likewise.
1023 (neon_vrecps<mode>): Likewise.
1024 (neon_vrsqrts<mode>): Likewise.
1025 (neon_vqabs<mode>): Likewise.
1026 (neon_vqneg<mode>): Likewise.
1027 (neon_vcls<mode>): Likewise.
1028 (clz<mode>2): Likewise.
1029 (popcount<mode>2): Likewise.
1030 (neon_vrecpe<mode>): Likewise.
1031 (neon_vrsqrte<mode>): Likewise.
1032 (neon_vget_lane<mode>_sext_internal): Likewise.
1033 (neon_vget_lane<mode>_zext_internal): Likewise.
1034 (neon_vdup_n<mode>): Likewise.
1035 (neon_vdup_n<mode>): Likewise.
1036 (neon_vdup_nv2di): Likewise.
1037 (neon_vdup_lane<mode>_interal): Likewise.
1038 (*neon_vswp<mode>): Likewise.
1039 (neon_vcombine<mode>): Likewise.
1040 (float<mode><V_cvtto>2): Likewise.
1041 (floatuns<mode><V_cvtto>2): Likewise.
1042 (fix_trunc<mode><V_cvtto>2): Likewise.
1043 (fixuns_trunc<mode><V_cvtto>2
1044 (neon_vcvt<mode>): Likewise.
1045 (neon_vcvt<mode>): Likewise.
1046 (neon_vcvtv4sfv4hf): Likewise.
1047 (neon_vcvtv4hfv4sf): Likewise.
1048 (neon_vcvt_n<mode>): Likewise.
1049 (neon_vcvt_n<mode>): Likewise.
1050 (neon_vmovn<mode>): Likewise.
1051 (neon_vqmovn<mode>): Likewise.
1052 (neon_vqmovun<mode>): Likewise.
1053 (neon_vmovl<mode>): Likewise.
1054 (neon_vmul_lane<mode>): Likewise.
1055 (neon_vmul_lane<mode>): Likewise.
1056 (neon_vmull_lane<mode>): Likewise.
1057 (neon_vqdmull_lane<mode>): Likewise.
1058 (neon_vqdmulh_lane<mode>): Likewise.
1059 (neon_vqdmulh_lane<mode>): Likewise.
1060 (neon_vmla_lane<mode>): Likewise.
1061 (neon_vmla_lane<mode>): Likewise.
1062 (neon_vmlal_lane<mode>): Likewise.
1063 (neon_vqdmlal_lane<mode>): Likewise.
1064 (neon_vmls_lane<mode>): Likewise.
1065 (neon_vmls_lane<mode>): Likewise.
1066 (neon_vmlsl_lane<mode>): Likewise.
1067 (neon_vqdmlsl_lane<mode>): Likewise.
1068 (neon_vext<mode>): Likewise.
1069 (neon_vrev64<mode>): Likewise.
1070 (neon_vrev32<mode>): Likewise.
1071 (neon_vrev16<mode>): Likewise.
1072 (neon_vbsl<mode>_internal): Likewise.
1073 (neon_vshl<mode>): Likewise.
1074 (neon_vqshl<mode>): Likewise.
1075 (neon_vshr_n<mode>): Likewise.
1076 (neon_vshrn_n<mode>): Likewise.
1077 (neon_vqshrn_n<mode>): Likewise.
1078 (neon_vqshrun_n<mode>): Likewise.
1079 (neon_vshl_n<mode>): Likewise.
1080 (neon_vqshl_n<mode>): Likewise.
1081 (neon_vqshlu_n<mode>): Likewise.
1082 (neon_vshll_n<mode>): Likewise.
1083 (neon_vsra_n<mode>): Likewise.
1084 (neon_vsri_n<mode>): Likewise.
1085 (neon_vsli_n<mode>): Likewise.
1086 (neon_vtbl1v8qi): Likewise.
1087 (neon_vtbl2v8qi): Likewise.
1088 (neon_vtbl3v8qi): Likewise.
1089 (neon_vtbl4v8qi): Likewise.
1090 (neon_vtbl1v16qi): Likewise.
1091 (neon_vtbl2v16qi): Likewise.
1092 (neon_vcombinev16qi): Likewise.
1093 (neon_vtbx1v8qi): Likewise.
1094 (neon_vtbx2v8qi): Likewise.
1095 (neon_vtbx3v8qi): Likewise.
1096 (neon_vtbx4v8qi): Likewise.
1097 (*neon_vtrn<mode>_insn): Likewise.
1098 (*neon_vzip<mode>_insn): Likewise.
1099 (*neon_vuzp<mode>_insn): Likewise.
1100 (neon_vld1<mode>): Likewise.
1101 (neon_vld1_lane<mode>): Likewise.
1102 (neon_vld1_lane<mode>): Likewise.
1103 (neon_vld1_dup<mode>): Likewise.
1104 (neon_vld1_dup<mode>): Likewise.
1105 (neon_vld1_dupv2di): Likewise.
1106 (neon_vst1<mode>): Likewise.
1107 (neon_vst1_lane<mode>): Likewise.
1108 (neon_vst1_lane<mode>): Likewise.
1109 (neon_vld2<mode>): Likewise.
1110 (neon_vld2<mode>): Likewise.
1111 (neon_vld2_lane<mode>): Likewise.
1112 (neon_vld2_lane<mode>): Likewise.
1113 (neon_vld2_dup<mode>): Likewise.
1114 (neon_vst2<mode>): Likewise.
1115 (neon_vst2<mode>): Likewise.
1116 (neon_vst2_lane<mode>): Likewise.
1117 (neon_vst2_lane<mode>): Likewise.
1118 (neon_vld3<mode>): Likewise.
1119 (neon_vld3qa<mode>): Likewise.
1120 (neon_vld3qb<mode>): Likewise.
1121 (neon_vld3_lane<mode>): Likewise.
1122 (neon_vld3_lane<mode>): Likewise.
1123 (neon_vld3_dup<mode>): Likewise.
1124 (neon_vst3<mode>): Likewise.
1125 (neon_vst3qa<mode>): Likewise.
1126 (neon_vst3qb<mode>): Likewise.
1127 (neon_vst3_lane<mode>): Likewise.
1128 (neon_vst3_lane<mode>): Likewise.
1129 (neon_vld4<mode>): Likewise.
1130 (neon_vld4qa<mode>): Likewise.
1131 (neon_vld4qb<mode>): Likewise.
1132 (neon_vld4_lane<mode>): Likewise.
1133 (neon_vld4_lane<mode>): Likewise.
1134 (neon_vld4_dup<mode>): Likewise.
1135 (neon_vst4<mode>): Likewise.
1136 (neon_vst4qa<mode>): Likewise.
1137 (neon_vst4qb<mode>): Likewise.
1138 (neon_vst4_lane<mode>): Likewise.
1139 (neon_vst4_lane<mode>): Likewise.
1140 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1141 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1142 (neon_vec_<US>mult_lo_<mode>): Likewise.
1143 (neon_vec_<US>mult_hi_<mode>): Likewise.
1144 (neon_vec_<US>shiftl_<mode>): Likewise.
1145 (neon_unpack<US>_<mode>): Likewise.
1146 (neon_vec_<US>mult_<mode>): Likewise.
1147 (vec_pack_trunc_<mode>): Likewise.
1148 (neon_vec_pack_trunc_<mode>): Likewise.
1149 (neon_vabd<mode>_2): Likewise.
1150 (neon_vabd<mode>_3): Likewise.
1152 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1154 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1155 (load_pair): Update type attribute.
1156 (store_pair): Update type attribute.
1157 * config/aarch64/iterators.md (q): New.
1159 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1161 * config/arm/types.md: Add new types for Neon insns.
1163 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1164 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1165 Sergey Lega <sergey.s.lega@intel.com>
1166 Anna Tikhonova <anna.tikhonova@intel.com>
1167 Ilya Tocar <ilya.tocar@intel.com>
1168 Andrey Turetskiy <andrey.turetskiy@intel.com>
1169 Ilya Verbin <ilya.verbin@intel.com>
1170 Kirill Yukhin <kirill.yukhin@intel.com>
1171 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1173 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1174 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1175 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1177 (srcp14<mode>): Ditto.
1178 (rsqrt14<mode>): Ditto.
1179 (rsqrt14<mode>): Ditto.
1180 (avx512f_vmscalef<mode>): Ditto.
1181 (avx512f_scalef<mode>): Ditto.
1182 (avx512f_getexp<mode>): Ditto.
1183 (avx512f_sgetexp<mode>): Ditto.
1184 (avx512f_fixupimm<mode>): Ditto.
1185 (avx512f_sfixupimm<mode>): Ditto.
1186 (avx512f_rndscale<mode>): Ditto.
1187 (*avx512er_exp2<mode>): Ditto.
1188 (*avx512er_rcp28<mode>): Ditto.
1189 (avx512er_rsqrt28<mode>): Ditto.
1190 (avx512f_getmant<mode>): Ditto.
1191 (avx512f_getmant<mode>): Ditto.
1192 (avx512f_rndscale<mode>): Fix formatting.
1194 2013-10-15 Martin Jambor <mjambor@suse.cz>
1196 * ipa-utils.h (ipa_edge_within_scc): Declare.
1197 * ipa-cp.c (edge_within_scc): Moved...
1198 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1200 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1201 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1202 Sergey Lega <sergey.s.lega@intel.com>
1203 Anna Tikhonova <anna.tikhonova@intel.com>
1204 Ilya Tocar <ilya.tocar@intel.com>
1205 Andrey Turetskiy <andrey.turetskiy@intel.com>
1206 Ilya Verbin <ilya.verbin@intel.com>
1207 Kirill Yukhin <kirill.yukhin@intel.com>
1208 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1210 * config/i386/predicates.md (const_8_to_15_operand): New.
1211 (const_16_to_31_operand): Ditto.
1212 * config/i386/sse.md (V8FI): New.
1214 (reduc_splus_v8df): Ditto.
1215 (reduc_splus_v16sf): Ditto.
1216 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1217 (vec_extract_hi_<mode>): Ditto.
1218 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1219 (vec_set_lo_<mode>): Ditto.
1220 (vec_set_hi_<mode>): Ditto.
1221 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1222 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1223 (avx512f_pshufd_1): Ditto.
1224 (avx512f_broadcast<mode>): Ditto.
1225 (avx512f_broadcast<mode>): Ditto.
1226 (define_split): Split vec_extract_lo into move.
1227 (ssequartermode): Ditto.
1228 (ssedoublemode): Extened with wider modes.
1229 (vec_extract_lo_<mode>): Ditto.
1231 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1232 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1233 Sergey Lega <sergey.s.lega@intel.com>
1234 Anna Tikhonova <anna.tikhonova@intel.com>
1235 Ilya Tocar <ilya.tocar@intel.com>
1236 Andrey Turetskiy <andrey.turetskiy@intel.com>
1237 Ilya Verbin <ilya.verbin@intel.com>
1238 Kirill Yukhin <kirill.yukhin@intel.com>
1239 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1241 * config/i386/predicates.md (register_or_constm1_operand): New.
1242 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1243 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1244 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1245 UNSPEC_SCATTER_PREFETCH
1247 (avx512f_ucmp<mode>3): Ditto.
1248 (avx512f_vternlog<mode>): Ditto.
1249 (avx512f_align<mode>): Ditto.
1250 (<shift_insn><mode>3): Ditto.
1251 (avx512f_<rotate>v<mode>): Ditto.
1252 (avx512f_<rotate><mode>): Ditto.
1253 (avx512f_eq<mode>3): Ditto.
1254 (avx512f_eq<mode>3_1): Ditto.
1255 (avx512f_gt<mode>3): Ditto.
1256 (avx512f_testm<mode>3): Ditto.
1257 (avx512f_testnm<mode>3): Ditto.
1258 (avx512pf_gatherpf<mode>): Ditto.
1259 (*avx512pf_gatherpf<mode>_mask): Ditto.
1260 (*avx512pf_gatherpf<mode>): Ditto.
1261 (avx512pf_scatterpf<mode>): Ditto.
1262 (*avx512pf_scatterpf<mode>_mask): Ditto.
1263 (*avx512pf_scatterpf<mode>): Ditto.
1264 (avx512f_vec_dup_gpr<mode>): Ditto.
1265 (clz<mode>2): Ditto.
1266 (conflict<mode>): Ditto.
1267 (REDUC_SMINMAX_MODE): Extened with wider modes.
1268 (reduc_<code>_<mode>): Ditto.
1269 (vlshr<mode>3): Ditto.
1270 (vashl<mode>3): Ditto.
1272 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1273 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1274 Sergey Lega <sergey.s.lega@intel.com>
1275 Anna Tikhonova <anna.tikhonova@intel.com>
1276 Ilya Tocar <ilya.tocar@intel.com>
1277 Andrey Turetskiy <andrey.turetskiy@intel.com>
1278 Ilya Verbin <ilya.verbin@intel.com>
1279 Kirill Yukhin <kirill.yukhin@intel.com>
1280 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1282 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1285 (avx512fmaskmode): Ditto.
1286 (bcstscalarsuff): Ditto.
1287 (avx512f_blendm<mode>): Ditto.
1288 (cmp_imm_predicate): Ditto.
1289 (avx512f_cmp<mode>3): Ditto.
1290 (avx512f_vec_dup<mode>): Ditto.
1291 (avx512f_vec_dup_mem<mode>): Ditto.
1292 (avx512f_vpermi2var<mode>3): Ditto.
1293 (avx512f_vpermt2var<mode>3): Ditto.
1294 (vec_init<mode>): Ditto.
1295 (avx512f_gathersi<mode>): Ditto.
1296 (*avx512f_gathersi<mode>): Ditto.
1297 (*avx512f_gathersi<mode>_2): Ditto.
1298 (avx512f_gatherdi<mode>): Ditto.
1299 (*avx512f_gatherdi<mode>): Ditto.
1300 (*avx512f_gatherdi<mode>_2): Ditto.
1301 (avx512f_scattersi<mode>): Ditto.
1302 (*avx512f_scattersi<mode>): Ditto.
1303 (avx512f_scatterdi<mode>): Ditto.
1304 (*avx512f_scatterdi<mode>): Ditto.
1305 (sseintprefix): Extened with wider modes.
1306 (VEC_GATHER_IDXSI): Ditto.
1307 (VEC_GATHER_IDXDI): Ditto.
1308 (VEC_GATHER_SRCDI): Ditto.
1310 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1311 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1313 * config/arm/t-aprofile: New file.
1314 * config.gcc: Handle --with-multilib-list option.
1316 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1318 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1321 2013-10-15 Richard Biener <rguenther@suse.de>
1323 * tree-loop-distribution.c (build_empty_rdg): Inline into
1325 (rdg_flag_vertex): Inline into single user.
1326 (rdg_flag_vertex_and_dependent): Likewise.
1327 (build_rdg_partition_for_vertex): Remove processed bitmap.
1328 (rdg_build_partitions): Simplify.
1330 2013-10-15 Richard Biener <rguenther@suse.de>
1332 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1333 Restructure forwarding through conversions and copies to
1334 avoid performing copy-propagation the wrong way. Adjust
1335 recursion invocations.
1336 (forward_propagate_addr_expr): Add argument stating if we
1337 are recursing from a single-use.
1338 (ssa_forward_propagate_and_combine): Adjust.
1340 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1342 * dumpfile.h (gcc::dump_manager): New class, to hold state
1343 relating to dumpfile management.
1344 (get_dump_file_name): Remove in favor of method of dump_manager.
1345 (dump_initialized_p): Likewise.
1346 (dump_start): Likewise.
1347 (dump_finish): Likewise.
1348 (dump_switch_p): Likewise.
1349 (dump_register): Likewise.
1350 (get_dump_file_info): Likewise.
1351 * context.c (gcc::context::context): Construct the dump_manager
1353 * context.h (gcc::context::get_dumps): New.
1354 (gcc::context::m_dumps): New.
1355 * coverage.c (coverage_init): Port to dump_manager API.
1356 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1357 (extra_dump_files_in_use): Likewise.
1358 (extra_dump_files_alloced): Likewise.
1359 (gcc::dump_manager::dump_manager): New.
1360 (dump_register): Convert to...
1361 (gcc::dump_manager::dump_register): ...method, replacing
1362 function-static next_dump with m_next_dump field.
1363 (get_dump_file_info): Convert to...
1364 (gcc::dump_manager::get_dump_file_info): ...method.
1365 (get_dump_file_name): Convert to...
1366 (gcc::dump_manager::get_dump_file_name): ...method.
1367 (dump_start): Convert to...
1368 (gcc::dump_manager::dump_start): ...method.
1369 (dump_finish): Convert to...
1370 (gcc::dump_manager::dump_finish): ...method.
1371 (dump_begin): Replace body with...
1372 (gcc::dump_manager::dump_begin): ...new method.
1373 (dump_phase_enabled_p): Convert to...
1374 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1375 (dump_phase_enabled_p): Convert to...
1376 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1377 (dump_initialized_p): Convert to...
1378 (gcc::dump_manager::dump_initialized_p): ...method.
1379 (dump_flag_name): Replace body with...
1380 (gcc::dump_manager::dump_flag_name): ...new method.
1381 (dump_enable_all): Convert to...
1382 (gcc::dump_manager::dump_enable_all): ...new method.
1383 (opt_info_enable_passes): Convert to...
1384 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1385 (dump_switch_p_1): Convert to...
1386 (gcc::dump_manager::dump_switch_p_1): ...new method.
1387 (dump_switch_p): Convert to...
1388 (gcc::dump_manager::dump_switch_p): ...new method.
1389 (opt_info_switch_p): Port to dump_manager API.
1390 (enable_rtl_dump_file): Likewise.
1391 * opts-global.c (handle_common_deferred_options): Port to new
1393 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1394 (pass_manager::register_one_dump_file): Likewise.
1395 (pass_manager::register_pass): Likewise.
1396 (pass_init_dump_file): Likewise.
1397 (pass_fini_dump_file): Likewise.
1398 * statistics.c (statistics_early_init): Likewise.
1400 2013-10-14 Richard Biener <rguenther@suse.de>
1402 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1403 iterative_hash_canonical_type, gimple_canonical_type_hash,
1404 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1405 gimple_register_canonical_type, print_gimple_types_stats,
1406 free_gimple_type_tables): Move to lto/lto.c
1407 (gt-gimple.h): Do not include.
1408 * gimple.h (gimple_register_canonical_type,
1409 print_gimple_types_stats, free_gimple_type_tables): Remove.
1410 * Makefile.in (GTFILES): Remove gimple.c.
1412 2013-10-14 Travis Snoozy <quandary@remstate.com>
1415 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1416 scanning for msp430x targets.
1418 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1421 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1422 (registers_ok_for_ldd_peep): Move around.
1423 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1424 * config/sparc/sparc.md (widening peepholes): Use it.
1426 2013-10-14 Richard Biener <rguenther@suse.de>
1430 * gimple.c (iterative_hash_canonical_type): Make sure to
1431 record the hash into the correct hashtable slot.
1433 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1435 PR rtl-optimization/58662
1436 * combine.c (try_combine): Take into account death nodes on I2 when
1437 splitting a PARALLEL of two independent SETs. Fix dump message.
1439 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1442 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1443 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1444 * config/sh/t-sh (sh_treg_combine.o): New entry.
1445 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1446 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1447 (register_sh_passes): New function. Register sh_treg_combine pass.
1448 (sh_option_override): Invoke it.
1449 (sh_canonicalize_comparison): Handle op0_preserve_value.
1450 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1451 opportunities. Canonicalize branch condition.
1452 (nott): Allow only if pseudos can be created for non-SH2A.
1454 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1457 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1458 (ix86_expand_movmem): Replace copy_addr_to_reg with
1459 ix86_copy_addr_to_reg.
1460 (ix86_expand_setmem): Likewise.
1462 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1464 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1465 provided by ix86_fp_compare_mode instead of CCFPUmode.
1467 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1469 * config/aarch64/arm_neon.h
1470 (vtbx<1,3>_<psu>8): Fix register constriants.
1472 2013-10-11 Jeff Law <law@redhat.com>
1474 PR tree-optimization/58640
1475 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1476 threading paths that cross over two loop entry points.
1478 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1480 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1481 handle vector float as well.
1482 (*vsx_le_perm_load_v4si): Likewise.
1483 (*vsx_le_perm_store_v2di): Likewise.
1484 (*vsx_le_perm_store_v4si): Likewise.
1486 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1488 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1489 directly to circumvent subtract from splat{31} workaround.
1490 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1492 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1493 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1494 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1495 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1496 separate big and little endian logic.
1497 (*altivec_vperm_<mode>_internal): New define_insn.
1498 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1499 separate big and little endian logic.
1500 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1501 (vec_permv16qi): Add little endian logic.
1503 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1505 * doc/extend.texi (returns_nonnull): Remove arguments.
1507 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1508 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1509 Sergey Lega <sergey.s.lega@intel.com>
1510 Anna Tikhonova <anna.tikhonova@intel.com>
1511 Ilya Tocar <ilya.tocar@intel.com>
1512 Andrey Turetskiy <andrey.turetskiy@intel.com>
1513 Ilya Verbin <ilya.verbin@intel.com>
1514 Kirill Yukhin <kirill.yukhin@intel.com>
1515 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1517 * config/i386/sse.md (VI48F_256_512): New.
1518 (avx2_permvar<mode>): Change to ...
1519 (<avx2_avx512f>_permvar<mode>): This.
1521 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1522 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1523 Sergey Lega <sergey.s.lega@intel.com>
1524 Anna Tikhonova <anna.tikhonova@intel.com>
1525 Ilya Tocar <ilya.tocar@intel.com>
1526 Andrey Turetskiy <andrey.turetskiy@intel.com>
1527 Ilya Verbin <ilya.verbin@intel.com>
1528 Kirill Yukhin <kirill.yukhin@intel.com>
1529 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1531 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1532 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1533 * config/i386/sse.md (VI4_AVX): New.
1534 (sf2simodelower): Ditto.
1535 (sse2_cvtps2dq): Change to ...
1536 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1538 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1539 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1540 Sergey Lega <sergey.s.lega@intel.com>
1541 Anna Tikhonova <anna.tikhonova@intel.com>
1542 Ilya Tocar <ilya.tocar@intel.com>
1543 Andrey Turetskiy <andrey.turetskiy@intel.com>
1544 Ilya Verbin <ilya.verbin@intel.com>
1545 Kirill Yukhin <kirill.yukhin@intel.com>
1546 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1548 * config/i386/sse.md (V_512): New.
1550 (vcond<V_512:mode><VF_512:mode>): Ditto.
1551 (vcond<V_512:mode><VI_512:mode>): Ditto.
1552 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1554 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1555 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1556 Sergey Lega <sergey.s.lega@intel.com>
1557 Anna Tikhonova <anna.tikhonova@intel.com>
1558 Ilya Tocar <ilya.tocar@intel.com>
1559 Andrey Turetskiy <andrey.turetskiy@intel.com>
1560 Ilya Verbin <ilya.verbin@intel.com>
1561 Kirill Yukhin <kirill.yukhin@intel.com>
1562 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1564 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1565 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1567 (fma<mode>4): Removed condition.
1568 (fms<mode>4): Ditto.
1569 (fnma<mode>4): Ditto.
1570 (fnms<mode>4): Ditto.
1571 (fma4i_fmadd_<mode>): Ditto.
1572 (*fma_fmadd_<mode>): Ditto.
1573 (*fma_fmsub_<mode>): Ditto.
1574 (*fma_fnmadd_<mode>): Ditto.
1575 (*fma_fnmsub_<mode>): Ditto.
1576 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1577 (*fma_fmaddsub_<mode>): Ditto.
1578 (*fma_fmsubadd_<mode>): Ditto.
1579 (*fmai_fmadd_<mode>): Ditto.
1580 (*fmai_fmsub_<mode>): Ditto.
1581 (*fmai_fnmadd_<mode>): Ditto.
1582 (*fmai_fnmsub_<mode>): Ditto.
1584 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1585 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1586 Sergey Lega <sergey.s.lega@intel.com>
1587 Anna Tikhonova <anna.tikhonova@intel.com>
1588 Ilya Tocar <ilya.tocar@intel.com>
1589 Andrey Turetskiy <andrey.turetskiy@intel.com>
1590 Ilya Verbin <ilya.verbin@intel.com>
1591 Kirill Yukhin <kirill.yukhin@intel.com>
1592 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1594 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1595 (VI124_256): Changed to ...
1596 (VI124_256_48_512): This.
1597 (ssepackmode): Extended with wider modes.
1598 (<code><mode>3): Changed iterator.
1599 (*avx2_<code><mode>3): Ditto.
1600 (vec_pack_trunc_<mode>): Ditto.
1602 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1603 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1604 Sergey Lega <sergey.s.lega@intel.com>
1605 Anna Tikhonova <anna.tikhonova@intel.com>
1606 Ilya Tocar <ilya.tocar@intel.com>
1607 Andrey Turetskiy <andrey.turetskiy@intel.com>
1608 Ilya Verbin <ilya.verbin@intel.com>
1609 Kirill Yukhin <kirill.yukhin@intel.com>
1610 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1612 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1613 (VI8F_256_512): Ditto.
1614 (abs<mode>2): Changed iterator.
1615 (avx2_perm<mode>): Changed to ...
1616 (<avx2_avx512f>_perm<mode>): This.
1617 (avx2_perm<mode>_1): Changed to ...
1618 (<avx2_avx512f>_perm<mode>_1): This.
1620 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1621 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1622 Sergey Lega <sergey.s.lega@intel.com>
1623 Anna Tikhonova <anna.tikhonova@intel.com>
1624 Ilya Tocar <ilya.tocar@intel.com>
1625 Andrey Turetskiy <andrey.turetskiy@intel.com>
1626 Ilya Verbin <ilya.verbin@intel.com>
1627 Kirill Yukhin <kirill.yukhin@intel.com>
1628 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1630 * config/i386/sse.md (VI48_AVX512F): New.
1631 (VI48_AVX2): Changed to ...
1632 (VI48_AVX2_48_AVX512F): This.
1633 (avx2_ashrv<mode>): Changed to ...
1634 (<avx2_avx512f>_ashrv<mode>): This.
1635 (avx2_<shift_insn>v<mode>): Changed to ...
1636 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1638 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1639 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1640 Sergey Lega <sergey.s.lega@intel.com>
1641 Anna Tikhonova <anna.tikhonova@intel.com>
1642 Ilya Tocar <ilya.tocar@intel.com>
1643 Andrey Turetskiy <andrey.turetskiy@intel.com>
1644 Ilya Verbin <ilya.verbin@intel.com>
1645 Kirill Yukhin <kirill.yukhin@intel.com>
1646 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1648 * config/i386/sse.md (VI4_AVX512F): New.
1649 (VI8_AVX2_AVX512F): Ditto.
1650 (mul<mode>3): Extended with wider modes.
1651 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1652 (mul<mode>3): Ditto.
1653 (vec_widen_<s>mult_odd_<mode>): Ditto.
1655 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1656 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1657 Sergey Lega <sergey.s.lega@intel.com>
1658 Anna Tikhonova <anna.tikhonova@intel.com>
1659 Ilya Tocar <ilya.tocar@intel.com>
1660 Andrey Turetskiy <andrey.turetskiy@intel.com>
1661 Ilya Verbin <ilya.verbin@intel.com>
1662 Kirill Yukhin <kirill.yukhin@intel.com>
1663 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1665 * config/i386/sse.md (VI2_AVX512F): New.
1666 (VI124_AVX512F): Ditto.
1667 (sseunpackmode): Extended with wider modes.
1668 (sseunpackfltmode): Ditto.
1669 (vec_unpacks_float_hi_<mode>): Ditto.
1670 (vec_unpacks_float_lo_<mode>): Ditto.
1671 (vec_unpacku_float_hi_<mode>): Ditto.
1672 (vec_unpacku_float_lo_<mode>): Ditto.
1673 (vec_unpacks_lo_<mode>): Ditto.
1674 (vec_unpacks_hi_<mode>): Ditto.
1675 (vec_unpacku_lo_<mode>): Ditto.
1676 (vec_unpacku_hi_<mode>): Ditto.
1678 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1679 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1680 Sergey Lega <sergey.s.lega@intel.com>
1681 Anna Tikhonova <anna.tikhonova@intel.com>
1682 Ilya Tocar <ilya.tocar@intel.com>
1683 Andrey Turetskiy <andrey.turetskiy@intel.com>
1684 Ilya Verbin <ilya.verbin@intel.com>
1685 Kirill Yukhin <kirill.yukhin@intel.com>
1686 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1688 * config/i386/i386.md (multdiv): New.
1689 (multdiv_mnemonic): Ditto.
1690 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1691 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1692 (<sse>_vmdiv<mode>3): Removed.
1694 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1695 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1696 Sergey Lega <sergey.s.lega@intel.com>
1697 Anna Tikhonova <anna.tikhonova@intel.com>
1698 Ilya Tocar <ilya.tocar@intel.com>
1699 Andrey Turetskiy <andrey.turetskiy@intel.com>
1700 Ilya Verbin <ilya.verbin@intel.com>
1701 Kirill Yukhin <kirill.yukhin@intel.com>
1702 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1704 * config/i386/sse.md (V): Extended with wider modes.
1706 (ssehalfvecmode): Ditto.
1708 (ssepackfltmode): Ditto.
1709 (avx_vec_concat<mode>): Ditto.
1710 (V_256_512): New iterator.
1711 (VF2_512_256): Ditto.
1712 (si2dfmode): New attribute.
1713 (si2dfmodelower): Ditto.
1715 (concat_tg_mode): Ditto.
1716 (floatv4siv4df2): Changed to ...
1717 (float<si2dfmodelower><mode>2): This.
1718 (avx_cvtps2pd256): Changed to ...
1719 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1720 (vec_pack_trunc_v4df): Changed to ...
1721 (vec_pack_trunc_<mode>): This.
1722 (avx_vpermil<mode>): Changed to ...
1723 (<sse2_avx_avx512f>_vpermil<mode>): This.
1724 (<fixsuffix>fix_truncv8dfv8si2): New.
1725 (vec_pack_sfix_trunc_v8df): Ditto.
1726 (avx512f_rndscale<mode>): Ditto.
1727 (avx512f_roundpd512): Ditto.
1728 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1730 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1731 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1732 Sergey Lega <sergey.s.lega@intel.com>
1733 Anna Tikhonova <anna.tikhonova@intel.com>
1734 Ilya Tocar <ilya.tocar@intel.com>
1735 Andrey Turetskiy <andrey.turetskiy@intel.com>
1736 Ilya Verbin <ilya.verbin@intel.com>
1737 Kirill Yukhin <kirill.yukhin@intel.com>
1738 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1740 * config/i386/i386.md (any_fix): New iterator.
1741 (fixsuffix): New attribute.
1742 * config/i386/sse.md (VF1): Extened with wider modes.
1746 (sseintvecmodelower): Ditto.
1747 (ssescalarmode): Ditto.
1748 (ssescalarnum): Ditto.
1750 (ssexmmmode): Ditto.
1751 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1752 (<sse>_rcp<mode>2): Change iterator.
1753 (rsqrt<mode>2): Ditto.
1754 (<sse>_rsqrt<mode>2): Ditto.
1755 (avx2_vec_dup<mode>): Ditto.
1756 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1757 (round<mode>2_sfix): Ditto.
1758 (avx2_pbroadcast<mode>): Ditto.
1759 (*andnot<mode>3): Handle XI mode.
1760 (*<code><mode>3): Ditto.
1761 (AVXTOSSEMODE): Removed.
1762 (avx_vpermil<mode>): Changed to ...
1763 (<sse2_avx_avx512f>_vpermil<mode>): This.
1765 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1766 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1767 Sergey Lega <sergey.s.lega@intel.com>
1768 Anna Tikhonova <anna.tikhonova@intel.com>
1769 Ilya Tocar <ilya.tocar@intel.com>
1770 Andrey Turetskiy <andrey.turetskiy@intel.com>
1771 Ilya Verbin <ilya.verbin@intel.com>
1772 Kirill Yukhin <kirill.yukhin@intel.com>
1773 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1775 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1776 (<sse>_comi): Ditto.
1777 (<sse>_ucomi): Ditto.
1778 (sse_cvtss2siq_2): Ditto.
1779 (sse2_cvtsd2si): Ditto.
1780 (sse2_cvtsd2siq): Ditto.
1781 (sse2_cvttsd2si): Ditto.
1782 (sse2_cvttsd2siq): Ditto.
1783 (<shift_insn><mode>3): Ditto.
1784 (sse2_cvtsi2sdq): Update constraint and prefix.
1785 (sse_cvtsi2ss): Update prefix.
1786 (sse_cvtsi2ssq): Ditto.
1788 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1790 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1791 ignore internal calls.
1793 2013-10-11 Richard Biener <rguenther@suse.de>
1795 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1796 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1797 TREE_OVERFLOW is set.
1799 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1801 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1803 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1804 * gimple.h: Likewise.
1806 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1808 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1811 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1813 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1814 * configure: Regenerate.
1816 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1818 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1819 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1820 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1821 similarly pp_right_paren.
1822 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1823 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1824 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1825 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1827 (omp_declare_simd_clauses_equal,
1828 omp_remove_redundant_declare_simd_attrs): New functions.
1829 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1830 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1831 * tree.h (OMP_LOOP_CHECK): Define.
1832 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1833 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1834 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1835 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1836 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1837 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1838 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1839 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1840 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1841 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1842 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1843 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1844 (omp_remove_redundant_declare_simd_attrs): New prototype.
1845 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1846 GIMPLE_OMP_TEAMS): New codes.
1847 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1848 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1850 (target_nesting_level): New variable.
1851 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1852 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1853 for collapse > 1 static schedule unless ordered.
1854 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1855 (determine_parallel_type): Adjust get_ws_args_for caller.
1856 (install_var_field): Handle mask & 4 for double indirection.
1857 (scan_sharing_clauses): Ignore shared clause on teams construct.
1858 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1859 (create_omp_child_function): If inside target or declare target
1860 constructs, set "omp declare target" attribute on the child function.
1861 (find_combined_for): New function.
1862 (scan_omp_parallel): Handle combined loops.
1863 (scan_omp_target, scan_omp_teams): New functions.
1864 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1865 restrictions and set ctx->cancellable for cancellable constructs.
1866 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1867 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1868 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1869 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1870 (omp_clause_aligned_alignment): New function.
1871 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
1872 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1873 on teams constructs. Handle user defined reductions and new
1875 (lower_reduction_clauses): Don't set placeholder to address of ref
1876 if it has already the right type.
1877 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1878 (expand_parallel_call): Use the new non-_start suffixed builtins,
1879 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1880 and GOMP_parallel_end after the call.
1881 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1882 (expand_omp_for_init_counts): Handle combined loops.
1883 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1885 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1886 end of cancellable loops.
1887 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1888 Likewise. Handle collapse > 1 loops.
1889 (expand_omp_simd): Handle combined loops.
1890 (expand_omp_for): Add inner_stmt argument, adjust callers of
1891 expand_omp_for* functions, use expand_omp_for_static*chunk even
1892 for collapse > 1 unless ordered.
1893 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1894 of cancellable sections.
1895 (expand_omp_single): Remove need_barrier variable, just rely on
1896 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1897 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1898 (expand_omp_atomic_load, expand_omp_atomic_store,
1899 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1900 (expand_omp_target): New function.
1901 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1902 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1903 (build_omp_regions_1): Immediately close region for
1904 GF_OMP_TARGET_KIND_UPDATE.
1905 (maybe_add_implicit_barrier_cancel): New function.
1906 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1908 (lower_omp_single): Likewise. Add clobber after the barrier.
1909 (lower_omp_taskgroup): New function.
1910 (lower_omp_for): Handle combined loops. Adjust
1911 lower_rec_input_clauses caller. Handle cancellation.
1912 (lower_depend_clauses): New function.
1913 (lower_omp_taskreg): Lower depend clauses. Adjust
1914 lower_rec_input_clauses caller. Add clobber after the call. Handle
1916 (lower_omp_target, lower_omp_teams): New functions.
1917 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1918 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1919 and GOMP_cancellation_point calls.
1920 (lower_omp): Fold stmts inside of target region.
1921 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1922 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1923 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1924 (BT_FN_VOID_OMPFN_PTR_UINT,
1925 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1926 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1927 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1928 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1929 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1930 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1931 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1932 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1933 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1934 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1935 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1936 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1937 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1938 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1939 BUILT_IN_GOMP_PARALLEL_END.
1940 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1941 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1942 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1943 GF_OMP_FOR_KIND_DISTRIBUTE.
1944 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1945 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1946 (dump_gimple_omp_return): Print lhs if it has any.
1947 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1948 gimple_omp_atomic_seq_cst_p.
1949 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1950 and GIMPLE_OMP_TEAMS.
1951 * langhooks.c (lhd_omp_mappable_type): New function.
1952 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1953 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1955 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1956 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1957 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
1958 (struct gimplify_omp_ctx): Add combined_loop field.
1959 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1960 on stmts inside of target region.
1961 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1962 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1963 ORT_TARGET and ORT_TARGET_DATA.
1964 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1966 (omp_notice_threadprivate_variable): Complain about threadprivate
1967 variables in target region.
1968 (omp_notice_variable): Complain about vars with non-mappable type
1969 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1970 (omp_check_private): Ignore ORT_TARGET* regions.
1971 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1972 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1973 (find_combined_omp_for): New function.
1974 (gimplify_omp_for): Handle gimplification of combined loops.
1975 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1977 (gimplify_omp_target_update): New function.
1978 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1979 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1980 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1981 (gimplify_body): If fndecl has "omp declare target" attribute, add
1982 implicit ORT_TARGET context around it.
1983 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1984 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1985 * tree-nested.c (convert_nonlocal_reference_stmt,
1986 convert_local_reference_stmt, convert_gimple_call): Handle
1987 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1988 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1989 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1990 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1991 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1992 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1993 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1994 BUILT_IN_GOMP_LOOP_END_CANCEL,
1995 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1996 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1997 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1998 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1999 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2000 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2001 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2002 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2003 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2004 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2005 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2006 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2007 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2008 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2009 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2010 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2011 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2012 gimple_build_omp_teams): New functions.
2013 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2014 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2015 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2016 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2017 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2018 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2019 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2020 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2021 GF_OMP_ATOMIC_SEQ_CST): New.
2022 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2023 gimple_build_omp_teams): New prototypes.
2024 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2025 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2026 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2027 GIMPLE_OMP_SINGLE as end of range.
2028 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2029 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2030 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2031 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2032 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2033 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2034 gimple_omp_target_kind, gimple_omp_target_set_kind,
2035 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2036 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2037 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2038 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2039 gimple_omp_teams_set_clauses): New inlines.
2040 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2041 and GIMPLE_OMP_TASKGROUP.
2042 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2043 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2044 enum omp_clause_proc_bind_kind): New.
2045 (union omp_clause_subcode): Add depend_kind, map_kind and
2046 proc_bind_kind fields.
2047 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2048 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2049 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2050 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2051 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2053 2013-10-10 Teresa Johnson <tejohnson@google.com>
2055 * predict.c (tree_estimate_probability): Add new parameter
2056 for estimate_bb_frequencies.
2057 (estimate_bb_frequencies): Add new parameter to force estimation.
2058 (rebuild_frequencies): When max frequency in function is small,
2059 recompute counts from frequencies.
2060 * predict.h (estimate_bb_frequencies): New parameter.
2062 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2064 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2065 optimizations are disabled.
2067 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2069 * coverage.c (coverage_finish): Fix leak of da_file_name.
2071 2013-10-10 Jan Hubicka <jh@suse.cz>
2073 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2074 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2077 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2080 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2081 if any labels are in FALLTHRU_BB, use a special label emitted
2082 immediately after the asm goto insn rather than label_rtx
2084 (expand_asm_stmt): Adjust caller.
2085 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2086 edge if the last insn in predecessor is a jump with single successor,
2087 but it isn't simplejump_p.
2089 2013-10-10 Richard Biener <rguenther@suse.de>
2091 PR tree-optimization/58656
2092 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2094 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2096 * gimplify.c: Include expr.h and tm_p.h for targets with special
2097 va-arg padding requirements.
2099 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2101 * tree-flow.h: Move some prototypes to gimple.h.
2102 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2103 * gimple.h: Relocate some prototypes from tree-flow.h
2104 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2106 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2107 (build_va_arg_indirect_ref): Relocate and make static.
2108 (std_gimplify_va_arg_expr): Relocate here.
2109 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2110 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2112 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2114 * doc/md.texi: Document the mnemonic attribute.
2116 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2119 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2120 (set_attr_alternative x ...) when searching for user defined
2123 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2125 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2127 2013-10-09 Easwaran Raman <eraman@google.com>
2129 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2130 * cfgexpand.c (defer_stack_allocation): ...use here
2131 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2133 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2135 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2136 (operand_equal_for_value_replacement): New function, extracted from
2137 value_replacement and enhanced to catch more cases.
2138 (value_replacement): Use operand_equal_for_value_replacement.
2140 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2142 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2143 get_max_loop_iterations.
2145 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2147 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2148 Place comment above function.
2150 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2152 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2153 are not related to tree-cfg.c.
2154 * tree-ssa-address.h: New file. Relocate prototypes.
2155 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2156 (addr_for_mem_ref): New. Combine call to get_address_description and
2157 return addr_for_mem_ref.
2158 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2159 * tree-ssa-live.h: Adjust prototypes.
2160 * passes.c: Include tree-ssa-live.h.
2161 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2162 * graphite.c (graphite_transform_loops): Make static.
2163 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2164 make_pass_graphite, pass_data_graphite_transforms,
2165 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2166 * ipa-pure-const.c (warn_function_noreturn): Make static.
2167 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2168 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2169 Relocate from tree-cfg.c
2170 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2172 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2173 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2174 Move to ipa-pure-const.c.
2175 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2176 Relocate from tree-optimize.c.
2177 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2178 make_pass_fixup_cfg): Move to tree-cfg.c.
2179 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2180 Relocate some prototypes.
2181 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2182 * tree-dump.c (dump_function_to_file): Remove prototype.
2183 Add tree-flow.h to the include file.
2184 * tree-dump.h: Remove prototype.
2185 * tree-parloops.h: New File. Add prototypes.
2186 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2187 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2188 from tree-ssa-loop.c.
2189 * tree-predcom.c (run_tree_predictive_commoning,
2190 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2191 Relocate here from tree-ssa-loop.c.
2192 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2193 ssa_name_values.release ().
2194 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2195 (ssa_name_values): Relocate from tree-flow.h.
2196 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2197 * tree-ssa-loop.c (run_tree_predictive_commoning,
2198 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2199 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2200 make_pass_graphite, pass_data_graphite_transforms,
2201 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2202 tree_parallelize_loops, pass_data_parallelize_loops,
2203 make_pass_parallelize_loops): Move to other files.
2204 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2206 * tree.h: Remove prototypes from tree-address.c.
2208 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2210 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2211 (struct int_tree_map): Move to tree-hasher.h
2212 (SCALE, LABEL, PERCENT): Move to gimple.h
2213 * tree-flow-inline.h: Delete. Move functions to other files.
2214 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2215 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2216 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2217 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2218 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2219 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2220 * tree-hasher.h: Don't include tree-flow.h.
2221 (struct int_tree_map): Relocate from tree-flow.h.
2222 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2223 tree-flow-inline.h and make static.
2224 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2226 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2227 tree-flow-inline.h and make static.
2228 * tree.h (is_global_var, may_be_aliased): Relocate from
2230 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2231 * value-prof.c: No longer include tree-flow-inline.h.
2232 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2234 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2236 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2237 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2238 (enum move_pos): Move to tree-ssa-loop-im.h
2239 * cfgloop.h: Move some prototypes.
2240 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2241 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2242 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2243 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2244 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2245 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2246 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2247 (enum move_pos): Relocate here.
2248 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2250 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2251 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2252 make_pass_lim): Relocate here from tree-ssa-loop.c.
2253 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2255 (loop_edge_to_cancel, unloop_loops): Make static.
2256 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2257 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2258 (tree_complete_unroll, gate_tree_complete_unroll,
2259 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2260 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2261 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2262 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2263 (stmt_invariant_in_loop_p): Remove unused function.
2264 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2265 * tree-ssa-loop-manip.h: New file. Add prototypes.
2266 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2267 (gcov_type_to_double_int): Move to cfgloop.h.
2268 (double_int_cmp, bound_index,
2269 estimate_numbers_of_iterations_loop): Make static.
2270 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2271 (max_loop_iterations): Factor out get_max_loop_iterations.
2272 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2274 * tree-ssa-loop-niter.h: New file. Add prototypes.
2275 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2276 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2277 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2278 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2279 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2280 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2281 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2282 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2283 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2284 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2285 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2286 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2287 pass_data_complete_unroll, make_pass_complete_unroll,
2288 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2289 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2290 tree-ssa-loop-ivcanon.c.
2291 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2292 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2293 tree-ssa-loop-prefetch.c.
2294 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2296 (get_lsm_tmp_name): Relocate and add suffix parameter.
2297 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2298 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2299 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2300 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2301 (get_estimated_loop_iterations): Factor out accessor from
2302 estimated_loop_iterations in tree-ssa-loop-niter.c.
2303 (get_max_loop_iterations): Factor out accessor from
2304 _max_loop_iterations in tree-ssa-niter.c.
2305 * loop-unroll.c (decide_unroll_constant_iterations,
2306 decide_unroll_runtime_iterations, decide_peel_simple,
2307 decide_unroll_stupid): Use new get_* accessors.
2309 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2311 PR tree-optimization/20318
2312 * doc/extend.texi (returns_nonnull): New function attribute.
2313 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2315 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2316 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2318 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2321 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2322 false if both components are bitfields.
2324 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2326 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2327 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2328 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2329 * config/aarch64/aarch64.h
2330 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2331 * config/aarch64/aarch64-simd-builtins.def
2332 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2334 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2336 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2338 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2340 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2342 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2344 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2345 (vneg_f64): New intrinsic.
2346 (vneg_s8): Asm replaced with C.
2347 (vneg_s16): Likewise.
2348 (vneg_s32): Likewise.
2349 (vneg_s64): New intrinsic.
2350 (vnegq_f32): Asm replaced with C.
2351 (vnegq_f64): Likewise.
2352 (vnegq_s8): Likewise.
2353 (vnegq_s16): Likewise.
2354 (vnegq_s32): Likewise.
2355 (vnegq_s64): Likewise.
2357 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2359 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2361 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2363 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2364 packed stack special handling.
2365 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2366 back to fixed stack slots for FPRs saved due to stdarg.
2368 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2370 * config/s390/s390.c (s390_frame_info): Restructure function.
2372 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2374 * config/s390/s390.c (struct s390_frame_layout): New field
2376 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2377 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2378 regs_ever_clobbered to char*.
2379 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2380 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2381 (s390_register_info): Enable FPR save slots. Move/Copy some
2382 functionality into ...
2383 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2384 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2386 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2387 not imply a stack frame.
2388 (s390_init_frame_layout): Remove variable clobbered_regs.
2389 (s390_update_register_info): Remove function.
2390 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2391 cannot be used for register renaming.
2392 (s390_hard_regno_scratch_ok): New function.
2393 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2394 (s390_initial_elimination_offset): Change offset calculation of
2395 the return address pointer.
2396 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2397 (restore_gprs): Set frame related flag.
2398 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2399 (s390_emit_prologue): Call s390_register_info instead of
2400 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2401 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2402 (s390_optimize_prologue): Call s390_optimize_register_info.
2403 Try to remove also FPR slot save/restore INSNs. Remove frame
2404 related flags from restore INSNs.
2406 2013-10-08 DJ Delorie <dj@redhat.com>
2408 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2411 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2412 avoid conflict with the MI use of %c.
2413 * config/rl78/rl78-real.md: change %c to %C throughout.
2414 * config/rl78/rl78-virt.md: Likewise.
2416 2013-10-08 Jan Hubicka <jh@suse.cz>
2418 * config/i386/i386.c (ix86_option_override_internal): Switch
2419 to SSE math for -ffast-math when target ISA supports SSE2.
2421 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2423 * tree-flow.h: Remove some prototypes.
2424 * tree.h: Remove some protypes, add a couple.
2425 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2426 using_eh_for_cleanups_p): Add interface routines for front ends.
2427 * tree-eh.h: New file. Add protoptyes.
2428 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2429 (add_stmt_to_eh_lp_fn): Make static.
2430 (lower_try_finally): Use new using_eh_for_cleanups_p.
2431 * emit-rtl.c: Include tree-eh.h.
2432 * gimple.h: Include tree-eh.h.
2434 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2436 PR tree-optimization/58480
2437 * tree-vrp.c (infer_nonnull_range): New function.
2438 (infer_value_range): Call infer_nonnull_range.
2440 2013-10-08 Dehao Chen <dehao@google.com>
2442 PR tree-optimization/58619
2443 * tree-inline.c (copy_phis_for_bb): Combine location data
2446 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2449 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2450 RTX_FRAME_RELATED_P on INSN.
2452 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2454 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2455 (altivec_expand_vec_perm_const): Call it.
2457 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2459 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2460 sequences for LE VSX loads and stores at expand time.
2461 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2463 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2464 (rs6000_gen_le_vsx_permute): New.
2465 (rs6000_gen_le_vsx_load): New.
2466 (rs6000_gen_le_vsx_store): New.
2467 (rs6000_gen_le_vsx_move): New.
2468 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2469 (*vsx_le_perm_load_v4si): New.
2470 (*vsx_le_perm_load_v8hi): New.
2471 (*vsx_le_perm_load_v16qi): New.
2472 (*vsx_le_perm_store_v2di): New.
2473 (*vsx_le_perm_store_v4si): New.
2474 (*vsx_le_perm_store_v8hi): New.
2475 (*vsx_le_perm_store_v16qi): New.
2476 (*vsx_xxpermdi2_le_<mode>): New.
2477 (*vsx_xxpermdi4_le_<mode>): New.
2478 (*vsx_xxpermdi8_le_V8HI): New.
2479 (*vsx_xxpermdi16_le_V16QI): New.
2480 (*vsx_lxvd2x2_le_<mode>): New.
2481 (*vsx_lxvd2x4_le_<mode>): New.
2482 (*vsx_lxvd2x8_le_V8HI): New.
2483 (*vsx_lxvd2x16_le_V16QI): New.
2484 (*vsx_stxvd2x2_le_<mode>): New.
2485 (*vsx_stxvd2x4_le_<mode>): New.
2486 (*vsx_stxvd2x8_le_V8HI): New.
2487 (*vsx_stxvd2x16_le_V16QI): New.
2489 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2491 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2493 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2495 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2496 loop to work also for 31bit ABI.
2497 Save the stack pointer for frame_size > 0.
2499 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2501 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2502 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2503 constraint letters from expanders.
2504 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2505 retry count to general_operand.
2506 ("tabort"): Give operand 0 a mode.
2507 ("tabort_1"): Add mode and constraint letter for operand 0.
2508 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2510 2013-10-04 Jeff Law <law@redhat.com>
2512 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2514 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2516 (thread_across_edge): Here. Call it.
2518 2013-10-04 Cary Coutant <ccoutant@google.com>
2520 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2521 discarding a location list expression (or a piece of one).
2523 2013-10-03 Jan Hubicka <jh@suse.cz>
2525 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2526 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2527 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2529 2013-10-03 Jan Hubicka <jh@suse.cz>
2531 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2532 accumulate-outgoing-args when producing unwind info.
2534 2013-10-03 Wei Mi <wmi@google.com>
2536 * lra-constraints.c (insert_move_for_subreg): New function
2537 extracted from simplify_operand_subreg.
2538 (simplify_operand_subreg): Add reload for paradoxical subreg.
2540 2013-10-03 Rong Xu <xur@google.com>
2542 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2543 the candidate of builtin_expect such that we should fix the
2544 size/time estimation.
2545 (estimate_function_body_sizes): Do the acutally size/time fix-up
2548 2013-10-03 Rong Xu <xur@google.com>
2550 * predict.c (tree_predict_by_opcode): Get the probability
2551 for builtin_expect from param builtin_expect_probability.
2552 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2553 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2554 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2556 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2559 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2560 * calls.c (alloca_call_p): Use get_callee_fndecl.
2561 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2562 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2564 (vrp_visit_stmt): Remove duplicated code.
2566 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2568 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2569 ceildf2, btruncdf2, instead of vsx_* name.
2571 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2572 iterators to only do V2DF and V4SF here. Move the DF code to
2573 rs6000.md where it is combined with SF mode. Replace <VSv> with
2574 just 'v' since only vector operations are handled with these insns
2575 after moving the DF support to rs6000.md.
2576 (vsx_sub<mode>3): Likewise.
2577 (vsx_mul<mode>3): Likewise.
2578 (vsx_div<mode>3): Likewise.
2579 (vsx_fre<mode>2): Likewise.
2580 (vsx_neg<mode>2): Likewise.
2581 (vsx_abs<mode>2): Likewise.
2582 (vsx_nabs<mode>2): Likewise.
2583 (vsx_smax<mode>3): Likewise.
2584 (vsx_smin<mode>3): Likewise.
2585 (vsx_sqrt<mode>2): Likewise.
2586 (vsx_rsqrte<mode>2): Likewise.
2587 (vsx_fms<mode>4): Likewise.
2588 (vsx_nfma<mode>4): Likewise.
2589 (vsx_copysign<mode>3): Likewise.
2590 (vsx_btrunc<mode>2): Likewise.
2591 (vsx_floor<mode>2): Likewise.
2592 (vsx_ceil<mode>2): Likewise.
2593 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2594 (vsx_sminsf3): Likewise.
2595 (vsx_fmadf4): Likewise.
2596 (vsx_fmsdf4): Likewise.
2597 (vsx_nfmadf4): Likewise.
2598 (vsx_nfmsdf4): Likewise.
2599 (vsx_cmpdf_internal1): Likewise.
2601 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2602 simpler to select whether a target has SPE or traditional floating
2603 point support in iterators.
2604 (TARGET_DF_SPE): Likewise.
2605 (TARGET_SF_FPR): Likewise.
2606 (TARGET_DF_FPR): Likewise.
2607 (TARGET_SF_INSN): Macros to say whether floating point support
2608 exists for a given operation for expanders.
2609 (TARGET_DF_INSN): Likewise.
2611 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2612 combining of SF/DF mode operations, using both traditional and VSX
2620 (abs<mode>2): Combine SF/DF modes using traditional floating point
2621 instructions. Add support for using the upper DF registers with
2622 VSX support, and SF registers with power8-vector support. Update
2623 expanders for operations supported by both the SPE and traditional
2624 floating point units.
2625 (abs<mode>2_fpr): Likewise.
2626 (nabs<mode>2): Likewise.
2627 (nabs<mode>2_fpr): Likewise.
2628 (neg<mode>2): Likewise.
2629 (neg<mode>2_fpr): Likewise.
2630 (add<mode>3): Likewise.
2631 (add<mode>3_fpr): Likewise.
2632 (sub<mode>3): Likewise.
2633 (sub<mode>3_fpr): Likewise.
2634 (mul<mode>3): Likewise.
2635 (mul<mode>3_fpr): Likewise.
2636 (div<mode>3): Likewise.
2637 (div<mode>3_fpr): Likewise.
2638 (sqrt<mode>3): Likewise.
2639 (sqrt<mode>3_fpr): Likewise.
2640 (fre<Fs>): Likewise.
2641 (rsqrt<mode>2): Likewise.
2642 (cmp<mode>_fpr): Likewise.
2643 (smax<mode>3): Likewise.
2644 (smin<mode>3): Likewise.
2645 (smax<mode>3_vsx): Likewise.
2646 (smin<mode>3_vsx): Likewise.
2647 (negsf2): Delete SF operations that are merged with DF.
2654 (fmasf4_fpr): Likewise.
2655 (fmssf4_fpr): Likewise.
2656 (nfmasf4_fpr): Likewise.
2657 (nfmssf4_fpr): Likewise.
2658 (sqrtsf2): Likewise.
2659 (rsqrtsf_internal1): Likewise.
2660 (smaxsf3): Likewise.
2661 (sminsf3): Likewise.
2662 (cmpsf_internal1): Likewise.
2663 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2664 (negdf2): Delete DF operations that are merged with SF.
2666 (nabsdf2): Likewise.
2672 (rsqrtdf_internal1): Likewise.
2673 (fmadf4_fpr): Likewise.
2674 (fmsdf4_fpr): Likewise.
2675 (nfmadf4_fpr): Likewise.
2676 (nfmsdf4_fpr): Likewise.
2677 (sqrtdf2): Likewise.
2678 (smaxdf3): Likewise.
2679 (smindf3): Likewise.
2680 (cmpdf_internal1): Likewise.
2681 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2682 (btrunc<mode>2): Delete separate expander, and combine with the
2683 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2684 (btrunc<mode>2_fpr): Likewise.
2685 (ceil<mode>2): Likewise.
2686 (ceil<mode>2_fpr): Likewise.
2687 (floor<mode>2): Likewise.
2688 (floor<mode>2_fpr): Likewise.
2689 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2690 Add support for using the upper registers with VSX and
2691 power8-vector. Move insns to be closer to the define_expands. On
2692 VSX systems, prefer the traditional form of FMA over the VSX
2693 version, since the traditional form allows the target not to
2694 overlap with the inputs.
2695 (fms<mode>4_fpr): Likewise.
2696 (nfma<mode>4_fpr): Likewise.
2697 (nfms<mode>4_fpr): Likewise.
2699 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2700 Richard Earnshaw <richard.earnshaw@arm.com>
2702 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2703 (struct mult_cost_table): Likewise.
2704 (struct mem_cost_table): Likewise.
2705 (struct fp_cost_table): Likewise.
2706 (struct vector_cost_table): Likewise.
2707 (cpu_cost_table): Likewise.
2708 * config/arm/arm.opt (mold-rts-costs): New option.
2709 (mnew-generic-costs): Likewise.
2710 * config/arm/arm.c (generic_extra_costs): New table.
2711 (cortexa15_extra_costs): Likewise.
2712 (arm_slowmul_tune): Use NULL as new costs.
2713 (arm_fastmul_tune): Likewise.
2714 (arm_strongarm_tune): Likewise.
2715 (arm_xscale_tune): Likewise.
2716 (arm_9e_tune): Likewise.
2717 (arm_v6t2_tune): Likewise.
2718 (arm_cortex_a5_tune): Likewise.
2719 (arm_cortex_a9_tune): Likewise.
2720 (arm_v6m_tune): Likewise.
2721 (arm_fa726te_tune): Likewise.
2722 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2723 (arm_cortex_tune): Use generict_extra_costs.
2724 (shifter_op_p): New function.
2725 (arm_unspec_cost): Likewise.
2726 (LIBCALL_COST): Define.
2727 (arm_new_rtx_costs): New function.
2728 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2729 table is available. Use old costs otherwise unless mnew-generic-costs
2731 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2732 (cpu_cost_table): Declare.
2734 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2737 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2738 (*subs_mul_imm_<mode>)
2739 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2740 (*sub_<shift>_<mode>)
2741 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2742 Remove k constraint.
2744 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2746 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2748 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2750 2013-10-02 Teresa Johnson <tejohnson@google.com>
2752 * predict.c (probably_never_executed): New function.
2753 (probably_never_executed_bb_p): Invoke probably_never_executed.
2754 (probably_never_executed_edge_p): Ditto.
2755 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2756 Treat profile insanities conservatively.
2758 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2760 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2762 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2764 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2765 use it. Use smaller increase for scratch. Don't increase reject
2766 for early clobber scratch.
2767 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2768 setting eliminated regs except setting fp from hfp.
2769 (lra_eliminate): Check lra_insn_recog_data on NULL.
2771 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2774 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2775 setting -mvsx-timode by default until the underlying problem is fixed.
2776 (RS6000_CPU, power7 defaults): Likewise.
2778 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2780 * config/x-linux (host-linux.o): Remove header dependencies.
2781 Use $(COMPILE) and $(POSTCOMPILE).
2782 * config/t-linux-android (linux-android.o): Ditto.
2784 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2786 * Makefile.in (expmed.o-warn): Remove.
2788 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2790 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2791 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2793 2013-10-02 Teresa Johnson <tejohnson@google.com>
2795 * dojump.c (do_jump_1): Divide probability between
2796 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2798 2013-10-02 Tom Tromey <tromey@redhat.com>
2800 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2802 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2804 * tree-flow.h: Remove some prototypes.
2805 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2806 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2807 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2808 mark_virtual_phi_result_for_renaming): Relocate here.
2809 * tree-into-ssa.h: Add prototypes.
2810 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2811 single_pred_before_succ_order.
2812 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2813 (nonfreeing_call_p) Move to gimple.c.
2814 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2816 * basic-block.h (single_pred_before_succ_order): Add prototype.
2817 * gimple.c (nonfreeing_call_p): Relocate here.
2818 * gimple.h: Add prototype.
2819 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2820 * tree-ssa-dom.h: New file. Relocate prototypes here.
2821 * tree-ssa.h: Include tree-ssa-dom.h.
2823 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2825 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2826 Use $(COMPILE) and $(POSTCOMPILE).
2828 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2830 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2832 * tree-flow.h: Remove some prototypes.
2833 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2834 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2835 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2836 propagate_tree_value*): Move from here to...
2837 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2838 propagate_tree_value*): Relocate here.
2839 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2840 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2841 * gimple-fold.c: Remove gimple-fold.h from include list.
2842 * tree-vrp.c: Remove gimple-fold.h from include list.
2843 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2844 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2845 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2846 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2847 * sese.c: Add tree-ssa-propagate.h to include list.
2849 2013-10-02 Richard Biener <rguenther@suse.de>
2851 * tree-loop-distribution.c: Include tree-vectorizer.h for
2853 (enum partition_kind): Remove PKIND_REDUCTION.
2854 (struct partition_s): Remove has_writes member, add reduction_p member.
2855 (partition_alloc): Adjust.
2856 (partition_builtin_p): Likewise.
2857 (partition_has_writes): Remove.
2858 (partition_reduction_p): New function.
2859 (partition_merge_into): Likewise.
2860 (generate_code_for_partition): Commonize builtin partition
2862 (rdg_cannot_recompute_vertex_p): Remove.
2863 (already_processed_vertex_p): Likewise.
2864 (rdg_flag_vertex): Do not set has_writes.
2865 (classify_partition): Adjust.
2866 (rdg_build_partitions): Do not set has_writes, treat all
2867 partitions as useful.
2868 (distribute_loop): Record number of library calls generated. Adjust.
2869 (tree_loop_distribution): Report number of loops and library
2870 calls generated as opt-info.
2872 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2874 * tree-flow.h: Include new .h files. Move prototypes.
2875 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2876 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2877 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2878 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2879 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2880 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2881 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2882 * tree.h (get_ref_base_and_extent): Move prototype out.
2883 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2885 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2886 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2887 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2888 * tree-scalar-evolution.c: Include tree.h.
2889 * sese.c: Include tree.h.
2890 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2891 * dwarf2out.c: Include tree-dfa.h.
2892 * tree-chrec.c: Include tree.h.
2893 * tree-data-ref.c: Include tree.h.
2895 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2897 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2900 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2902 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2903 Use $(COMPILE) and $(POSTCOMPILE).
2904 (sol2-cxx.o): Likewise.
2905 (sol2-stubs.o): Likewise.
2907 * config/x-solaris (host-solaris.o): Likewise.
2909 * config/sparc/t-sparc (sparc.o): Remove.
2910 (sparc-c.o): Remove header dependencies.
2911 Use $(COMPILE) and $(POSTCOMPILE).
2912 * config/sparc/x-sparc: Likewise.
2914 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2916 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2917 * config/arc/arc700.md: Likewise.
2918 * config/arc/arc-modes.def: Likewise.
2919 * config/arc/arc-simd.h: Likewise.
2920 * config/arc/t-arc-uClibc: Likewise.
2921 * config/arc/t-arc-newlib: Likewise.
2923 2013-10-02 Renlin Li <renlin.li@arm.com>
2925 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2927 (aarch64_expand_epilogue): Likewise.
2929 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2930 Yufeng Zhang <yufeng.zhang@arm.com>
2932 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2934 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2935 'base_in' represent a conversion and legal_cast_p_1 holds; set
2936 'base_in' with the returned value from get_unwidened.
2938 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2940 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2941 plus_constant is not used.
2943 2013-10-01 Wei Mi <wmi@google.com>
2945 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2946 * config/i386/i386.md: Add define_peephole2 to
2947 break partial reg stall for cvtss2sd/cvtsd2ss.
2949 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2951 * config/arc/arc.c (pass_arc_ifcvt::clone):
2952 Update for ctxt_ -> m_ctxt change.
2954 2013-10-01 Jeff Law <law@redhat.com>
2956 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2957 outgoing_edge and intermediate_edge fields. Instead store the path.
2958 (redirection_data::hash): Hash on the last edge's destination index.
2959 (redirection_data::equal): Check the entire thread path.
2960 (lookup_redirectio_data): Corresponding changes.
2961 (create_edge_and_update_destination_phis): Likewise.
2962 (thread_single_edge): Likewise.
2964 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2965 Diego Novillo <dnovillo@google.com>
2967 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2968 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2969 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2970 versions of these patterns.
2972 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2973 (__builtin_arc_aligned): Likewise.
2975 * config/arc/arc.md: Expand adc_0 comment stating the intended
2976 purpose and why it isn't ready.
2977 Replace commented out call_value_via_label_mixed with a
2978 plain comment about bl_s.
2980 * config/arc/arc.c (stdio.h): Don't include directly.
2981 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2982 Remove if (1) condition.
2983 (arc_encode_section_info): Fix comment.
2985 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2987 * config/arc/arc.c (arc_conditional_register_usage):
2988 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2989 Also set reg_alloc_order for DMA config regs.
2991 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2992 Jeremy Bennett <jeremy.bennett@embecosm.com>
2994 * doc/install.texi (--with-cpu): Mention ARC.
2995 (arc-*-elf32): New paragraph.
2996 (arc-linux-uclibc): Likewise.
2997 * doc/md.texi (Machine Constraints): Add ARC part.
2998 * doc/invoke.texi: (menu): Add ARC Options.
2999 (Machine Dependent Options) <ARC Options>: Add synopsis.
3000 (node ARC Options): Add.
3001 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3002 (ARC Built-in Functions): New section defining
3003 generic ARC built-in functions.
3004 (ARC SIMD Built-in Functions): New section defining SIMD specific
3006 (Declaring Attributes of Functions): Extended
3007 description of short_call and long_call attributes for ARC and
3008 added index entries.
3010 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3011 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3012 Joern Rennecke <joern.rennecke@embecosm.com>
3013 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3014 Brendan Kehoe <brendan@zen.org>
3015 Michael Eager <eager@eagercon.com>
3016 Simon Cook <simon.cook@embecosm.com>
3017 Jeremy Bennett <jeremy.bennett@embecosm.com>
3019 * config/arc, common/config/arc: New directories.
3021 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3022 Brendan Kehoe <brendan@zen.org>
3023 Simon Cook <simon.cook@embecosm.com>
3025 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3027 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3029 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3030 * tree-ssa-coalesce.h: New. Move prototype to here.
3031 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3032 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3033 (gimple_can_coalesce_p): Move to...
3034 * gimple.c (gimple_can_coalesce_p): Here.
3036 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3038 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3039 (dump_decl_set): Move to gimple.c.
3040 * gimple.h: Don't include tree-ssa-operands.h.
3041 (dump_decl_set): Add prototype.
3042 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3043 Move to gimple-ssa.h.
3044 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3045 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3046 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3047 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3048 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3050 (dump_decl_set): Relocate here.
3051 * gimple-ssa.h: New file.
3052 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3053 Relocate from gimple.h.
3054 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3055 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3056 swap_tree_operands and remove non-ssa path.
3057 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3058 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3060 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3061 vect_is_simple_reduction_1): Use swap_ssa_operands.
3062 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3063 (enum need_phi_state): Move to tree-into-ssa.c.
3064 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3065 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3066 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3067 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3068 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3069 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3070 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3071 num_imm_uses): Move to ssa-iterators.h.
3072 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3073 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3075 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3076 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3077 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3078 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3079 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3080 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3081 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3082 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3083 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3084 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3085 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3086 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3087 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3088 (set_phi_nodes): Move to tree-phinodes.h.
3089 * tree-ssa-operands.h (enum ssa_op_iter_type,
3090 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3091 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3092 (dump_decl_set): Remove prototype.
3093 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3094 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3095 (set_phi_nodes): Relocate from tree-flow-inline.h.
3096 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3098 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3099 include list. Temporarily add gimple.h to include list.
3100 * ssa-iterators.h: New file.
3101 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3102 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3103 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3104 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3105 Relocate from tree-ssa-operands.h.
3106 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3107 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3108 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3109 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3110 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3111 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3112 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3113 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3114 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3115 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3116 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3117 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3118 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3119 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3120 Relocate from tree-flow-inline.h.
3121 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3123 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3126 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3127 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3128 Insert '\t' to output template.
3129 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3130 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3131 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3132 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3134 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3136 * doc/install.texi (Host/target specific installation notes for GCC):
3137 Put @anchor before @heading.
3138 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3139 Use @email for email addresses.
3141 2013-10-01 Jeff Law <law@redhat.com>
3143 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3144 a vec. Only delete the path if we create one without successfully
3145 registering a jump thread.
3146 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3148 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3149 (paths): New vector of jump threading paths.
3150 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3151 (THREAD_PATH): New accessor macro for the entire thread path.
3152 (lookup_redirection_data): Get intermediate and final outgoing edge
3153 from the thread path.
3154 (create_edge_and_update_destination_phis): Copy the threading path.
3155 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3156 jump threading path.
3157 (ssa_redirect_edges): Get edges and block types from the jump threading
3158 path. Free the path vector.
3159 (thread_block): Get edges from the jump threading path. Look at the
3160 entire path to see if we thread to a loop exit. If we cancel a jump
3161 thread request, then free the path vector.
3162 (thread_single_edge): Get edges and block types from the jump threading
3163 path. Free the path vector.
3164 (thread_through_loop_header): Get edges and block types from the jump
3165 threading path. Free the path vector.
3166 (mark_threaded_blocks): Iterate over the vector of paths and store
3167 the path on the appropriate edge. Get edges and block types from the
3168 jump threading path.
3169 (mark_threaded_blocks): Get edges and block types from the jump
3170 threading path. Free the path vector.
3171 (thread_through_all_blocks): Use the vector of paths rather than
3172 a vector of 3-edge sets.
3173 (register_jump_thread): Accept pointer to a path vector rather
3174 than the path vector itself. Store the path vector for later use.
3177 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3178 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3181 * config/s390/s390.c (s390_split_branches): Modify check for table
3183 (s390_chunkify_start): Rearrange table jump insn check in order to
3184 deal with compare and branch insns correctly.
3186 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3190 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3191 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3192 define_insn_and_split.
3193 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3194 (shiftsi3_compare): New pattern.
3196 * config/arm/unspecs.md (UNSPEC_RRX): New.
3198 2013-10-01 Alan Modra <amodra@gmail.com>
3200 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3201 casing inout operands.
3203 2013-10-01 Richard Biener <rguenther@suse.de>
3205 PR tree-optimization/58553
3206 * tree-loop-distribution.c (struct partition_s): Add niter member.
3207 (classify_partition): Populate niter member for the partition
3208 and properly identify whether the relevant store happens before
3209 or after the loop exit.
3210 (generate_memset_builtin): Use niter member from the partition.
3211 (generate_memcpy_builtin): Likewise.
3213 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3215 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3216 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3218 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3220 * basic-block.h (edge_list): Prefix member names with "m_".
3221 * context.h (context): Likewise.
3222 * domwalk.h (dom_walker): Likewise.
3223 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3224 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3225 * hash-table.h (hash_table): Likewise.
3226 * machmode.h (bit_field_mode_iterator): Likewise.
3227 * pass_manager.h (pass_list): Likewise.
3228 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3229 * tree-pass.h (pass_data): Likewise.
3230 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3231 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3232 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3233 * asan.c (pass_data_asan): Update accordingly.
3234 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3235 (control_dependences::control_dependences): Likewise.
3236 (control_dependences::~control_dependences): Likewise.
3237 (control_dependences::~control_dependences): Likewise.
3238 (control_dependences::get_edges_dependent_on): Likewise.
3239 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3240 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3241 * context.c (gcc::context::context): Likewise.
3242 * cprop.c (pass_rtl_cprop::clone): Likewise.
3243 * domwalk.c (dom_walker::walk): Likewise.
3244 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3245 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3246 * mode-switching.c (pass_mode_switching::clone): Likewise.
3247 * passes.c (opt_pass::opt_pass): Likewise.
3248 (pass_manager::pass_manager): Likewise.
3249 * predict.c (pass_strip_predict_hints::clone): Likewise.
3250 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3251 (pass_split_all_insns::clone): Likewise.
3252 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3254 (bit_field_mode_iterator::next_mode): Likewise.
3255 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3256 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3257 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3258 * tree-complex.c (pass_lower_complex::clone): Likewise.
3259 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3260 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3261 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3262 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3263 (pass_fold_builtins::clone): Likewise.
3264 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3265 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3266 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3267 (pass_cd_dce::clone): Likewise.
3268 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3269 (pass_phi_only_cprop::clone): Likewise.
3270 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3271 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3272 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3273 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3274 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3275 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3276 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3277 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3278 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3279 * tree-vrp.c (pass_vrp::clone): Likewise.
3280 * tsan.c (pass_tsan::clone): Likewise.
3282 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3285 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3286 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3289 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3290 optimization, punt if sign_bit_p looked through any zero extension.
3292 2013-09-30 Teresa Johnson <tejohnson@google.com>
3294 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3295 Update redirected out edge count in joiner case.
3296 (ssa_redirect_edges): Common the joiner and non-joiner cases
3297 so that joiner case gets profile updates.
3299 2013-09-30 Richard Biener <rguenther@suse.de>
3301 PR tree-optimization/58554
3302 * tree-loop-distribution.c (classify_partition): Require
3303 unconditionally executed stores for memcpy and memset recognition.
3304 (tree_loop_distribution): Calculate dominance info.
3306 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3308 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3309 (NO_PROFILE_COUNTERS): Likewise.
3310 (PROFILE_HOOK): Likewise.
3311 (FUNCTION_PROFILER): Likewise.
3312 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3314 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3316 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3317 calls and defines in TARGET_MACHO conditional.
3318 (load_macho_picbase_di): Likewise.
3319 (reload_macho_picbase): Likewise.
3320 (reload_macho_picbase_si): Likewise.
3321 (reload_macho_picbase_di): Likewise.
3322 (nonlocal_goto_receiver): Likewise.
3324 2013-09-30 Nick Clifton <nickc@redhat.com>
3326 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3327 that use the MSP430X ISA.
3328 (msp430_option_override): Scan -mmcu command line option for any
3329 MCU name that supports the MSP430X ISA.
3330 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3331 -mmcu options which enable the MSP430X ISA.
3333 2013-09-30 Richard Biener <rguenther@suse.de>
3336 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3337 before looking for setjmp-like calls.
3339 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3342 * config/darwin-protos.h (machopic_get_function_picbase): New.
3343 * config/darwin.c (machopic_get_function_picbase): New.
3344 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3345 label for a new func. (load_macho_picbase_di): Likewise.
3346 (reload_macho_picbase): New expand.
3347 (reload_macho_picbase_si): New insn.
3348 (reload_macho_picbase_di): New insn.
3349 (nonlocal_goto_receiver): New define and split.
3350 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3351 (unspecv enum): Add UNSPECV_NLGR.
3353 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3355 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3356 that altivec registers are correctly sized on Darwin.
3358 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3360 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3361 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3362 * config/x-darwin (host-darwin.o): Likewise.
3363 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3364 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3365 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3367 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3369 * doc/invoke.texi: Fix usage of @tie{} command.
3371 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3373 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3375 2013-09-28 Jan Hubicka <jh@suse.cz>
3377 * config/i386/x86-tune.def: Add documentation for each of the options;
3380 2013-09-28 Jan Hubicka <jh@suse.cz>
3382 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3384 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3385 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3386 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3387 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3388 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3390 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3392 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3393 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3394 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3395 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3396 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3397 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3398 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3399 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3400 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3401 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3402 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3403 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3404 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3405 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3406 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3407 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3408 gimple.h, godump.c, graphite-clast-to-gimple.c,
3409 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3410 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3411 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3412 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3413 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3414 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3415 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3416 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3417 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3418 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3419 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3420 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3421 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3422 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3423 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3424 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3425 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3426 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3427 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3428 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3429 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3430 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3431 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3432 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3433 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3434 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3435 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3436 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3437 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3438 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3439 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3440 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3441 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3442 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3443 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3444 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3445 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3446 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3447 whitespace before "(".
3449 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3451 * expr.h (extract_bit_field): Remove packedp parameter.
3452 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3453 from forward declaration.
3454 (store_split_bit_field): Remove packedp arg from calls to
3455 extract_fixed_bit_field.
3456 (extract_bit_field_1): Remove packedp parameter and packedp
3457 argument from recursive calls and calls to extract_fixed_bit_field.
3458 (extract_bit_field): Remove packedp parameter and corresponding
3459 arg to extract_bit_field_1.
3460 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3462 (extract_split_bit_field): Remove packedp arg from call to
3463 extract_fixed_bit_field.
3464 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3465 (copy_blkmode_from_reg): Likewise.
3466 (copy_blkmode_to_reg): Likewise.
3467 (read_complex_part): Likewise.
3468 (store_field): Likewise.
3469 (expand_expr_real_1): Likewise.
3470 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3471 to extract_bit_field.
3472 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3473 call to extract_bit_field.
3474 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3475 call to extract_bit_field.
3476 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3477 and special packedp behavior from -fstrict-volatile-bitfields
3480 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3482 * lra-eliminations.c (init_elim_table): Guard value_p.
3484 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3486 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3487 DFmode, DImode, and SFmode in the upper VSX registers based on the
3488 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3489 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3490 of the VSX register handling.
3492 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3494 (zero_extendsidi2_lfiwzx): Add support for loading into the
3495 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3496 only do VSX memory options on Altivec registers.
3497 (extendsidi2_lfiwax): Likewise.
3498 (extendsfdf2_fpr): Likewise.
3499 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3500 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3501 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3502 (movdi_internal64): Likewise.
3504 2013-09-27 Xinliang David Li <davidxl@google.com>
3506 * opts.c (finish_options): Adjust parameters
3507 according to vect cost model.
3508 (common_handle_option): Set dynamic vect cost
3510 targhooks.c (default_add_stmt_cost): Compute stmt cost
3512 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3513 Use helper function.
3514 * tree-vectorizer.h (unlimited_cost_model): New function.
3515 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3516 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3518 (vect_enhance_data_refs_alignment): Ditto.
3519 * flag-types.h: New enum.
3520 * common/config/i386/i386-common.c (ix86_option_init_struct):
3521 No need to initialize vect_cost_model flag.
3522 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3525 2013-09-27 Diego Novillo <dnovillo@google.com>
3527 * gimple.h (enum ssa_mode): Remove.
3529 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3531 * cfgloop.h (number_of_loops): Fix typo in check for null.
3533 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3536 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3537 are children of outermost saved_cfun's loop, and set it up to
3538 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3539 if loop != loop0 and SESE region contains bbs that belong to loop0.
3541 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3543 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3544 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3545 (set_address_segment, set_address_base, set_address_index)
3546 (set_address_disp): Accept the argument unconditionally.
3547 (baseness): Remove must_be_base_p and must_be_index_p checks.
3548 (decompose_normal_address): Classify as much as possible in the
3551 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3553 * cse.c (count_reg_usage): Handle INT_LIST.
3554 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3555 * reginfo.c (reg_scan_mark_refs): Likewise.
3556 * reload1.c (eliminate_regs_1): Likewise.
3558 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3561 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3564 2013-09-27 Richard Biener <rguenther@suse.de>
3566 PR tree-optimization/58459
3567 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3568 restriction not propagating into loops.
3570 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3572 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3573 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3574 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3576 2013-09-26 Richard Biener <rguenther@suse.de>
3578 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3580 2013-09-26 Richard Biener <rguenther@suse.de>
3582 * alias.h (component_uses_parent_alias_set): Rename to ...
3583 (component_uses_parent_alias_set_from): ... this.
3584 * alias.c (component_uses_parent_alias_set): Rename to ...
3585 (component_uses_parent_alias_set_from): ... this and return
3587 (reference_alias_ptr_type_1): Use the result from
3588 component_uses_parent_alias_set_from instead of stripping
3589 components one at a time.
3590 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3592 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3594 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3595 Move prototypes to...
3596 * tree-ssa-ter.h: New File. Move prototypes here.
3597 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3598 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3599 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3600 from is_replaceable_p.
3601 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3602 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3603 (process_replaceable): Use ter_is_replaceable_p.
3604 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3605 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3606 newly refactored ssa_is_replaceable_p.
3607 * cfgexpand.c: Include tree-outof-ssa.h.
3608 * ssaexpand.h: Delete.
3610 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3612 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3613 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3615 (create_gimple_tmp): Delete.
3616 (get_expr_type, build_assign, build_type_cast): Move to...
3617 * gimple-builder.c: New File.
3618 (get_expr_type): Relocate from gimple.c.
3619 (build_assign, build_type_cast): Change to only create ssanames.
3620 * gimple.h: Move prototypes to...
3621 * gimple-builder.h: New File. Here.
3622 * tree-ssa.h: And here.
3623 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3624 count_uses_and_derefs): Relocate from gimple.c.
3625 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3626 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3627 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3628 gimple_replace_ssa_lhs.
3629 * asan.c: Include gimple-builder.h.
3630 * Makefile.in: Add gimple-builder.o.
3632 2013-09-26 Richard Biener <rguenther@suse.de>
3634 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3635 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3636 (loe_visit_block): Use gcc_checking_assert.
3637 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3638 gimple_assign_ssa_name_copy_p.
3639 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3642 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3644 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3645 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3647 2013-09-26 Richard Biener <rguenther@suse.de>
3649 PR tree-optimization/58539
3650 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3651 the fact that debug statements are not taking part in loop-closed
3654 2013-09-26 Nick Clifton <nickc@redhat.com>
3656 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3657 time warning message.
3658 (msp430_print_operand_raw): Delete unused letter parameter.
3659 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3660 (msp430_print_operand_address): New function.
3661 (msp430_print_operand): Move address printing code from here to
3663 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3665 (zero_extendpsisi2): Likewise.
3666 (extendpsisi2): New pattern.
3667 (andneghi3): New pattern.
3669 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3671 * config/aarch64/aarch64.opt (mlra): New option.
3672 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3673 (TARGET_LRA_P): Define.
3675 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3677 * expr.c (expand_assignment): Remove obsolete comment.
3679 2013-09-25 Jeff Law <law@redhat.com>
3681 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3682 tree-ssa-threadupdate.h.
3683 (register_jump_thread): Similarly.
3684 * tree-ssa-threadupdate.h: New header file.
3685 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3686 * tree-vrp.c: Likewise.
3687 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3688 (thread_around_empty_blocks): Change type of path vector argument to
3689 an edge,type pair from just an edge. Initialize both elements when
3690 appending to a jump threading path. Tweak references to elements
3692 (thread_across_edge): Similarly. Release memory for the elements
3694 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3695 (dump_jump_thread_path): New function broken out from
3696 register_jump_thread.
3697 (register_jump_thread): Use dump_jump_thread_path. Change type of
3698 path vector entries. Search the path for NULL edges and dump
3699 the path if one is found. Tweak the conversion of path to 3-edge
3700 form to use the block copy type information embedded in the path.
3702 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3704 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3706 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3707 Vladimir Makarov <vmakarov@redhat.com>
3709 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3710 from the least significant bit.
3711 (strip_address_mutations): Add bitfield operations handling.
3712 (must_be_index_p): Add shifting and rotate operations handling.
3713 (set_address_base): Use must_be_base_p predicate.
3714 (set_address_index): Use must_be_index_p predicate.
3716 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3717 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3718 Sergey Lega <sergey.s.lega@intel.com>
3719 Anna Tikhonova <anna.tikhonova@intel.com>
3720 Ilya Tocar <ilya.tocar@intel.com>
3721 Andrey Turetskiy <andrey.turetskiy@intel.com>
3722 Ilya Verbin <ilya.verbin@intel.com>
3723 Kirill Yukhin <kirill.yukhin@intel.com>
3724 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3726 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3728 (ix86_expand_vector_move_misalign): Support new unaligned load and
3729 stores and use new names.
3730 (CODE_FOR_sse2_storedqu): Rename to ...
3731 (CODE_FOR_sse2_storedquv16qi): ... this.
3732 (CODE_FOR_sse2_loaddqu): Rename to ...
3733 (CODE_FOR_sse2_loaddquv16qi): ... this.
3734 (CODE_FOR_avx_loaddqu256): Rename to ...
3735 (CODE_FOR_avx_loaddquv32qi): ... this.
3736 (CODE_FOR_avx_storedqu256): Rename to ...
3737 (CODE_FOR_avx_storedquv32qi): ... this.
3738 * config/i386/i386.md (fpint_logic): New.
3739 * config/i386/sse.md (VMOVE): Extend for AVX512.
3743 (VI_UNALIGNED_LOADSTORE): Ditto.
3744 (sse2_avx_avx512f): Ditto.
3745 (sse2_avx2): Extend for AVX512.
3746 (sse4_1_avx2): Ditto.
3747 (avx2_avx512f): New.
3748 (sse): Extend for AVX512.
3751 (avxsizesuffix): Ditto.
3752 (sseintvecmode): Ditto.
3754 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3755 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3756 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3757 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3758 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3759 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3760 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3761 (STORENT_MODE): Extend for AVX512.
3762 (*absneg<mode>2): Replace constraint "x" with "v".
3763 (*mul<mode>3): Ditto.
3764 (*ieee_smin<mode>3): Ditto.
3765 (*ieee_smax<mode>3): Ditto.
3766 (avx_cmp<mode>3): Replace VF with VF_128_256.
3767 (*<sse>_maskcmp<mode>3_comm): Ditto.
3768 (<sse>_maskcmp<mode>3): Ditto.
3769 (<sse>_andnot<mode>3): Extend for AVX512.
3770 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3771 (<code><mode>3, fpint_logic): New.
3772 (*<code><mode>3): Extend for AVX512.
3773 (avx512flogicsuff): New.
3774 (avx512f_<logic><mode>): Ditto.
3775 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3777 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3778 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3779 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3780 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3781 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3782 (xop_vpermil2<mode>3): Ditto.
3783 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3784 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3785 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3786 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3788 2013-09-25 Tom Tromey <tromey@redhat.com>
3790 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3791 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3792 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3793 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3794 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3795 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3796 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3797 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3798 (GRAPHITE_HTAB_H): Remove.
3800 2013-09-25 Tom Tromey <tromey@redhat.com>
3802 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3804 2013-09-25 Tom Tromey <tromey@redhat.com>
3806 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3808 2013-09-25 Tom Tromey <tromey@redhat.com>
3810 * config/i386/t-i386 (i386.o): Remove.
3811 (i386-c.o): Use COMPILE and POSTCOMPILE.
3813 2013-09-25 Tom Tromey <tromey@redhat.com>
3815 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3817 2013-09-25 Tom Tromey <tromey@redhat.com>
3819 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3820 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3821 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3822 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3823 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3824 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3825 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3826 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3827 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3828 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3829 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3830 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3831 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3832 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3833 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3834 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3835 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3836 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3837 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3838 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3839 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3840 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3841 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3842 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3843 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3844 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3845 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3846 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3847 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3848 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3849 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3850 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3851 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3852 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3853 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3854 (graphite-clast-to-gimple.o, graphite-dependences.o)
3855 (graphite-interchange.o, graphite-poly.o)
3856 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3857 (graphite-optimize-isl.o, tree-vect-loop.o)
3858 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3859 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3860 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3861 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3862 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3863 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3864 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3865 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3866 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3867 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3868 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3869 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3870 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3871 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3872 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3873 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3874 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3875 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3876 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3877 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3878 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3879 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3880 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3881 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3882 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3883 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3884 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3885 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3886 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3887 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3888 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3889 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3890 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3891 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3892 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3893 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3894 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3895 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3896 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3897 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3898 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3899 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3900 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3901 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3902 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3903 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3904 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3905 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3906 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3907 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3908 (gcov-dump.o): Remove.
3909 (default-c.o): Use COMPILE and POSTCOMPILE.
3910 (CFLAGS-gcc.o): New variable.
3911 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3913 2013-09-25 Tom Tromey <tromey@redhat.com>
3915 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3916 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3917 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3918 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3919 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3920 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3921 (c-family/array-notation-common.o, c-family/stub-objc.o)
3922 (c-family/c-ubsan.o): Remove.
3924 2013-09-25 Tom Tromey <tromey@redhat.com>
3926 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3928 2013-09-25 Tom Tromey <tromey@redhat.com>
3930 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3931 to add -DENABLE_SHARED_LIBGCC.
3932 (gcc.o): Don't use subshell.
3934 2013-09-25 Tom Tromey <tromey@redhat.com>
3936 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3937 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3938 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3939 * configure, config.in: Rebuild.
3941 2013-09-25 Tom Tromey <tromey@redhat.com>
3943 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3944 (COMPILE, POSTCOMPILE): New variables.
3945 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3946 (DEPFILES): New variable.
3947 Include ".Po" files.
3948 * configure.ac: Add checks for dependency checking.
3949 * configure, aclocal.m4: Regenerate.
3951 2013-09-25 Tom Tromey <tromey@redhat.com>
3953 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3954 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3956 2013-09-25 Tom Tromey <tromey@redhat.com>
3958 * Makefile.in (generated_files): Add options.h,
3959 target-hooks-def.h, insn-opinit.h,
3960 common/common-target-hooks-def.h, pass-instances.def,
3961 c-family/c-target-hooks-def.h.
3963 2013-09-25 Jeff Law <law@redhat.com>
3965 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3966 than foo[foo.length () - 1] to access last member in a vec.
3967 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3969 2013-09-25 Richard Biener <rguenther@suse.de>
3972 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3974 2013-09-25 Jan Hubicka <jh@suse.cz>
3976 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3979 2013-09-25 Marek Polacek <polacek@redhat.com>
3982 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3983 when determining the type name.
3985 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3987 * config/sh/sh.md: Fix formatting.
3989 2013-09-24 Xinliang David Li <davidxl@google.com>
3991 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3992 max peel iterations parameter.
3993 * param.def: New parameter.
3994 * doc/invoke.texi: Document New parameter.
3996 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3998 * gimple-pretty-print.c: Various whitespace tweaks.
3999 * tree-core.h: Likewise.
4000 * tree-pretty-print.c: Likewise.
4001 * tree-ssa-alias.c: Likewise.
4002 * tree-ssa-copy.c: Likewise.
4003 * tree-ssanames.c: Likewise.
4004 * tree-ssanames.h: Likewise.
4005 * tree-vrp.c: Likewise.
4007 2013-09-24 Alan Modra <amodra@gmail.com>
4011 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4012 for output operands that disallow regs. Don't use EXPAND_WRITE on
4015 2013-09-24 Richard Biener <rguenther@suse.de>
4018 * tree.c (reference_alias_ptr_type): Move ...
4019 * alias.c (reference_alias_ptr_type): ... here and implement
4020 in terms of the new reference_alias_ptr_type_1.
4021 (ref_all_alias_ptr_type_p): New helper.
4022 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4023 use ref_all_alias_ptr_type_p.
4024 (get_deref_alias_set): Add flag_strict_aliasing check here.
4025 (reference_alias_ptr_type_1): New function, split out from ...
4026 (get_alias_set): ... here.
4027 (alias_ptr_types_compatible_p): New function.
4028 * alias.h (reference_alias_ptr_type): Declare.
4029 (alias_ptr_types_compatible_p): Likewise.
4030 * tree.h (reference_alias_ptr_type): Remove.
4031 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4032 to compare MEM_REF alias types.
4034 2013-09-24 Richard Biener <rguenther@suse.de>
4036 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4038 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4040 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4041 reload helper function arrays into a single array reg_addr.
4042 (reload_fpr_gpr): Likewise.
4043 (reload_gpr_vsx): Likewise.
4044 (reload_vsx_gpr): Likewise.
4045 (struct rs6000_reg_addr): Likewise.
4046 (reg_addr): Likewise.
4047 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4048 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4049 (rs6000_init_hard_regno_mode_ok): Likewise.
4050 (rs6000_secondary_reload_direct_move): Likewise.
4051 (rs6000_secondary_reload): Likewise.
4053 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4054 constraints: wu, ww, and wy. Repurpose wv constraint added during
4055 power8 changes. Put wg constraint in alphabetical order.
4057 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4058 for future work to add ISA 2.07 VSX single precision support.
4059 (-mvsx-scalar-double): Change default from -1 to 1, update
4060 documentation comment.
4061 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4062 (-mupper-regs-df): New debug switch to control whether DF values
4063 can go in the traditional Altivec registers.
4064 (-mupper-regs-sf): New debug switch to control whether SF values
4065 can go in the traditional Altivec registers.
4067 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4069 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4070 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4071 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4072 (rs6000_debug_legitimate_address_p): Print if we are running
4073 before, during, or after reload.
4074 (rs6000_secondary_reload): Add a comment.
4075 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4077 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4078 constraints. Update documentation string.
4079 (wd constraint): Likewise.
4080 (wf constraint): Likewise.
4081 (wg constraint): Likewise.
4082 (wn constraint): Likewise.
4083 (ws constraint): Likewise.
4084 (wt constraint): Likewise.
4085 (wx constraint): Likewise.
4086 (wz constraint): Likewise.
4087 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4089 (ww constraint): Likewise.
4090 (wy constraint): Likewise.
4091 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4092 in the previous submissions.
4093 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4095 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4097 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4098 (REG_BR_PROB): Say that the probability is stored in an int_list.
4099 * reg-notes.def: Update commentary to mention INT_LIST.
4100 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4101 (INT_LIST): New rtx.
4102 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4103 * rtlanal.c (int_reg_note_p): New function.
4104 (alloc_reg_note): Assert that the note does not have an int argument.
4105 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4106 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4107 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4108 rather than an INSN_LIST. Handle INT_LIST.
4109 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4110 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4111 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4112 Manipulate them as ints rather than rtxes.
4113 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4114 * regmove.c (copy_src_to_dest): Likewise.
4115 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4117 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4118 into the cases that need it.
4119 * config/arm/arm.c (arm_unwind_emit): Likewise.
4121 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4122 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4123 * loop-doloop.c (add_test, doloop_modify): Likewise.
4124 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4125 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4126 * predict.c (combine_predictions_for_insn): Likewise.
4127 * print-rtl.c (print_rtx): Handle INT_LIST.
4128 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4129 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4130 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4131 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4132 (ix86_split_fp_branch, predict_jump): Likewise.
4133 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4134 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4135 * config/spu/spu.c (ea_load_store_inline): Likewise.
4137 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4138 value of a REG_BR_PROB note.
4139 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4140 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4141 * emit-rtl.c (try_split): Likewise.
4142 * predict.c (br_prob_note_reliable_p): Likewise.
4143 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4144 * reorg.c (mostly_true_jump): Likewise.
4145 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4146 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4147 * config/i386/i386.c (ix86_print_operand): Likewise.
4148 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4149 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4150 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4151 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4152 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4153 * config/sparc/sparc.c (output_cbranch): Likewise.
4154 * config/spu/spu.c (get_branch_target): Likewise.
4155 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4156 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4158 2013-09-23 Jan Hubicka <jh@suse.cz>
4160 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4162 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4163 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4165 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4166 (gimple_fold_call): Dump inconsistent devirtualizations; add
4167 sanity check for type based devirtualizations.
4168 * ipa-prop.c: Include ipa-utils.h
4169 (ipa_intraprocedural_devirtualization): Add sanity check.
4170 (try_make_edge_direct_virtual_call): Likewise.
4172 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4174 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4175 assignment statements.
4177 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4179 * gimple-pretty-print.c (dump_ssaname_info): New function.
4180 (dump_gimple_phi): Call it.
4181 (pp_gimple_stmt_1): Likewise.
4182 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4183 (range_info_def): Declare.
4184 * tree-pretty-print.c (pp_double_int): New function.
4185 (dump_generic_node): Call it.
4186 * tree-pretty-print.h (pp_double_int): Declare.
4187 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4188 * tree-ssanames.h (range_info_def): New structure.
4189 (value_range_type): Move definition here.
4190 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4192 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4194 (set_range_info): New function.
4195 (get_range_info): Likewise.
4196 (duplicate_ssa_name_range_info): Likewise.
4197 (duplicate_ssa_name_fn): Check pointer type and call
4198 duplicate_ssa_name_range_info.
4199 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4200 * tree-vrp.c (value_range_type): Remove definition, now in
4202 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4203 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4204 (SSA_NAME_RANGE_INFO): New macro.
4206 2013-09-23 Richard Biener <rguenther@suse.de>
4208 PR tree-optimization/58464
4209 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4210 (phi_trans_add): Change to add conditionally on being not present.
4211 (phi_translate_1): Remove recursion detection here.
4212 (phi_translate): Pre-seed the cache with NULL to catch
4213 recursion here in a more generic way.
4214 (bitmap_find_leader): Adjust comment.
4215 (get_representative_for): Dump value-numbers.
4216 (create_expression_by_pieces): Likewise.
4217 (insert_into_preds_of_block): Likewise.
4219 2013-09-23 Christian Bruel <christian.bruel@st.com>
4222 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4223 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4225 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4228 2013-09-20 Renlin Li <renlin.li@arm.com>
4230 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4232 (aarch64_expand_epilogue): Likewise.
4233 (aarch64_legitimize_reload_address): Likewise.
4235 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4237 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4238 NULL_TREE before pushing them onto the vector. Likewise for labels.
4240 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4242 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4243 comparison operators when -fno-trapping-math is in effect.
4244 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4245 comparison operators in TFmode and assert that unsupported operators
4247 (ia64_print_operand): Likewise.
4249 2013-09-21 Jan Hubicka <jh@suse.cz>
4251 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4252 (sse_typeless_stores): Enable for core
4253 (sse_load0_by_pxor): Likewise.
4254 (four_jump_limit): Disable for core.
4255 (pad_returns): Likewise.
4256 (avoid_vector_decode): Likewise.
4257 (fuse_cmp_and_branch): Enable for cores.
4258 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4260 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4263 * config/pa/pa.c (pa_option_override): Disable auto increment and
4264 decrement instructions until reload is completed.
4266 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4267 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4268 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4270 2013-09-20 DJ Delorie <dj@redhat.com>
4271 Nick Clifton <nickc@redhat.com>
4273 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4274 (need_to_save): Save bank 0 on interrupts.
4275 (characterize_address): Strip far address wrappers.
4276 (rl78_as_legitimate_address): Likewise.
4277 (transcode_memory_rtx): Likewise.
4278 (rl78_peep_movhi_p): Disable this peephole after devirt.
4279 (rl78_propogate_register_origins): Forget all origins when a
4281 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4282 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4284 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4285 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4286 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4288 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4290 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4292 * config/pa/pa.md: In "scc" insn patterns, change output template to
4293 handle const0_rtx in reg_or_0_operand operands.
4295 2013-09-20 Martin Husemann <martin@NetBSD.org>
4298 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4299 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4301 2013-09-20 Richard Biener <rguenther@suse.de>
4304 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4305 remembering SSA name version and block index.
4306 (new_scev_info_str): Adjust.
4307 (hash_scev_info): Likewise. Also hash the block index.
4308 (eq_scev_info): Adjust.
4309 (find_var_scev_info): Likewise.
4310 (struct instantiate_cache_entry): Remove.
4311 (struct instantiate_cache_type): Use a htab to map name, block
4313 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4314 (get_instantiated_value_entry): Likewise.
4315 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4316 (instantiate_scev_name): Adjust.
4318 2013-09-20 Jeff Law <law@redhat.com>
4320 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4322 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4324 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4325 Call aarch64_simd_expand_args to update op[argc].
4327 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4329 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4332 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4334 * gengtype.c (file_rules): Added rule for *.cc files.
4335 (get_output_file_with_visibility): Give fatal message when no
4338 2013-09-20 Renlin Li <renlin.li@arm.com>
4340 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4341 (aarch64_expand_epilogue): Likewise.
4342 (aarch64_legitimize_reload_address): Likewise.
4344 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4347 * expr.c (expand_assignment): Remove misalignp code path.
4349 2013-09-20 Marek Polacek <polacek@redhat.com>
4352 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4353 TYPE_PRECISION. Add asserts.
4355 2013-09-20 Richard Biener <rguenther@suse.de>
4357 PR tree-optimization/58453
4358 * tree-loop-distribution.c (distribute_loop): Apply the cost
4359 model for -ftree-loop-distribute-patterns, too.
4361 2013-09-20 Richard Biener <rguenther@suse.de>
4364 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4365 make type comparison less strict.
4367 2013-09-20 Alan Modra <amodra@gmail.com>
4369 * configure: Regenerate.
4370 * aclocal.m4: Regenerate.
4372 2013-09-20 Marek Polacek <polacek@redhat.com>
4375 * doc/extend.texi: Document that attribute used is meant to be used
4376 on variables with static storage duration.
4378 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4380 PR tree-optimization/58472
4381 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4382 simd_lane_access set inv_p = false.
4383 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4384 the simduid magic VAR_DECL.
4386 2013-09-19 Jan Hubicka <jh@suse.cz>
4388 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4390 2013-09-17 Jeff Law <law@redhat.com>
4392 * tree-ssa-dom.c (record_temporary_equivalences): New function
4393 split out of dom_opt_dom_walker::after_dom_children.
4394 (dom_opt_dom_walker::thread_across_edge): Move common code
4395 in here from dom_opt_dom_walker::after_dom_children.
4396 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4398 2013-09-19 Jan Hubicka <jh@suse.cz>
4400 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4401 (TARGET_GENERIC): Use PROCESOR_GENERIC
4402 (enum processor_type): Unify generic32 and 64.
4403 * i386.md (cpu): Likewise.
4404 * x86-tune.def (use_leave): Enable for generic32.
4405 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4406 * athlon.md: Change generic64 to generic in all occurences.
4407 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4408 (ix86_target_macros_internal): Likewise.
4409 * driver-i386.c (host_detect_local_cpu): Likewise.
4410 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4412 (generic_memcpy, generic_memset, generic_cost): This one.
4413 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4414 (m_GENERIC32, m_GENERIC64): Remove.
4415 (m_GENERIC): Turn into one flag.
4416 (processor_target): Unify generic tunnings.
4417 (ix86_option_override_internal): Replace generic32/64 by generic.
4418 (ix86_issue_rate): Likewise.
4419 (ix86_adjust_cost): Likewise.
4421 2013-09-19 Jan Hubicka <jh@suse.cz>
4423 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4424 of speculative flag.
4426 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4428 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4429 GOMP_sections_start, even if !exit_reachable.
4431 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4433 * lra-constraints.c (need_for_all_save_p): Use macro
4434 HARD_REGNO_CALL_PART_CLOBBERED.
4435 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4436 set up pseudo conflict hard regs.
4438 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4441 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4444 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4446 PR rtl-optimization/58438
4447 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4448 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4449 from previous subpasses.
4451 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4453 * arm.c (arm_get_frame_offsets): Validate architecture supports
4454 LDRD/STRD before accepting the tuning preference.
4455 (arm_expand_prologue): Likewise.
4456 (arm_expand_epilogue): Likewise.
4458 2013-09-18 Richard Biener <rguenther@suse.de>
4460 PR tree-optimization/58417
4461 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4462 have chrecs with symbols defined in the loop as operands.
4463 (chrec_fold_multiply): Likewise.
4464 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4465 parameters before folding binary operations.
4466 (struct instantiate_cache_entry_hasher): Remove.
4467 (struct instantiate_cache_type): Use a pointer-map.
4468 (instantiate_cache_type::instantiate_cache_type): New function.
4469 (instantiate_cache_type::get): Likewise.
4470 (instantiate_cache_type::set): Likewise.
4471 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4472 (get_instantiated_value_entry): Likewise.
4473 (global_cache): New global.
4474 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4475 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4476 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4477 (instantiate_scev_name): Adjust.
4478 (instantiate_scev): Construct global instead of local cache.
4479 (resolve_mixers): Likewise.
4481 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4482 Paolo Carlini <paolo.carlini@oracle.com>
4485 * doc/implement-cxx.texi: Fix references to the C++ standards.
4487 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4489 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4490 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4491 simd_lane_access drs, update also DR_ALIGNED_TO.
4493 2013-09-18 Marek Polacek <polacek@redhat.com>
4496 * doc/extend.texi: Document no_sanitize_undefined attribute.
4497 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4498 no_sanitize_undefined attribute.
4500 2013-09-18 Nick Clifton <nickc@redhat.com>
4502 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4503 (ASM_DECLARE_FUNCTION_NAME): Define.
4505 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4507 * compare-elim.c (find_comparison_dom_walker): New class
4508 (find_comparisons_in_bb): Rename to
4509 find_comparison_dom_walker::before_dom_children
4510 (find_comparisons): Adjust
4511 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4513 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4514 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4515 (init_walk_dominator_tree): Remove declaration.
4516 (fini_walk_dominator_tree): Remove declaration.
4517 * fwprop.c (single_def_use_dom_walker): New class
4518 (single_def_use_enter_block): Convert to
4519 single_def_use_dom_walker::before_dom_children.
4520 (single_def_use_leave_block): Convert to
4521 single_def_use_dom_walker::after_dom_children.
4522 (build_single_def_use_links): Adjust.
4523 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4525 (find_candidates_in_block): Convert to
4526 find_candidates_dom_walker::before_dom_children.
4527 (execute_strength_reduction): Adjust.
4528 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4529 (sese_dom_walker): New class.
4530 (sese_dom_walker::sese_dom_walker): New constructor.
4531 (sese_dom_walker::~sese_dom_walker): New destructor.
4532 (build_sese_conditions_before): Convert to
4533 sese_dom_walker::before_dom_children.
4534 (build_sese_conditions_after): Convert to
4535 sese_dom_walker::after_dom_children.
4536 (build_poly_scop): Adjust
4537 * tree-into-ssa.c (rewrite_dom_walker): New class
4538 (rewrite_enter_block): Convert to
4539 rewrite_dom_walker::before_dom_children.
4540 (rewrite_leave_block): Convert to
4541 rewrite_dom_walker::after_dom_children.
4542 (rewrite_update_dom_walker): New class.
4543 (rewrite_update_enter_block): Convert to
4544 rewrite_update_dom_walker::before_dom_children.
4545 (rewrite_update_leave_block): Convert to
4546 rewrite_update_dom_walker::after_dom_children.
4547 (rewrite_blocks, rewrite_into_ssa): Adjust.
4548 (mark_def_dom_walker): New class.
4549 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4550 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4551 (mark_def_sites_blocks): Convert to
4552 mark_def_dom_walker::before_dom_children.
4553 (mark_def_site_blocks): Remove.
4554 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4555 (tree_ssa_dominator_optimize): Adjust.
4556 (dom_thread_across_edge): Convert to method
4557 dom_opt_dom_walker::thread_across_edge.
4558 (dom_opt_enter_block): Convert to member function
4559 dom_opt_dom_walker::before_dom_children.
4560 (dom_opt_leave_block): Convert to member function
4561 dom_opt_dom_walker::after_dom_children.
4562 * tree-ssa-dse.c (dse_dom_walker): New class.
4563 (dse_enter_block): Convert to member function
4564 dse_dom_walker::before_dom_children.
4565 (tree_ssa_dse): Adjust.
4566 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4567 (determine_invariantness_stmt): Convert to method
4568 invariantness_dom_walker::before_dom_children.
4569 (determine_invariantness): Remove
4570 (move_computations_dom_walker): New class.
4571 (move_computations_stmt): Convert to method
4572 move_computations_dom_walker::before_dom_children.
4573 (move_computations, tree_ssa_lim): Adjust.
4574 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4575 (nt_init_block): Make method
4576 notrappping_dom_walker::before_dom_children.
4577 (nt_fini_block): Make
4578 method nontrapping_dom_walker::after_dom_children.
4579 (get_non_trapping): Adjust.
4580 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4581 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4582 (eliminate_leave_block): Make method.
4583 eliminate_dom_walker::after_dom_children.
4585 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4586 (strlen_enter_block): Make method
4587 strlen_dom_walker::before_dom_children.
4588 (strlen_leave_block): Make
4589 method strlen_dom_walker::after_dom_children.
4590 (tree_ssa_strlen): Adjust.
4591 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4592 (tree_ssa_uncprop): Adjust.
4593 (uncprop_leave_block): Make method
4594 uncprop_dom_walker::after_dom_children.
4595 (uncprop_leave_block): Make method
4596 uncprop_dom_walker::before_dom_children.
4598 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4600 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4601 branch in same basic block. Check both src and dest of the move insn.
4603 2013-09-17 Nick Clifton <nickc@redhat.com>
4605 * config/rl78/rl78-real.md (bf): New pattern.
4607 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4608 (rl78_print_operand): Do not put a # before a %B.
4609 * config/rl78/rl78.opt: Tweak doc strings.
4611 2013-09-17 DJ Delorie <dj@redhat.com>
4613 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4614 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4615 (rl78_option_override): Likewise, if -mallregs.
4616 (is_virtual_register): Likewise.
4617 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4618 (REGNO_OK_FOR_BASE_P): Likewise.
4619 * config/rl78/rl78.opt (-mallregs): New.
4621 2013-09-17 Nick Clifton <nickc@redhat.com>
4623 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4624 For interrupt functions: save all call clobbered registers if the
4625 interrupt handler is not a leaf function.
4626 (rl78_expand_prologue): Always recompute the frame information.
4627 For interrupt functions: only select bank 0 if one of the bank 0
4628 registers is going to be psuhed.
4630 2013-09-17 DJ Delorie <dj@redhat.com>
4632 * config/rl78/constraints.md: For each W* constraint, rename to C*
4633 and create a W* constraint that checks for an optional ES: prefix
4635 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4636 (es_addr): New. Used to wrap far addresses.
4637 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4638 (rl78_es_base): New.
4639 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4640 wrapped far addresses.
4641 (rl78_print_operand_1): Unwrap far addresses before processing.
4642 (rl78_lo16): Wrap far addresses in unspecs.
4643 (rl78_es_addr): New.
4644 (rl78_es_base): New.
4645 (insn_ok_now): Check for not-yet-wrapped far addresses.
4646 (transcode_memory_rtx): Properly re-wrap far addresses.
4648 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4650 * config/sparc/t-rtems: Add leon3 multilibs.
4652 2013-09-17 Cong Hou <congh@google.com>
4654 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4655 when checking the dot production pattern. The type of rhs operand
4656 of multiply is now checked correctly.
4658 2013-09-17 Jeff Law <law@redhat.com>
4660 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4661 edge implied equivalences into successor phis.
4662 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4663 here from tree-ssa-threadedge.c.
4664 (mark_threaded_blocks): When threading through a joiner, if both
4665 successors of the joiner's clone reach the same block, verify the
4666 PHI arguments are equal. If not, cancel the jump threading request.
4667 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4668 tree-ssa-threadupdate.c
4669 (thread_across_edge): Don't check PHI argument equality when
4670 threading through joiner block here.
4672 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4674 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4675 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4676 (warn_uninit, warn_uninitialized_vars,
4677 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4678 Move to tree-ssa-uninit.c.
4679 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4680 (has_undefined_value_p): New. Pass dependant parts of
4681 ssa_undefined_value_p.
4682 (uninit_undefined_value_p): Use has_undefined_value_p.
4683 (warn_uninit, warn_uninitialized_vars,
4684 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4685 Move from tree-ssa.c.
4686 * tree-ssa.h: Adjust prototypes.
4688 2013-09-17 Jan Hubicka <jh@suse.cz>
4691 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4692 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4693 FUNCTION_NOT_OPTIMIZED.
4694 * ipa-inline-analysis.c (compute_inline_parameters): Function
4695 not optimized is not inlinable unless it is alwaysinline.
4696 (inline_analyze_function): Force calls in not optimized
4697 function not inlinable.
4699 2013-09-17 Jan Hubicka <jh@suse.cz>
4702 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4704 * ipa.c (function_and_variable_visibility): Likewise.
4705 * ipa-profile.c (ipa_profile): Likewise.
4707 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4710 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4711 attribute, and don't inline the resolver in this case.
4713 2013-09-17 Teresa Johnson <tejohnson@google.com>
4715 * coverage.c (get_coverage_counts): Add missing newline.
4717 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4719 PR tree-optimization/58088
4720 * fold-const.c (mask_with_trailing_zeros): New function.
4721 (fold_binary_loc): Make sure we don't recurse infinitely
4722 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4723 Use mask_with_trailing_zeros where appropriate.
4725 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4727 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4728 of 'prev' var to get better distance estimation.
4730 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4732 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4733 * tree-inline.c (is_parameter_of): New predicate.
4734 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4735 a parameter has been remapped.
4736 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4737 (optimize_inline_calls): Initialize transform_parameter.
4738 (copy_gimple_seq_and_replace_locals): Likewise.
4739 (tree_function_versioning): Likewise.
4740 (maybe_inline_call_in_expr): Likewise.
4742 2013-09-17 Nick Clifton <nickc@redhat.com>
4744 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4745 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4747 (is_attr_func): New function.
4748 (msp430_is_interrupt_func): New function.
4749 (is_naked_func): New function.
4750 (is_reentrant_func): New function.
4751 (is_critical_func): New function.
4752 (msp430_start_function): Add annotations for function attributes.
4753 (msp430_attr): New function.
4754 (msp430_attribute_table): New.
4755 (msp430_function_section): New function.
4756 (TARGET_ASM_FUNCTION_SECTION): Define.
4757 (msp430_builtin): New enum.
4758 (msp430_init_builtins): New function.
4759 (msp430_builtin_devl): New function.
4760 (msp430_expand_builtin): New function.
4761 (TARGET_INIT_BUILTINS): Define.
4762 (TARGET_EXPAND_BUILTINS): Define.
4763 (TARGET_BUILTIN_DECL): Define.
4764 (msp430_expand_prologue): Add support for naked, interrupt,
4765 critical and reentrant functions.
4766 (msp430_expand_epilogue): Likewise.
4767 (msp430_print_operand): Handle 'O' character.
4768 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4770 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4771 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4772 (pushm): Use a 'n' rather than an 'i' constraint.
4773 (msp_return): Add generation of the interrupt return instruction.
4774 (disable_interrupts): New pattern.
4775 (enable_interrupts): New pattern.
4776 (push_intr_state): New pattern.
4777 (pop_intr_state): New pattern.
4778 (bic_SR): New pattern.
4779 (bis_SR): New pattern.
4780 * doc/extend.texi: Document MSP430 function attributes and builtin
4783 2013-09-17 Richard Biener <rguenther@suse.de>
4785 PR tree-optimization/58432
4786 * tree-loop-distribution.c (tree_loop_distribution): Also
4787 scan PHIs for outside loop uses and seed a partition from them.
4789 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4791 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4792 (restructure_reference): Call backtrace_base_for_ref.
4794 2013-09-17 Alan Modra <amodra@gmail.com>
4797 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4800 2013-09-16 DJ Delorie <dj@redhat.com>
4802 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4803 vregs location for RL78/G10.
4804 (rl78_expand_prologue): Avoid SEL on G10.
4805 (rl78_expand_epilogue): Likewise.
4806 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4807 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4808 __RL78_G10__ when appropriate.
4809 (ASM_SPEC): Pass -mg10 along to the assembler.
4810 * config/rl78/rl78.md (sel_rb): Disable for G10.
4811 * config/rl78/rl78.opt: Add -mg10 option.
4812 * config/rl78/t-rl78: Add -mg10 multilib.
4814 2013-09-16 Xinliang David Li <davidxl@google.com>
4816 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4817 * omp-low.c (omp_max_vf): Ditto.
4818 (expand_omp_simd): Ditto.
4819 * tree-vectorizer.c (vectorize_loops): Ditto.
4820 (gate_vect_slp): Ditto.
4821 (gate_increase_alignment): Ditto.
4822 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4823 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4824 (gate_tree_vectorize): Name change.
4825 (tree_vectorize): Ditto.
4826 (pass_vectorize::gate): Call new function.
4827 (pass_vectorize::execute): Ditto.
4828 * opts.c: O3 default setting change.
4829 (finish_options): Check new flag.
4830 * doc/invoke.texi: Document new flags.
4831 * common.opt: New flags.
4833 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4835 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4836 (Condition Code): Don't reference it.
4837 * doc/tm.texi: Regenerate.
4839 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4842 * lra-constraints.c (undo_optional_reloads): Consider all optional
4843 reload even if it did not get a hard reg.
4845 2013-09-16 Teresa Johnson <tejohnson@google.com>
4847 * dumpfile.c (dump_loc): Remove newline emission.
4848 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4849 emission to dump_printf_loc calls where missing.
4850 (vect_mark_for_runtime_alias_test): Ditto.
4851 (vect_analyze_data_ref_dependence): Ditto.
4852 (vect_analyze_data_ref_dependences): Ditto.
4853 (vect_slp_analyze_data_ref_dependence): Ditto.
4854 (vect_slp_analyze_data_ref_dependences): Ditto.
4855 (vect_compute_data_ref_alignment): Ditto.
4856 (vect_update_misalignment_for_peel): Ditto.
4857 (vect_verify_datarefs_alignment): Ditto.
4858 (vector_alignment_reachable_p): Ditto.
4859 (vect_get_data_access_cost): Ditto.
4860 (vect_enhance_data_refs_alignment): Ditto.
4861 (vect_find_same_alignment_drs): Ditto.
4862 (vect_analyze_data_refs_alignment): Ditto.
4863 (vect_analyze_group_access): Ditto.
4864 (vect_analyze_data_ref_access): Ditto.
4865 (vect_analyze_data_ref_accesses): Ditto.
4866 (vect_prune_runtime_alias_test_list): Ditto.
4867 (vect_analyze_data_refs): Ditto.
4868 (vect_create_addr_base_for_vector_ref): Ditto.
4869 (vect_create_data_ref_ptr): Ditto.
4870 (vect_grouped_store_supported): Ditto.
4871 (vect_grouped_load_supported): Ditto.
4872 * value-prof.c (check_counter): Ditto.
4873 (check_ic_target): Ditto.
4874 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4875 (vect_recog_widen_mult_pattern): Ditto.
4876 (vect_recog_widen_sum_pattern): Ditto.
4877 (vect_recog_over_widening_pattern): Ditto.
4878 (vect_recog_widen_shift_pattern): Ditto.
4879 (vect_recog_rotate_pattern): Ditto.
4880 (vect_recog_vector_vector_shift_pattern): Ditto.
4881 (vect_recog_divmod_pattern): Ditto.
4882 (vect_recog_mixed_size_cond_pattern): Ditto.
4883 (vect_recog_bool_pattern): Ditto.
4884 (vect_pattern_recog_1): Ditto.
4885 (vect_pattern_recog): Ditto.
4886 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4887 (vect_is_simple_iv_evolution): Ditto.
4888 (vect_analyze_scalar_cycles_1): Ditto.
4889 (vect_get_loop_niters): Ditto.
4890 (vect_analyze_loop_1): Ditto.
4891 (vect_analyze_loop_form): Ditto.
4892 (vect_analyze_loop_operations): Ditto.
4893 (vect_analyze_loop_2): Ditto.
4894 (vect_analyze_loop): Ditto.
4895 (report_vect_op): Ditto.
4896 (vect_is_slp_reduction): Ditto.
4897 (vect_is_simple_reduction_1): Ditto.
4898 (vect_get_known_peeling_cost): Ditto.
4899 (vect_estimate_min_profitable_iters): Ditto.
4900 (vect_model_reduction_cost): Ditto.
4901 (vect_model_induction_cost): Ditto.
4902 (get_initial_def_for_induction): Ditto.
4903 (vect_create_epilog_for_reduction): Ditto.
4904 (vectorizable_reduction): Ditto.
4905 (vectorizable_induction): Ditto.
4906 (vectorizable_live_operation): Ditto.
4907 (vect_loop_kill_debug_uses): Ditto.
4908 (vect_transform_loop): Ditto.
4909 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4910 (vect_stmt_relevant_p): Ditto.
4911 (process_use): Ditto.
4912 (vect_mark_stmts_to_be_vectorized): Ditto.
4913 (vect_model_simple_cost): Ditto.
4914 (vect_model_promotion_demotion_cost): Ditto.
4915 (vect_model_store_cost): Ditto.
4916 (vect_get_store_cost): Ditto.
4917 (vect_model_load_cost): Ditto.
4918 (vect_get_load_cost): Ditto.
4919 (vect_init_vector_1): Ditto.
4920 (vect_get_vec_def_for_operand): Ditto.
4921 (vect_finish_stmt_generation): Ditto.
4922 (vectorizable_call): Ditto.
4923 (vectorizable_conversion): Ditto.
4924 (vectorizable_assignment): Ditto.
4925 (vectorizable_shift): Ditto.
4926 (vectorizable_operation): Ditto.
4927 (vectorizable_store): Ditto.
4928 (vectorizable_load): Ditto.
4929 (vectorizable_condition): Ditto.
4930 (vect_analyze_stmt): Ditto.
4931 (vect_transform_stmt): Ditto.
4932 (vect_is_simple_use): Ditto.
4933 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4934 (vect_can_advance_ivs_p): Ditto.
4935 (vect_update_ivs_after_vectorizer): Ditto.
4936 (vect_do_peeling_for_loop_bound): Ditto.
4937 (vect_gen_niters_for_prolog_loop): Ditto.
4938 (vect_update_inits_of_drs): Ditto.
4939 (vect_create_cond_for_alias_checks): Ditto.
4940 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4941 (vect_build_slp_tree_1): Ditto.
4942 (vect_supported_load_permutation_p): Ditto.
4943 (vect_analyze_slp_instance): Ditto.
4944 (vect_analyze_slp): Ditto.
4945 (vect_make_slp_decision): Ditto.
4946 (vect_detect_hybrid_slp): Ditto.
4947 (vect_bb_vectorization_profitable_p): Ditto.
4948 (vect_slp_analyze_bb_1): Ditto.
4949 (vect_update_slp_costs_according_to_vf): Ditto.
4950 (vect_get_mask_element): Ditto.
4951 (vect_transform_slp_perm_load): Ditto.
4952 (vect_schedule_slp_instance): Ditto.
4953 (vect_schedule_slp): Ditto.
4954 (vect_slp_transform_bb): Ditto.
4955 * profile.c (read_profile_edge_counts): Ditto.
4956 (compute_branch_probabilities): Ditto.
4957 * coverage.c (get_coverage_counts): Ditto.
4959 2013-09-16 Diego Novillo <dnovillo@google.com>
4961 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4963 2013-09-16 Jan Hubicka <jh@suse.cz>
4965 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4966 abstract functions; for static functions check the presence of body.
4968 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4970 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4971 * config/aarch64/aarch64-simd.md
4972 (aarch64_mla_elt<mode>): New.
4973 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4974 (aarch64_mls_elt<mode>): Likewise.
4975 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4976 (aarch64_fma4_elt<mode>): Likewise.
4977 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4978 (aarch64_fma4_elt_to_128v2df): Likewise.
4979 (aarch64_fma4_elt_to_64df): Likewise.
4980 (fnma<mode>4): Likewise.
4981 (aarch64_fnma4_elt<mode>): Likewise.
4982 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4983 (aarch64_fnma4_elt_to_128v2df): Likewise.
4984 (aarch64_fnma4_elt_to_64df): Likewise.
4985 * config/aarch64/iterators.md (VDQSF): New.
4986 * config/aarch64/arm_neon.h
4987 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4988 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4990 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4992 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4993 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4994 (aarch64_mul3_elt_to_128df): Likewise.
4995 (aarch64_mul3_elt_to_64v2df): Likewise.
4996 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4998 (VMUL_CHANGE_NLANES) Likewise.
5001 * config/aarch64/arm_neon.h
5002 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5004 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5006 * config/aarch64/arm_neon.h
5007 (vcvtx_high_f32_f64): Fix parameters.
5009 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5010 Uros Bizjak <ubizjak@gmail.com>
5012 * config/alpha.c: Include tree-ssanames.h.
5014 2013-09-16 Richard Biener <rguenther@suse.de>
5016 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5017 (dot_rdg_1): Handle control_dd.
5018 (create_edge_for_control_dependence): New function.
5019 (create_rdg_edges): Add control dependences if asked for.
5020 (build_rdg): Likewise.
5021 (generate_loops_for_partition): If there are not necessary
5022 control stmts remove all their dependencies.
5023 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5024 (distribute_loop): Pass on control dependences.
5025 (tree_loop_distribution): Compute control dependences and remove
5026 restriction on number of loop nodes.
5028 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5030 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5033 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5035 * cse.c (try_const_anchors): Punt on CC modes.
5037 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5039 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5041 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5044 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5047 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5050 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5052 (darwin_objc1_section): Likewise.
5053 (darwin_file_end): Emit Image Info section when required.
5055 2013-09-14 Jan Hubicka <jh@suse.cz>
5057 * tree-into-ssa.c (gate_into_ssa): New.
5058 (pass_data_build_ssa): Use it.
5059 * cgraph.h (expand_thunk): Update prototype.
5060 * cgraphunit.c (analyze_function): Expand thunks early.
5061 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5062 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5063 set lowered flag; do not add new function.
5064 (assemble_thunks_and_aliases): Update.
5065 * tree-ssa.c (gate_init_datastructures): New gate.
5066 (pass_data_init_datastructures): Use it.
5068 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5071 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5072 xmm register set described in the psABI.
5074 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5076 * dwarf2out.c (should_emit_struct_debug): Add check
5077 for type_decl variable is not NULL.
5079 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5081 * config.gcc: Use new winnt-c.c target hooks
5082 * config/t-winnt: New file
5083 * config/winnt-c.c: New file
5084 * doc/tm.texi.in: Document new hook
5085 * doc/tm.texi: Regenerated
5087 2013-09-13 Jan Hubicka <jh@suse.cz>
5090 * ipa-inline.c (check_callers): New function.
5091 (check_caller_edge): Remove.
5092 (want_inline_function_to_all_callers_p): Also permit alises that are
5094 (inline_to_all_callers): Terminate the walk when devirtualization
5095 introduce new calls.
5097 2013-09-13 Jan Hubicka <jh@suse.cz>
5099 * ipa-inline-analysis.c (struct growth_data): Add node.
5100 (do_estimate_growth_1): Fix detection of recursion.
5102 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5104 PR tree-optimization/58392
5105 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5106 to avoid shadowing of outer loop variable. If
5107 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5108 replace_by_duplicate_decl simduid of loops that have it set and
5109 set dest_cfun->has_simduid_loops and/or
5110 dest_cfun->has_force_vect_loops.
5111 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5112 instead of maybe_lookup_decl.
5113 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5114 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5115 call. Copy over force_vect and copy and remap simduid. Set
5116 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5117 (copy_cfg_body): Remove blocks_to_copy argument. Use
5118 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5119 caller. Don't set cfun->has_simduid_loops and/or
5120 cfun->has_force_vect_loops here.
5121 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5123 (expand_call_inline, tree_function_versioning): Adjust copy_body
5126 2013-09-13 Martin Jambor <mjambor@suse.cz>
5129 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5130 the assert if the edge was a speculative one.
5132 2013-09-13 Richard Biener <rguenther@suse.de>
5134 * tree-data-ref.h (known_dependences_p): Move here ...
5135 * tree-loop-distribution.c (known_dependences_p): ... from here.
5136 (dump_rdg_component, debug_rdg_component): Remove.
5138 (generate_loops_for_partition): Use gimple_uid instead of
5139 relying on matching stmt visit order.
5140 (rdg_build_partitions): Take starting stmt vector.
5141 (ldist_gen): Merge into ...
5142 (distribute_loop): ... this function. Do not compute starting
5144 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5146 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5148 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5149 Set type attribute correctly. Set predicable_short_it attribute.
5150 (cmpsi_shiftsi): Remove %? from output template.
5152 2013-09-13 Richard Biener <rguenther@suse.de>
5154 * tree-loop-distribution.c (struct rdg_component,
5155 rdg_defs_used_in_other_loops_p, free_rdg_components,
5156 rdg_build_components): Remove.
5157 (stmts_from_loop): Do not record virtual PHIs.
5158 (generate_loops_for_partition): Skip virtual PHIs.
5159 (build_rdg_partition_for_component): Rename to ...
5160 (build_rdg_partition_for_vertex): ... this and adjust.
5161 (rdg_build_partitions): Take a vector of starting vertices
5162 instead of components. Remove unnecessary leftover handling.
5163 (ldist_gen): Do not build components or record other stores.
5164 (distribute_loop): Do not distribute loops containing stmts
5167 2013-09-13 Christian Bruel <christian.bruel@st.com>
5170 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5172 2013-09-13 Kai Tietz <ktietz@redhat.com>
5174 * config.gcc: Separate cases for mingw and cygwin targets,
5175 and add 64-bit cygwin target case.
5177 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5178 dll-export inline-functions.
5179 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5182 2013-09-13 Jeff Law <law@redhat.com>
5186 2013-09-06 Jeff Law <law@redhat.com>
5188 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5189 edge implied equivalences into successor phis.
5191 2013-09-12 DJ Delorie <dj@redhat.com>
5193 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5196 2013-09-12 Brooks Moses <bmoses@google.com>
5199 * Makefile.in: Do not install driver binaries in $(target)/bin.
5201 2013-09-12 DJ Delorie <dj@redhat.com>
5203 * config/rl78/rl78.opt (mrelax): New.
5204 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5205 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5207 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5209 (rl78_expand_epilogue): Likewise.
5211 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5214 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5215 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5216 pointer elimination with using remove_reg_equal_offset_note.
5218 2013-09-12 DJ Delorie <dj@redhat.com>
5220 * config/msp430/: New port.
5221 * config.gcc (msp430): Added.
5222 * doc/invoke.texi: Document MSP430 options.
5223 * doc/install.texi: Document msp430-elf
5224 * doc/md.texi: Document msp430-elf
5225 * doc/contrib.texi: Document msp430-elf
5227 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5230 2013-09-12 Martin Jambor <mjambor@suse.cz>
5233 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5234 reference descriptor is NULL.
5235 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5238 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5240 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5241 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5242 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5243 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5244 make_temp_ssa_name): move to tree-ssanames.h
5245 * tree-ssa-alias.h: Move prototype.
5246 * tree-ssa.h: Include tree-ssanames.h.
5247 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5248 * tree-ssanames.h: New. Move items from tree-flow*.h
5249 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5251 2013-09-12 Richard Biener <rguenther@suse.de>
5253 PR tree-optimization/58404
5254 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5255 propagate non-invariant addresses into dereferences wrapped
5256 in component references.
5258 2013-09-12 Richard Biener <rguenther@suse.de>
5260 PR tree-optimization/58402
5261 * passes.def: Move pass_late_warn_uninitialized later.
5263 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5265 * tree-ssa.h: New. Move content from tree-flow.h and
5267 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5268 Move prototypes belonging to tree-ssa.c.
5269 * tree-flow-inline.h (redirect_edge_var_map_def,
5270 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5272 * gimple.h: Adjust prototypes.
5273 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5275 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5276 * tree.h: Move prototype to tree-ssa.h.
5277 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5278 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5279 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5280 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5281 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5282 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5283 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5284 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5285 graphite-blocking.c, graphite-clast-to-gimple.c,
5286 graphite-dependences.c, graphite-interchange.c,
5287 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5288 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5289 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5290 ipa-reference.c, ipa-split.c, ipa-utils.c,
5291 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5292 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5293 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5294 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5295 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5296 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5297 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5298 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5299 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5300 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5301 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5302 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5303 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5304 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5305 value-prof.c, var-tracking.c,
5306 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5308 2013-09-12 Richard Biener <rguenther@suse.de>
5310 PR tree-optimization/58396
5311 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5312 (build_rdg): Take a loop-nest parameter, fix memleaks.
5313 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5315 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5317 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5320 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5322 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5323 The FMA instruction names should have a 'v' prefix.
5325 2013-09-12 Richard Biener <rguenther@suse.de>
5327 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5328 (dot_rdg): Use popen instead of system in optional code.
5329 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5330 (already_processed_vertex_p): Adjust.
5331 (has_anti_or_output_dependence, predecessor_has_mem_write,
5332 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5333 rdg_flag_uses): Remove.
5334 (rdg_flag_vertex): Simplify.
5335 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5337 (build_rdg_partition_for_component): Process the first vertex
5338 of a component only.
5339 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5341 2013-09-12 Alan Modra <amodra@gmail.com>
5343 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5345 2013-09-11 DJ Delorie <dj@redhat.com>
5346 Nick Clifton <nickc@redhat.com>
5348 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5349 (rl78_stack_based_mem): New.
5350 * config/rl78/constraints.md (Iv08): New.
5359 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5361 (movsi): Change from expand to insn-and-split.
5362 (ashrsi3): Clobber AX.
5366 * config/rl78/rl78.md (CC_REG): Fix.
5367 (addsi3): Allow memory and immediate operands.
5368 (addsi3_internal): Split into...
5369 (addsi3_internal_virt): ...new, and ...
5370 (addsi3_internal_real): ...new.
5372 (subsi3_internal_virt): New.
5373 (subsi3_internal_real): New.
5374 (mulsi3): Add memory operand.
5375 (mulsi3_rl78): Likewise.
5376 (mulsi3_g13): Likewise.
5377 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5378 (cbranchqi4_real): Add more constraint options.
5379 (cbranchhi4_real): Expand pattern.
5380 (cbranchhi4_real_signed): New.
5381 (cbranchhi4_real_inverted): New.
5382 (cbranchsi4_real_lt): New.
5383 (cbranchsi4_real_ge): New.
5384 (cbranchsi4_real_signed): New.
5385 (cbranchsi4_real): New.
5387 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5389 (lshrsi3_virt): Likewise.
5390 (ashlsi3_virt): Likewise.
5391 (cbranchqi4_virt_signed): New.
5392 (cbranchhi4_virt_signed): New.
5393 (cbranchsi4_virt): New.
5394 * config/rl78/rl78.c: Whitespace fixes throughout.
5395 (move_elim_pass): New.
5396 (pass_data_rl78_move_elim): New.
5397 (pass_rl78_move_elim): New.
5398 (make_pass_rl78_move_elim): New.
5399 (rl78_devirt_info): Run devirt earlier.
5400 (rl78_move_elim_info): New.
5401 (rl78_asm_file_start): Register it.
5402 (rl78_split_movsi): New.
5403 (rl78_as_legitimate_address): Allow virtual base registers when
5405 (rl78_addr_space_convert): Remove spurious debug stuff.
5406 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5407 (rl78_print_operand): More cases for not printing '#'.
5408 (rl78_expand_compare): Remove most of the logic.
5409 (content_memory): New.
5410 (clear_content_memory): New.
5411 (get_content_index): New.
5412 (get_content_name): New.
5413 (display_content_memory): New.
5414 (update_content): New.
5415 (record_content): New.
5416 (already_contains): New.
5417 (insn_ok_now): Re-recog insns with virtual registers.
5418 (add_postponed_content_update): New.
5419 (process_postponed_content_update): New.
5420 (gen_and_emit_move): New.
5421 (transcode_memory_rtx): Record new location content.
5422 Use gen_and_emit_move.
5423 (force_into_acc): New.
5424 (move_to_acc): Use gen_and_emit_move.
5425 (move_from_acc): Likewise.
5426 (move_acc_to_reg): Likewise.
5427 (move_to_x): Likewise.
5428 (move_to_hl): Likewise.
5429 (move_to_de): Likewise.
5430 (rl78_alloc_physical_registers_op1): Record location content.
5431 (has_constraint): New.
5432 (rl78_alloc_physical_registers_op2): Record location content.
5434 (rl78_alloc_physical_registers_ro1): Likewise.
5435 (rl78_alloc_physical_registers_cmp): Likewise.
5436 (rl78_alloc_physical_registers_umul): Likewise.
5437 (rl78_alloc_address_registers_macax): New.
5438 (rl78_alloc_physical_registers): Initialize and set location
5439 content memory as needed.
5440 (rl78_reorg): Make sure split2 is called.
5441 (rl78_rtx_costs): New.
5443 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5445 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5446 for (not (neg ...)) and (neg (not ...)) cases.
5448 2013-09-11 Richard Biener <rguenther@suse.de>
5451 * passes.def: Split critical edges before late uninit warning passes.
5452 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5454 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5456 PR tree-optimization/58385
5457 * fold-const.c (build_range_check): If both low and high are NULL,
5458 use omit_one_operand_loc to preserve exp side-effects.
5460 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5462 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5464 2013-09-11 Richard Biener <rguenther@suse.de>
5466 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5467 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5468 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5469 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5470 stmts_from_loop, known_dependences_p, build_empty_rdg,
5471 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5472 * tree-loop-distribution.c: ... here.
5473 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5474 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5475 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5476 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5477 * tree-loop-distribution.c: ... here.
5478 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5479 (struct partition_s): Add loops member.
5480 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5481 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5482 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5484 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5485 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5486 Sergey Lega <sergey.s.lega@intel.com>
5487 Anna Tikhonova <anna.tikhonova@intel.com>
5488 Ilya Tocar <ilya.tocar@intel.com>
5489 Andrey Turetskiy <andrey.turetskiy@intel.com>
5490 Ilya Verbin <ilya.verbin@intel.com>
5491 Kirill Yukhin <kirill.yukhin@intel.com>
5492 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5494 * config/i386/constraints.md (k): New.
5496 * config/i386/i386.c (const regclass_map): Add new mask registers.
5497 (dbx_register_map): Ditto.
5498 (dbx64_register_map): Ditto.
5499 (svr4_dbx_register_map): Ditto.
5500 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5502 (ix86_preferred_reload_class): Disable constants for mask registers.
5503 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5504 (ix86_hard_regno_mode_ok): Support new mask registers.
5505 (x86_order_regs_for_local_alloc): Ditto.
5506 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5507 (FIXED_REGISTERS): Add new mask registers.
5508 (CALL_USED_REGISTERS): Ditto.
5509 (REG_ALLOC_ORDER): Ditto.
5510 (VALID_MASK_REG_MODE): New.
5511 (FIRST_MASK_REG): Ditto.
5512 (LAST_MASK_REG): Ditto.
5513 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5514 (MAYBE_MASK_CLASS_P): New.
5515 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5516 (REG_CLASS_CONTENTS): Ditto.
5517 (MASK_REGNO_P): New.
5518 (ANY_MASK_REG_P): Ditto.
5519 (HI_REGISTER_NAMES): Add new mask registers.
5520 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5521 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5523 (attribute "type"): Add mskmov, msklog.
5524 (attribute "length_immediate"): Support them.
5525 (attribute "memory"): Ditto.
5526 (attribute "prefix_0f"): Ditto.
5527 (*movhi_internal): Support new mask registers.
5528 (*movqi_internal): Ditto.
5529 (define_split): Split out clobber pattern is a logic
5530 insn on mask registers.
5531 (*k<logic><mode>): New.
5532 (*andhi_1): Extend to support mask regs.
5533 (*andqi_1): Extend to support mask regs.
5535 (define_split): Split and-not to and and not if operands
5537 (*<code><mode>_1): Separate HI mode to new pattern...
5538 (*<code>hi_1): This.
5539 (*<code>qi_1): Extend to support mask regs.
5541 (kortestzhi): Ditto.
5542 (kortestchi): Ditto.
5544 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5545 (*one_cmplhi2_1): ...Here, now with mask registers support.
5546 (*one_cmplqi2_1): Support new mask registers.
5547 (HI/QImode arithmetics splitter): Don't split if mask registers
5549 (HI/QImode not splitter): Ditto.
5550 * config/i386/predicated.md (mask_reg_operand): New.
5551 (general_reg_operand): Ditto.
5553 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5555 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5556 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5558 2013-09-10 Jeff Law <law@redhat.com>
5560 PR tree-optimization/58380
5561 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5562 of threading through a buried loop header.
5564 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5565 return value for single successor case.
5567 2013-09-10 Jan Hubicka <jh@suse.cz>
5569 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5571 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5574 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5575 support conditional execution.
5576 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5578 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5580 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5582 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5583 Check destination too to check move insn.
5584 (undo_optional_reloads): Add check that the original peudo did not
5585 changed its allocation and the optional reload was inherited on last
5586 inheritance pass. Break loop after deciding to keep optional reload.
5587 (lra_undo_inheritance): Add check that inherited pseudo still in
5590 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5592 * config/aarch64/aarch64.md (generic_sched): New.
5593 * config/aarch64/aarch64-generic.md (load): Make conditional
5594 on generic_sched attribute.
5595 (nonload): Likewise.
5597 2013-09-10 Jan Hubicka <jh@suse.cz>
5599 * lto-cgraph.c: Include ipa-utils.h.
5600 (compute_ltrans_boundary): Also add possible targets into the boundary.
5602 2013-09-10 Jan Hubicka <jh@suse.cz>
5604 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5605 VAR_DECL of vtable rather than full expression.
5607 2013-09-10 Jan Hubicka <jh@suse.cz>
5608 Paolo Carlini <paolo.carlini@oracle.com>
5610 * cgraphunit.c (analyze_functions): Save input_location, set it
5611 to UNKNOWN_LOCATION and restore it at the end.
5613 2013-09-10 Martin Jambor <mjambor@suse.cz>
5615 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5616 represented by a thunk.
5618 2013-09-10 Jeff Law <law@redhat.com>
5620 PR tree-optimization/58343
5621 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5622 jump threading requests through loop headers buried in the middle
5623 of a jump threading path.
5625 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5626 in return value/type.
5628 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5630 PR rtl-optimization/58365
5631 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5632 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5635 2013-09-10 Richard Biener <rguenther@suse.de>
5637 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5638 * tree-data-ref.c (create_rdg_vertices): Collect all data
5639 references, signal failure to the caller, use data-ref API.
5640 (build_rdg): Compute data references only once. Maintain lifetime
5641 of data references and data dependences from within RDG.
5642 (free_rdg): Free dependence relations.
5643 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5644 inventing extra dependences.
5645 (distribute_loop): Update for RDG API changes.
5647 2013-09-10 Kai Tietz <ktietz@redhat.com>
5649 * doc/invoke.texi (fms-extensions): Document changed
5650 behavior for ms-abi targets.
5651 * config/i386/i386.c (ix86_option_override_internal):
5652 Set default value of option -fms-extension for ms-abi targets.
5654 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5656 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5658 2013-09-10 Alan Modra <amodra@gmail.com>
5661 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5663 2013-09-10 Alan Modra <amodra@gmail.com>
5665 * config/rs6000/predicates.md (add_cint_operand): New.
5666 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5667 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5668 using add_cint_operand.
5669 (largetoc_high_plus_aix): Likewise.
5671 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5673 PR tree-optimization/58364
5674 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5675 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5676 the current range can't be an unconditional true or false.
5678 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5680 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5682 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5684 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5686 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5689 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5691 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5693 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5694 NO_REGS for immediate that can't be moved directly into FP_REGS.
5696 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5698 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5699 comparison with negated operand.
5700 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5703 2013-09-09 Richard Biener <rguenther@suse.de>
5706 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5707 of a subloop record all its block as affecting loop-closed SSA form.
5709 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5711 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5712 of an rtx/bitpos pair.
5713 (store_fixed_bit_field): Update accordingly.
5715 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5717 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5719 * builtins.c (expand_errno_check): Likewise.
5720 * dwarf2cfi.c (init_return_column_size): Likewise.
5721 * except.c (sjlj_mark_call_sites): Likewise.
5722 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5723 * lra-constraints.c (emit_inc): Likewise.
5724 * ree.c (combine_set_extension): Likewise.
5725 * regmove.c (fixup_match_2): Likewise.
5726 * reload1.c (inc_for_reload): Likewise.
5728 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5730 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5731 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5732 Use gen_int_mode with the mode of the associated simplify_* call.
5733 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5734 * expmed.c (expand_shift_1): Likewise.
5735 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5736 * loop-iv.c (iv_number_of_iterations): Likewise.
5737 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5738 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5740 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5742 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5743 of the associated expand_* call.
5744 (asan_emit_stack_protection): Likewise.
5745 * builtins.c (round_trampoline_addr): Likewise.
5746 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5747 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5748 (emit_store_flag): Likewise.
5749 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5750 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5752 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5753 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5754 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5755 * modulo-sched.c (generate_prolog_epilog): Likewise.
5756 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5757 (expand_ctz, expand_ffs, expand_unop): Likewise.
5759 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5761 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5762 of the associated gen_rtx_* call.
5763 * caller-save.c (init_caller_save): Likewise.
5764 * combine.c (find_split_point, make_extraction): Likewise.
5765 (make_compound_operation): Likewise.
5766 * dwarf2out.c (mem_loc_descriptor): Likewise.
5767 * explow.c (plus_constant, probe_stack_range): Likewise.
5768 * expmed.c (expand_mult_const): Likewise.
5769 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5770 * reload1.c (init_reload): Likewise.
5771 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5772 * var-tracking.c (adjust_mems): Likewise.
5773 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5774 rather than gen_rtx_fmt_ee.
5776 2013-09-09 Jan Hubicka <jh@suse.cz>
5779 * value-prof.c (gimple_ic): Copy also abnormal edges.
5781 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5783 * asan.c (asan_shadow_cst): Use gen_int_mode.
5785 2013-09-08 Jan Hubicka <jh@suse.cz>
5787 * ipa-profile.c: Add toplevel comment.
5788 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5789 (contains_hot_call_p): New function.
5790 (ipa_propagate_frequency): Set frequencies based on counts when
5792 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5793 profile; do not tamper with profile after inlining if it is read.
5795 2013-09-08 Jan Hubicka <jh@suse.cz>
5797 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5800 2013-09-08 Jan Hubicka <jh@suse.cz>
5802 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5805 2013-09-08 Jeff Law <law@redhat.com>
5808 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5811 2013-09-08 Andi Kleen <ak@linux.intel.com>
5813 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5815 2013-09-08 Jan Hubicka <jh@suse.cz>
5817 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5819 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5821 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5822 for non-debug insns.
5823 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5824 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5825 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5828 2013-09-08 Jan Hubicka <jh@suse.cz>
5830 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5831 targets and devirtualize to BUILT_IN_UNREACHABLE.
5832 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5833 * ipa.c (walk_polymorphic_call_targets): New function.
5834 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5835 functions; use the new timevar.
5836 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5837 was removed from the program.
5838 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5839 not consider it in the walk when its vtable is dead.
5840 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5842 (devirt_variable_node_removal_hook): New function.
5843 (possible_polymorphic_call_targets): Also register
5844 devirt_variable_node_removal_hook.
5845 (ipa_devirt): Do not do non-speculative devirtualization.
5846 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5848 2013-09-08 Jan Hubicka <jh@suse.cz>
5850 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5851 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5852 varpool_remove_variable_insertion_hook): Declare.
5853 * varpool.c (varpool_node_hook_list): New structure.
5854 (first_varpool_node_removal_hook,
5855 first_varpool_variable_insertion_hook): New variables.
5856 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5857 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5858 varpool_remove_variable_insertion_hook,
5859 varpool_call_variable_insertion_hooks): New functions.
5860 (varpool_remove_node): Use it.
5862 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5865 * diagnostic.c (diagnostic_build_prefix): When s.file is
5866 "<built-in>" don't output line and column numbers.
5868 2013-09-06 Jan Hubicka <jh@suse.cz>
5870 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5871 * lto-streamer-out.c: Stream thunks, too.
5872 * lto-streamer-in.c (input_function): Pop cfun here
5873 (lto_read_body): Instead of here.
5875 2013-09-06 Caroline Tice <cmtice@google.com>
5877 * doc/install.texi: Add documentation for the --enable-vtable-verify
5878 and the --disable-libvtv configure options.
5880 2013-09-06 Jeff Law <law@redhat.com>
5882 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5883 edge implied equivalences into successor phis.
5885 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5887 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5889 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5891 * resource.c (mark_target_live_regs): Compute resources taking
5892 into account if a call is predicated or not.
5894 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5896 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5897 the compiler in the function names.
5899 2013-09-06 Jan Hubicka <jh@suse.cz>
5902 * ipa-inline.c (has_caller_p): New function.
5903 (want_inline_function_to_all_callers_p): Use it.
5904 (sum_callers, inline_to_all_callers): Break out from ...
5905 (ipa_inline): ... here.
5907 2013-09-06 Jan Hubicka <jh@suse.cz>
5909 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5910 only when AVX is enabled.
5912 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5914 * config/aarch64/aarch64.md
5915 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5916 is fpsimd_<load/store>2.
5917 (load_pair<mode>): Likewise.
5918 (store_pair<mode>): Likewise.
5920 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5922 * config/arm/types.md (type): Add "mrs" type.
5923 * config/aarch64/aarch64.md
5924 (aarch64_load_tp_hard): Make type "mrs".
5926 (load_tp_hard): Make type "mrs".
5927 * config/arm/cortex-a15.md: Update with new attributes.
5928 * config/arm/cortex-a5.md: Update with new attributes.
5929 * config/arm/cortex-a53.md: Update with new attributes.
5930 * config/arm/cortex-a7.md: Update with new attributes.
5931 * config/arm/cortex-a8.md: Update with new attributes.
5932 * config/arm/cortex-a9.md: Update with new attributes.
5933 * config/arm/cortex-m4.md: Update with new attributes.
5934 * config/arm/cortex-r4.md: Update with new attributes.
5935 * config/arm/fa526.md: Update with new attributes.
5936 * config/arm/fa606te.md: Update with new attributes.
5937 * config/arm/fa626te.md: Update with new attributes.
5938 * config/arm/fa726te.md: Update with new attributes.
5940 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5942 * config/aarch64/aarch64.md
5943 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5944 (*movtf_aarch64): Likewise.
5946 (thumb1_movdi_insn): Use "multiple" for type where more than one
5947 instruction is used for a move.
5948 (*arm32_movhf): Likewise.
5949 (*thumb_movdf_insn): Likewise.
5951 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5953 * config/arm/types.md (type): Rename fcpys to fmov.
5955 (*arm_movsi_vfp): Rename type fcpys as fmov.
5956 (*thumb2_movsi_vfp): Likewise
5957 (*movhf_vfp_neon): Likewise
5958 (*movhf_vfp): Likewise
5959 (*movsf_vfp): Likewise
5960 (*thumb2_movsf_vfp): Likewise
5961 (*movsfcc_vfp): Likewise
5962 (*thumb2_movsfcc_vfp): Likewise
5963 * config/aarch64/aarch64-simd.md
5964 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5965 * config/aarch64/aarch64.md
5966 (*movsi_aarch64): Replace type mov_reg with fmovs.
5967 (*movdi_aarch64): Likewise
5968 (*movsf_aarch64): Likewise
5969 (*movdf_aarch64): Likewise
5971 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5972 * config/arm/iwmmxt.md
5973 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5974 * config/arm/arm1020e.md: Update with new attributes.
5975 * config/arm/cortex-a15-neon.md: Update with new attributes.
5976 * config/arm/cortex-a5.md: Update with new attributes.
5977 * config/arm/cortex-a53.md: Update with new attributes.
5978 * config/arm/cortex-a7.md: Update with new attributes.
5979 * config/arm/cortex-a8-neon.md: Update with new attributes.
5980 * config/arm/cortex-a9.md: Update with new attributes.
5981 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5982 * config/arm/cortex-r4f.md: Update with new attributes.
5983 * config/arm/marvell-pj4.md: Update with new attributes.
5984 * config/arm/vfp11.md: Update with new attributes.
5986 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5988 * config/aarch64/aarch64.md
5989 (*madd<mode>): Fix type attribute.
5990 (*maddsi_uxtw): Likewise.
5991 (*msub<mode>): Likewise.
5992 (*msubsi_uxtw): Likewise.
5993 (<su_optab>maddsidi4): Likewise.
5994 (<su_optab>msubsidi4): Likewise.
5996 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5998 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5999 * config/arm/arm.md (core_cycles): Remove fdiv.
6000 * config/arm/vfp.md:
6001 (*sqrtsf2_vfp): Update for attribute changes.
6002 (*sqrtdf2_vfp): Likewise.
6003 * config/aarch64/aarch64.md:
6004 (sqrt<mode>2): Update for attribute changes.
6005 * config/arm/arm1020e.md: Update with new attributes.
6006 * config/arm/cortex-a15-neon.md: Update with new attributes.
6007 * config/arm/cortex-a5.md: Update with new attributes.
6008 * config/arm/cortex-a53.md: Update with new attributes.
6009 * config/arm/cortex-a7.md: Update with new attributes.
6010 * config/arm/cortex-a8-neon.md: Update with new attributes.
6011 * config/arm/cortex-a9.md: Update with new attributes.
6012 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6013 * config/arm/cortex-r4f.md: Update with new attributes.
6014 * config/arm/marvell-pj4.md: Update with new attributes.
6015 * config/arm/vfp11.md: Update with new attributes.
6017 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6019 * config/arm/types.md
6020 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6021 * config/aarch64/aarch64.md
6022 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6024 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6025 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6026 (float<GPI:mode><GPF:mode>2): Likewise.
6028 (*truncsisf2_vfp): Update with new attributes.
6029 (*truncsidf2_vfp): Likewise.
6030 (fixuns_truncsfsi2): Likewise.
6031 (fixuns_truncdfsi2): Likewise.
6032 (*floatsisf2_vfp): Likewise.
6033 (*floatsidf2_vfp): Likewise.
6034 (floatunssisf2): Likewise.
6035 (floatunssidf2): Likewise.
6036 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6037 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6038 * config/arm/arm1020e.md: Update with new attributes.
6039 * config/arm/cortex-a15-neon.md: Update with new attributes.
6040 * config/arm/cortex-a5.md: Update with new attributes.
6041 * config/arm/cortex-a53.md: Update with new attributes.
6042 * config/arm/cortex-a7.md: Update with new attributes.
6043 * config/arm/cortex-a8-neon.md: Update with new attributes.
6044 * config/arm/cortex-a9.md: Update with new attributes.
6045 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6046 * config/arm/cortex-r4f.md: Update with new attributes.
6047 * config/arm/marvell-pj4.md: Update with new attributes.
6048 * config/arm/vfp11.md: Update with new attributes.
6050 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6052 * config/aarch64/arm_neon.h
6053 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6054 (vqtbx<1,2,3,4><q>_s8): Likewise.
6056 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6058 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6059 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6060 (add<mode>3): Add type attribute.
6061 (add<mode>3): Likewise.
6062 (usadd<mode>3): Likewise.
6063 (ssadd<mode>3): Likewise.
6064 (sub<mode>3): Likewise.
6065 (sub<mode>3): Likewise.
6066 (ussub<mode>3): Likewise.
6067 (sssub<mode>3): Likewise.
6068 (ssmulsa3): Likewise.
6069 (usmulusa3): Likewise.
6070 (arm_usatsihi): Likewise.
6072 (*movdi_vfp): Add types for all instructions.
6073 (*movdi_vfp_cortexa8): Likewise.
6074 (*movhf_vfp_neon): Likewise.
6075 (*movhf_vfp): Likewise.
6076 (*movdf_vfp): Likewise.
6077 (*thumb2_movdf_vfp): Likewise.
6078 (*thumb2_movdfcc_vfp): Likewise.
6079 * config/arm/arm.md: Add type attribute to all insn patterns.
6080 (*thumb1_adddi3): Add type attribute.
6081 (*arm_adddi3): Likewise.
6082 (*adddi_sesidi_di): Likewise.
6083 (*adddi_zesidi_di): Likewise.
6084 (*thumb1_addsi3): Likewise.
6085 (addsi3_compare0): Likewise.
6086 (*addsi3_compare0_scratch): Likewise.
6087 (*compare_negsi_si): Likewise.
6088 (cmpsi2_addneg): Likewise.
6089 (*addsi3_carryin_<optab>): Likewise.
6090 (*addsi3_carryin_alt2_<optab>): Likewise.
6091 (*addsi3_carryin_clobercc_<optab>): Likewise.
6092 (*subsi3_carryin): Likewise.
6093 (*subsi3_carryin_const): Likewise.
6094 (*subsi3_carryin_compare): Likewise.
6095 (*subsi3_carryin_compare_const): Likewise.
6096 (*arm_subdi3): Likewise.
6097 (*thumb_subdi3): Likewise.
6098 (*subdi_di_zesidi): Likewise.
6099 (*subdi_di_sesidi): Likewise.
6100 (*subdi_zesidi_di): Likewise.
6101 (*subdi_sesidi_di): Likewise.
6102 (*subdi_zesidi_ze): Likewise.
6103 (thumb1_subsi3_insn): Likewise.
6104 (*arm_subsi3_insn): Likewise.
6105 (*anddi3_insn): Likewise.
6106 (*anddi_zesidi_di): Likewise.
6107 (*anddi_sesdi_di): Likewise.
6108 (*ne_zeroextracts): Likewise.
6109 (*ne_zeroextracts): Likewise.
6110 (*ite_ne_zeroextr): Likewise.
6111 (*ite_ne_zeroextr): Likewise.
6112 (*anddi_notdi_di): Likewise.
6113 (*anddi_notzesidi): Likewise.
6114 (*anddi_notsesidi): Likewise.
6115 (andsi_notsi_si): Likewise.
6116 (thumb1_bicsi3): Likewise.
6117 (*iordi3_insn): Likewise.
6118 (*iordi_zesidi_di): Likewise.
6119 (*iordi_sesidi_di): Likewise.
6120 (*thumb1_iorsi3_insn): Likewise.
6121 (*xordi3_insn): Likewise.
6122 (*xordi_zesidi_di): Likewise.
6123 (*xordi_sesidi_di): Likewise.
6124 (*arm_xorsi3): Likewise.
6125 (*andsi_iorsi3_no): Likewise.
6126 (*smax_0): Likewise.
6127 (*smax_m1): Likewise.
6128 (*arm_smax_insn): Likewise.
6129 (*smin_0): Likewise.
6130 (*arm_smin_insn): Likewise.
6131 (*arm_umaxsi3): Likewise.
6132 (*arm_uminsi3): Likewise.
6133 (*minmax_arithsi): Likewise.
6134 (*minmax_arithsi_): Likewise.
6135 (*satsi_<SAT:code>): Likewise.
6136 (arm_ashldi3_1bit): Likewise.
6137 (arm_ashrdi3_1bit): Likewise.
6138 (arm_lshrdi3_1bit): Likewise.
6139 (*arm_negdi2): Likewise.
6140 (*thumb1_negdi2): Likewise.
6141 (*arm_negsi2): Likewise.
6142 (*thumb1_negsi2): Likewise.
6143 (*negdi_extendsid): Likewise.
6144 (*negdi_zero_extend): Likewise.
6145 (*arm_abssi2): Likewise.
6146 (*thumb1_abssi2): Likewise.
6147 (*arm_neg_abssi2): Likewise.
6148 (*thumb1_neg_abss): Likewise.
6149 (one_cmpldi2): Likewise.
6150 (extend<mode>di2): Likewise.
6151 (*compareqi_eq0): Likewise.
6152 (*arm_extendhisi2addsi): Likewise.
6153 (*arm_movdi): Likewise.
6154 (*thumb1_movdi_insn): Likewise.
6155 (*arm_movt): Likewise.
6156 (*thumb1_movsi_insn): Likewise.
6157 (pic_add_dot_plus_four): Likewise.
6158 (pic_add_dot_plus_eight): Likewise.
6159 (tls_load_dot_plus_eight): Likewise.
6160 (*thumb1_movhi_insn): Likewise.
6161 (*thumb1_movsf_insn): Likewise.
6162 (*movdf_soft_insn): Likewise.
6163 (*thumb_movdf_insn): Likewise.
6164 (cbranchsi4_insn): Likewise.
6165 (cbranchsi4_scratch): Likewise.
6166 (*negated_cbranchsi4): Likewise.
6167 (*tbit_cbranch): Likewise.
6168 (*tlobits_cbranch): Likewise.
6169 (*tstsi3_cbranch): Likewise.
6170 (*cbranchne_decr1): Likewise.
6171 (*addsi3_cbranch): Likewise.
6172 (*addsi3_cbranch_scratch): Likewise.
6173 (*arm_cmpdi_insn): Likewise.
6174 (*arm_cmpdi_unsig): Likewise.
6175 (*arm_cmpdi_zero): Likewise.
6176 (*thumb_cmpdi_zero): Likewise.
6177 (*deleted_compare): Likewise.
6178 (*mov_scc): Likewise.
6179 (*mov_negscc): Likewise.
6180 (*mov_notscc): Likewise.
6181 (*cstoresi_eq0_thumb1_insn): Likewise.
6182 (cstoresi_nltu_thumb1): Likewise.
6183 (cstoresi_ltu_thu): Likewise.
6184 (thumb1_addsi3_addgeu): Likewise.
6185 (*arm_jump): Likewise.
6186 (*thumb_jump): Likewise.
6187 (*check_arch2): Likewise.
6188 (arm_casesi_internal): Likewise.
6189 (thumb1_casesi_dispatch): Likewise.
6190 (*arm_indirect_jump): Likewise.
6191 (*thumb1_indirect_jump): Likewise.
6193 (*and_scc): Likewise.
6194 (*ior_scc): Likewise.
6195 (*compare_scc): Likewise.
6196 (*cond_move): Likewise.
6197 (*cond_arith): Likewise.
6198 (*cond_sub): Likewise.
6199 (*cmp_ite0): Likewise.
6200 (*cmp_ite1): Likewise.
6201 (*cmp_and): Likewise.
6202 (*cmp_ior): Likewise.
6203 (*ior_scc_scc): Likewise.
6204 (*ior_scc_scc_cmp): Likewise.
6205 (*and_scc_scc): Likewise.
6206 (*and_scc_scc_cmp): Likewise.
6207 (*and_scc_scc_nod): Likewise.
6208 (*negscc): Likewise.
6209 (movcond_addsi): Likewise.
6210 (movcond): Likewise.
6211 (*ifcompare_plus_move): Likewise.
6212 (*if_plus_move): Likewise.
6213 (*ifcompare_move_plus): Likewise.
6214 (*if_move_plus): Likewise.
6215 (*ifcompare_arith_arith): Likewise.
6216 (*if_arith_arith): Likewise.
6217 (*ifcompare_arith_move): Likewise.
6218 (*if_arith_move): Likewise.
6219 (*ifcompare_move_arith): Likewise.
6220 (*if_move_arith): Likewise.
6221 (*ifcompare_move_not): Likewise.
6222 (*if_move_not): Likewise.
6223 (*ifcompare_not_move): Likewise.
6224 (*if_not_move): Likewise.
6225 (*ifcompare_shift_move): Likewise.
6226 (*if_shift_move): Likewise.
6227 (*ifcompare_move_shift): Likewise.
6228 (*if_move_shift): Likewise.
6229 (*ifcompare_shift_shift): Likewise.
6230 (*ifcompare_not_arith): Likewise.
6231 (*ifcompare_arith_not): Likewise.
6232 (*if_arith_not): Likewise.
6233 (*ifcompare_neg_move): Likewise.
6234 (*if_neg_move): Likewise.
6235 (*ifcompare_move_neg): Likewise.
6236 (*if_move_neg): Likewise.
6237 (prologue_thumb1_interwork): Likewise.
6238 (*cond_move_not): Likewise.
6239 (*sign_extract_onebit): Likewise.
6240 (*not_signextract_onebit): Likewise.
6241 (stack_tie): Likewise.
6242 (align_4): Likewise.
6243 (align_8): Likewise.
6244 (consttable_end): Likewise.
6245 (consttable_1): Likewise.
6246 (consttable_2): Likewise.
6247 (consttable_4): Likewise.
6248 (consttable_8): Likewise.
6249 (consttable_16): Likewise.
6250 (*thumb1_tablejump): Likewise.
6251 (prefetch): Likewise.
6252 (force_register_use): Likewise.
6253 (thumb_eh_return): Likewise.
6254 (load_tp_hard): Likewise.
6255 (load_tp_soft): Likewise.
6256 (tlscall): Likewise.
6257 (*arm_movtas_ze): Likewise.
6258 (*arm_rev): Likewise.
6259 (*arm_revsh): Likewise.
6260 (*arm_rev16): Likewise.
6261 * config/arm/thumb2.md
6262 (*thumb2_smaxsi3): Likewise.
6263 (*thumb2_sminsi3): Likewise.
6264 (*thumb32_umaxsi3): Likewise.
6265 (*thumb2_uminsi3): Likewise.
6266 (*thumb2_negdi2): Likewise.
6267 (*thumb2_abssi2): Likewise.
6268 (*thumb2_neg_abss): Likewise.
6269 (*thumb2_movsi_insn): Likewise.
6270 (tls_load_dot_plus_four): Likewise.
6271 (*thumb2_movhi_insn): Likewise.
6272 (*thumb2_mov_scc): Likewise.
6273 (*thumb2_mov_negs): Likewise.
6274 (*thumb2_mov_negs): Likewise.
6275 (*thumb2_mov_nots): Likewise.
6276 (*thumb2_mov_nots): Likewise.
6277 (*thumb2_movsicc_): Likewise.
6278 (*thumb2_movsfcc_soft_insn): Likewise.
6279 (*thumb2_indirect_jump): Likewise.
6280 (*thumb2_and_scc): Likewise.
6281 (*thumb2_ior_scc): Likewise.
6282 (*thumb2_ior_scc_strict_it): Likewise.
6283 (*thumb2_cond_move): Likewise.
6284 (*thumb2_cond_arith): Likewise.
6285 (*thumb2_cond_ari): Likewise.
6286 (*thumb2_cond_sub): Likewise.
6287 (*thumb2_negscc): Likewise.
6288 (*thumb2_movcond): Likewise.
6289 (thumb2_casesi_internal): Likewise.
6290 (thumb2_casesi_internal_pic): Likewise.
6291 (*thumb2_alusi3_short): Likewise.
6292 (*thumb2_mov<mode>_shortim): Likewise.
6293 (*thumb2_addsi_short): Likewise.
6294 (*thumb2_subsi_short): Likewise.
6295 (thumb2_addsi3_compare0): Likewise.
6296 (*thumb2_cbz): Likewise.
6297 (*thumb2_cbnz): Likewise.
6298 (*thumb2_one_cmplsi2_short): Likewise.
6299 (*thumb2_negsi2_short): Likewise.
6300 (*orsi_notsi_si): Likewise.
6301 * config/arm/arm1020e.md: Update with new attributes.
6302 * config/arm/arm1026ejs.md: Update with new attributes.
6303 * config/arm/arm1136jfs.md: Update with new attributes.
6304 * config/arm/arm926ejs.md: Update with new attributes.
6305 * config/arm/cortex-a15.md: Update with new attributes.
6306 * config/arm/cortex-a5.md: Update with new attributes.
6307 * config/arm/cortex-a53.md: Update with new attributes.
6308 * config/arm/cortex-a7.md: Update with new attributes.
6309 * config/arm/cortex-a8.md: Update with new attributes.
6310 * config/arm/cortex-a9.md: Update with new attributes.
6311 * config/arm/cortex-m4.md: Update with new attributes.
6312 * config/arm/cortex-r4.md: Update with new attributes.
6313 * config/arm/fa526.md: Update with new attributes.
6314 * config/arm/fa606te.md: Update with new attributes.
6315 * config/arm/fa626te.md: Update with new attributes.
6316 * config/arm/fa726te.md: Update with new attributes.
6318 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6320 * config/aarch64/aarch64-simd.md
6321 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6322 <vwx> iterator to ensure correct register choice.
6323 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6324 (aarch64_sqdmull_n<mode>): Likewise.
6325 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6326 * config/aarch64/arm_neon.h
6327 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6328 (vml<as><q>_n_<su>16): Likewise.
6329 (vml<as>l_high_lane<q>_<su>16): Likewise.
6330 (vml<as>l_high_n_<su>16): Likewise.
6331 (vml<as>l_lane<q>_<su>16): Likewise.
6332 (vml<as>l_n_<su>16): Likewise.
6333 (vmul<q>_lane<q>_<su>16): Likewise.
6334 (vmul<q>_n_<su>16): Likewise.
6335 (vmull_lane<q>_<su>16): Likewise.
6336 (vmull_n_<su>16): Likewise.
6337 (vmull_high_lane<q>_<su>16): Likewise.
6338 (vmull_high_n_<su>16): Likewise.
6339 (vqrdmulh<q>_n_s16): Likewise.
6341 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6343 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6344 have the correct lane parameter.
6346 2013-09-06 Richard Biener <rguenther@suse.de>
6348 * cfganal.c (control_dependences::~control_dependences):
6349 Properly free all of the vector.
6351 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6354 * config/i386/i386.c (ix86_conditional_register_usage):
6355 Proper initialize extended SSE registers.
6357 2013-09-06 Jan Hubicka <jh@suse.cz>
6359 PR tree-optimization/58311
6360 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6362 2013-09-06 Jan Hubicka <jh@suse.cz>
6364 * Makefile.in (tree-sra.o): Update dependencies.
6365 * tree-sra.c: Include ipa-utils.h
6366 (scan_function): Use recursive_call_p.
6367 (has_caller_p): New function.
6368 (cgraph_for_node_and_aliases): Count also callers of aliases.
6370 2013-09-06 Jan Hubicka <jh@suse.cz>
6373 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6374 * tree-tailcall.c: Include ipa-utils.h.
6375 (find_tail_calls): Use it.
6376 * ipa-pure-const.c (check_call): Likewise.
6377 * ipa-utils.c (recursive_call_p): New function.
6378 * ipa-utils.h (recursive_call_p): Dclare.
6379 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6380 (symtab_semantically_equivalent_p): New function.
6381 * Makefile.in (tree-tailcall.o): Update dependencies.
6383 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6385 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6388 2013-09-06 Richard Biener <rguenther@suse.de>
6390 * lto-streamer.h (lto_global_var_decls): Remove.
6391 * Makefile.in (OBJS): Remove lto-symtab.o.
6392 (lto-symtab.o): Remove.
6393 (GTFILES): Remove lto-symtab.c
6394 * lto-symtab.c: Move to lto/
6396 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6398 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6399 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6400 (UNSPEC_FPINT_RINT): New constant definitions.
6401 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6402 definition with 2 attributes.
6403 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6404 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6407 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6409 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6410 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6411 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6412 attribute to "z196_alone".
6413 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6416 2013-09-06 Richard Biener <rguenther@suse.de>
6418 * basic-block.h (class control_dependences): New.
6419 * tree-ssa-dce.c (control_dependence_map): Remove.
6421 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6422 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6423 find_pdom, find_control_dependence, find_all_control_dependences):
6425 (mark_control_dependent_edges_necessary,
6426 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6427 tree_dce_done, perform_tree_ssa_dce): Adjust.
6428 * cfganal.c (set_control_dependence_map_bit,
6429 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6430 find_all_control_dependences): Move from tree-ssa-dce.c and
6431 implement as methods of control_dependences class.
6432 (control_dependences::control_dependences): New.
6433 (control_dependences::~control_dependences): Likewise.
6434 (control_dependences::get_edges_dependent_on): Likewise.
6435 (control_dependences::get_edge): Likewise.
6437 2013-09-04 Jan Hubicka <jh@suse.cz>
6439 * tree.c (types_same_for_odr): Drop overactive check.
6440 * ipa-devirt.c (hash_type_name): Likewise.
6442 2013-09-04 Jan Hubicka <jh@suse.cz>
6444 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6445 (analyze_functions): ... here.
6447 2013-09-04 Jan Hubicka <jh@suse.cz>
6450 * cgraphunit.c (analyze_functions): Clear AUX fields
6451 after processing; initialize assembler name has.
6453 2013-09-05 Jeff Law <law@redhat.com>
6455 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6456 from thread_around_empty_block. Record threading path into PATH.
6457 Recurse if threading through the initial block is successful.
6458 (thread_across_edge): Corresponding changes to slightly simplify.
6460 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6462 * config/aarch64/aarch64.md
6463 (type): Remove frecpe, frecps, frecpx.
6464 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6465 fix to be a TARGET_SIMD instruction.
6466 (aarch64_frecps): Remove.
6467 * config/aarch64/aarch64-simd.md
6468 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6469 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6471 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6472 Sofiane Naci <sofiane.naci@arm.com>
6474 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6475 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6476 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6477 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6478 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6479 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6480 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6481 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6482 into "clz, "rbit". Rename "shift" to "shift_imm".
6483 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6484 changes. Update for attribute changes all occurrences of arlo_* and
6486 * config/arm/arm-fixed.md: Update for attribute changes
6487 all occurrences of arlo_* types.
6488 * config/arm/thumb2.md: Update for attribute changes all occurrences
6490 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6491 (cortexa7_older_only): Likewise.
6492 (cortexa7_younger): Likewise.
6493 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6494 (1020alu_shift_op): Likewise.
6495 (1020alu_shift_reg_op): Likewise.
6496 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6497 (alu_shift_op): Likewise.
6498 (alu_shift_reg_op): Likewise.
6499 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6500 (11_alu_shift_op): Likewise.
6501 (11_alu_shift_reg_op): Likewise.
6502 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6503 (9_alu_shift_reg_op): Likewise.
6504 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6506 (cortex_a15_alu_shift): Likewise.
6507 (cortex_a15_alu_shift_reg): Likewise.
6508 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6510 (cortex_a5_alu_shift): Likewise.
6511 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6513 (cortex_a53_alu_shift): Likewise.
6514 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6516 (cortex_a7_alu_reg): Likewise.
6517 (cortex_a7_alu_shift): Likewise.
6518 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6520 (cortex_a8_alu_shift): Likewise.
6521 (cortex_a8_alu_shift_reg): Likewise.
6522 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6523 (cortex_a9_dp_shift): Likewise.
6524 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6526 * config/arm/cortex-r4.md
6527 (cortex_r4_alu): Update for attribute changes.
6528 (cortex_r4_mov): Likewise.
6529 (cortex_r4_alu_shift_reg): Likewise.
6530 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6531 (526_alu_shift_op): Likewise.
6532 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6533 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6534 (626te_alu_shift_op): Likewise.
6535 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6536 (726te_alu_shift_op): Likewise.
6537 (726te_alu_shift_reg_op): Likewise.
6538 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6539 (mp626_alu_shift_op): Likewise.
6540 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6541 (pj4_alu_conds): Likewise.
6542 (pj4_shift): Likewise.
6543 (pj4_shift_conds): Likewise.
6544 (pj4_alu_shift): Likewise.
6545 (pj4_alu_shift_conds): Likewise.
6546 * config/aarch64/aarch64.md: Update for attribute change
6547 all occurrences of arlo_* and shift* types.
6549 2013-09-05 Mike Stump <mikestump@comcast.net>
6551 * tree.h: Move documentation for tree_function_decl to tree-core.h
6552 with the declaration.
6554 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6557 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6558 looking for widest mode.
6560 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6562 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6564 2013-09-05 Richard Biener <rguenther@suse.de>
6566 PR tree-optimization/58137
6567 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6568 Do not create vectors of pointers.
6569 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6570 types for the components of the vector initializer.
6571 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6572 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6574 2013-09-05 Martin Jambor <mjambor@suse.cz>
6576 * ipa-prop.c (remove_described_reference): Accept missing references,
6577 return false if that hppens, otherwise return true.
6578 (cgraph_node_for_jfunc): New function.
6579 (try_decrement_rdesc_refcount): Likewise.
6580 (try_make_edge_direct_simple_call): Use them.
6581 (ipa_edge_removal_hook): Remove references from rdescs.
6582 (ipa_edge_duplication_hook): Clone rdescs and their references
6583 when the new edge has the same caller as the old one.
6584 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6585 reference before removing any edges.
6587 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6589 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6591 * thumb2.md (thumb2_storewb_parisi): New pattern.
6593 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6595 * config/aarch64/aarch64-option-extensions.def: Add
6596 AARCH64_OPT_EXTENSION of 'crc'.
6597 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6598 (AARCH64_ISA_CRC): Ditto.
6599 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6600 description of the CRC extension.
6602 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6604 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6605 * config/rs6000/linux.h: Ditto.
6606 * alpha/linux.h: Ditto.
6607 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6608 no_c99_libc_has_function.
6609 * config/c6x/uclinux-elf.h: Ditto.
6610 * config/lm32/uclinux-elf.h: Ditto.
6611 * config/m68k/uclinux.h: Ditto.
6612 * config/moxie/uclinux.h: Ditto.
6613 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6614 (crisv32-*-linux*, cris-*-linux*): Ditto.
6615 * config/bfin/bfin.c: Include "tm_p.h".
6617 2013-09-05 Richard Biener <rguenther@suse.de>
6619 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6620 check for a definition without a basic-block.
6622 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6623 Sofiane Naci <sofiane.naci@arm.com>
6625 * config/aarch64/aarch64.md
6626 (*movti_aarch64): Rename r_2_f and f_2_r.
6627 (*movsf_aarch64): Likewise.
6628 (*movdf_aarch64): Likewise.
6629 (*movtf_aarch64): Likewise.
6630 (aarch64_movdi_<mode>low): Likewise.
6631 (aarch64_movdi_<mode>high): Likewise.
6632 (aarch64_mov<mode>high_di): Likewise.
6633 (aarch64_mov<mode>low_di): Likewise.
6634 (aarch64_movtilow_tilow): Likewise.
6635 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6636 values to config/arm/types.md
6637 (attribute "conds"): Update for attribute change.
6638 (anddi3_insn): Likewise.
6639 (iordi3_insn): Likewise.
6640 (xordi3_insn): Likewise.
6641 (one_cmpldi2): Likewise.
6642 * config/arm/types.md (type): Add Neon types.
6643 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6644 use "type" attribute.
6645 (movmisalign<mode>_neon_store): Likewise.
6646 (movmisalign<mode>_neon_load): Likewise.
6647 (vec_set<mode>_internal): Likewise.
6648 (vec_setv2di_internal): Likewise.
6649 (vec_extract<mode>): Likewise.
6650 (vec_extractv2di): Likewise.
6651 (add<mode>3_neon): Likewise.
6652 (adddi3_neon): Likewise.
6653 (sub<mode>3_neon): Likewise.
6654 (subdi3_neon): Likewise.
6655 (mul<mode>3_neon): Likewise.
6656 (mul<mode>3add<mode>_neon): Likewise.
6657 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6658 (fma<VCVTF:mode>4)): Likewise.
6659 (fma<VCVTF:mode>4_intrinsic): Likewise.
6660 (fmsub<VCVTF:mode>4)): Likewise.
6661 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6662 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6663 (ior<mode>3): Likewise.
6664 (and<mode>3): Likewise.
6665 (anddi3_neon): Likewise.
6666 (orn<mode>3_neon): Likewise.
6667 (orndi3_neon): Likewise.
6668 (bic<mode>3_neon): Likewise.
6669 (bicdi3_neon): Likewise.
6670 (xor<mode>3): Likewise.
6671 (one_cmpl<mode>2): Likewise.
6672 (abs<mode>2): Likewise.
6673 (neg<mode>2): Likewise.
6674 (umin<mode>3_neon): Likewise.
6675 (umax<mode>3_neon): Likewise.
6676 (smin<mode>3_neon): Likewise.
6677 (smax<mode>3_neon): Likewise.
6678 (vashl<mode>3): Likewise.
6679 (vashr<mode>3_imm): Likewise.
6680 (vlshr<mode>3_imm): Likewise.
6681 (ashl<mode>3_signed): Likewise.
6682 (ashl<mode>3_unsigned): Likewise.
6683 (neon_load_count): Likewise.
6684 (ashldi3_neon_noclobber): Likewise.
6685 (signed_shift_di3_neon): Likewise.
6686 (unsigned_shift_di3_neon): Likewise.
6687 (ashrdi3_neon_imm_noclobber): Likewise.
6688 (lshrdi3_neon_imm_noclobber): Likewise.
6689 (widen_ssum<mode>3): Likewise.
6690 (widen_usum<mode>3): Likewise.
6691 (quad_halves_<code>v4si): Likewise.
6692 (quad_halves_<code>v4sf): Likewise.
6693 (quad_halves_<code>v8hi): Likewise.
6694 (quad_halves_<code>v16qi): Likewise.
6695 (reduc_splus_v2di): Likewise.
6696 (neon_vpadd_internal<mode>): Likewise.
6697 (neon_vpsmin<mode>): Likewise.
6698 (neon_vpsmax<mode>): Likewise.
6699 (neon_vpumin<mode>): Likewise.
6700 (neon_vpumax<mode>): Likewise.
6701 (ss_add<mode>_neon): Likewise.
6702 (us_add<mode>_neon): Likewise.
6703 (ss_sub<mode>_neon): Likewise.
6704 (us_sub<mode>_neon): Likewise.
6705 (neon_vadd<mode>_unspec): Likewise.
6706 (neon_vaddl<mode>): Likewise.
6707 (neon_vaddw<mode>): Likewise.
6708 (neon_vhadd<mode>): Likewise.
6709 (neon_vqadd<mode>): Likewise.
6710 (neon_vaddhn<mode>): Likewise.
6711 (neon_vmul<mode>): Likewise.
6712 (neon_vmla<mode>): Likewise.
6713 (neon_vmlal<mode>): Likewise.
6714 (neon_vmls<mode>): Likewise.
6715 (neon_vmlsl<mode>): Likewise.
6716 (neon_vqdmulh<mode>): Likewise.
6717 (neon_vqdmlal<mode>): Likewise.
6718 (neon_vqdmlsl<mode>): Likewise.
6719 (neon_vmull<mode>): Likewise.
6720 (neon_vqdmull<mode>): Likewise.
6721 (neon_vsub<mode>_unspec): Likewise.
6722 (neon_vsubl<mode>): Likewise.
6723 (neon_vsubw<mode>): Likewise.
6724 (neon_vqsub<mode>): Likewise.
6725 (neon_vhsub<mode>): Likewise.
6726 (neon_vsubhn<mode>): Likewise.
6727 (neon_vceq<mode>): Likewise.
6728 (neon_vcge<mode>): Likewise.
6729 (neon_vcgeu<mode>): Likewise.
6730 (neon_vcgt<mode>): Likewise.
6731 (neon_vcgtu<mode>): Likewise.
6732 (neon_vcle<mode>): Likewise.
6733 (neon_vclt<mode>): Likewise.
6734 (neon_vcage<mode>): Likewise.
6735 (neon_vcagt<mode>): Likewise.
6736 (neon_vtst<mode>): Likewise.
6737 (neon_vabd<mode>): Likewise.
6738 (neon_vabdl<mode>): Likewise.
6739 (neon_vaba<mode>): Likewise.
6740 (neon_vabal<mode>): Likewise.
6741 (neon_vmax<mode>): Likewise.
6742 (neon_vmin<mode>): Likewise.
6743 (neon_vpaddl<mode>): Likewise.
6744 (neon_vpadal<mode>): Likewise.
6745 (neon_vpmax<mode>): Likewise.
6746 (neon_vpmin<mode>): Likewise.
6747 (neon_vrecps<mode>): Likewise.
6748 (neon_vrsqrts<mode>): Likewise.
6749 (neon_vqabs<mode>): Likewise.
6750 (neon_vqneg<mode>): Likewise.
6751 (neon_vcls<mode>): Likewise.
6752 (clz<mode>2): Likewise.
6753 (popcount<mode>2): Likewise.
6754 (neon_vrecpe): Likewise.
6755 (neon_vrsqrte): Likewise.
6756 (neon_vget_lane<mode>_sext_internal): Likewise.
6757 (neon_vget_lane<mode>_zext_internal): Likewise.
6758 (neon_vdup_n<mode>): Likewise.
6759 (neon_vdup_nv2di): Likewise.
6760 (neon_vdpu_lane<mode>_internal): Likewise.
6761 (neon_vswp<mode>): Likewise.
6762 (float<mode><V_cvtto>2): Likewise.
6763 (floatuns<mode><V_cvtto>2): Likewise.
6764 (fix_trunc<mode><V_cvtto>)2): Likewise
6765 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6766 (neon_vcvt<mode>): Likewise.
6767 (neon_vcvtv4sfv4hf): Likewise.
6768 (neon_vcvtv4hfv4sf): Likewise.
6769 (neon_vcvt_n<mode>): Likewise.
6770 (neon_vmovn<mode>): Likewise.
6771 (neon_vqmovn<mode>): Likewise.
6772 (neon_vqmovun<mode>): Likewise.
6773 (neon_vmovl<mode>): Likewise.
6774 (neon_vmul_lane<mode>): Likewise.
6775 (neon_vmull_lane<mode>): Likewise.
6776 (neon_vqdmull_lane<mode>): Likewise.
6777 (neon_vqdmulh_lane<mode>): Likewise.
6778 (neon_vmla_lane<mode>): Likewise.
6779 (neon_vmlal_lane<mode>): Likewise.
6780 (neon_vqdmlal_lane<mode>): Likewise.
6781 (neon_vmls_lane<mode>): Likewise.
6782 (neon_vmlsl_lane<mode>): Likewise.
6783 (neon_vqdmlsl_lane<mode>): Likewise.
6784 (neon_vext<mode>): Likewise.
6785 (neon_vrev64<mode>): Likewise.
6786 (neon_vrev32<mode>): Likewise.
6787 (neon_vrev16<mode>): Likewise.
6788 (neon_vbsl<mode>_internal): Likewise.
6789 (neon_vshl<mode>): Likewise.
6790 (neon_vqshl<mode>): Likewise.
6791 (neon_vshr_n<mode>): Likewise.
6792 (neon_vshrn_n<mode>): Likewise.
6793 (neon_vqshrn_n<mode>): Likewise.
6794 (neon_vqshrun_n<mode>): Likewise.
6795 (neon_vshl_n<mode>): Likewise.
6796 (neon_vqshl_n<mode>): Likewise.
6797 (neon_vqshlu_n<mode>): Likewise.
6798 (neon_vshll_n<mode>): Likewise.
6799 (neon_vsra_n<mode>): Likewise.
6800 (neon_vsri_n<mode>): Likewise.
6801 (neon_vsli_n<mode>): Likewise.
6802 (neon_vtbl1v8qi): Likewise.
6803 (neon_vtbl2v8qi): Likewise.
6804 (neon_vtbl3v8qi): Likewise.
6805 (neon_vtbl4v8qi): Likewise.
6806 (neon_vtbx1v8qi): Likewise.
6807 (neon_vtbx2v8qi): Likewise.
6808 (neon_vtbx3v8qi): Likewise.
6809 (neon_vtbx4v8qi): Likewise.
6810 (neon_vtrn<mode>_internal): Likewise.
6811 (neon_vzip<mode>_internal): Likewise.
6812 (neon_vuzp<mode>_internal): Likewise.
6813 (neon_vld1<mode>): Likewise.
6814 (neon_vld1_lane<mode>): Likewise.
6815 (neon_vld1_dup<mode>): Likewise.
6816 (neon_vld1_dupv2di): Likewise.
6817 (neon_vst1<mode>): Likewise.
6818 (neon_vst1_lane<mode>): Likewise.
6819 (neon_vld2<mode>): Likewise.
6820 (neon_vld2_lane<mode>): Likewise.
6821 (neon_vld2_dup<mode>): Likewise.
6822 (neon_vst2<mode>): Likewise.
6823 (neon_vst2_lane<mode>): Likewise.
6824 (neon_vld3<mode>): Likewise.
6825 (neon_vld3qa<mode>): Likewise.
6826 (neon_vld3qb<mode>): Likewise.
6827 (neon_vld3_lane<mode>): Likewise.
6828 (neon_vld3_dup<mode>): Likewise.
6829 (neon_vst3<mode>): Likewise.
6830 (neon_vst3qa<mode>): Likewise.
6831 (neon_vst3qb<mode>): Likewise.
6832 (neon_vst3_lane<mode>): Likewise.
6833 (neon_vld4<mode>): Likewise.
6834 (neon_vld4qa<mode>): Likewise.
6835 (neon_vld4qb<mode>): Likewise.
6836 (neon_vld4_lane<mode>): Likewise.
6837 (neon_vld4_dup<mode>): Likewise.
6838 (neon_vst4<mode>): Likewise.
6839 (neon_vst4qa<mode>): Likewise.
6840 (neon_vst4qb<mode>): Likewise.
6841 (neon_vst4_lane<mode>): Likewise.
6842 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6843 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6844 (neon_vec_<US>mult_lo_<mode>): Likewise.
6845 (neon_vec_<US>mult_hi_<mode>): Likewise.
6846 (neon_vec_<US>shiftl_<mode>): Likewise.
6847 (neon_unpack<US>_<mode>): Likewise.
6848 (neon_vec_<US>mult_<mode>): Likewise.
6849 (vec_pack_trunc_<mode>): Likewise.
6850 (neon_vec_pack_trunk_<mode>): Likewise.
6851 (neon_vabd<mode>_2): Likewise.
6852 (neon_vabd<mode>_3): Likewise.
6853 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6854 (thumb2_movsi_vfp): Likewise.
6855 (movdi_vfp): Likewise.
6856 (movdi_vfp_cortexa8): Likewise.
6857 (movhf_vfp_neon): Likewise.
6858 (movhf_vfp): Likewiwse.
6859 (movsf_vfp): Likewiwse.
6860 (thumb2_movsf_vfp): Likewiwse.
6861 (movdf_vfp): Likewise.
6862 (thumb2_movdf_vfp): Likewise.
6863 (movsfcc_vfp): Likewise.
6864 (thumb2_movsfcc_vfp): Likewise.
6865 (movdfcc_vfp): Likewise.
6866 (thumb2_movdfcc_vfp): Likewise.
6867 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6868 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6869 (v10_v2c): Likewise.
6870 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6872 (cortex_a15_neon_int_2): Likewise.
6873 (cortex_a15_neon_int_3): Likewise.
6874 (cortex_a15_neon_int_4): Likewise.
6875 (cortex_a15_neon_int_5): Likewise.
6876 (cortex_a15_neon_vqneg_vqabs): Likewise.
6877 (cortex_a15_neon_vmov): Likewise.
6878 (cortex_a15_neon_vaba): Likewise.
6879 (cortex_a15_neon_vaba_qqq): Likewise.
6880 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6881 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6882 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6883 scalar_64_32_long_scalar): Likewise.
6884 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6885 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6886 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6887 lotype_qdd_64_32_long): Likewise.
6888 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6889 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6890 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6891 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6892 (cortex_a15_neon_shift_1): Likewise.
6893 (cortex_a15_neon_shift_2): Likewise.
6894 (cortex_a15_neon_shift_3): Likewise.
6895 (cortex_a15_neon_vshl_ddd): Likewise.
6896 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6897 (cortex_a15_neon_vsra_vrsra): Likewise.
6898 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6899 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6900 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6901 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6902 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6903 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6904 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6905 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6906 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6907 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6908 (cortex_a15_neon_bp_simple): Likewise.
6909 (cortex_a15_neon_bp_2cycle): Likewise.
6910 (cortex_a15_neon_bp_3cycle): Likewise.
6911 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6912 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6913 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6914 (cortex_a15_neon_vld2_4_regs): Likewise.
6915 (cortex_a15_neon_vld3_vld4): Likewise.
6916 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6917 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6918 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6919 (cortex_a15_neon_vst3_vst4): Likewise.
6920 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6921 (cortex_a15_neon_vld3_vld4_lane" 10
6922 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6923 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6924 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6925 (cortex_a15_neon_ldm_2): Likewise.0
6926 (cortex_a15_neon_stm_2): Likewise.
6927 (cortex_a15_neon_mcr): Likewise.
6928 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6929 (cortex_a15_neon_mrc): Likewise.
6930 (cortex_a15_neon_mrrc): Likewise.
6931 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6933 (cortex_a15_alu_shift): Likewise.
6934 (cortex_a15_alu_shift_reg): Likewise.
6935 (cortex_a15_mult32): Likewise.
6936 (cortex_a15_mult64): Likewise.
6937 (cortex_a15_block): Likewise.
6938 (cortex_a15_branch): Likewise.
6939 (cortex_a15_load1): Likewise.
6940 (cortex_a15_load3): Likewise.
6941 (cortex_a15_store1): Likewise.
6942 (cortex_a15_store3): Likewise.
6943 (cortex_a15_call): Likewise.
6944 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6945 (cortex_a5_f2r): Likewise.
6946 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6948 (cortex_a53_f2r): Likewise.
6949 * config/arm/cortex-a7.md
6950 (cortex_a7_branch): Update for attribute change.
6951 (cortex_a7_call): Likewise.
6952 (cortex_a7_alu_imm): Likewise.
6953 (cortex_a7_alu_reg): Likewise.
6954 (cortex_a7_alu_shift): Likewise.
6955 (cortex_a7_mul): Likewise.
6956 (cortex_a7_load1): Likewise.
6957 (cortex_a7_store1): Likewise.
6958 (cortex_a7_load2): Likewise.
6959 (cortex_a7_store2): Likewise.
6960 (cortex_a7_load3): Likewise.
6961 (cortex_a7_store3): Likewise.
6962 (cortex_a7_load4): Likewise.
6963 (cortex_a7_store4): Likewise.
6964 (cortex_a7_fpalu): Likewise.
6965 (cortex_a7_fconst): Likewise.
6966 (cortex_a7_fpmuls): Likewise.
6967 (cortex_a7_neon_mul): Likewise.
6968 (cortex_a7_fpmacs): Likewise.
6969 (cortex_a7_neon_mla: Likewise.
6970 (cortex_a7_fpmuld: Likewise.
6971 (cortex_a7_fpmacd: Likewise.
6972 (cortex_a7_fpfmad: Likewise.
6973 (cortex_a7_fdivs: Likewise.
6974 (cortex_a7_fdivd: Likewise.
6975 (cortex_a7_r2f: Likewise.
6976 (cortex_a7_f2r: Likewise.
6977 (cortex_a7_f_flags: Likewise.
6978 (cortex_a7_f_loads: Likewise.
6979 (cortex_a7_f_loadd: Likewise.
6980 (cortex_a7_f_stores: Likewise.
6981 (cortex_a7_f_stored: Likewise.
6982 (cortex_a7_neon): Likewise.
6983 * config/arm/cortex-a8-neon.md
6984 (cortex_a8_neon_mrc): Update for attribute change.
6985 (cortex_a8_neon_mrrc): Likewise.
6986 (cortex_a8_neon_int_1): Likewise.
6987 (cortex_a8_neon_int_2): Likewise.
6988 (cortex_a8_neon_int_3): Likewise.
6989 (cortex_a8_neon_int_4): Likewise.
6990 (cortex_a8_neon_int_5): Likewise.
6991 (cortex_a8_neon_vqneg_vqabs): Likewise.
6992 (cortex_a8_neon_vmov): Likewise.
6993 (cortex_a8_neon_vaba): Likewise.
6994 (cortex_a8_neon_vaba_qqq): Likewise.
6995 (cortex_a8_neon_vsma): Likewise.
6996 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6997 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6998 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6999 long_scalar): Likewise.
7000 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7001 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7002 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7003 long_scalar_qdd_64_32_long): Likewise.
7004 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7005 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7006 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7007 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7008 (cortex_a8_neon_shift_1): Likewise.
7009 (cortex_a8_neon_shift_2): Likewise.
7010 (cortex_a8_neon_shift_3): Likewise.
7011 (cortex_a8_neon_vshl_ddd): Likewise.
7012 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7013 (cortex_a8_neon_vsra_vrsra): Likewise.
7014 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7015 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7016 (cortex_a8_neon_fp_vsum): Likewise.
7017 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7018 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7019 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7020 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7021 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7022 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7023 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7024 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7025 (cortex_a8_neon_bp_simple): Likewise.
7026 (cortex_a8_neon_bp_2cycle): Likewise.
7027 (cortex_a8_neon_bp_3cycle): Likewise.
7028 (cortex_a8_neon_ldr): Likewise.
7029 (cortex_a8_neon_str): Likewise.
7030 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7031 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7032 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7033 (cortex_a8_neon_vld2_4_regs): Likewise.
7034 (cortex_a8_neon_vld3_vld4): Likewise.
7035 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7036 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7037 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7038 (cortex_a8_neon_vst3_vst4): Likewise.
7039 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7040 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7041 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7042 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7043 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7044 (cortex_a8_neon_mcr): Likewise.
7045 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7046 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7047 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7049 (ca9_neon_mrrc): Likewise.
7050 (cortex_a9_neon_int_1): Likewise.
7051 (cortex_a9_neon_int_2): Likewise.
7052 (cortex_a9_neon_int_3): Likewise.
7053 (cortex_a9_neon_int_4): Likewise.
7054 (cortex_a9_neon_int_5): Likewise.
7055 (cortex_a9_neon_vqneg_vqabs): Likewise.
7056 (cortex_a9_neon_vmov): Likewise.
7057 (cortex_a9_neon_vaba): Likewise.
7058 (cortex_a9_neon_vaba_qqq): Likewise.
7059 (cortex_a9_neon_vsma): Likewise.
7060 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7061 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7062 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7063 long_scalar): Likewise.
7064 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7065 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7066 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7067 long_scalar_qdd_64_32_long): Likewise.
7068 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7069 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7070 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7071 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7072 (cortex_a9_neon_shift_1): Likewise.
7073 (cortex_a9_neon_shift_2): Likewise.
7074 (cortex_a9_neon_shift_3): Likewise.
7075 (cortex_a9_neon_vshl_ddd): Likewise.
7076 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7077 (cortex_a9_neon_vsra_vrsra): Likewise.
7078 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7079 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7080 (cortex_a9_neon_fp_vsum): Likewise.
7081 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7082 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7083 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7084 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7085 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7086 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7087 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7088 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7089 (cortex_a9_neon_bp_simple): Likewise.
7090 (cortex_a9_neon_bp_2cycle): Likewise.
7091 (cortex_a9_neon_bp_3cycle): Likewise.
7092 (cortex_a9_neon_ldr): Likewise.
7093 (cortex_a9_neon_str): Likewise.
7094 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7095 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7096 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7097 (cortex_a9_neon_vld2_4_regs): Likewise.
7098 (cortex_a9_neon_vld3_vld4): Likewise.
7099 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7100 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7101 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7102 (cortex_a9_neon_vst3_vst4): Likewise.
7103 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7104 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7105 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7106 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7107 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7108 (cortex_a9_neon_mcr): Likewise.
7109 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7110 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7111 (cortex_a9_fps): Likewise.
7112 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7114 (cortex_m4_fmuls): Likewise.
7115 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7117 (cortex_r4_mrc): Likewise.
7118 * config/arm/iterators.md: Update comment referring to neon_type.
7119 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7120 (iwmmxt_movsi_insn): Likewise.
7121 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7123 (pj4_core_to_vfp): Likewise.
7124 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7126 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7127 (vfp_fstore): Likewise.
7128 * doc/md.texi: Change references to neon_type to refer to type.
7130 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7132 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7134 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7136 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7139 2013-09-04 Easwaran Raman <eraman@google.com>
7142 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7143 (build_and_add_sum): Use it.
7144 (appears_later_in_bb): Simplify code.
7146 2013-09-04 Teresa Johnson <tejohnson@google.com>
7148 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7150 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7152 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7154 2013-09-04 Jan Hubicka <jh@suse.cz>
7156 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7157 * ipa-devirt.c: Include diganostic.h
7158 (odr_type_d): Add types and types_set.
7159 (hash_type_name): Work for types with vtables during LTO.
7160 (odr_hasher::remove): Fix comment; destroy types_set.
7161 (add_type_duplicate): New function,
7162 (get_odr_type): Use it.
7163 (dump_type_inheritance_graph): Dump type duplicates.
7164 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7166 * tree.c (types_same_for_odr): Give exact answers on types with
7169 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7171 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7172 explaining their differences.
7174 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7176 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7178 2013-09-03 Jeff Law <law@redhat.com>
7180 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7181 when not threading through a joiner block. Pass joiner/no joiner
7182 state to register_jump_thread.
7183 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7184 state from argument rather than implying on path length.
7185 Dump the entire jump thread path into debugging dump.
7186 * tree-flow.h (register_jump_thread): Update prototype.
7188 2013-08-29 Xinliang David Li <davidxl@google.com>
7190 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7191 Remove a trivial gcc_assert.
7193 2013-08-29 Xinliang David Li <davidxl@google.com>
7195 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7196 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7197 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7198 Delay base decl alignment adjustment.
7199 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7200 * tree-vectorizer.h: New data structure.
7201 (set_dr_misalignment): New function.
7202 (dr_misalignment): Ditto.
7203 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7204 (vectorizable_load): Ditto.
7205 (ensure_base_align): New function.
7206 (vectorize_loops): Add dbg_cnt support.
7207 (execute_vect_slp): Ditto.
7208 * dbgcnt.def: New debug counter.
7209 * Makefile: New dependency.
7211 2013-09-03 Meador Inge <meadori@codesourcery.com>
7215 2013-08-30 Meador Inge <meadori@codesourcery.com>
7217 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7219 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7221 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7222 function descriptor.
7224 2013-09-03 Richard Biener <rguenther@suse.de>
7226 * tree-affine.c (add_elt_to_tree): Fix association issue,
7227 avoid useless converts and make sure to always return a
7228 properly typed result.
7230 2013-09-03 Richard Biener <rguenther@suse.de>
7233 * fold-const.c (negate_expr_p): Fix division case.
7234 (negate_expr): Likewise.
7236 2013-09-03 Richard Biener <rguenther@suse.de>
7239 * tree-streamer-out.c: Include tm.h.
7240 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7242 2013-09-03 Jan Hubicka <jh@suse.cz>
7244 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7246 2013-09-03 Alan Modra <amodra@gmail.com>
7248 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7249 multilibs to big-endian target and vice versa.
7250 * config/rs6000/t-linux64: Use := assignment on all vars.
7251 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7252 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7253 * config/rs6000/t-linux64le: New file.
7254 * config/rs6000/t-linux64bele: New file.
7255 * config/rs6000/t-linux64lebe: New file.
7257 2013-09-02 Jan Hubicka <jh@suse.cz>
7259 * ipa-inline-transform.c (inline_transform): Do not
7260 optimize_inline_calls when not optimizing.
7262 2013-09-02 Jan Hubicka <jh@suse.cz>
7264 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7265 duplicated nodes for assembler names.
7266 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7267 hard registers from assembler name hash.
7269 2013-09-02 Jan Hubicka <jh@suse.cz>
7271 * ipa-split.c (execute_split_functions): Split externally visible
7272 functions called once.
7274 2013-09-02 Martin Jambor <mjambor@suse.cz>
7277 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7278 linked list. When finding the correct duplicate, also consider also
7279 the caller in additon to its inlined_to node.
7281 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7283 * config/aarch64/aarch64-simd-builtins.def
7284 (dup_lane_scalar): Remove.
7285 * config/aarch64/aarch64-simd.md
7286 (aarch64_simd_dup): Add 'w->w' alternative.
7287 (aarch64_dup_lane<mode>): Allow for VALL.
7288 (aarch64_dup_lane_scalar<mode>): Remove.
7289 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7290 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7291 (aarch64_get_lane_unsigned<mode>): Likewise.
7292 (aarch64_get_lane<mode>): Likewise.
7293 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7294 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7295 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7296 (VCON): Change container of V2SF.
7297 (vswap_width_name): Likewise.
7298 * config/aarch64/arm_neon.h
7299 (__aarch64_vdup_lane_any): New.
7300 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7301 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7302 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7304 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7307 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7308 the source or the destination is a single hard register.
7310 2013-09-02 Richard Biener <rguenther@suse.de>
7313 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7316 2013-09-02 Richard Biener <rguenther@suse.de>
7318 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7319 arithmetic to sizetype.
7321 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7323 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7324 Find auto-increment use both before and after candidate.
7326 2013-09-02 Marek Polacek <polacek@redhat.com>
7328 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7330 2013-09-01 Jan Hubicka <jh@suse.cz>
7332 * Makefile.in: Add ipa-profile.o
7333 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7334 * cgraph.c (struct cgraph_propagate_frequency_data,
7335 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7336 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7337 * cgraph.h (cgraph_propagate_frequency): Remove.
7338 * ipa-inline-analysis.c: Include ipa-utils.h;
7339 drop duplicated cfgloop.h.
7340 (inline_update_callee_summaries): Update.
7341 * ipa-profile.c: New file.
7342 * ipa-utils.h (ipa_propagate_frequency): Declare.
7343 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7344 data-streamer.h, value-prof.h.
7345 (symtab_remove_unreachable_nodes): Update profile.
7346 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7347 account_time_size, cmp_counts, dump_histogram,
7348 ipa_profile_generate_summary, ipa_profile_write_summary,
7349 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7350 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7351 Move to ipa-profile.c.
7353 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7355 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7357 2013-09-01 Jan Hubicka <jh@suse.cz>
7359 * common.opt (fdevirtualize-speculatively): New function.
7360 * invoke.texi (fdevirtualize-speculatively): Document.
7361 * ipa-devirt.c: Include ipa-inline.h
7362 (likely_target_p): New function.
7363 (ipa_devirt): New function.
7364 (gate_ipa_devirt): New function.
7365 (pass_data_ipa_devirt): New static var.
7366 (pass_ipa_devirt): Likewise.
7367 (make_pass_ipa_devirt): New function.
7368 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7369 (common_handle_option): Disable devirtualization when
7370 value range profiling is available.
7371 * passes.def (pass_ipa_devirt): Add.
7372 * timever.def (TV_IPA_DEVIRT): New timevar.
7373 * tree-pass.h (make_pass_ipa_devirt):
7375 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7377 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7378 include sanitize(undefined).
7380 2013-08-31 Diego Novillo <dnovillo@google.com>
7382 * Makefile.in (TREE_CORE_H): Define.
7384 (GTFILES): Add tree-core.h.
7385 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7387 * tree-core.h: New file.
7388 Move all data structures, enum, typedefs, global
7389 declarations and constants from ...
7392 2013-08-31 Jan Hubicka <jh@suse.cz>
7394 * bulitins.c (expand_builtin): Do not early exit for gcov
7395 instrumented functions.
7397 2013-08-31 Marek Polacek <polacek@redhat.com>
7399 * ubsan.c: Include tm_p.h.
7401 2013-08-31 Jan Hubicka <jh@suse.cz>
7403 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7406 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7407 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7408 * tree-cfg.c (verify_gimple_label): ... here.
7409 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7410 (ipa_merge_profiles): New function.
7411 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7412 (lto_input_function_body): Likewise.
7413 * ipa-utils.h (ipa_merge_profiles): Declare.
7414 * lto-streamer.h (lto_input_function_body): Update prototype.
7415 (emit_label_in_global_context_p): Remove.
7416 * lto-symtab.c: Include ipa-utils.h
7417 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7419 2013-08-31 Jan Hubicka <jh@suse.cz>
7421 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7423 2013-08-31 Jan Hubicka <jh@suse.cz>
7425 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7427 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7429 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7430 "cmp" RTX before signed_comparison_operator check to account
7433 2013-08-30 Jan Hubicka <jh@suse.cz>
7435 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7436 (detect_type_change_1): Rename to ...
7437 (detect_type_change): ... this one; early return on non-polymorphic
7439 (detect_type_change_ssa): Add comp_type parameter; update
7440 use of detect_type_change.
7441 (compute_complex_assign_jump_func): Add param_type parameter;
7442 update use of detect_type_change_ssa.
7443 (compute_complex_ancestor_jump_func): Likewise.
7444 (ipa_get_callee_param_type): New function.
7445 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7446 update calls to the jump function computation functions.
7448 2013-08-30 Teresa Johnson <tejohnson@google.com>
7449 Steven Bosscher <steven@gcc.gnu.org>
7451 * cfgrtl.c (fixup_new_cold_bb): New routine.
7452 (commit_edge_insertions): Invoke fixup_partitions.
7453 (find_partition_fixes): New routine.
7454 (fixup_partitions): Ditto.
7455 (verify_hot_cold_block_grouping): Update comments.
7456 (rtl_verify_edges): Invoke find_partition_fixes.
7457 (rtl_verify_bb_pointers): Update comments.
7458 (rtl_verify_bb_layout): Ditto.
7459 * basic-block.h (probably_never_executed_edge_p): Declare.
7460 (fixup_partitions): Ditto.
7461 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7462 * bb-reorder.c (sanitize_hot_paths): New function.
7463 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7465 * predict.c (probably_never_executed_edge_p): New routine.
7466 * cfg.c (check_bb_profile): Add partition insanity warnings.
7468 2013-08-30 Meador Inge <meadori@codesourcery.com>
7470 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7472 2013-08-30 Marek Polacek <polacek@redhat.com>
7474 * Makefile.in (ubsan.o): Add.
7475 (c-family/c-ubsan.o): Add.
7476 (builtins.o): Add ubsan.h dependency.
7477 * ubsan.h: New file.
7478 * ubsan.c: New file.
7479 * common.opt: Add -fsanitize=undefined option.
7480 (flag_sanitize): Add variable.
7481 (fsanitize=): Add option. Add Driver.
7482 (fsanitize=thread): Remove option.
7483 (fsanitize=address): Likewise.
7484 (static-libubsan): New option.
7485 * doc/invoke.texi: Document the new flag and -static-libubsan.
7486 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7487 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7488 * builtin-attrs.def (ATTR_COLD): Define.
7489 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7490 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7491 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7492 * flag-types.h (sanitize_code): New enum.
7493 * opts.c (common_handle_option): Parse command line arguments
7494 of -fsanitize=. Add -fsanitize=unreachable option.
7495 * varasm.c (get_variable_section): Adjust.
7496 (assemble_noswitch_variable): Likewise.
7497 (assemble_variable): Likewise.
7498 (output_constant_def_contents): Likewise.
7499 (categorize_decl_for_section): Likewise.
7500 (place_block_symbol): Likewise.
7501 (output_object_block): Likewise.
7502 * builtins.def: Likewise.
7503 * toplev.c (compile_file): Likewise.
7504 (process_options): Likewise.
7505 * cppbuiltin.c: Likewise.
7506 * tsan.c (tsan_pass): Likewise.
7507 (tsan_gate): Likewise.
7508 (tsan_gate_O0): Likewise.
7509 * cfgexpand.c (partition_stack_vars): Likewise.
7510 (expand_stack_vars): Likewise.
7511 (defer_stack_allocation): Likewise.
7512 (expand_used_vars): Likewise.
7513 * cfgcleanup.c (old_insns_match_p): Likewise.
7514 * asan.c (asan_finish_file): Likewise.
7515 (asan_instrument): Likewise.
7516 (gate_asan): Likewise.
7517 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7518 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7519 (asan_global_struct): Use pointer_sized_int_node instead
7520 calling build_nonstandard_integer_type.
7521 (initialize_sanitizer_builtins): Likewise.
7522 (asan_finish_file): Likewise.
7523 * gcc.c: Document %{%:function(args):X}.
7524 (static_spec_functions): Add sanitize.
7525 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7526 store funcval != NULL there.
7527 (do_spec_1): Adjust handle_spec_function caller.
7528 (handle_braces): Allow %:function(args) as condition.
7529 (sanitize_spec_function): New function.
7530 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7531 (LIBUBSAN_SPEC): Likewise.
7532 (LIBUBSAN_EARLY_SPEC): Likewise.
7533 (SANITIZER_SPEC): Handle libubsan.
7534 (SANITIZER_EARLY_SPEC): Likewise.
7535 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7536 instead of fsanitize=address.
7537 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7538 instead of fsanitize=address*.
7539 * builtins.c: Include ubsan.h.
7540 (fold_builtin_0): Instrument __builtin_unreachable.
7541 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7542 instead of flag_asan.
7543 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7544 (pointer_sized_int_node): Define.
7545 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7547 2013-08-30 Mike Stump <mikestump@comcast.net>
7549 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7552 2013-08-29 Jan Hubicka <jh@suse.cz>
7554 * cgraph.c (cgraph_function_body_availability): Handle weakref
7556 * passes.def: Remove pass_fixup_cfg.
7557 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7558 track when we need to remove functions.
7559 (gate_ipa_inline): Execute inlining always; add comment why.
7560 (pass_data_ipa_inline): Remove TODO_remove_functions.
7561 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7562 do not produce summaries.
7563 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7564 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7565 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7566 constant pool and vtable.
7568 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7570 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7571 New arm_neon.h's internal macros to specify 64-bit constants.
7572 Avoid using stdint.h's macros.
7574 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7576 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7577 before applying REG_P.
7579 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7581 PR tree-optimization/58277
7582 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7583 after seeing too many stmts with vdef in between dombb and current
7584 bb, invalidate everything.
7586 2013-08-30 Richard Biener <rguenther@suse.de>
7588 * fold-const.c (fold_single_bit_test): Fix overflow test.
7590 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7592 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7593 and which can live in a register, always retrieve the value on entry.
7594 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7595 passed by invisible reference specially.
7596 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7597 (vt_add_function_parameter): Correctly deal with a parameter passed by
7598 invisible reference.
7600 2013-08-30 Jan Hubicka <jh@suse.cz>
7602 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7604 2013-08-30 Richard Biener <rguenther@suse.de>
7606 PR tree-optimization/58228
7607 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7608 allow invariant loads in nested loop vectorization.
7610 2013-08-30 Richard Biener <rguenther@suse.de>
7612 PR tree-optimization/58223
7613 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7614 (has_anti_or_output_dependence): ... this and adjust to also
7615 look for output dependences.
7616 (mark_nodes_having_upstream_mem_writes): Adjust.
7617 (rdg_flag_uses): Likewise.
7619 2013-08-30 Richard Biener <rguenther@suse.de>
7621 PR tree-optimization/58010
7622 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7623 assert that we have a loop-closed PHI.
7625 2013-08-29 Jan Hubicka <jh@suse.cz>
7627 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7628 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7629 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7630 (lto_free_function_in_decl_state_for_node): New function.
7632 2013-08-29 Xinliang David Li <davidxl@google.com>
7634 * loop-unroll.c (report_unroll_peel): Minor message change.
7635 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7636 Emit alignment peeling message with default -fopt-info.
7637 (vect_loop_versioning): Emit loop version info message.
7638 * tree-vectorizer.c (vectorize_loops): Minor message change.
7639 (execute_vect_slp): Ditto.
7641 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7643 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7644 of the clone from the DECL_NAME of the original function.
7646 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7648 * passes.c (register_pass): Add overload.
7649 * tree-pass.h (register_pass): Forward declare it. Add comment.
7651 2013-08-29 Jan Hubicka <jh@suse.cz>
7653 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7654 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7655 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7656 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7657 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7658 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7659 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7660 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7662 2013-08-29 Teresa Johnson <tejohnson@google.com>
7664 * dumpfile.c (dump_loc): Output column number.
7665 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7666 * doc/invoke.texi: Document optall -fopt-info flag.
7667 * profile.c (read_profile_edge_counts): Use new dump framework.
7668 (compute_branch_probabilities): Ditto.
7669 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7670 when pass not in any opt group.
7671 * pass_manager.h (pass_manager::get_pass_profile): New method.
7672 * value-prof.c (check_counter): Use new dump framework.
7673 (check_ic_target): Ditto.
7674 * coverage.c (get_coverage_counts): Ditto.
7675 (coverage_init): Setup new dump framework.
7677 2013-08-29 Richard Biener <rguenther@suse.de>
7679 PR tree-optimization/58246
7680 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7681 handle the dominance check inside a basic-block.
7683 2013-08-29 Richard Biener <rguenther@suse.de>
7686 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7687 of default defs that appear in abnormal PHI nodes.
7689 2013-08-29 Richard Biener <rguenther@suse.de>
7691 PR tree-optimization/57685
7692 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7693 single-use operands to avoid exponential complexity.
7695 2013-08-28 Dehao Chen <dehao@google.com>
7697 * ipa-inline.c (edge_badness): Fix integer underflow.
7699 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7701 * gtm-builtins.def (_ITM_free): Declare leaf.
7703 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7706 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7707 (*tls_local_dynamic_base_64_largepic): Likewise.
7708 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7709 Remove predicate from call operand.
7710 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7711 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7713 2013-08-28 Jeff Law <law@redhat.com>
7715 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7716 checks for the number of predecessors and successors allowed.
7717 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7718 which require copying a joiner block if there is a request which
7719 is a subpath that requires no joiner block copying.
7721 2013-08-28 Jan Hubicka <jh@suse.cz>
7723 * lto-streamer-out.c (DFS_write_tree_body): Drop
7724 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7725 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7726 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7727 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7728 Do not read DECL_ERROR_ISSUED.
7729 (unpack_ts_decl_with_vis_value_fields): Do not read
7731 (lto_input_ts_binfo_tree_pointers): Do not read
7732 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7733 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7734 write DECL_ERROR_ISSUED..
7735 (pack_ts_decl_with_vis_value_fields): Do not write
7737 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7738 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7739 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7740 * tree.h (tree_decl_common): Update comment.
7741 (DECL_ERROR_ISSUED): Remove.
7743 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7746 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7748 2013-08-28 Jan Hubicka <jh@suse.cz>
7750 * builtins.def (free): Declare leaf.
7752 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7755 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7756 * configure: Regenerate.
7758 2013-08-27 Martin Jambor <mjambor@suse.cz>
7760 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7761 (ipa_ancestor_jf_data): Likewise.
7762 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7763 (ipa_get_jf_pass_through_type_preserved): New function.
7764 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7765 (ipa_get_jf_ancestor_type_preserved): New function.
7766 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7767 (ipa_get_jf_ancestor_result): Likewise.
7768 (propagate_vals_accross_pass_through): Use
7769 ipa_get_jf_pass_through_result to do all the value mappings.
7770 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7771 type_preserved flag.
7772 (ipa_set_jf_cst_copy): New function.
7773 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7774 (ipa_set_jf_arith_pass_through): Likewise.
7775 (ipa_set_ancestor_jf): Likewise.
7776 (compute_complex_assign_jump_func): Set type_preserved instead of
7778 (ipa_compute_jump_functions_for_edge): Likewise.
7779 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7780 (update_jump_functions_after_inlining): Update type_preserved.
7781 Explicitely create jump functions when combining one with pass_through.
7782 (ipa_write_jump_function): Stream the type_preserved flags.
7783 (ipa_read_jump_function): Likewise.
7785 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7786 Aldy Hernandez <aldyh@redhat.com>
7788 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7789 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7790 * function.h (struct function): Add has_force_vect_loops and
7792 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7793 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7795 * gimple.def: Update CLAUSES comments.
7796 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7797 (gimple_build_omp_for): Add argument to prototype.
7798 (gimple_omp_for_kind): New.
7799 (gimple_omp_for_set_kind): New.
7800 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7801 GOVD_DATA_SHARE_CLASS.
7802 (enum omp_region_type): Add ORT_SIMD.
7803 (gimple_add_tmp_var): Handle ORT_SIMD.
7804 (gimplify_var_or_parm_decl): Same.
7805 (is_gimple_stmt): Same.
7806 (omp_firstprivatize_variable): Same.
7807 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7808 (omp_notice_variable): Handle ORT_SIMD.
7809 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7810 (omp_check_private): Handle ORT_SIMD.
7811 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7813 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7814 Handle OMP_CLAUSE_LASTPRIVATE.
7815 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7817 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7818 (gimplify_expr): Handle OMP_SIMD.
7819 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7820 (expand_GOMP_SIMD_VF): New.
7821 (expand_GOMP_SIMD_LAST_LANE): New.
7822 * internal-fn.def (GOMP_SIMD_LANE): New.
7823 (GOMP_SIMD_VF): New.
7824 (GOMP_SIMD_LAST_LANE): New.
7825 * omp-low.c: Include target.h.
7826 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7828 (check_omp_nesting_restrictions): Same.
7830 (lower_rec_simd_input_clauses): New.
7831 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7833 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7834 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7835 (expand_omp_build_assign): New.
7836 (expand_omp_for_init_counts): New.
7837 (expand_omp_for_init_vars): New.
7838 (extract_omp_for_update_vars): New.
7839 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7840 and rewrite accordingly.
7841 (expand_omp_simd): New.
7842 (expand_omp_for): Use expand_omp_simd.
7843 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7844 (lower_omp_for): Do not lower the body.
7845 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7847 * tree-flow.h (find_omp_clause): Remove prototype.
7848 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7849 forcing vectorization of the loop, or if flag_tree_vectorize.
7850 (gate_tree_if_conversion): Similarly.
7851 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7852 gimple_build_omp_for.
7853 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7854 * tree-parloops (create_parallel_loop): Pass kind argument to
7855 gimple_build_omp_for.
7856 * tree-pretty-print.c (dump_omp_clause): Add cases for
7857 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7858 OMP_CLAUSE__SIMDUID_.
7859 (dump_generic_node): Handle OMP_SIMD.
7860 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7861 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7862 unroll OMP_SIMD loops here.
7863 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7864 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7866 (vect_analyze_data_refs): Handle simd loops.
7867 * tree-vect-loop.c (vectorizable_live_operation): Handle
7869 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7870 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7871 (vectorizable_load): Same.
7872 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7873 (struct simduid_to_vf): New.
7874 (simduid_to_vf::hash): New.
7875 (simduid_to-vf::equal): New.
7876 (struct simd_array_to_simduid): New.
7877 (simd_array_to_simduid::hash): New.
7878 (simd_array_to_simduid::equal): New.
7879 (adjust_simduid_builtins): New.
7880 (struct note_simd_array_uses_struct): New.
7881 (note_simd_array_uses_cb): New.
7882 (note_simd_array_uses): New.
7883 (vectorize_loops): Handle simd hints and adjust simd builtins
7885 * tree-vectorizer.h (struct _stmt_vec_info): Add
7886 simd_lane_access_p field.
7887 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7888 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7889 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7890 (omp_clause_code_name): Same.
7891 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7892 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7893 * tree.def (OMP_SIMD): New entry.
7894 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7895 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7896 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7897 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7898 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7899 (OMP_CLAUSE_LINEAR_STEP): New.
7900 (OMP_CLAUSE_SAFELEN_EXPR): New.
7901 (OMP_CLAUSE__SIMDUID__DECL): New.
7902 (find_omp_clause): New prototype.
7904 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7906 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7907 Haswell processor detection.
7909 2013-08-27 Christian Widmer <shadow@umbrox.de>
7912 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7913 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7914 AVX2 capable processors.
7916 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7918 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7919 of vget_low_* with implementations in terms of other intrinsics.
7921 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7924 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7925 values to -1, 0 and 1.
7927 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7929 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7930 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7931 (<optab><mode>3_insn): Remove.
7932 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7933 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7934 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7935 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7936 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7937 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7938 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7939 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7940 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7941 (ror<mode>3_insn): Likewise.
7942 * config/aarch64/predicates.md (aarch64_simd_register): New.
7944 2013-08-27 Richard Biener <rguenther@suse.de>
7946 PR tree-optimization/57521
7947 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7948 one edge is non-critical.
7949 (find_phi_replacement_condition): Make sure to use a non-critical
7950 edge. Cleanup and remove old bug workarounds.
7951 (bb_postdominates_preds): Remove.
7952 (if_convertible_loop_p_1): Do not compute post-dominators.
7953 (combine_blocks): Do not free post-dominators.
7954 (main_tree_if_conversion): Likewise.
7955 (pass_data_if_conversion): Add TODO_verify_ssa.
7957 2013-08-27 DJ Delorie <dj@redhat.com>
7959 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7961 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7963 * function.c (assign_parm_find_data_types): Set passed_mode and
7964 nominal_mode to the TYPE_MODE of nominal_type for the built
7965 pointer type in case of the struct-pass-by-reference.
7967 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7969 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7970 (UINT16_TYPE): Change default to "unsigned int".
7972 * config/avr/avr.opt (mfract-convert-truncate): New option.
7973 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7974 is set, round negative fractional integers according to n1169
7975 when converting to integer types.
7977 2013-08-26 Jan Hubicka <jh@suse.cz>
7979 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7980 methods can not be called indirectly when their address is not taken.
7982 2013-08-26 Jan Hubicka <jh@suse.cz>
7984 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7987 2013-08-26 Jan Hubicka <jh@suse.cz>
7989 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7992 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7994 * reload.c (find_valid_class): Allow classes that do not include
7995 FIRST_PSEUDO_REGISTER - 1.
7997 2013-08-26 Jan Hubicka <jh@suse.cz>
7999 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8000 fix edge count/frequency when speculation failed; fix type check
8001 for the direct call.
8003 2013-08-26 Jan Hubicka <jh@suse.cz>
8005 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8007 2013-08-26 Jan Hubicka <jh@suse.cz>
8009 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8010 to be changed by edge redirection.
8012 2013-08-26 Jan Hubicka <jh@suse.cz>
8014 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8015 formating; add sanity check.
8016 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8017 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8018 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8019 (resolve_noninline_speculation): Update callee keys, too.
8021 2013-08-26 Jan Hubicka <jh@suse.cz>
8023 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8024 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8026 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8028 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8031 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8033 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8034 (x86_64_elf_select_section): Ditto.
8035 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8037 (ix86_pass_by_reference): Ditto.
8038 (output_set_got): Ditto.
8039 (ix86_unary_operator_ok): Ditto.
8040 (ix86_expand_builtin): Ditto.
8042 2013-08-23 Jan Hubicka <jh@suse.cz>
8044 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8046 2013-08-23 Jan Hubicka <jh@suse.cz>
8048 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8049 (tree_decl_with_vis): Add FINAL field.
8051 2013-08-23 Jeff Law <law@redhat.com>
8053 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8054 the debugging dump when the expression is fully redundant.
8056 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8058 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8059 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8060 * pretty-print.c (pp_formatted_text_data): Likewise.
8061 (pp_write_text_to_stream): Likewise.
8062 (pp_write_text_as_dot_label_to_stream): Likewise.
8063 (pp_append_r): Likewise.
8064 (pp_format): Likewise.
8065 (pp_flush): Likewise.
8066 (pp_clear_output_area): Likewise.
8067 (pp_append_text): Likewise.
8068 (pp_formatted_text): Likewise.
8069 (pp_remaining_character_count_for_line): Likewise.
8070 (pp_newline): Likewise.
8071 (pp_character): Likewise.
8072 (output_buffer::~output_buffer): Define.
8073 (pretty_printer::~pretty_printer): Destruct output buffer.
8074 * pretty-print.h (output_buffer::~output_buffer): Declare.
8075 (pretty_printer::~pretty_printer): Declare virtual.
8077 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8080 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8081 HOST_WIDE_INT_M1U): New macros.
8082 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8083 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8084 unsigned -1 for lshift.
8085 * cse.c (cse_insn): Likewise.
8086 * double-int.c (rshift_double, lshift_double): Likewise.
8087 * builtins.c (fold_builtin_bitop): Likewise.
8088 * combine.c (force_to_mode): Likewise.
8089 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8090 * simplify-rtx.c (simplify_const_unary_operation,
8091 simplify_const_binary_operation): Likewise.
8092 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8093 check_va_list_escapes): Likewise.
8094 * rtlanal.c (nonzero_bits1): Likewise.
8095 * expmed.c (expand_smod_pow2): Likewise.
8096 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8098 2013-08-23 Jan Hubicka <jh@suse.cz>
8100 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8101 as having address taken.
8103 2013-08-23 Jan Hubicka <jh@suse.cz>
8105 * ipa-utils.h (method_class_type): Declare.
8106 * ipa-devirt.c (method_class_type): Export.
8108 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8109 do not walk base classes of anonymous types.
8111 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8113 PR rtl-optimization/58220
8115 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8116 handle SEQUENCE insns properly.
8118 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8120 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8122 (pp_newline_and_indent): Likewise.
8123 (pp_separate_with): Likewise.
8124 * pretty-print.c (pp_newline_and_flush): Define.
8125 (pp_newline_and_indent): Likewise.
8126 (pp_separate_with): Likewise.
8128 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8131 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8132 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8134 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8136 * config/i386/predicates.md (ext_sse_reg_operand): New.
8137 * config/i386/i386.md (*movti_internal): Use
8138 predicate to determine if EVEX is needed.
8139 (*movsi_internal): Ditto.
8140 (*movdf_internal): Ditto.
8141 (*movsf_internal): Ditto.
8142 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8144 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8146 PR tree-optimization/58209
8147 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8148 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8149 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8150 emit POINTER_PLUS_EXPR.
8151 (create_tailcall_accumulator): For pointer result type accumulate in
8154 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8156 * configure.ac: Add backslashes missing from the last change.
8157 * configure: Regenerate.
8159 2013-08-22 Jan Hubicka <jh@susue.cz>
8161 * ipa.c (function_and_variable_visibility): First remember function
8162 was global and then make it local.
8164 2013-08-22 Julian Brown <julian@codesourcery.com>
8166 * configure.ac: Add aarch64 to list of arches which use "nop" in
8168 * configure: Regenerate.
8170 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8172 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8173 gnu_libc_has_function.
8174 * config/s390/tpf.h: Likewise.
8176 2013-08-22 Jan Hubicka <jh@susue.cz>
8178 * timevar.c (validate_phases): Add cast.
8180 2013-08-22 Jan Hubicka <jh@susue.cz>
8182 * timevar.c (validate_phases): Use size_t for memory.
8183 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8185 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8187 * pretty-print.h (output_buffer::output_buffer): Declare.
8188 (pretty_printer::pretty_printer): Likewise.
8189 (pp_construct): Remove.
8190 * pretty-print.c (output_buffer::output_buffer): Define.
8191 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8192 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8193 (print_gimple_expr): Likewise.
8194 (print_gimple_seq): Likewise.
8195 (gimple_dump_bb): Likewise.
8196 * sched-vis.c (dump_value_slim): Likewise.
8197 (dump_insn_slim): Likewise.
8198 (dump_rtl_slim): Likewise.
8199 (str_pattern_slim): Likewise.
8200 * tree-mudflap.c (mf_varname_tree): Likewise.
8201 * graph.c (print_graph_cfg): Likewise.
8202 (start_graph_dump): Likewise.
8203 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8205 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8206 pretty printer initialization.
8207 * coretypes.h (diagnostic_context): Remove superflous type alias
8209 (pretty_printer): Likewise. Declare directly as a class.
8210 (pretty_print_info): Remove declaration as class.
8211 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8212 and pp_clear_output_area.
8213 (asan_add_global): Likewise.
8215 2013-08-22 Jan Hubicka <jh@suse.cz>
8217 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8218 * ipa-utils.h (update_type_inheritance_graph): Declare.
8219 (possible_polymorphic_call_target_p): Declare.
8220 (possible_polymorphic_call_target_p): New.
8221 * ipa-devirt.c: Update toplevel comments.
8222 (cached_polymorphic_call_targets): Move up.
8223 (odr_type_d): Move ID down.
8224 (polymorphic_type_binfo_p): Update comment.
8225 (odr_hasher::remove): Likewise;
8226 (get_odr_type): Set anonymous_namespace.
8227 (dump_odr_type): Dump it.
8228 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8229 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8230 (record_binfo): Add comment.
8231 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8233 (devirt_node_removal_hook): Do not iCE when cache is freed.
8234 (possible_polymorphic_call_target_p): New predicate.
8235 (update_type_inheritance_graph): New function.
8237 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8238 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8239 Sergey Lega <sergey.s.lega@intel.com>
8240 Anna Tikhonova <anna.tikhonova@intel.com>
8241 Ilya Tocar <ilya.tocar@intel.com>
8242 Andrey Turetskiy <andrey.turetskiy@intel.com>
8243 Ilya Verbin <ilya.verbin@intel.com>
8244 Kirill Yukhin <kirill.yukhin@intel.com>
8245 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8247 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8248 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8249 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8250 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8251 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8252 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8253 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8254 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8255 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8256 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8257 OPT_mavx512pf, OPT_mavx512er cases.
8258 * config/i386/constraints.md (v): New constraint.
8259 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8260 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8261 (bit_AVX512CD): New.
8262 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8263 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8264 * config/i386/i386-c.c (ix86_target_macros_internal):
8265 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8267 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8268 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8269 * config/i386/i386.c (regclass_map, dbx_register_map)
8270 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8271 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8272 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8274 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8275 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8276 -mavx512pf options. Fix formatting.
8277 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8278 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8279 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8280 -mavx512cd, -mavx512pf options.
8281 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8282 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8283 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8285 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8286 (ix86_set_reg_reg_cost): Ditto.
8287 (x86_order_regs_for_local_alloc): Ditto.
8288 (MAX_VECT_LEN): Extend to 64-byte.
8289 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8290 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8291 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8292 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8293 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8294 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8295 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8296 (SSE_REG_MODE_P): Support new modes.
8297 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8298 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8299 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8300 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8301 (REG_CLASS_CONTENTS): Add new registers.
8302 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8303 (EXT_REX_SSE_REGNO_P): New.
8304 (HI_REGISTER_NAMES): Add new registers.
8305 * config/i386/i386.md: Define constants for new registers.
8306 (mode): Add new 512-bit modes.
8307 (prefix): Support evex prefix.
8308 (isa): Support avx512f, noavx512f, fma_avx512f.
8309 (ssemodesuffix): Add new 512-bit modes.
8311 (*movxi_internal_avx512f): Ditto.
8312 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8314 (*movsi_internal): Likewise.
8315 (*movdf_internal): Likewise.
8316 (*movsf_internal): Likewise.
8317 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8318 (<code><mode>3): Likewise.
8319 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8321 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8322 with the new constraint "v".
8323 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8325 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8326 with the new constraint "v".
8327 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8328 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8329 (*<plusminus_insn><mode>3): Likewise.
8330 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8331 (*mul<mode>3): Likewise.
8332 (<sse>_vmmul<mode>3): Likewise.
8333 (<sse>_div<mode>3): Likewise.
8334 (<sse>_vmdiv<mode>3): Likewise.
8335 (<sse>_sqrt<mode>2): Likewise.
8336 (<sse>_vmsqrt<mode>2): Likewise.
8337 (*<code><mode>3_finite): Likewise.
8338 (*<code><mode>3) <smaxmin>: Likewise.
8339 (<sse>_vm<code><mode>3): Likewise.
8340 (*<code><mode>3) <any_logic>: Likewise.
8341 (*fma_fmadd_<mode>): Likewise.
8342 (*fma_fmsub_<mode>): Likewise.
8343 (*fma_fnmadd_<mode>): Likewise.
8344 (*fma_fnmsub_<mode>): Likewise.
8345 (*fma_fmaddsub_<mode>): Likewise.
8346 (*fma_fmsubadd_<mode>): Likewise.
8347 (*fmai_fmadd_<mode>): Likewise.
8348 (*fmai_fmsub_<mode>): Likewise.
8349 (*fmai_fnmadd_<mode>): Likewise.
8350 (*fmai_fnmsub_<mode>): Likewise.
8351 (sse_cvtsi2ss): Likewise.
8352 (sse_cvtsi2ssq): Likewise.
8353 (sse_cvtss2si): Likewise.
8354 (sse_cvtss2si_2): Likewise.
8355 (sse_cvtss2siq): Likewise.
8356 (sse_cvtss2siq_2): Likewise.
8357 (sse_cvttss2si): Likewise.
8358 (sse_cvtss2siq_2): Likewise.
8359 (float<sseintvecmodelower><mode>2): Likewise.
8360 (sse2_cvtsd2si_2): Likewise.
8361 (sse2_cvtsd2siq_2): Likewise.
8362 (*<plusminus_insn><mode>3): Likewise.
8363 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8364 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8365 (ashr<mode>3): Likewise.
8366 (<shift_insn><mode>3): Likewise.
8367 (avx2_<code><mode>3): Likewise.
8368 (*avx2_<code><mode>3): Likewise.
8369 (*andnot<mode>3): Likewise.
8370 (*<code><mode>3) <any_logic>: Likewise.
8371 (abs<mode>2): Likewise.
8372 (avx2_permvar<mode>): Likewise.
8373 (avx2_perm<mode>_1): Likewise.
8374 (*avx_vpermilp<mode>): Likewise.
8375 (avx_vpermilvar<mode>3): Likewise.
8376 (avx2_ashrv<mode>): Likewise.
8377 (avx2_<shift_insn>v<mode>): Likewise.
8378 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8380 * doc/rtl.texi: Document XImode.
8382 2013-08-21 Jeff Law <law@redhat.com>
8384 * tree-flow.h (register_jump_thread): Pass vector of edges
8385 instead of each important edge.
8386 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8387 thread path into a vector and pass that to register_jump_thread.
8388 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8389 passed in edge vector to the current 3-edge form.
8392 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8394 * dce.c (fini_dce): Call df_analyze again just in case
8395 delete_unmarked_insns removed anything.
8397 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8399 * reload.h (struct reg_equivs): Rename to ..
8400 (struct reg_equivs_s): .. this.
8402 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8404 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8406 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8408 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8410 2013-08-21 Jeff Law <law@redhat.com>
8412 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8413 simplify assignments too. If the RHS collapses to a singleton
8414 range, then return the value for the range.
8416 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8418 * config/i386/sse.md (V16): Rename to...
8420 (mov<mode>): Update iterator name.
8421 (*mov<mode>_internal): Ditto.
8422 (push<mode>1): Ditto.
8423 (movmisalign<mode>): Ditto.
8425 2013-08-20 Jan Hubicka <jh@suse.cz>
8428 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8429 entry for direct edges.
8430 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8432 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8434 Revert my last two changes, r201865 and r201864:
8437 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8439 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8440 instances can own GC refs.
8442 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8443 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8444 (gcc::context::gt_pch_nx): Likewise.
8445 (gcc::context::gt_pch_nx): Likewise.
8446 * ggc.h (gt_ggc_mx <T>): New.
8447 (gt_pch_nx_with_op <T>): New.
8448 (gt_pch_nx <T>): New.
8449 * passes.c (opt_pass::gt_ggc_mx): New.
8450 (opt_pass::gt_pch_nx): New.
8451 (opt_pass::gt_pch_nx_with_op): New.
8452 (pass_manager::gt_ggc_mx): New.
8453 (pass_manager::gt_pch_nx): New.
8454 (pass_manager::gt_pch_nx_with_op): New.
8455 (pass_manager::operator new): Use
8456 ggc_internal_cleared_alloc_stat rather than xcalloc.
8457 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8458 (pass_manager::gt_ggc_mx): New.
8459 (pass_manager::gt_pch_nx): New.
8460 (pass_manager::gt_pch_nx_with_op): New.
8461 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8462 (opt_pass::operator new): New.
8463 (opt_pass::gt_ggc_mx): New.
8464 (opt_pass::gt_pch_nx): New.
8465 (opt_pass::gt_pch_nx_with_op): New.
8468 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8470 * Makefile.in (GTFILES): Add context.h.
8471 * context.c (gcc::context::operator new): New.
8472 (gcc::context::gt_ggc_mx): New.
8473 (gcc::context::gt_pch_nx): New.
8474 (gcc::context::gt_pch_nx): New.
8475 * context.h (gcc::context): Add GTY((user)) marking.
8476 (gcc::context::operator new): New.
8477 (gcc::context::gt_ggc_mx): New.
8478 (gcc::context::gt_pch_nx): New.
8479 (gcc::context::gt_pch_nx): New.
8480 (g): Add GTY marking.
8481 (gt_ggc_mx (gcc::context *)): New.
8482 (gt_pch_nx (gcc::context *)): New.
8483 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8484 void *cookie)): New.
8485 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8487 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8489 * dce.c (fini_dce): Call df_analyze again just in case
8490 delete_unmarked_insns removed anything.
8492 2013-08-20 Teresa Johnson <tejohnson@google.com>
8494 PR rtl-optimizations/57451
8495 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8496 from crossing split section boundaries.
8498 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8500 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8501 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8502 with MULTLIB_DEFAULTS.
8504 2013-08-20 Nick Clifton <nickc@redhat.com>
8506 * target.def (narrow_volatile_bitfield): Note that the default
8507 value is false, not !TARGET_STRICT_ALIGN.
8508 * doc/tm.texi: Regenerate.
8510 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8512 Fix LIB_SPEC for systems without libpthread.
8514 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8515 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8517 * config/i386/linux-common.h: Likewise.
8518 * config/mips/linux-common.h: Likewise.
8520 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8522 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8525 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8527 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8528 instances can own GC refs.
8530 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8531 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8532 (gcc::context::gt_pch_nx): Likewise.
8533 (gcc::context::gt_pch_nx): Likewise.
8534 * ggc.h (gt_ggc_mx <T>): New.
8535 (gt_pch_nx_with_op <T>): New.
8536 (gt_pch_nx <T>): New.
8537 * passes.c (opt_pass::gt_ggc_mx): New.
8538 (opt_pass::gt_pch_nx): New.
8539 (opt_pass::gt_pch_nx_with_op): New.
8540 (pass_manager::gt_ggc_mx): New.
8541 (pass_manager::gt_pch_nx): New.
8542 (pass_manager::gt_pch_nx_with_op): New.
8543 (pass_manager::operator new): Use
8544 ggc_internal_cleared_alloc_stat rather than xcalloc.
8545 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8546 (pass_manager::gt_ggc_mx): New.
8547 (pass_manager::gt_pch_nx): New.
8548 (pass_manager::gt_pch_nx_with_op): New.
8549 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8550 (opt_pass::operator new): New.
8551 (opt_pass::gt_ggc_mx): New.
8552 (opt_pass::gt_pch_nx): New.
8553 (opt_pass::gt_pch_nx_with_op): New.
8555 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8557 * Makefile.in (GTFILES): Add context.h.
8558 * context.c (gcc::context::operator new): New.
8559 (gcc::context::gt_ggc_mx): New.
8560 (gcc::context::gt_pch_nx): New.
8561 (gcc::context::gt_pch_nx): New.
8562 * context.h (gcc::context): Add GTY((user)) marking.
8563 (gcc::context::operator new): New.
8564 (gcc::context::gt_ggc_mx): New.
8565 (gcc::context::gt_pch_nx): New.
8566 (gcc::context::gt_pch_nx): New.
8567 (g): Add GTY marking.
8568 (gt_ggc_mx (gcc::context *)): New.
8569 (gt_pch_nx (gcc::context *)): New.
8570 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8571 void *cookie)): New.
8572 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8574 2013-08-20 Alan Modra <amodra@gmail.com>
8577 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8578 (rs6000_emit_epilogue): Likewise.
8580 2013-08-19 Dehao Chen <dehao@google.com>
8582 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8584 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8585 Jakub Jelinek <jakub@redhat.com>
8587 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8588 (BUILT_IN_FABSD64): Likewise.
8589 (BUILT_IN_FABSD128): Likewise.
8590 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8591 (fold_builtin_1): Likewise.
8592 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8593 destination and source operands.
8594 (*abstd2_fpr): Likewise.
8595 (*nabstd2_fpr): Likewise.
8597 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8599 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8602 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8604 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8607 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8609 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8610 * builtins.c (default_libc_has_function): New.
8611 (gnu_libc_has_function): Ditto.
8612 (no_c99_libc_has_function): Ditto.
8613 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8614 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8615 (fold_builtin_sincos): Likewise.
8616 (fold_builtin_cexp): Likewise.
8617 * builtins.def (DEF_C94_BUILTIN): Likewise.
8618 (DEF_C99_BUILTIN): Likewise.
8619 (DEF_C99_C90RES_BUILTIN): Likewise.
8620 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8621 definitions to using this define.
8622 * config/darwin-protos.h (darwin_libc_has_function): New.
8623 * config/darwin.c (darwin_libc_has_function): Ditto.
8624 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8625 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8626 * config/darwin.h: Ditto.
8627 * config/elfos.h: Ditto.
8628 * config/freebsd.h: Ditto.
8629 * config/i386/cygming.h: Ditto.
8630 * config/i386/djgpp.h: Ditto.
8631 * config/i386/i386-interix.h: Ditto.
8632 * config/microblaze/microblaze.h: Ditto.
8633 * config/mmix/mmix.h: Ditto.
8634 * config/gnu-user.h: Ditto.
8635 * config/ia64/hpux.h: Ditto.
8636 * config/pa/pa-hpux.h: Ditto.
8637 * config/pdp11/pdp11.h: Ditto.
8638 * config/picochip/picochip.h: Ditto.
8639 * config/linux.h: Ditto.
8640 * config/netbsd.h: Ditto.
8641 * config/openbsd.h: Ditto.
8642 * config/rs6000/aix43.h: Ditto.
8643 * config/rs6000/aix51.h: Ditto.
8644 * config/rs6000/aix52.h: Ditto.
8645 * config/rs6000/aix53.h: Ditto.
8646 * config/rs6000/aix61.h: Ditto.
8647 * config/rs6000/darwin.h: Ditto.
8648 * config/rs6000/linux.h: Ditto.
8649 * config/rs6000/linux64.h: Ditto.
8650 * config/s390/tpf.h: Ditto.
8651 * config/sol2-10.h: Ditto.
8652 * config/sol2.h: Ditto.
8653 * config/vms/vms.h: Ditto.
8654 * config/vxworks.h: Ditto.
8655 * config/linux-android.c (linux_android_libc_has_function):
8656 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8657 * config/linux-protos.h (linux_android_libc_has_function):
8659 * config/i386/i386.c (ix86_libc_has_function): New.
8660 * config/i386/i386-protos.h
8661 (ix86_libc_has_function): New declaration.
8662 * config/i386/i386.md
8663 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8664 ("isinf<mode>2): Likewise.
8665 * convert.c (convert_to_integer): Using new target hook
8666 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8667 TARGET_C99_FUNCTIONS.
8668 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8669 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8670 * coretypes.h (function_class): New enum for different
8671 classes of functions.
8672 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8673 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8674 (TARGET_HAS_SINCOS): Likewise.
8675 (TARGET_LIBC_HAS_FUNCTION): New.
8676 * doc/tm.texi: Regenerated.
8677 * targhooks.h (default_libc_has_function): New declaration.
8678 (no_c99_libc_has_function): Ditto.
8679 (gnu_libc_has_function): Ditto.
8680 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8681 and TARGET_HAS_SINCOS.
8683 2013-08-18 Jan Hubicka <jh@suse.cz>
8685 * Makeifle-in (ipa-devirt.o): New.
8686 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8687 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8688 (analyze_functions): Look into possible targets of polymorphic call.
8689 * dumpfile.c (dump_files): Add type-inheritance dump.
8690 * dumpfile.h (TDI_inheritance): New.
8691 * ipa-devirt.c: New file.
8692 * ipa-utils.h (odr_type_d): Forward declare.
8693 (odr_type): New type.
8694 (build_type_inheritance_graph): Declare.
8695 (possible_polymorphic_call_targets): Declare and introduce inline
8696 variant when only edge is pased.
8697 (dump_possible_polymorphic_call_targets): Likewise.
8698 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8699 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8700 (types_same_for_odr): ... here.
8701 * tree.h (type_in_anonymous_namespace_p): Declare.
8703 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8705 PR tree-optimization/58006
8706 * tree-parloops.c (take_address_of): Don't ICE if get_name
8708 (eliminate_local_variables_stmt): Remove clobber stmts.
8710 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8712 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8713 error message is issued for an alias to undefined symbol.
8715 2013-08-18 Jan Hubicka <jh@suse.cz>
8717 * cgraph.c (cgraph_create_indirect_edge): Discover
8718 polymorphic calls and record basic info into indirect_info.
8719 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8720 devirtualization, ignore objc function calls.
8721 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8722 call with no parm index info.
8723 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8724 * tree.c (virtual_method_call_p): New function.
8725 * tree.h (virtual_method_call_p): Declare.
8727 2013-08-16 Jan Hubicka <jh@suse.cz>
8730 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8732 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8734 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8737 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8739 * gengtype.c (type_for_name): Add special-case support for
8740 locating types within the "gcc::" namespace.
8741 (open_base_files): Emit a "using namespace gcc" directive.
8743 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8746 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8747 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8749 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8750 array instead of each individual operand as a separate argument.
8751 (emit_fusion_gpr_load): Likewise.
8752 (expand_fusion_gpr_load): Add new function declaration.
8754 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8755 signature to have the operands passed as an array, instead of as
8756 separate arguments. Allow ZERO_EXTEND to be in the memory
8757 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8758 depend on the register live/dead flags when peepholes are run.
8759 (expand_fusion_gpr_load): New function to be called from the
8760 peephole2 pass, to change the register that addis sets to be the
8762 (emit_fusion_gpr_load): Change the calling signature to have the
8763 operands passed as an array, instead of as separate arguments.
8764 Allow ZERO_EXTEND to be in the memory address, and also
8765 SIGN_EXTEND if -mpower8-fusion-sign.
8767 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8769 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8770 adjust the register addis loads up in the peephole2 pass. Do not
8771 depend on the register live/dead state when the peephole pass is done.
8773 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8775 * gengtype.c (create_user_defined_type): Ensure that the kind
8776 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8777 declaration is seen before the GTY((user)) marking.
8779 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8782 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8784 2013-08-16 Jan Hubicka <jh@suse.cz>
8786 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8787 arugment expected_type.
8788 (gimple_fold_call): Use it.
8789 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8790 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8791 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8792 (try_make_edge_direct_virtual_call): Likewise.
8793 * tree.c (obj_type_ref_class): New.
8794 * tree.h (obj_type_ref_class): Use it.
8796 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8798 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8799 (rtl_slim_pp): Likewise.
8800 (init_rtl_slim_pretty_print): Likewise.
8801 (dump_value_slim): Don't call it. Use local pretty printer.
8802 (dump_insn_slim): Likewise.
8803 (dump_rtl_slim): Likewise.
8804 (str_pattern_slim): Likewise.
8805 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8808 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8810 PR tree-optimization/58164
8811 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8812 walk gimple_goto_dest of GIMPLE_GOTO.
8814 PR tree-optimization/58165
8815 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8816 bi_call must be the last stmt in a bb, don't split_block, instead
8817 use fallthru edge from it and give up if there is none.
8818 Release conds vector when returning early.
8820 2013-08-14 Xinliang David Li <davidxl@google.com>
8822 * config/i386/i386.c (ix86_option_override_internal):
8823 Remove unused variable and field.
8825 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8828 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8829 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8830 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8831 and Linux, correct BLKmode alignment when 128-bit alignment is
8832 required and compatibility flag is not set.
8833 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8834 for zero-size arguments when compatibility flag is not set.
8836 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8838 PR tree-optimization/58145
8839 * tree-sra.c (build_ref_for_offset): If prev_base has
8840 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8842 2013-08-14 Xinliang David Li <davidxl@google.com>
8844 * config/i386/i386.c (ix86_option_override_internal):
8845 Fix uninitialized variable error.
8847 2013-08-14 Xinliang David Li <davidxl@google.com>
8849 * config/i386/i386.opt: Define two new options.
8850 * config/i386/x86-tune.def: Add arch selector field in macros.
8851 * config/i386/i386.h: Adjust macro definition.
8852 * config/i386/i386.c (ix86_option_override_internal):
8854 (parse_mtune_ctrl_str): New function.
8855 (set_ix86_tune_features): New function.
8856 (ix86_function_specific_restore): Call the new helper function.
8858 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8860 PR rtl-optimization/57662
8861 * sel-sched.c (code_motion_process_successors): When the current insn
8862 is removed after the recursive traversal, break from the loop.
8863 Add comments and debug printouts.
8865 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8866 Alexandre Oliva <aoliva@redhat.com>
8869 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8870 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8871 there also UNSPEC_PLTOFF.
8873 2013-08-14 Marek Polacek <polacek@redhat.com>
8875 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8876 NUM_CONDITIONS bit positions.
8878 2013-08-13 Cary Coutant <ccoutant@google.com>
8880 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8881 (attr_checksum): Hash vector contents instead of pointer.
8882 (attr_checksum_ordered): Likewise.
8884 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8886 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8887 when Pmode != word_mode. Add length_address attribute.
8888 (sse3_monitor_<mode>): Merge from sse3_monitor and
8889 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8890 Pmode != word_mode. Update insn length attribute.
8891 * config/i386/i386.c (ix86_option_override_internal): Update
8892 ix86_gen_monitor selection for merged sse3_monitor insn.
8894 2013-08-13 Julian Brown <julian@codesourcery.com>
8896 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8897 perform invalid legitimization on greater-than-word-size modes for
8900 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8902 * ira.c (setup_class_translate_array): Use aclass instead of cl
8903 for classes not fully covered by allocno classes.
8905 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8907 PR tree-optimization/57661
8908 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8909 * tree-inline.c (tree_function_versioning): Initialize it.
8910 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8911 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8912 that is not being copied.
8915 * asan.c (instrument_strlen_call): Fix typo in comment.
8916 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8918 2013-08-13 Steve Ellcey <sellcey@mips.com>
8920 * config/mips/mips.md (prefetch): Use lw instead of ld on
8921 loongson in 32bit mode.
8923 2013-08-13 Nick Clifton <nickc@redhat.com>
8925 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8927 2013-08-13 Jan Hubicka <jh@suse.cz>
8929 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8930 introduced edge; fix typo in sanity check.
8931 (cgraph_resolve_speculation): Export; improve diagnostic.
8932 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8933 speculation at type mismatch.
8934 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8935 (cgraph_resolve_speculation): Declare.
8936 (symtab_can_be_discarded): New function.
8937 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8938 * ipa-inline-transform.c (speculation_removed): New global var.
8939 (clone_inlined_nodes): See if speculation can be removed.
8940 (inline_call): If speculations was removed, we growths may not match.
8941 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8942 (speculation_useful_p): New function.
8943 (resolve_noninline_speculation): New function.
8944 (inline_small_functions): Resolve useless speculations.
8945 * ipa-inline.h (speculation_useful_p): Declare
8946 * ipa.c (can_replace_by_local_alias): Simplify.
8947 (ipa_profile): Produce speculative calls in non-lto, too;
8948 add simple cost model; produce local aliases.
8950 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8952 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8955 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8957 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8959 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8961 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8962 iterator instead of X87MODEF.
8964 2013-08-12 Perez Read <netfirewall@gmail.com>
8967 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8968 operand 0 for intel asm alternative.
8969 (*movabs<mode>_2): Ditto for operand 1.
8971 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8973 * config/aarch64/arm_none.h
8974 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8976 2013-08-12 Nick Clifton <nickc@redhat.com>
8978 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8980 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8982 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8983 expand for QI/HImode operand to produce more effictive code for
8984 unsigned char(short) --> float(double) conversion.
8986 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8988 * doc/invoke.texi: Mention that -ftls-model does not force the final
8991 2013-08-12 Marek Polacek <polacek@redhat.com>
8992 Marc Glisse <marc.glisse@inria.fr>
8994 PR tree-optimization/57980
8995 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8996 when creating -1 constant.
8998 2013-08-10 Jan Hubicka <jh@suse.cz>
9000 Workaround binutils PR14342.
9001 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9002 (gimple_init_edge_profiler): Likewise.
9003 (gimple_gen_ic_func_profiler): Likewise.
9005 2013-08-09 Jan Hubicka <jh@suse.cz>
9007 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9009 2013-08-09 Xinliang David Li <davidxl@google.com>
9011 * config/i386/stringop.def: New file.
9012 * config/i386/stringop.opt: New file.
9013 * config/i386/i386-opts.h: Include stringopt.def.
9014 * config/i386/i386.opt: Include stringopt.opt.
9015 * config/i386/i386.c (ix86_option_override_internal):
9016 Override default size based stringop inline strategies with options.
9017 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9020 2013-08-09 Jan Hubicka <jh@suse.cz>
9022 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9024 2013-08-09 Jan Hubicka <jh@suse.cz>
9026 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9028 (dump_cgraph_node): Dump profile-id.
9029 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9030 and common_target_probability.
9031 * lto-cgraph.c (lto_output_edge): Stream common targets.
9032 (lto_output_node): Stream profile ids.
9033 (input_node): Stream profile ids.
9034 (input_edge): Stream common targets.
9035 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9036 * ipa.c: Include value-prof.h
9037 (ipa_profile_generate_summary): Turn indirect call statement histograms
9038 into common targets.
9039 (ipa_profile): Turn common targets into speculative edges.
9041 2013-08-09 Jan Hubicka <jh@suse.cz>
9043 * cgraph.h (cgraph_node): Add profile_id.
9044 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9045 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9046 (del_node_map): Update.
9047 (find_func_by_funcdef_no): Replace by ...
9048 (find_func_by_profile_id): ... this one.
9049 (gimple_ic_transform): Do not remove useful histograms when
9050 speculation is not done; dump info when indirect call removal
9052 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9053 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9054 (__gcov_indirect_call_profiler_v2): .. this one.
9055 * profile.h (init_node_map): Update.
9056 * coverage.c (coverage_compute_profile_id): New function.
9057 * coverage.h (coverage_compute_profile_id): Declare.
9058 * tree-profile.c (init_ic_make_global_vars): Make
9059 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9060 (gimple_init_edge_profiler): Update prototype of
9061 __gcov_indirect_call_profiler.
9062 (gimple_gen_ic_func_profiler): Simplify.
9063 (tree_profiling): Use init_node_map
9065 2013-08-09 Jan Hubicka <jh@suse.cz>
9067 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9068 non-speculative refs.
9069 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9070 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9071 (cgraph_set_call_stmt): Likewise.
9072 (cgraph_create_edge_1): Fix release checking compilatoin;
9074 (cgraph_free_edge): Free indirect info.
9075 (cgraph_turn_edge_to_speculative): New function.
9076 (cgraph_speculative_call_info): New function.
9077 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9078 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9079 (dump_cgraph_node): Dump speculation.
9080 (verify_edge_count_and_frequency): Accept speculative edges.
9081 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9082 (verify_cgraph_node): Handle speculation.
9083 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9084 (cgraph_set_call_stmt): Update prototype.
9085 (cgraph_make_edge_direct): Update prototype.
9086 (cgraph_speculative_call_info): Declare.
9087 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9088 to change; update call of ipa_find_references.
9089 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9090 lto_stmt_uid and speculative flags.
9091 (ipa_dump_references): Dump speculation.
9092 (ipa_clone_references): Clone speculative flag.
9093 (ipa_clone_referring): Likewise.
9094 (ipa_clone_ref): New function.
9095 (ipa_find_reference): Look into lto_stmt_uids
9096 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9097 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9098 (ipa_find_reference): Update declaration.
9099 (ipa_clone_ref): Declare.
9100 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9101 stream speculative flag.
9102 (lto_output_ref): Stream statements uids and speculation.
9103 (input_ref): Likewise.
9104 (input_edge): Stream speuclation.
9105 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9106 (cgraph_set_call_stmt_including_clones): Handle speculation.
9107 * ipa-inline.c (heap_edge_removal_hook): New function.
9108 (inline_small_functions): Register it.
9109 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9110 also initialize refs.
9111 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9113 (try_make_edge_direct_simple_call): Likewise.
9114 (try_make_edge_direct_simple_call): Likewise.
9115 (update_indirect_edges_after_inlining): Likewise.
9116 (remove_described_reference): Look proper lto_stmt_uid.
9117 (propagate_controlled_uses): Likewise.
9118 (propagate_controlled_uses): Liekwise.
9119 * tree-inline.c (copy_bb): Copy speculative edges.
9120 (redirect_all_calls): New function.
9121 (copy_cfg_body): Do redirection after loop info is updated.
9122 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9124 2013-08-09 Jan Hubicka <jh@suse.cz>
9126 * lto-streamer-out.c (output_function): Renumber PHIs.
9127 * lto-streamer-in.c (input_function): Likewise.
9129 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9131 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9132 (get_lane_unsigned): Likewise.
9133 (dup_lane_scalar): Likewise.
9134 (get_lane): enable for VALL.
9135 * config/aarch64/aarch64-simd.md
9136 (aarch64_dup_lane_scalar<mode>): Remove.
9137 (aarch64_get_lane_signed<mode>): Likewise.
9138 (aarch64_get_lane_unsigned<mode>): Likewise.
9139 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9140 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9141 (aarch64_get_lane<mode>): Enable for all vector modes.
9142 (aarch64_get_lanedi): Remove misleading constraints.
9143 * config/aarch64/arm_neon.h
9144 (__aarch64_vget_lane_any): Define.
9145 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9146 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9147 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9148 * config/aarch64/iterators.md (VDQQH): New.
9152 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9154 * configure.ac: Add GAS check for LEON instructions on SPARC.
9155 * configure: Regenerate.
9156 * config.in: Likewise.
9157 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9159 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9160 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9162 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9163 (AS_LEON_FLAG): New macro.
9164 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9165 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9166 Deal with LEON and LEON3 for the memory model.
9167 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9168 (atomic_compare_and_swap<mode>_1): Likewise.
9169 (*atomic_compare_and_swap<mode>_1): Likewise.
9171 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9173 * config/arm/neon.md (vcond): Fix floating-point vector
9174 comparisons against 0.
9176 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9178 * lra-constraints.c (emit_spill_move): Remove assert.
9179 (process_alt_operands): Add more debugging
9180 output. Increase reject for spilling into memory. Decrease
9181 reject for reloading scratch.
9182 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9184 2013-08-08 Steve Ellcey <sellcey@mips.com>
9186 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9187 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9188 micromips incompatible. Add nan2008.
9189 (MULTILIB_DIRNAMES): Add nan2008.
9190 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9191 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9192 and micromips incompatible. Add nan2008.
9193 (MULTILIB_DIRNAMES): Add nan2008.
9194 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9196 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9198 PR rtl-optimization/58079
9199 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9200 simplify_comparison has widened a comparison with an integer.
9202 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9204 * config/arm/neon.md (movmisalign<mode>): Disable when we
9205 don't allow unaligned accesses.
9206 (*movmisalign<mode>_neon_store): Likewise.
9207 (*movmisalign<mode>_neon_load): Likewise.
9208 (*movmisalign<mode>_neon_store): Likewise.
9209 (*movmisalign<mode>_neon_load): Likewise.
9211 2013-08-08 Jan Hubicka <jh@suse.cz>
9213 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9214 (make_pass_rebuild_cgraph_edges): Also clear references.
9215 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9216 * ipa-inline-transform.c (inline_transform): Remove all references
9218 * cgraphunit.c (expand_function): Remove all references after
9220 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9221 (ipa_find_reference): Rewrite to iterator.
9222 (remove_stmt_references): Likewise.
9223 (ipa_clear_stmts_in_references): New function.
9224 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9225 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9227 * ipa-split.c (split_function): Remove references in split function.
9229 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9232 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9233 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9235 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9238 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9239 suggested mode for the assignment isn't compatible with the
9242 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9245 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9247 2013-08-07 Xinliang David Li <davidxl@google.com>
9249 * config/i386/i386.opt: New option -mtune-ctrl=.
9250 * config/i386/x86-tune.def: New file.
9251 * config/i386/i386.h: include x86-tune.def.
9252 * config/i386/i386.c (ix86_option_override_internal):
9253 Parsing -mtune-ctrl= option and set tune features.
9255 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9258 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9259 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9262 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9264 * rtl.h (update_alignments): Declare.
9265 * final.c (grow_label_align): New function extracted from...
9266 (shorten_branches): ...here. Call it.
9267 (update_alignments): New function.
9268 * reorg.c (sibling_labels): New variable.
9269 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9270 the new label along with it onto the sibling_labels vector.
9271 (fill_simple_delay_slots): Adjust call to get_label_before.
9272 (fill_slots_from_thread): Likewise.
9273 (relax_delay_slots): Likewise.
9274 (make_return_insns): Likewise.
9275 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9277 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9279 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9280 document its semantics.
9281 (diagnostic_report_diagnostic): Adjust accordingly.
9283 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9285 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9286 (sparc_option_override): ...and port to new C++ pass API.
9287 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9289 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9291 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9293 2013-08-06 Caroline Tice <cmtice@google.com>
9295 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9296 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9297 * tree-pass.h: Add pass_vtable_verify.
9298 * varasm.c (assemble_variable): Add code to properly set the comdat
9299 section and name for the .vtable_map_vars section.
9300 (assemble_vtyv_preinit_initializer): New function.
9301 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9303 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9304 * vtable-verify.c: New file.
9305 * vtable-verify.h: New file.
9306 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9307 initialiation levels.
9308 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9309 * passes.def: Insert pass_vtable_verify.
9310 * aclocal.m4: Reorder includes.
9311 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9312 -fvtv-counts options.
9313 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9314 as appropriate, if -fvtable-verify=... is used.
9315 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9316 -fvtable-verify=... is used.
9317 * Makefile.in (OBJS): Add vtable-verify.o to list.
9318 (vtable-verify.o): Add new build rule.
9319 (GTFILES): Add vtable-verify.c to list.
9320 * common.opt (fvtable-verify=): New flag.
9321 (vtv_priority): Values for fvtable-verify= flag.
9322 (fvtv-counts): New flag.
9323 (fvtv-debug): New flag.
9324 * tree.h (save_vtable_map_decl): New extern function decl.
9326 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9328 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9329 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9330 (pass_data_rl78_devirt): ...new pass_data instance and...
9331 (make_pass_rl78_devirt): ...new function.
9332 (rl78_asm_file_start): Port pass registration to new C++ API.
9334 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9336 * coretypes.h (rtl_opt_pass): Add.
9337 (gcc::context): Add.
9338 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9339 (epiphany_init): Port to new C++ pass API.
9340 (epiphany_optimize_mode_switching): Likewise.
9341 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9342 (pass_manager::get_pass_mode_switching): New.
9343 (pass_manager::get_pass_peephole2): New.
9344 * mode-switching.c (pass_mode_switching): Add clone method.
9345 * recog.c (pass_peephole2): Add clone method.
9346 (pass_split_all_insns): Add clone method.
9348 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9350 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9351 (mips_option_override): ...here, porting to new C++ API for passes.
9353 2013-08-06 Jan Hubicka <jh@suse.cz>
9355 * cgraph.c (cgraph_get_body): New function based on lto.c
9357 * cgraph.h (cgraph_get_body): Declare.
9358 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9360 * cgraphunit.c (expand_function): Get body prior expanding.
9361 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9362 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9364 * passes.c (do_per_function_toporder): Get body.
9365 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9366 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9368 2013-08-06 Martin Jambor <mjambor@suse.cz>
9371 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9372 is not re-finalized. Rename second parameter to no_collect.
9374 2013-08-06 Martin Jambor <mjambor@suse.cz>
9377 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9378 MEM_REF has proper alignment information.
9380 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9383 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9385 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9386 rtx (*) (rtx, ...) with insn_gen_fn.
9387 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9388 insn_gen_fn::stored_funcptr. Add initializer braces.
9390 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9392 Rewrite how instances of passes are cloned to remove assumptions
9393 about their sizes (thus allowing pass subclasses to have
9394 additional data fields, albeit non-GC-managed ones at this point).
9396 * passes.c (make_pass_instance): Now that passes have clone
9397 methods, rewrite this function to eliminate XNEW and memcpy
9398 calls that used hardcoded sizes. Since this function no longer
9399 creates pass instances, rename it to...
9400 (add_pass_instance): ...this. Document the old way that passes were
9401 numbered and flagged, and rework this function to continue using it.
9402 (next_pass_1): Add an initial_pass argument for use by
9404 (position_pass): When adding multiple instances of a pass, use
9405 the pass's clone method, rather than relying on the XNEW/memcpy
9406 within the former make_pass_instance (now add_pass_instance).
9407 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9408 the initial instance of the current pass within the pass manager.
9410 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9412 This is the automated part of the conversion of passes from C
9413 structs to C++ classes.
9415 Patch autogenerated by refactor_passes.py from
9416 https://github.com/davidmalcolm/gcc-refactoring-scripts
9417 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9419 * asan.c (pass_asan): Convert from a global struct to a subclass of
9420 gimple_opt_pass along with...
9421 (pass_data_asan): ...new pass_data instance and...
9422 (make_pass_asan): ...new function.
9423 (pass_asan_O0): Convert from a global struct to a subclass of
9424 gimple_opt_pass along with...
9425 (pass_data_asan_O0): ...new pass_data instance and...
9426 (make_pass_asan_O0): ...new function.
9427 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9428 subclass of rtl_opt_pass along with...
9429 (pass_data_inc_dec): ...new pass_data instance and...
9430 (make_pass_inc_dec): ...new function.
9431 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9432 a subclass of rtl_opt_pass along with...
9433 (pass_data_reorder_blocks): ...new pass_data instance and...
9434 (make_pass_reorder_blocks): ...new function.
9435 (pass_duplicate_computed_gotos): Convert from a global struct to a
9436 subclass of rtl_opt_pass along with...
9437 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9438 (make_pass_duplicate_computed_gotos): ...new function.
9439 (pass_partition_blocks): Convert from a global struct to a subclass of
9440 rtl_opt_pass along with...
9441 (pass_data_partition_blocks): ...new pass_data instance and...
9442 (make_pass_partition_blocks): ...new function.
9443 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9444 struct to a subclass of rtl_opt_pass along with...
9445 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9447 (make_pass_branch_target_load_optimize1): ...new function.
9448 (pass_branch_target_load_optimize2): Convert from a global struct to a
9449 subclass of rtl_opt_pass along with...
9450 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9452 (make_pass_branch_target_load_optimize2): ...new function.
9453 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9454 of rtl_opt_pass along with...
9455 (pass_data_jump): ...new pass_data instance and...
9456 (make_pass_jump): ...new function.
9457 (pass_jump2): Convert from a global struct to a subclass of
9458 rtl_opt_pass along with...
9459 (pass_data_jump2): ...new pass_data instance and...
9460 (make_pass_jump2): ...new function.
9461 * cfgexpand.c (pass_expand): Convert from a global struct to a
9462 subclass of rtl_opt_pass along with...
9463 (pass_data_expand): ...new pass_data instance and...
9464 (make_pass_expand): ...new function.
9465 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9466 of rtl_opt_pass along with...
9467 (pass_data_free_cfg): ...new pass_data instance and...
9468 (make_pass_free_cfg): ...new function.
9469 (pass_into_cfg_layout_mode): Convert from a global struct to a
9470 subclass of rtl_opt_pass along with...
9471 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9472 (make_pass_into_cfg_layout_mode): ...new function.
9473 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9474 subclass of rtl_opt_pass along with...
9475 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9476 (make_pass_outof_cfg_layout_mode): ...new function.
9477 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9478 struct to a subclass of gimple_opt_pass along with...
9479 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9480 (make_pass_build_cgraph_edges): ...new function.
9481 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9482 subclass of gimple_opt_pass along with...
9483 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9484 (make_pass_rebuild_cgraph_edges): ...new function.
9485 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9486 subclass of gimple_opt_pass along with...
9487 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9489 (make_pass_remove_cgraph_callee_edges): ...new function.
9490 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9491 struct to a subclass of rtl_opt_pass along with...
9492 (pass_data_stack_adjustments): ...new pass_data instance and...
9493 (make_pass_stack_adjustments): ...new function.
9494 * combine.c (pass_combine): Convert from a global struct to a subclass
9495 of rtl_opt_pass along with...
9496 (pass_data_combine): ...new pass_data instance and...
9497 (make_pass_combine): ...new function.
9498 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9499 global struct to a subclass of rtl_opt_pass along with...
9500 (pass_data_compare_elim_after_reload): ...new pass_data instance
9502 (make_pass_compare_elim_after_reload): ...new function.
9503 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9504 of rtl_opt_pass along with...
9505 (pass_data_rtl_cprop): ...new pass_data instance and...
9506 (make_pass_rtl_cprop): ...new function.
9507 * cse.c (pass_cse): Convert from a global struct to a subclass of
9508 rtl_opt_pass along with...
9509 (pass_data_cse): ...new pass_data instance and...
9510 (make_pass_cse): ...new function.
9511 (pass_cse2): Convert from a global struct to a subclass of
9512 rtl_opt_pass along with...
9513 (pass_data_cse2): ...new pass_data instance and...
9514 (make_pass_cse2): ...new function.
9515 (pass_cse_after_global_opts): Convert from a global struct to a
9516 subclass of rtl_opt_pass along with...
9517 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9518 (make_pass_cse_after_global_opts): ...new function.
9519 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9520 of rtl_opt_pass along with...
9521 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9522 (make_pass_ud_rtl_dce): ...new function.
9523 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9524 rtl_opt_pass along with...
9525 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9526 (make_pass_fast_rtl_dce): ...new function.
9527 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9528 a subclass of rtl_opt_pass along with...
9529 (pass_data_df_initialize_opt): ...new pass_data instance and...
9530 (make_pass_df_initialize_opt): ...new function.
9531 (pass_df_initialize_no_opt): Convert from a global struct to a
9532 subclass of rtl_opt_pass along with...
9533 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9534 (make_pass_df_initialize_no_opt): ...new function.
9535 (pass_df_finish): Convert from a global struct to a subclass of
9536 rtl_opt_pass along with...
9537 (pass_data_df_finish): ...new pass_data instance and...
9538 (make_pass_df_finish): ...new function.
9539 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9540 rtl_opt_pass along with...
9541 (pass_data_rtl_dse1): ...new pass_data instance and...
9542 (make_pass_rtl_dse1): ...new function.
9543 (pass_rtl_dse2): Convert from a global struct to a subclass of
9544 rtl_opt_pass along with...
9545 (pass_data_rtl_dse2): ...new pass_data instance and...
9546 (make_pass_rtl_dse2): ...new function.
9547 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9548 subclass of rtl_opt_pass along with...
9549 (pass_data_dwarf2_frame): ...new pass_data instance and...
9550 (make_pass_dwarf2_frame): ...new function.
9551 * except.c (pass_set_nothrow_function_flags): Convert from a global
9552 struct to a subclass of rtl_opt_pass along with...
9553 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9555 (make_pass_set_nothrow_function_flags): ...new function.
9556 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9557 subclass of rtl_opt_pass along with...
9558 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9560 (make_pass_convert_to_eh_region_ranges): ...new function.
9561 * final.c (pass_compute_alignments): Convert from a global struct to a
9562 subclass of rtl_opt_pass along with...
9563 (pass_data_compute_alignments): ...new pass_data instance and...
9564 (make_pass_compute_alignments): ...new function.
9565 (pass_final): Convert from a global struct to a subclass of
9566 rtl_opt_pass along with...
9567 (pass_data_final): ...new pass_data instance and...
9568 (make_pass_final): ...new function.
9569 (pass_shorten_branches): Convert from a global struct to a subclass of
9570 rtl_opt_pass along with...
9571 (pass_data_shorten_branches): ...new pass_data instance and...
9572 (make_pass_shorten_branches): ...new function.
9573 (pass_clean_state): Convert from a global struct to a subclass of
9574 rtl_opt_pass along with...
9575 (pass_data_clean_state): ...new pass_data instance and...
9576 (make_pass_clean_state): ...new function.
9577 * function.c (pass_instantiate_virtual_regs): Convert from a global
9578 struct to a subclass of rtl_opt_pass along with...
9579 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9580 (make_pass_instantiate_virtual_regs): ...new function.
9581 (pass_leaf_regs): Convert from a global struct to a subclass of
9582 rtl_opt_pass along with...
9583 (pass_data_leaf_regs): ...new pass_data instance and...
9584 (make_pass_leaf_regs): ...new function.
9585 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9586 subclass of rtl_opt_pass along with...
9587 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9589 (make_pass_thread_prologue_and_epilogue): ...new function.
9590 (pass_match_asm_constraints): Convert from a global struct to a
9591 subclass of rtl_opt_pass along with...
9592 (pass_data_match_asm_constraints): ...new pass_data instance and...
9593 (make_pass_match_asm_constraints): ...new function.
9594 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9595 subclass of rtl_opt_pass along with...
9596 (pass_data_rtl_fwprop): ...new pass_data instance and...
9597 (make_pass_rtl_fwprop): ...new function.
9598 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9599 rtl_opt_pass along with...
9600 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9601 (make_pass_rtl_fwprop_addr): ...new function.
9602 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9603 rtl_opt_pass along with...
9604 (pass_data_rtl_pre): ...new pass_data instance and...
9605 (make_pass_rtl_pre): ...new function.
9606 (pass_rtl_hoist): Convert from a global struct to a subclass of
9607 rtl_opt_pass along with...
9608 (pass_data_rtl_hoist): ...new pass_data instance and...
9609 (make_pass_rtl_hoist): ...new function.
9610 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9611 subclass of gimple_opt_pass along with...
9612 (pass_data_lower_cf): ...new pass_data instance and...
9613 (make_pass_lower_cf): ...new function.
9614 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9615 from a global struct to a subclass of gimple_opt_pass along with...
9616 (pass_data_strength_reduction): ...new pass_data instance and...
9617 (make_pass_strength_reduction): ...new function.
9618 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9619 of rtl_opt_pass along with...
9620 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9621 (make_pass_rtl_ifcvt): ...new function.
9622 (pass_if_after_combine): Convert from a global struct to a subclass of
9623 rtl_opt_pass along with...
9624 (pass_data_if_after_combine): ...new pass_data instance and...
9625 (make_pass_if_after_combine): ...new function.
9626 (pass_if_after_reload): Convert from a global struct to a subclass of
9627 rtl_opt_pass along with...
9628 (pass_data_if_after_reload): ...new pass_data instance and...
9629 (make_pass_if_after_reload): ...new function.
9630 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9631 a subclass of rtl_opt_pass along with...
9632 (pass_data_initialize_regs): ...new pass_data instance and...
9633 (make_pass_initialize_regs): ...new function.
9634 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9635 of ipa_opt_pass_d along with...
9636 (pass_data_ipa_cp): ...new pass_data instance and...
9637 (make_pass_ipa_cp): ...new function.
9638 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9639 global struct to a subclass of gimple_opt_pass along with...
9640 (pass_data_inline_parameters): ...new pass_data instance and...
9641 (make_pass_inline_parameters): ...new function.
9642 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9643 subclass of gimple_opt_pass along with...
9644 (pass_data_early_inline): ...new pass_data instance and...
9645 (make_pass_early_inline): ...new function.
9646 (pass_ipa_inline): Convert from a global struct to a subclass of
9647 ipa_opt_pass_d along with...
9648 (pass_data_ipa_inline): ...new pass_data instance and...
9649 (make_pass_ipa_inline): ...new function.
9650 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9651 struct to a subclass of gimple_opt_pass along with...
9652 (pass_data_local_pure_const): ...new pass_data instance and...
9653 (make_pass_local_pure_const): ...new function.
9654 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9655 ipa_opt_pass_d along with...
9656 (pass_data_ipa_pure_const): ...new pass_data instance and...
9657 (make_pass_ipa_pure_const): ...new function.
9658 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9659 to a subclass of ipa_opt_pass_d along with...
9660 (pass_data_ipa_reference): ...new pass_data instance and...
9661 (make_pass_ipa_reference): ...new function.
9662 * ipa-split.c (pass_split_functions): Convert from a global struct to
9663 a subclass of gimple_opt_pass along with...
9664 (pass_data_split_functions): ...new pass_data instance and...
9665 (make_pass_split_functions): ...new function.
9666 (pass_feedback_split_functions): Convert from a global struct to a
9667 subclass of gimple_opt_pass along with...
9668 (pass_data_feedback_split_functions): ...new pass_data instance and...
9669 (make_pass_feedback_split_functions): ...new function.
9670 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9671 global struct to a subclass of simple_ipa_opt_pass along with...
9672 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9674 (make_pass_ipa_function_and_variable_visibility): ...new function.
9675 (pass_ipa_free_inline_summary): Convert from a global struct to a
9676 subclass of simple_ipa_opt_pass along with...
9677 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9678 (make_pass_ipa_free_inline_summary): ...new function.
9679 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9680 subclass of ipa_opt_pass_d along with...
9681 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9683 (make_pass_ipa_whole_program_visibility): ...new function.
9684 (pass_ipa_profile): Convert from a global struct to a subclass of
9685 ipa_opt_pass_d along with...
9686 (pass_data_ipa_profile): ...new pass_data instance and...
9687 (make_pass_ipa_profile): ...new function.
9688 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9689 ipa_opt_pass_d along with...
9690 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9691 (make_pass_ipa_cdtor_merge): ...new function.
9692 * ira.c (pass_ira): Convert from a global struct to a subclass of
9693 rtl_opt_pass along with...
9694 (pass_data_ira): ...new pass_data instance and...
9695 (make_pass_ira): ...new function.
9696 (pass_reload): Convert from a global struct to a subclass of
9697 rtl_opt_pass along with...
9698 (pass_data_reload): ...new pass_data instance and...
9699 (make_pass_reload): ...new function.
9700 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9701 subclass of rtl_opt_pass along with...
9702 (pass_data_cleanup_barriers): ...new pass_data instance and...
9703 (make_pass_cleanup_barriers): ...new function.
9704 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9705 of rtl_opt_pass along with...
9706 (pass_data_loop2): ...new pass_data instance and...
9707 (make_pass_loop2): ...new function.
9708 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9709 rtl_opt_pass along with...
9710 (pass_data_rtl_loop_init): ...new pass_data instance and...
9711 (make_pass_rtl_loop_init): ...new function.
9712 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9713 rtl_opt_pass along with...
9714 (pass_data_rtl_loop_done): ...new pass_data instance and...
9715 (make_pass_rtl_loop_done): ...new function.
9716 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9717 subclass of rtl_opt_pass along with...
9718 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9719 (make_pass_rtl_move_loop_invariants): ...new function.
9720 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9721 rtl_opt_pass along with...
9722 (pass_data_rtl_unswitch): ...new pass_data instance and...
9723 (make_pass_rtl_unswitch): ...new function.
9724 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9725 subclass of rtl_opt_pass along with...
9726 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9728 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9729 (pass_rtl_doloop): Convert from a global struct to a subclass of
9730 rtl_opt_pass along with...
9731 (pass_data_rtl_doloop): ...new pass_data instance and...
9732 (make_pass_rtl_doloop): ...new function.
9733 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9734 a subclass of rtl_opt_pass along with...
9735 (pass_data_lower_subreg): ...new pass_data instance and...
9736 (make_pass_lower_subreg): ...new function.
9737 (pass_lower_subreg2): Convert from a global struct to a subclass of
9738 rtl_opt_pass along with...
9739 (pass_data_lower_subreg2): ...new pass_data instance and...
9740 (make_pass_lower_subreg2): ...new function.
9741 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9742 struct to a subclass of ipa_opt_pass_d along with...
9743 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9744 (make_pass_ipa_lto_gimple_out): ...new function.
9745 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9746 of ipa_opt_pass_d along with...
9747 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9748 (make_pass_ipa_lto_finish_out): ...new function.
9749 * mode-switching.c (pass_mode_switching): Convert from a global struct
9750 to a subclass of rtl_opt_pass along with...
9751 (pass_data_mode_switching): ...new pass_data instance and...
9752 (make_pass_mode_switching): ...new function.
9753 * modulo-sched.c (pass_sms): Convert from a global struct to a
9754 subclass of rtl_opt_pass along with...
9755 (pass_data_sms): ...new pass_data instance and...
9756 (make_pass_sms): ...new function.
9757 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9758 subclass of gimple_opt_pass along with...
9759 (pass_data_expand_omp): ...new pass_data instance and...
9760 (make_pass_expand_omp): ...new function.
9761 (pass_lower_omp): Convert from a global struct to a subclass of
9762 gimple_opt_pass along with...
9763 (pass_data_lower_omp): ...new pass_data instance and...
9764 (make_pass_lower_omp): ...new function.
9765 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9766 of gimple_opt_pass along with...
9767 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9768 (make_pass_diagnose_omp_blocks): ...new function.
9769 * passes.c (pass_early_local_passes): Convert from a global struct to
9770 a subclass of simple_ipa_opt_pass along with...
9771 (pass_data_early_local_passes): ...new pass_data instance and...
9772 (make_pass_early_local_passes): ...new function.
9773 (pass_all_early_optimizations): Convert from a global struct to a
9774 subclass of gimple_opt_pass along with...
9775 (pass_data_all_early_optimizations): ...new pass_data instance and...
9776 (make_pass_all_early_optimizations): ...new function.
9777 (pass_all_optimizations): Convert from a global struct to a subclass
9778 of gimple_opt_pass along with...
9779 (pass_data_all_optimizations): ...new pass_data instance and...
9780 (make_pass_all_optimizations): ...new function.
9781 (pass_all_optimizations_g): Convert from a global struct to a subclass
9782 of gimple_opt_pass along with...
9783 (pass_data_all_optimizations_g): ...new pass_data instance and...
9784 (make_pass_all_optimizations_g): ...new function.
9785 (pass_rest_of_compilation): Convert from a global struct to a subclass
9786 of rtl_opt_pass along with...
9787 (pass_data_rest_of_compilation): ...new pass_data instance and...
9788 (make_pass_rest_of_compilation): ...new function.
9789 (pass_postreload): Convert from a global struct to a subclass of
9790 rtl_opt_pass along with...
9791 (pass_data_postreload): ...new pass_data instance and...
9792 (make_pass_postreload): ...new function.
9793 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9794 subclass of rtl_opt_pass along with...
9795 (pass_data_gcse2): ...new pass_data instance and...
9796 (make_pass_gcse2): ...new function.
9797 * postreload.c (pass_postreload_cse): Convert from a global struct to
9798 a subclass of rtl_opt_pass along with...
9799 (pass_data_postreload_cse): ...new pass_data instance and...
9800 (make_pass_postreload_cse): ...new function.
9801 * predict.c (pass_profile): Convert from a global struct to a subclass
9802 of gimple_opt_pass along with...
9803 (pass_data_profile): ...new pass_data instance and...
9804 (make_pass_profile): ...new function.
9805 (pass_strip_predict_hints): Convert from a global struct to a subclass
9806 of gimple_opt_pass along with...
9807 (pass_data_strip_predict_hints): ...new pass_data instance and...
9808 (make_pass_strip_predict_hints): ...new function.
9809 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9810 of rtl_opt_pass along with...
9811 (pass_data_peephole2): ...new pass_data instance and...
9812 (make_pass_peephole2): ...new function.
9813 (pass_split_all_insns): Convert from a global struct to a subclass of
9814 rtl_opt_pass along with...
9815 (pass_data_split_all_insns): ...new pass_data instance and...
9816 (make_pass_split_all_insns): ...new function.
9817 (pass_split_after_reload): Convert from a global struct to a subclass
9818 of rtl_opt_pass along with...
9819 (pass_data_split_after_reload): ...new pass_data instance and...
9820 (make_pass_split_after_reload): ...new function.
9821 (pass_split_before_regstack): Convert from a global struct to a
9822 subclass of rtl_opt_pass along with...
9823 (pass_data_split_before_regstack): ...new pass_data instance and...
9824 (make_pass_split_before_regstack): ...new function.
9825 (pass_split_before_sched2): Convert from a global struct to a subclass
9826 of rtl_opt_pass along with...
9827 (pass_data_split_before_sched2): ...new pass_data instance and...
9828 (make_pass_split_before_sched2): ...new function.
9829 (pass_split_for_shorten_branches): Convert from a global struct to a
9830 subclass of rtl_opt_pass along with...
9831 (pass_data_split_for_shorten_branches): ...new pass_data instance
9833 (make_pass_split_for_shorten_branches): ...new function.
9834 * ree.c (pass_ree): Convert from a global struct to a subclass of
9835 rtl_opt_pass along with...
9836 (pass_data_ree): ...new pass_data instance and...
9837 (make_pass_ree): ...new function.
9838 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9839 subclass of rtl_opt_pass along with...
9840 (pass_data_stack_regs): ...new pass_data instance and...
9841 (make_pass_stack_regs): ...new function.
9842 (pass_stack_regs_run): Convert from a global struct to a subclass of
9843 rtl_opt_pass along with...
9844 (pass_data_stack_regs_run): ...new pass_data instance and...
9845 (make_pass_stack_regs_run): ...new function.
9846 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9847 subclass of rtl_opt_pass along with...
9848 (pass_data_cprop_hardreg): ...new pass_data instance and...
9849 (make_pass_cprop_hardreg): ...new function.
9850 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9851 subclass of rtl_opt_pass along with...
9852 (pass_data_reginfo_init): ...new pass_data instance and...
9853 (make_pass_reginfo_init): ...new function.
9854 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9855 of rtl_opt_pass along with...
9856 (pass_data_regmove): ...new pass_data instance and...
9857 (make_pass_regmove): ...new function.
9858 * regrename.c (pass_regrename): Convert from a global struct to a
9859 subclass of rtl_opt_pass along with...
9860 (pass_data_regrename): ...new pass_data instance and...
9861 (make_pass_regrename): ...new function.
9862 * reorg.c (pass_delay_slots): Convert from a global struct to a
9863 subclass of rtl_opt_pass along with...
9864 (pass_data_delay_slots): ...new pass_data instance and...
9865 (make_pass_delay_slots): ...new function.
9866 (pass_machine_reorg): Convert from a global struct to a subclass of
9867 rtl_opt_pass along with...
9868 (pass_data_machine_reorg): ...new pass_data instance and...
9869 (make_pass_machine_reorg): ...new function.
9870 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9871 of rtl_opt_pass along with...
9872 (pass_data_sched): ...new pass_data instance and...
9873 (make_pass_sched): ...new function.
9874 (pass_sched2): Convert from a global struct to a subclass of
9875 rtl_opt_pass along with...
9876 (pass_data_sched2): ...new pass_data instance and...
9877 (make_pass_sched2): ...new function.
9878 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9879 to a subclass of rtl_opt_pass along with...
9880 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9881 (make_pass_stack_ptr_mod): ...new function.
9882 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9883 to a subclass of rtl_opt_pass along with...
9884 (pass_data_rtl_store_motion): ...new pass_data instance and...
9885 (make_pass_rtl_store_motion): ...new function.
9886 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9887 of gimple_opt_pass along with...
9888 (pass_data_tracer): ...new pass_data instance and...
9889 (make_pass_tracer): ...new function.
9890 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9891 to a subclass of gimple_opt_pass along with...
9892 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9893 (make_pass_diagnose_tm_blocks): ...new function.
9894 (pass_lower_tm): Convert from a global struct to a subclass of
9895 gimple_opt_pass along with...
9896 (pass_data_lower_tm): ...new pass_data instance and...
9897 (make_pass_lower_tm): ...new function.
9898 (pass_tm_init): Convert from a global struct to a subclass of
9899 gimple_opt_pass along with...
9900 (pass_data_tm_init): ...new pass_data instance and...
9901 (make_pass_tm_init): ...new function.
9902 (pass_tm_mark): Convert from a global struct to a subclass of
9903 gimple_opt_pass along with...
9904 (pass_data_tm_mark): ...new pass_data instance and...
9905 (make_pass_tm_mark): ...new function.
9906 (pass_tm_edges): Convert from a global struct to a subclass of
9907 gimple_opt_pass along with...
9908 (pass_data_tm_edges): ...new pass_data instance and...
9909 (make_pass_tm_edges): ...new function.
9910 (pass_tm_memopt): Convert from a global struct to a subclass of
9911 gimple_opt_pass along with...
9912 (pass_data_tm_memopt): ...new pass_data instance and...
9913 (make_pass_tm_memopt): ...new function.
9914 (pass_ipa_tm): Convert from a global struct to a subclass of
9915 simple_ipa_opt_pass along with...
9916 (pass_data_ipa_tm): ...new pass_data instance and...
9917 (make_pass_ipa_tm): ...new function.
9918 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9919 subclass of gimple_opt_pass along with...
9920 (pass_data_call_cdce): ...new pass_data instance and...
9921 (make_pass_call_cdce): ...new function.
9922 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9923 subclass of gimple_opt_pass along with...
9924 (pass_data_build_cfg): ...new pass_data instance and...
9925 (make_pass_build_cfg): ...new function.
9926 (pass_split_crit_edges): Convert from a global struct to a subclass of
9927 gimple_opt_pass along with...
9928 (pass_data_split_crit_edges): ...new pass_data instance and...
9929 (make_pass_split_crit_edges): ...new function.
9930 (pass_warn_function_return): Convert from a global struct to a
9931 subclass of gimple_opt_pass along with...
9932 (pass_data_warn_function_return): ...new pass_data instance and...
9933 (make_pass_warn_function_return): ...new function.
9934 (pass_warn_function_noreturn): Convert from a global struct to a
9935 subclass of gimple_opt_pass along with...
9936 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9937 (make_pass_warn_function_noreturn): ...new function.
9938 (pass_warn_unused_result): Convert from a global struct to a subclass
9939 of gimple_opt_pass along with...
9940 (pass_data_warn_unused_result): ...new pass_data instance and...
9941 (make_pass_warn_unused_result): ...new function.
9942 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9943 a subclass of gimple_opt_pass along with...
9944 (pass_data_merge_phi): ...new pass_data instance and...
9945 (make_pass_merge_phi): ...new function.
9946 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9947 a subclass of gimple_opt_pass along with...
9948 (pass_data_lower_complex): ...new pass_data instance and...
9949 (make_pass_lower_complex): ...new function.
9950 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9951 gimple_opt_pass along with...
9952 (pass_data_lower_complex_O0): ...new pass_data instance and...
9953 (make_pass_lower_complex_O0): ...new function.
9954 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9955 subclass of gimple_opt_pass along with...
9956 (pass_data_lower_eh): ...new pass_data instance and...
9957 (make_pass_lower_eh): ...new function.
9958 (pass_refactor_eh): Convert from a global struct to a subclass of
9959 gimple_opt_pass along with...
9960 (pass_data_refactor_eh): ...new pass_data instance and...
9961 (make_pass_refactor_eh): ...new function.
9962 (pass_lower_resx): Convert from a global struct to a subclass of
9963 gimple_opt_pass along with...
9964 (pass_data_lower_resx): ...new pass_data instance and...
9965 (make_pass_lower_resx): ...new function.
9966 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9967 of gimple_opt_pass along with...
9968 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9969 (make_pass_lower_eh_dispatch): ...new function.
9970 (pass_cleanup_eh): Convert from a global struct to a subclass of
9971 gimple_opt_pass along with...
9972 (pass_data_cleanup_eh): ...new pass_data instance and...
9973 (make_pass_cleanup_eh): ...new function.
9974 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9975 to a subclass of simple_ipa_opt_pass along with...
9976 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9977 (make_pass_ipa_lower_emutls): ...new function.
9978 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9979 a subclass of gimple_opt_pass along with...
9980 (pass_data_if_conversion): ...new pass_data instance and...
9981 (make_pass_if_conversion): ...new function.
9982 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9983 subclass of gimple_opt_pass along with...
9984 (pass_data_build_ssa): ...new pass_data instance and...
9985 (make_pass_build_ssa): ...new function.
9986 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9987 global struct to a subclass of gimple_opt_pass along with...
9988 (pass_data_loop_distribution): ...new pass_data instance and...
9989 (make_pass_loop_distribution): ...new function.
9990 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9991 subclass of gimple_opt_pass along with...
9992 (pass_data_mudflap_1): ...new pass_data instance and...
9993 (make_pass_mudflap_1): ...new function.
9994 (pass_mudflap_2): Convert from a global struct to a subclass of
9995 gimple_opt_pass along with...
9996 (pass_data_mudflap_2): ...new pass_data instance and...
9997 (make_pass_mudflap_2): ...new function.
9998 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9999 subclass of gimple_opt_pass along with...
10000 (pass_data_mudflap_1): ...new pass_data instance and...
10001 (make_pass_mudflap_1): ...new function.
10002 (pass_mudflap_2): Convert from a global struct to a subclass of
10003 gimple_opt_pass along with...
10004 (pass_data_mudflap_2): ...new pass_data instance and...
10005 (make_pass_mudflap_2): ...new function.
10006 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10007 gimple_opt_pass along with...
10008 (pass_data_nrv): ...new pass_data instance and...
10009 (make_pass_nrv): ...new function.
10010 (pass_return_slot): Convert from a global struct to a subclass of
10011 gimple_opt_pass along with...
10012 (pass_data_return_slot): ...new pass_data instance and...
10013 (make_pass_return_slot): ...new function.
10014 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10015 to a subclass of gimple_opt_pass along with...
10016 (pass_data_object_sizes): ...new pass_data instance and...
10017 (make_pass_object_sizes): ...new function.
10018 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10019 global struct to a subclass of gimple_opt_pass along with...
10020 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10022 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10023 (pass_fixup_cfg): Convert from a global struct to a subclass of
10024 gimple_opt_pass along with...
10025 (pass_data_fixup_cfg): ...new pass_data instance and...
10026 (make_pass_fixup_cfg): ...new function.
10027 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10028 (make_pass_mudflap_1): ...new function.
10029 (pass_mudflap_2): Replace declaration with that of...
10030 (make_pass_mudflap_2): ...new function.
10031 (pass_asan): Replace declaration with that of...
10032 (make_pass_asan): ...new function.
10033 (pass_asan_O0): Replace declaration with that of...
10034 (make_pass_asan_O0): ...new function.
10035 (pass_tsan): Replace declaration with that of...
10036 (make_pass_tsan): ...new function.
10037 (pass_tsan_O0): Replace declaration with that of...
10038 (make_pass_tsan_O0): ...new function.
10039 (pass_lower_cf): Replace declaration with that of...
10040 (make_pass_lower_cf): ...new function.
10041 (pass_refactor_eh): Replace declaration with that of...
10042 (make_pass_refactor_eh): ...new function.
10043 (pass_lower_eh): Replace declaration with that of...
10044 (make_pass_lower_eh): ...new function.
10045 (pass_lower_eh_dispatch): Replace declaration with that of...
10046 (make_pass_lower_eh_dispatch): ...new function.
10047 (pass_lower_resx): Replace declaration with that of...
10048 (make_pass_lower_resx): ...new function.
10049 (pass_build_cfg): Replace declaration with that of...
10050 (make_pass_build_cfg): ...new function.
10051 (pass_early_tree_profile): Replace declaration with that of...
10052 (make_pass_early_tree_profile): ...new function.
10053 (pass_cleanup_eh): Replace declaration with that of...
10054 (make_pass_cleanup_eh): ...new function.
10055 (pass_sra): Replace declaration with that of...
10056 (make_pass_sra): ...new function.
10057 (pass_sra_early): Replace declaration with that of...
10058 (make_pass_sra_early): ...new function.
10059 (pass_early_ipa_sra): Replace declaration with that of...
10060 (make_pass_early_ipa_sra): ...new function.
10061 (pass_tail_recursion): Replace declaration with that of...
10062 (make_pass_tail_recursion): ...new function.
10063 (pass_tail_calls): Replace declaration with that of...
10064 (make_pass_tail_calls): ...new function.
10065 (pass_tree_loop): Replace declaration with that of...
10066 (make_pass_tree_loop): ...new function.
10067 (pass_tree_loop_init): Replace declaration with that of...
10068 (make_pass_tree_loop_init): ...new function.
10069 (pass_lim): Replace declaration with that of...
10070 (make_pass_lim): ...new function.
10071 (pass_tree_unswitch): Replace declaration with that of...
10072 (make_pass_tree_unswitch): ...new function.
10073 (pass_predcom): Replace declaration with that of...
10074 (make_pass_predcom): ...new function.
10075 (pass_iv_canon): Replace declaration with that of...
10076 (make_pass_iv_canon): ...new function.
10077 (pass_scev_cprop): Replace declaration with that of...
10078 (make_pass_scev_cprop): ...new function.
10079 (pass_empty_loop): Replace declaration with that of...
10080 (make_pass_empty_loop): ...new function.
10081 (pass_record_bounds): Replace declaration with that of...
10082 (make_pass_record_bounds): ...new function.
10083 (pass_graphite): Replace declaration with that of...
10084 (make_pass_graphite): ...new function.
10085 (pass_graphite_transforms): Replace declaration with that of...
10086 (make_pass_graphite_transforms): ...new function.
10087 (pass_if_conversion): Replace declaration with that of...
10088 (make_pass_if_conversion): ...new function.
10089 (pass_loop_distribution): Replace declaration with that of...
10090 (make_pass_loop_distribution): ...new function.
10091 (pass_vectorize): Replace declaration with that of...
10092 (make_pass_vectorize): ...new function.
10093 (pass_slp_vectorize): Replace declaration with that of...
10094 (make_pass_slp_vectorize): ...new function.
10095 (pass_complete_unroll): Replace declaration with that of...
10096 (make_pass_complete_unroll): ...new function.
10097 (pass_complete_unrolli): Replace declaration with that of...
10098 (make_pass_complete_unrolli): ...new function.
10099 (pass_parallelize_loops): Replace declaration with that of...
10100 (make_pass_parallelize_loops): ...new function.
10101 (pass_loop_prefetch): Replace declaration with that of...
10102 (make_pass_loop_prefetch): ...new function.
10103 (pass_iv_optimize): Replace declaration with that of...
10104 (make_pass_iv_optimize): ...new function.
10105 (pass_tree_loop_done): Replace declaration with that of...
10106 (make_pass_tree_loop_done): ...new function.
10107 (pass_ch): Replace declaration with that of...
10108 (make_pass_ch): ...new function.
10109 (pass_ccp): Replace declaration with that of...
10110 (make_pass_ccp): ...new function.
10111 (pass_phi_only_cprop): Replace declaration with that of...
10112 (make_pass_phi_only_cprop): ...new function.
10113 (pass_build_ssa): Replace declaration with that of...
10114 (make_pass_build_ssa): ...new function.
10115 (pass_build_alias): Replace declaration with that of...
10116 (make_pass_build_alias): ...new function.
10117 (pass_build_ealias): Replace declaration with that of...
10118 (make_pass_build_ealias): ...new function.
10119 (pass_dominator): Replace declaration with that of...
10120 (make_pass_dominator): ...new function.
10121 (pass_dce): Replace declaration with that of...
10122 (make_pass_dce): ...new function.
10123 (pass_dce_loop): Replace declaration with that of...
10124 (make_pass_dce_loop): ...new function.
10125 (pass_cd_dce): Replace declaration with that of...
10126 (make_pass_cd_dce): ...new function.
10127 (pass_call_cdce): Replace declaration with that of...
10128 (make_pass_call_cdce): ...new function.
10129 (pass_merge_phi): Replace declaration with that of...
10130 (make_pass_merge_phi): ...new function.
10131 (pass_split_crit_edges): Replace declaration with that of...
10132 (make_pass_split_crit_edges): ...new function.
10133 (pass_pre): Replace declaration with that of...
10134 (make_pass_pre): ...new function.
10135 (pass_profile): Replace declaration with that of...
10136 (make_pass_profile): ...new function.
10137 (pass_strip_predict_hints): Replace declaration with that of...
10138 (make_pass_strip_predict_hints): ...new function.
10139 (pass_lower_complex_O0): Replace declaration with that of...
10140 (make_pass_lower_complex_O0): ...new function.
10141 (pass_lower_complex): Replace declaration with that of...
10142 (make_pass_lower_complex): ...new function.
10143 (pass_lower_vector): Replace declaration with that of...
10144 (make_pass_lower_vector): ...new function.
10145 (pass_lower_vector_ssa): Replace declaration with that of...
10146 (make_pass_lower_vector_ssa): ...new function.
10147 (pass_lower_omp): Replace declaration with that of...
10148 (make_pass_lower_omp): ...new function.
10149 (pass_diagnose_omp_blocks): Replace declaration with that of...
10150 (make_pass_diagnose_omp_blocks): ...new function.
10151 (pass_expand_omp): Replace declaration with that of...
10152 (make_pass_expand_omp): ...new function.
10153 (pass_expand_omp_ssa): Replace declaration with that of...
10154 (make_pass_expand_omp_ssa): ...new function.
10155 (pass_object_sizes): Replace declaration with that of...
10156 (make_pass_object_sizes): ...new function.
10157 (pass_strlen): Replace declaration with that of...
10158 (make_pass_strlen): ...new function.
10159 (pass_fold_builtins): Replace declaration with that of...
10160 (make_pass_fold_builtins): ...new function.
10161 (pass_stdarg): Replace declaration with that of...
10162 (make_pass_stdarg): ...new function.
10163 (pass_early_warn_uninitialized): Replace declaration with that of...
10164 (make_pass_early_warn_uninitialized): ...new function.
10165 (pass_late_warn_uninitialized): Replace declaration with that of...
10166 (make_pass_late_warn_uninitialized): ...new function.
10167 (pass_cse_reciprocals): Replace declaration with that of...
10168 (make_pass_cse_reciprocals): ...new function.
10169 (pass_cse_sincos): Replace declaration with that of...
10170 (make_pass_cse_sincos): ...new function.
10171 (pass_optimize_bswap): Replace declaration with that of...
10172 (make_pass_optimize_bswap): ...new function.
10173 (pass_optimize_widening_mul): Replace declaration with that of...
10174 (make_pass_optimize_widening_mul): ...new function.
10175 (pass_warn_function_return): Replace declaration with that of...
10176 (make_pass_warn_function_return): ...new function.
10177 (pass_warn_function_noreturn): Replace declaration with that of...
10178 (make_pass_warn_function_noreturn): ...new function.
10179 (pass_cselim): Replace declaration with that of...
10180 (make_pass_cselim): ...new function.
10181 (pass_phiopt): Replace declaration with that of...
10182 (make_pass_phiopt): ...new function.
10183 (pass_forwprop): Replace declaration with that of...
10184 (make_pass_forwprop): ...new function.
10185 (pass_phiprop): Replace declaration with that of...
10186 (make_pass_phiprop): ...new function.
10187 (pass_tree_ifcombine): Replace declaration with that of...
10188 (make_pass_tree_ifcombine): ...new function.
10189 (pass_dse): Replace declaration with that of...
10190 (make_pass_dse): ...new function.
10191 (pass_nrv): Replace declaration with that of...
10192 (make_pass_nrv): ...new function.
10193 (pass_rename_ssa_copies): Replace declaration with that of...
10194 (make_pass_rename_ssa_copies): ...new function.
10195 (pass_sink_code): Replace declaration with that of...
10196 (make_pass_sink_code): ...new function.
10197 (pass_fre): Replace declaration with that of...
10198 (make_pass_fre): ...new function.
10199 (pass_check_data_deps): Replace declaration with that of...
10200 (make_pass_check_data_deps): ...new function.
10201 (pass_copy_prop): Replace declaration with that of...
10202 (make_pass_copy_prop): ...new function.
10203 (pass_vrp): Replace declaration with that of...
10204 (make_pass_vrp): ...new function.
10205 (pass_uncprop): Replace declaration with that of...
10206 (make_pass_uncprop): ...new function.
10207 (pass_return_slot): Replace declaration with that of...
10208 (make_pass_return_slot): ...new function.
10209 (pass_reassoc): Replace declaration with that of...
10210 (make_pass_reassoc): ...new function.
10211 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10212 (make_pass_rebuild_cgraph_edges): ...new function.
10213 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10214 (make_pass_remove_cgraph_callee_edges): ...new function.
10215 (pass_build_cgraph_edges): Replace declaration with that of...
10216 (make_pass_build_cgraph_edges): ...new function.
10217 (pass_local_pure_const): Replace declaration with that of...
10218 (make_pass_local_pure_const): ...new function.
10219 (pass_tracer): Replace declaration with that of...
10220 (make_pass_tracer): ...new function.
10221 (pass_warn_unused_result): Replace declaration with that of...
10222 (make_pass_warn_unused_result): ...new function.
10223 (pass_diagnose_tm_blocks): Replace declaration with that of...
10224 (make_pass_diagnose_tm_blocks): ...new function.
10225 (pass_lower_tm): Replace declaration with that of...
10226 (make_pass_lower_tm): ...new function.
10227 (pass_tm_init): Replace declaration with that of...
10228 (make_pass_tm_init): ...new function.
10229 (pass_tm_mark): Replace declaration with that of...
10230 (make_pass_tm_mark): ...new function.
10231 (pass_tm_memopt): Replace declaration with that of...
10232 (make_pass_tm_memopt): ...new function.
10233 (pass_tm_edges): Replace declaration with that of...
10234 (make_pass_tm_edges): ...new function.
10235 (pass_split_functions): Replace declaration with that of...
10236 (make_pass_split_functions): ...new function.
10237 (pass_feedback_split_functions): Replace declaration with that of...
10238 (make_pass_feedback_split_functions): ...new function.
10239 (pass_strength_reduction): Replace declaration with that of...
10240 (make_pass_strength_reduction): ...new function.
10241 (pass_ipa_lower_emutls): Replace declaration with that of...
10242 (make_pass_ipa_lower_emutls): ...new function.
10243 (pass_ipa_function_and_variable_visibility): Replace declaration with
10245 (make_pass_ipa_function_and_variable_visibility): ...new function.
10246 (pass_ipa_tree_profile): Replace declaration with that of...
10247 (make_pass_ipa_tree_profile): ...new function.
10248 (pass_early_local_passes): Replace declaration with that of...
10249 (make_pass_early_local_passes): ...new function.
10250 (pass_ipa_whole_program_visibility): Replace declaration with that
10252 (make_pass_ipa_whole_program_visibility): ...new function.
10253 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10254 (make_pass_ipa_lto_gimple_out): ...new function.
10255 (pass_ipa_increase_alignment): Replace declaration with that of...
10256 (make_pass_ipa_increase_alignment): ...new function.
10257 (pass_ipa_inline): Replace declaration with that of...
10258 (make_pass_ipa_inline): ...new function.
10259 (pass_ipa_free_lang_data): Replace declaration with that of...
10260 (make_pass_ipa_free_lang_data): ...new function.
10261 (pass_ipa_free_inline_summary): Replace declaration with that of...
10262 (make_pass_ipa_free_inline_summary): ...new function.
10263 (pass_ipa_cp): Replace declaration with that of...
10264 (make_pass_ipa_cp): ...new function.
10265 (pass_ipa_reference): Replace declaration with that of...
10266 (make_pass_ipa_reference): ...new function.
10267 (pass_ipa_pure_const): Replace declaration with that of...
10268 (make_pass_ipa_pure_const): ...new function.
10269 (pass_ipa_pta): Replace declaration with that of...
10270 (make_pass_ipa_pta): ...new function.
10271 (pass_ipa_lto_finish_out): Replace declaration with that of...
10272 (make_pass_ipa_lto_finish_out): ...new function.
10273 (pass_ipa_tm): Replace declaration with that of...
10274 (make_pass_ipa_tm): ...new function.
10275 (pass_ipa_profile): Replace declaration with that of...
10276 (make_pass_ipa_profile): ...new function.
10277 (pass_ipa_cdtor_merge): Replace declaration with that of...
10278 (make_pass_ipa_cdtor_merge): ...new function.
10279 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10281 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10282 (pass_init_datastructures): Replace declaration with that of...
10283 (make_pass_init_datastructures): ...new function.
10284 (pass_fixup_cfg): Replace declaration with that of...
10285 (make_pass_fixup_cfg): ...new function.
10286 (pass_expand): Replace declaration with that of...
10287 (make_pass_expand): ...new function.
10288 (pass_instantiate_virtual_regs): Replace declaration with that of...
10289 (make_pass_instantiate_virtual_regs): ...new function.
10290 (pass_rtl_fwprop): Replace declaration with that of...
10291 (make_pass_rtl_fwprop): ...new function.
10292 (pass_rtl_fwprop_addr): Replace declaration with that of...
10293 (make_pass_rtl_fwprop_addr): ...new function.
10294 (pass_jump): Replace declaration with that of...
10295 (make_pass_jump): ...new function.
10296 (pass_jump2): Replace declaration with that of...
10297 (make_pass_jump2): ...new function.
10298 (pass_lower_subreg): Replace declaration with that of...
10299 (make_pass_lower_subreg): ...new function.
10300 (pass_cse): Replace declaration with that of...
10301 (make_pass_cse): ...new function.
10302 (pass_fast_rtl_dce): Replace declaration with that of...
10303 (make_pass_fast_rtl_dce): ...new function.
10304 (pass_ud_rtl_dce): Replace declaration with that of...
10305 (make_pass_ud_rtl_dce): ...new function.
10306 (pass_rtl_dce): Replace declaration with that of...
10307 (make_pass_rtl_dce): ...new function.
10308 (pass_rtl_dse1): Replace declaration with that of...
10309 (make_pass_rtl_dse1): ...new function.
10310 (pass_rtl_dse2): Replace declaration with that of...
10311 (make_pass_rtl_dse2): ...new function.
10312 (pass_rtl_dse3): Replace declaration with that of...
10313 (make_pass_rtl_dse3): ...new function.
10314 (pass_rtl_cprop): Replace declaration with that of...
10315 (make_pass_rtl_cprop): ...new function.
10316 (pass_rtl_pre): Replace declaration with that of...
10317 (make_pass_rtl_pre): ...new function.
10318 (pass_rtl_hoist): Replace declaration with that of...
10319 (make_pass_rtl_hoist): ...new function.
10320 (pass_rtl_store_motion): Replace declaration with that of...
10321 (make_pass_rtl_store_motion): ...new function.
10322 (pass_cse_after_global_opts): Replace declaration with that of...
10323 (make_pass_cse_after_global_opts): ...new function.
10324 (pass_rtl_ifcvt): Replace declaration with that of...
10325 (make_pass_rtl_ifcvt): ...new function.
10326 (pass_into_cfg_layout_mode): Replace declaration with that of...
10327 (make_pass_into_cfg_layout_mode): ...new function.
10328 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10329 (make_pass_outof_cfg_layout_mode): ...new function.
10330 (pass_loop2): Replace declaration with that of...
10331 (make_pass_loop2): ...new function.
10332 (pass_rtl_loop_init): Replace declaration with that of...
10333 (make_pass_rtl_loop_init): ...new function.
10334 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10335 (make_pass_rtl_move_loop_invariants): ...new function.
10336 (pass_rtl_unswitch): Replace declaration with that of...
10337 (make_pass_rtl_unswitch): ...new function.
10338 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10339 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10340 (pass_rtl_doloop): Replace declaration with that of...
10341 (make_pass_rtl_doloop): ...new function.
10342 (pass_rtl_loop_done): Replace declaration with that of...
10343 (make_pass_rtl_loop_done): ...new function.
10344 (pass_web): Replace declaration with that of...
10345 (make_pass_web): ...new function.
10346 (pass_cse2): Replace declaration with that of...
10347 (make_pass_cse2): ...new function.
10348 (pass_df_initialize_opt): Replace declaration with that of...
10349 (make_pass_df_initialize_opt): ...new function.
10350 (pass_df_initialize_no_opt): Replace declaration with that of...
10351 (make_pass_df_initialize_no_opt): ...new function.
10352 (pass_reginfo_init): Replace declaration with that of...
10353 (make_pass_reginfo_init): ...new function.
10354 (pass_inc_dec): Replace declaration with that of...
10355 (make_pass_inc_dec): ...new function.
10356 (pass_stack_ptr_mod): Replace declaration with that of...
10357 (make_pass_stack_ptr_mod): ...new function.
10358 (pass_initialize_regs): Replace declaration with that of...
10359 (make_pass_initialize_regs): ...new function.
10360 (pass_combine): Replace declaration with that of...
10361 (make_pass_combine): ...new function.
10362 (pass_if_after_combine): Replace declaration with that of...
10363 (make_pass_if_after_combine): ...new function.
10364 (pass_ree): Replace declaration with that of...
10365 (make_pass_ree): ...new function.
10366 (pass_partition_blocks): Replace declaration with that of...
10367 (make_pass_partition_blocks): ...new function.
10368 (pass_match_asm_constraints): Replace declaration with that of...
10369 (make_pass_match_asm_constraints): ...new function.
10370 (pass_regmove): Replace declaration with that of...
10371 (make_pass_regmove): ...new function.
10372 (pass_split_all_insns): Replace declaration with that of...
10373 (make_pass_split_all_insns): ...new function.
10374 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10375 (make_pass_fast_rtl_byte_dce): ...new function.
10376 (pass_lower_subreg2): Replace declaration with that of...
10377 (make_pass_lower_subreg2): ...new function.
10378 (pass_mode_switching): Replace declaration with that of...
10379 (make_pass_mode_switching): ...new function.
10380 (pass_sms): Replace declaration with that of...
10381 (make_pass_sms): ...new function.
10382 (pass_sched): Replace declaration with that of...
10383 (make_pass_sched): ...new function.
10384 (pass_ira): Replace declaration with that of...
10385 (make_pass_ira): ...new function.
10386 (pass_reload): Replace declaration with that of...
10387 (make_pass_reload): ...new function.
10388 (pass_clean_state): Replace declaration with that of...
10389 (make_pass_clean_state): ...new function.
10390 (pass_branch_prob): Replace declaration with that of...
10391 (make_pass_branch_prob): ...new function.
10392 (pass_value_profile_transformations): Replace declaration with that
10394 (make_pass_value_profile_transformations): ...new function.
10395 (pass_postreload_cse): Replace declaration with that of...
10396 (make_pass_postreload_cse): ...new function.
10397 (pass_gcse2): Replace declaration with that of...
10398 (make_pass_gcse2): ...new function.
10399 (pass_split_after_reload): Replace declaration with that of...
10400 (make_pass_split_after_reload): ...new function.
10401 (pass_branch_target_load_optimize1): Replace declaration with that
10403 (make_pass_branch_target_load_optimize1): ...new function.
10404 (pass_thread_prologue_and_epilogue): Replace declaration with that
10406 (make_pass_thread_prologue_and_epilogue): ...new function.
10407 (pass_stack_adjustments): Replace declaration with that of...
10408 (make_pass_stack_adjustments): ...new function.
10409 (pass_peephole2): Replace declaration with that of...
10410 (make_pass_peephole2): ...new function.
10411 (pass_if_after_reload): Replace declaration with that of...
10412 (make_pass_if_after_reload): ...new function.
10413 (pass_regrename): Replace declaration with that of...
10414 (make_pass_regrename): ...new function.
10415 (pass_cprop_hardreg): Replace declaration with that of...
10416 (make_pass_cprop_hardreg): ...new function.
10417 (pass_reorder_blocks): Replace declaration with that of...
10418 (make_pass_reorder_blocks): ...new function.
10419 (pass_branch_target_load_optimize2): Replace declaration with that
10421 (make_pass_branch_target_load_optimize2): ...new function.
10422 (pass_leaf_regs): Replace declaration with that of...
10423 (make_pass_leaf_regs): ...new function.
10424 (pass_split_before_sched2): Replace declaration with that of...
10425 (make_pass_split_before_sched2): ...new function.
10426 (pass_compare_elim_after_reload): Replace declaration with that of...
10427 (make_pass_compare_elim_after_reload): ...new function.
10428 (pass_sched2): Replace declaration with that of...
10429 (make_pass_sched2): ...new function.
10430 (pass_stack_regs): Replace declaration with that of...
10431 (make_pass_stack_regs): ...new function.
10432 (pass_stack_regs_run): Replace declaration with that of...
10433 (make_pass_stack_regs_run): ...new function.
10434 (pass_df_finish): Replace declaration with that of...
10435 (make_pass_df_finish): ...new function.
10436 (pass_compute_alignments): Replace declaration with that of...
10437 (make_pass_compute_alignments): ...new function.
10438 (pass_duplicate_computed_gotos): Replace declaration with that of...
10439 (make_pass_duplicate_computed_gotos): ...new function.
10440 (pass_variable_tracking): Replace declaration with that of...
10441 (make_pass_variable_tracking): ...new function.
10442 (pass_free_cfg): Replace declaration with that of...
10443 (make_pass_free_cfg): ...new function.
10444 (pass_machine_reorg): Replace declaration with that of...
10445 (make_pass_machine_reorg): ...new function.
10446 (pass_cleanup_barriers): Replace declaration with that of...
10447 (make_pass_cleanup_barriers): ...new function.
10448 (pass_delay_slots): Replace declaration with that of...
10449 (make_pass_delay_slots): ...new function.
10450 (pass_split_for_shorten_branches): Replace declaration with that of...
10451 (make_pass_split_for_shorten_branches): ...new function.
10452 (pass_split_before_regstack): Replace declaration with that of...
10453 (make_pass_split_before_regstack): ...new function.
10454 (pass_convert_to_eh_region_ranges): Replace declaration with that
10456 (make_pass_convert_to_eh_region_ranges): ...new function.
10457 (pass_shorten_branches): Replace declaration with that of...
10458 (make_pass_shorten_branches): ...new function.
10459 (pass_set_nothrow_function_flags): Replace declaration with that of...
10460 (make_pass_set_nothrow_function_flags): ...new function.
10461 (pass_dwarf2_frame): Replace declaration with that of...
10462 (make_pass_dwarf2_frame): ...new function.
10463 (pass_final): Replace declaration with that of...
10464 (make_pass_final): ...new function.
10465 (pass_rtl_seqabstr): Replace declaration with that of...
10466 (make_pass_rtl_seqabstr): ...new function.
10467 (pass_release_ssa_names): Replace declaration with that of...
10468 (make_pass_release_ssa_names): ...new function.
10469 (pass_early_inline): Replace declaration with that of...
10470 (make_pass_early_inline): ...new function.
10471 (pass_inline_parameters): Replace declaration with that of...
10472 (make_pass_inline_parameters): ...new function.
10473 (pass_update_address_taken): Replace declaration with that of...
10474 (make_pass_update_address_taken): ...new function.
10475 (pass_convert_switch): Replace declaration with that of...
10476 (make_pass_convert_switch): ...new function.
10477 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10478 to a subclass of simple_ipa_opt_pass along with...
10479 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10480 (make_pass_ipa_tree_profile): ...new function.
10481 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10482 subclass of gimple_opt_pass along with...
10483 (pass_data_sra_early): ...new pass_data instance and...
10484 (make_pass_sra_early): ...new function.
10485 (pass_sra): Convert from a global struct to a subclass of
10486 gimple_opt_pass along with...
10487 (pass_data_sra): ...new pass_data instance and...
10488 (make_pass_sra): ...new function.
10489 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10490 gimple_opt_pass along with...
10491 (pass_data_early_ipa_sra): ...new pass_data instance and...
10492 (make_pass_early_ipa_sra): ...new function.
10493 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10494 subclass of gimple_opt_pass along with...
10495 (pass_data_ccp): ...new pass_data instance and...
10496 (make_pass_ccp): ...new function.
10497 (pass_fold_builtins): Convert from a global struct to a subclass of
10498 gimple_opt_pass along with...
10499 (pass_data_fold_builtins): ...new pass_data instance and...
10500 (make_pass_fold_builtins): ...new function.
10501 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10502 subclass of gimple_opt_pass along with...
10503 (pass_data_copy_prop): ...new pass_data instance and...
10504 (make_pass_copy_prop): ...new function.
10505 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10506 global struct to a subclass of gimple_opt_pass along with...
10507 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10508 (make_pass_rename_ssa_copies): ...new function.
10509 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10510 subclass of gimple_opt_pass along with...
10511 (pass_data_dce): ...new pass_data instance and...
10512 (make_pass_dce): ...new function.
10513 (pass_dce_loop): Convert from a global struct to a subclass of
10514 gimple_opt_pass along with...
10515 (pass_data_dce_loop): ...new pass_data instance and...
10516 (make_pass_dce_loop): ...new function.
10517 (pass_cd_dce): Convert from a global struct to a subclass of
10518 gimple_opt_pass along with...
10519 (pass_data_cd_dce): ...new pass_data instance and...
10520 (make_pass_cd_dce): ...new function.
10521 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10522 subclass of gimple_opt_pass along with...
10523 (pass_data_dominator): ...new pass_data instance and...
10524 (make_pass_dominator): ...new function.
10525 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10526 gimple_opt_pass along with...
10527 (pass_data_phi_only_cprop): ...new pass_data instance and...
10528 (make_pass_phi_only_cprop): ...new function.
10529 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10530 subclass of gimple_opt_pass along with...
10531 (pass_data_dse): ...new pass_data instance and...
10532 (make_pass_dse): ...new function.
10533 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10534 a subclass of gimple_opt_pass along with...
10535 (pass_data_forwprop): ...new pass_data instance and...
10536 (make_pass_forwprop): ...new function.
10537 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10538 struct to a subclass of gimple_opt_pass along with...
10539 (pass_data_tree_ifcombine): ...new pass_data instance and...
10540 (make_pass_tree_ifcombine): ...new function.
10541 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10542 subclass of gimple_opt_pass along with...
10543 (pass_data_ch): ...new pass_data instance and...
10544 (make_pass_ch): ...new function.
10545 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10546 subclass of gimple_opt_pass along with...
10547 (pass_data_tree_loop): ...new pass_data instance and...
10548 (make_pass_tree_loop): ...new function.
10549 (pass_tree_loop_init): Convert from a global struct to a subclass of
10550 gimple_opt_pass along with...
10551 (pass_data_tree_loop_init): ...new pass_data instance and...
10552 (make_pass_tree_loop_init): ...new function.
10553 (pass_lim): Convert from a global struct to a subclass of
10554 gimple_opt_pass along with...
10555 (pass_data_lim): ...new pass_data instance and...
10556 (make_pass_lim): ...new function.
10557 (pass_tree_unswitch): Convert from a global struct to a subclass of
10558 gimple_opt_pass along with...
10559 (pass_data_tree_unswitch): ...new pass_data instance and...
10560 (make_pass_tree_unswitch): ...new function.
10561 (pass_predcom): Convert from a global struct to a subclass of
10562 gimple_opt_pass along with...
10563 (pass_data_predcom): ...new pass_data instance and...
10564 (make_pass_predcom): ...new function.
10565 (pass_vectorize): Convert from a global struct to a subclass of
10566 gimple_opt_pass along with...
10567 (pass_data_vectorize): ...new pass_data instance and...
10568 (make_pass_vectorize): ...new function.
10569 (pass_graphite): Convert from a global struct to a subclass of
10570 gimple_opt_pass along with...
10571 (pass_data_graphite): ...new pass_data instance and...
10572 (make_pass_graphite): ...new function.
10573 (pass_graphite_transforms): Convert from a global struct to a subclass
10574 of gimple_opt_pass along with...
10575 (pass_data_graphite_transforms): ...new pass_data instance and...
10576 (make_pass_graphite_transforms): ...new function.
10577 (pass_check_data_deps): Convert from a global struct to a subclass of
10578 gimple_opt_pass along with...
10579 (pass_data_check_data_deps): ...new pass_data instance and...
10580 (make_pass_check_data_deps): ...new function.
10581 (pass_iv_canon): Convert from a global struct to a subclass of
10582 gimple_opt_pass along with...
10583 (pass_data_iv_canon): ...new pass_data instance and...
10584 (make_pass_iv_canon): ...new function.
10585 (pass_scev_cprop): Convert from a global struct to a subclass of
10586 gimple_opt_pass along with...
10587 (pass_data_scev_cprop): ...new pass_data instance and...
10588 (make_pass_scev_cprop): ...new function.
10589 (pass_record_bounds): Convert from a global struct to a subclass of
10590 gimple_opt_pass along with...
10591 (pass_data_record_bounds): ...new pass_data instance and...
10592 (make_pass_record_bounds): ...new function.
10593 (pass_complete_unroll): Convert from a global struct to a subclass of
10594 gimple_opt_pass along with...
10595 (pass_data_complete_unroll): ...new pass_data instance and...
10596 (make_pass_complete_unroll): ...new function.
10597 (pass_complete_unrolli): Convert from a global struct to a subclass of
10598 gimple_opt_pass along with...
10599 (pass_data_complete_unrolli): ...new pass_data instance and...
10600 (make_pass_complete_unrolli): ...new function.
10601 (pass_parallelize_loops): Convert from a global struct to a subclass
10602 of gimple_opt_pass along with...
10603 (pass_data_parallelize_loops): ...new pass_data instance and...
10604 (make_pass_parallelize_loops): ...new function.
10605 (pass_loop_prefetch): Convert from a global struct to a subclass of
10606 gimple_opt_pass along with...
10607 (pass_data_loop_prefetch): ...new pass_data instance and...
10608 (make_pass_loop_prefetch): ...new function.
10609 (pass_iv_optimize): Convert from a global struct to a subclass of
10610 gimple_opt_pass along with...
10611 (pass_data_iv_optimize): ...new pass_data instance and...
10612 (make_pass_iv_optimize): ...new function.
10613 (pass_tree_loop_done): Convert from a global struct to a subclass of
10614 gimple_opt_pass along with...
10615 (pass_data_tree_loop_done): ...new pass_data instance and...
10616 (make_pass_tree_loop_done): ...new function.
10617 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10618 struct to a subclass of gimple_opt_pass along with...
10619 (pass_data_cse_reciprocals): ...new pass_data instance and...
10620 (make_pass_cse_reciprocals): ...new function.
10621 (pass_cse_sincos): Convert from a global struct to a subclass of
10622 gimple_opt_pass along with...
10623 (pass_data_cse_sincos): ...new pass_data instance and...
10624 (make_pass_cse_sincos): ...new function.
10625 (pass_optimize_bswap): Convert from a global struct to a subclass of
10626 gimple_opt_pass along with...
10627 (pass_data_optimize_bswap): ...new pass_data instance and...
10628 (make_pass_optimize_bswap): ...new function.
10629 (pass_optimize_widening_mul): Convert from a global struct to a
10630 subclass of gimple_opt_pass along with...
10631 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10632 (make_pass_optimize_widening_mul): ...new function.
10633 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10634 subclass of gimple_opt_pass along with...
10635 (pass_data_phiopt): ...new pass_data instance and...
10636 (make_pass_phiopt): ...new function.
10637 (pass_cselim): Convert from a global struct to a subclass of
10638 gimple_opt_pass along with...
10639 (pass_data_cselim): ...new pass_data instance and...
10640 (make_pass_cselim): ...new function.
10641 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10642 subclass of gimple_opt_pass along with...
10643 (pass_data_phiprop): ...new pass_data instance and...
10644 (make_pass_phiprop): ...new function.
10645 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10646 subclass of gimple_opt_pass along with...
10647 (pass_data_pre): ...new pass_data instance and...
10648 (make_pass_pre): ...new function.
10649 (pass_fre): Convert from a global struct to a subclass of
10650 gimple_opt_pass along with...
10651 (pass_data_fre): ...new pass_data instance and...
10652 (make_pass_fre): ...new function.
10653 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10654 subclass of gimple_opt_pass along with...
10655 (pass_data_reassoc): ...new pass_data instance and...
10656 (make_pass_reassoc): ...new function.
10657 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10658 subclass of gimple_opt_pass along with...
10659 (pass_data_sink_code): ...new pass_data instance and...
10660 (make_pass_sink_code): ...new function.
10661 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10662 subclass of gimple_opt_pass along with...
10663 (pass_data_strlen): ...new pass_data instance and...
10664 (make_pass_strlen): ...new function.
10665 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10666 struct to a subclass of gimple_opt_pass along with...
10667 (pass_data_build_alias): ...new pass_data instance and...
10668 (make_pass_build_alias): ...new function.
10669 (pass_build_ealias): Convert from a global struct to a subclass of
10670 gimple_opt_pass along with...
10671 (pass_data_build_ealias): ...new pass_data instance and...
10672 (make_pass_build_ealias): ...new function.
10673 (pass_ipa_pta): Convert from a global struct to a subclass of
10674 simple_ipa_opt_pass along with...
10675 (pass_data_ipa_pta): ...new pass_data instance and...
10676 (make_pass_ipa_pta): ...new function.
10677 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10678 subclass of gimple_opt_pass along with...
10679 (pass_data_uncprop): ...new pass_data instance and...
10680 (make_pass_uncprop): ...new function.
10681 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10682 global struct to a subclass of gimple_opt_pass along with...
10683 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10684 (make_pass_late_warn_uninitialized): ...new function.
10685 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10686 to a subclass of gimple_opt_pass along with...
10687 (pass_data_init_datastructures): ...new pass_data instance and...
10688 (make_pass_init_datastructures): ...new function.
10689 (pass_early_warn_uninitialized): Convert from a global struct to a
10690 subclass of gimple_opt_pass along with...
10691 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10692 (make_pass_early_warn_uninitialized): ...new function.
10693 (pass_update_address_taken): Convert from a global struct to a
10694 subclass of gimple_opt_pass along with...
10695 (pass_data_update_address_taken): ...new pass_data instance and...
10696 (make_pass_update_address_taken): ...new function.
10697 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10698 struct to a subclass of gimple_opt_pass along with...
10699 (pass_data_release_ssa_names): ...new pass_data instance and...
10700 (make_pass_release_ssa_names): ...new function.
10701 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10702 subclass of gimple_opt_pass along with...
10703 (pass_data_stdarg): ...new pass_data instance and...
10704 (make_pass_stdarg): ...new function.
10705 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10706 global struct to a subclass of gimple_opt_pass along with...
10707 (pass_data_convert_switch): ...new pass_data instance and...
10708 (make_pass_convert_switch): ...new function.
10709 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10710 to a subclass of gimple_opt_pass along with...
10711 (pass_data_tail_recursion): ...new pass_data instance and...
10712 (make_pass_tail_recursion): ...new function.
10713 (pass_tail_calls): Convert from a global struct to a subclass of
10714 gimple_opt_pass along with...
10715 (pass_data_tail_calls): ...new pass_data instance and...
10716 (make_pass_tail_calls): ...new function.
10717 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10718 struct to a subclass of gimple_opt_pass along with...
10719 (pass_data_lower_vector): ...new pass_data instance and...
10720 (make_pass_lower_vector): ...new function.
10721 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10722 gimple_opt_pass along with...
10723 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10724 (make_pass_lower_vector_ssa): ...new function.
10725 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10726 to a subclass of gimple_opt_pass along with...
10727 (pass_data_slp_vectorize): ...new pass_data instance and...
10728 (make_pass_slp_vectorize): ...new function.
10729 (pass_ipa_increase_alignment): Convert from a global struct to a
10730 subclass of simple_ipa_opt_pass along with...
10731 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10732 (make_pass_ipa_increase_alignment): ...new function.
10733 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10734 gimple_opt_pass along with...
10735 (pass_data_vrp): ...new pass_data instance and...
10736 (make_pass_vrp): ...new function.
10737 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10738 subclass of simple_ipa_opt_pass along with...
10739 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10740 (make_pass_ipa_free_lang_data): ...new function.
10741 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10742 gimple_opt_pass along with...
10743 (pass_data_tsan): ...new pass_data instance and...
10744 (make_pass_tsan): ...new function.
10745 (pass_tsan_O0): Convert from a global struct to a subclass of
10746 gimple_opt_pass along with...
10747 (pass_data_tsan_O0): ...new pass_data instance and...
10748 (make_pass_tsan_O0): ...new function.
10749 * var-tracking.c (pass_variable_tracking): Convert from a global
10750 struct to a subclass of rtl_opt_pass along with...
10751 (pass_data_variable_tracking): ...new pass_data instance and...
10752 (make_pass_variable_tracking): ...new function.
10753 * web.c (pass_web): Convert from a global struct to a subclass of
10754 rtl_opt_pass along with...
10755 (pass_data_web): ...new pass_data instance and...
10756 (make_pass_web): ...new function.
10757 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10758 declaration with that of...
10759 (make_pass_mode_switch_use): ...new function.
10760 (pass_resolve_sw_modes): Replace declaration with that of...
10761 (make_pass_resolve_sw_modes): ...new function.
10762 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10763 from a global struct to a subclass of rtl_opt_pass along with...
10764 (pass_data_mode_switch_use): ...new pass_data instance and...
10765 (make_pass_mode_switch_use): ...new function.
10766 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10767 from a global struct to a subclass of rtl_opt_pass along with...
10768 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10769 (make_pass_resolve_sw_modes): ...new function.
10770 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10771 struct to a subclass of rtl_opt_pass along with...
10772 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10773 (make_pass_insert_vzeroupper): ...new function.
10774 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10775 global struct to a subclass of rtl_opt_pass along with...
10776 (pass_data_work_around_errata): ...new pass_data instance and...
10777 (make_pass_work_around_errata): ...new function.
10778 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10779 struct to a subclass of rtl_opt_pass along with...
10780 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10781 (make_pass_mips_machine_reorg2): ...new function.
10783 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10785 * passes.c (pass_manager::operator new): New.
10787 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10789 Handwritten part of conversion of passes to C++ classes.
10791 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10792 (toplev.o): Add dep on PASS_MANAGER_H.
10793 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10794 of early local pases to reflect this moving from a global to a
10795 member of gcc::pass_manager.
10796 (cgraph_add_new_function): Likewise.
10797 * lto-cgraph.c (lto_output_node): Update for conversion of
10798 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10799 * passes.c (opt_pass::clone): New.
10800 (opt_pass::gate): New.
10801 (opt_pass::execute): New.
10802 (opt_pass::opt_pass): New.
10803 (pass_manager::execute_early_local_passes): New.
10804 (pass_manager::execute_pass_mode_switching): new.
10805 (finish_optimization_passes): Convert to...
10806 (pass_manager::finish_optimization_passes): ...this.
10807 (finish_optimization_passes): Update for conversion of passes to
10809 (register_dump_files_1): Use has_gate since we cannot portably
10810 check a vtable entry against NULL.
10811 (dump_one_pass): Likewise.
10812 (ipa_write_summaries_2): Likewise.
10813 (ipa_write_optimization_summaries_1): Likewise.
10814 (ipa_read_summaries_1): Likewise.
10815 (ipa_read_optimization_summaries_1): Likewise.
10816 (execute_ipa_stmt_fixups): Likewise.
10817 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10818 pass-creation functions rather than wiring up globals, and
10819 storing the results in fields of pass_manager generated using
10820 pass-instances.def.
10821 (pass_manager::dump_profile_report): Update for conversion of
10822 passes to C++ classes.
10823 (pass_manager::execute_ipa_summary_passes): Likewise.
10824 (execute_one_ipa_transform_pass): Likewise.
10825 (execute_one_pass): Use has_gate and has_execute since we cannot
10826 portably check a vtable entry against NULL.
10827 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10828 (pass_manager): Use pass-instances.def to add fields for the
10829 various pass instances.
10830 * toplev.c (finalize): Update for move of
10831 finish_optimization_passes to a method of gcc::pass_manager.
10832 * toplev.h (finish_optimization_passes): Move to method of class
10834 * tree-pass.h (struct pass_data): New.
10835 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10836 (opt_pass::gate): Convert to virtual function.
10837 (opt_pass::~opt_pass): New.
10838 (opt_pass::clone): New.
10839 (opt_pass::execute): Convert to virtual function.
10840 (opt_pass::opt_pass): New.
10841 (opt_pass::ctxt_): new.
10842 (gimple_opt_pass): Convert to subclass of opt_pass.
10843 (gimple_opt_pass::gimple_opt_pass): New.
10844 (rtl_opt_pass): Convert to subclass of opt_pass.
10845 (rtl_opt_pass::rtl_opt_pass): New.
10846 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10847 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10848 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10849 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10850 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10851 invocation of pass_mode_switching to reflect this moving from a
10852 global to a member of gcc::pass_manager.
10853 (ix86_option_override): Rework how pass_insert_vzeroupper is
10854 added to the pass_manager to reflect autogenerated changes.
10855 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10857 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10859 PR rtl-optimization/57708
10860 * recog.c (peep2_find_free_register): Validate all regs in a
10863 2013-08-05 Jan Hubicka <jh@suse.cz>
10866 * cgraph.c (verify_cgraph_node): Accept local flags from other
10868 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10869 (function_and_variable_visibility): Likewise.
10870 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10872 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10874 * graph.c (init_graph_slim_pretty_print): Remove.
10875 (print_graph_cfg): Do not call it. Use local pretty printer.
10876 (start_graph_dump): Likewise.
10878 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10880 * gimple-pretty-print.c (buffer): Remove.
10881 (initialized): Likewise.
10882 (maybe_init_pretty_print): Likewise.
10883 (print_gimple_stmt): Do not call it. Use non-static local
10884 pretty_printer variable.
10885 (print_gimple_expr): Likewise.
10886 (print_gimple_seq): Likewise.
10887 (gimple_dump_bb): Likewise.
10889 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10891 * asan.c (asan_pp): Remove.
10892 (asan_pp_initialized): Likewise.
10893 (asan_pp_initialize): Likewise.
10894 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10895 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10896 (asan_add_global): Likewise.
10898 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10900 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10901 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10902 * pretty-print.c (pp_formatted_text_data): Likewise.
10903 (pp_indent): Rename from pp_base_indent.
10904 (pp_format): Rename from pp_base_format.
10905 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10906 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10907 (pp_flush): Rename from pp_base_flush.
10908 (pp_set_line_maximum_length): Rename from
10909 pp_base_set_line_maximum_length.
10910 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10911 (pp_set_prefix): Rename from pp_base_set_prefix.
10912 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10913 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10914 (pp_append_text): Rename from pp_base_append_text.
10915 (pp_formatted_text): Rename from pp_base_formatted_text.
10916 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10917 (pp_remaining_character_count_for_line): Rename from
10918 pp_base_remaining_character_count_for_line.
10919 (pp_newline): Rename from pp_base_newline.
10920 (pp_character): Rename from pp_base_character.
10921 (pp_string): Rename from pp_base_string.
10922 (pp_maybe_space): Rename from pp_base_maybe_space.
10923 * asan.c (asan_pp_string): Adjust.
10924 (asan_emit_stack_protection): Likewise.
10925 (asan_add_global): Likewise.
10926 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10927 * tree-mudflap.c (mf_varname_tree): Likewise.
10928 * tree-pretty-print.c (pp_tree_identifier): Rename from
10929 pp_base_tree_identifier.
10930 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10931 Declare as function.
10933 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10935 * pretty-print.h (pp_bar_bar): New.
10936 (pp_ampersand_ampersand): Likewise.
10937 (pp_less_equal): Likewise.
10938 (pp_greater_equal): Likewise.
10939 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10940 printer functions instead of pp_string or operators and punctuators.
10941 (dump_gimple_call): Likewise.
10942 (dump_gimple_omp_for): Likewise.
10943 (dump_gimple_transaction): Likewise.
10944 (dump_gimple_phi): Likewise.
10945 (pp_gimple_stmt_1): Likewise.
10946 * sched-vis.c (print_insn): Likewise.
10947 * tree-mudflap.c (mf_varname_tree): Likewise.
10948 * tree-pretty-print.c (dump_block_node): Likewise.
10949 (dump_generic_node): Likewise.
10951 2013-08-02 Jan Hubicka <jh@suse.cz>
10953 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10955 * lto-streamer-out.c (tree_is_indexable): Results decls and
10956 parm decls are not indexable.
10957 (DFS_write_tree_body): Do not follow args and results.
10958 (hash_tree): Likewise.
10959 (output_functions): Rearrange so struct function is needed
10960 only when real body is output; be able to also ouptut abstract
10961 functions; output DECL_ARGUMENTS and DECL_RESULT.
10962 (lto_output): When not in WPA, ale store abstract functions.
10963 (write_symbol): Do not care about RESULT_DECL.
10964 (output_symbol_p): Handle correctly sbtract decls.
10965 * lto-streamer-in.c (input_function): Rearrange so struct
10966 function can be NULL at entry; allow streaming of
10967 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10968 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10969 sanity check during LTO.
10970 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10971 RESULT_DECl and DECL_ARGUMENTS.
10972 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10975 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10977 * pretty-print.h (pp_underscore): New.
10979 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10980 printer functions instead of pp_character.
10981 (dump_binary_rhs): Likewise.
10982 (dump_ternary_rhs): Likewise.
10983 (dump_gimple_call_args): Likewise.
10984 (pp_points_to_solution): Likewise.
10985 (dump_gimple_call): Likewise.
10986 (dump_gimple_switch): Likewise.
10987 (dump_gimple_cond): Likewise.
10988 (dump_gimple_bind): Likewise.
10989 (dump_gimple_try): Likewise.
10990 (dump_gimple_omp_for): Likewise.
10991 (dump_gimple_omp_continue): Likewise.
10992 (dump_gimple_omp_single): Likewise.
10993 (dump_gimple_omp_sections): Likewise.
10994 (dump_gimple_omp_block): Likewise.
10995 (dump_gimple_omp_critical): Likewise.
10996 (dump_gimple_transaction): Likewise.
10997 (dump_gimple_asm): Likewise.
10998 (dump_gimple_phi): Likewise.
10999 (dump_gimple_omp_parallel): Likewise.
11000 (dump_gimple_omp_task): Likewise.
11001 (dump_gimple_omp_atomic_load): Likewise.
11002 (dump_gimple_omp_atomic_store): Likewise.
11003 (dump_gimple_mem_ops): Likewise.
11004 (pp_gimple_stmt_1): Likewise.
11005 (pp_cfg_jump): Likewise.
11006 (dump_implicit_edges): Likewise.
11007 (gimple_dump_bb_for_graph): Likewise.
11008 * graph.c (draw_cfg_node): Likewise.
11009 * langhooks.c (lhd_print_error_function): Likewise.
11010 * sched-vis.c (print_exp): Likewise.
11011 (print_value): Likewise.
11012 (print_pattern): Likewise.
11013 (print_insn): Likewise.
11014 (rtl_dump_bb_for_graph): Likewise.
11015 * tree-pretty-print.c (dump_function_declaration): Likewise.
11016 (dump_array_domain): Likewise.
11017 (dump_omp_clause): Likewise.
11018 (dump_location): Likewise.
11019 (dump_generic_node): Likewise.
11020 (print_struct_decl): Likewise.
11021 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11023 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11025 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11026 candidate table when replacing a candidate statement.
11027 (replace_rhs_if_not_dup): Likewise.
11028 (replace_one_candidate): Likewise.
11030 2013-08-02 Jan Hubicka <jh@suse.cz>
11031 Martin Liska <marxin.liska@gmail.com>
11033 * cgraphunit.c (add_new_function): Fix logic when adding from
11035 (assemble_thunk): Rename to ...
11036 (expand_thunk); .. this one; export; get it working with
11037 general functions; make produced gimple valid.
11038 * cgraph.h (expand_thunk): Declare.
11040 2013-08-02 Jan Hubicka <jh@suse.cz>
11042 * ipa-cp.c (gather_context_independent_values): Use
11043 ipa_get_param_move_cost.
11044 (get_replacement_map): Remove PARAM; move parameter folding
11046 (create_specialized_node): Update.
11047 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11048 assert that we have gimple body; update move_cost.
11049 (count_formal_params): Assert that we have gimple body.
11050 (ipa_dump_param): New function.
11051 (ipa_alloc_node_params): Break out from ...
11052 (ipa_initialize_node_params): ... here.
11053 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11054 (ipa_write_node_info): Stream move costs.
11055 (ipa_read_node_info): Read move costs.
11056 (ipa_update_after_lto_read): Do not recompute node params.
11057 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11058 (ipa_get_param): Check we are not in WPA.
11059 (ipa_get_param_move_cost): New.
11060 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11061 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11062 parm numbers to be present.
11064 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11066 PR rtl-optimization/58048
11067 * lra-constraints.c (process_alt_operands): Don't check asm
11068 operand on register.
11070 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11072 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11073 the implied StoreLoad barrier for atomic operations if before.
11075 2013-08-02 Jan Hubicka <jh@suse.cz>
11076 Martin Liska <marxin.liska@gmail.com>
11078 * cgraph.c (cgraph_function_body_availability): Do not check
11080 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11081 symtab_node_availability): Declare.
11082 * ipa.c (can_replace_by_local_alias): New.
11083 (function_and_variable_visibility): Use it.
11084 * symtab.c (symtab_for_node_and_aliases,
11085 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11087 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11089 PR rtl-optimization/57963
11090 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11091 (lra_constraints): Use them.
11093 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11095 * config/arm/types.md (define_attr "type"): Add "load_acq"
11097 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11099 (cortex_a53_store1): Likewise.
11101 2013-08-01 Jan Hubicka <jh@suse.cz>
11103 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11104 partitions are not needed.
11106 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11108 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11109 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11110 MAYBE_NON_Q_CLASS_P where appropriate.
11112 2013-08-01 Jan Hubicka <jh@suse.cz>
11114 * cgraph.h (release_function_body): Declare.
11115 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11116 of unused delcarations.
11118 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11120 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11121 RTL form when subtracting a constant.
11123 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11125 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11126 Generate canonical plus rtx with negated immediate instead of minus
11128 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11130 2013-08-01 Jan Hubicka <jh@suse.cz>
11132 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11133 (cgraph_release_function_body): Likewise.
11134 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11135 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11136 to used_as_abstract_origin.
11137 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11138 symbols used as abstract origins.
11139 * cgraphunit.c (analyze_functions): Update.
11140 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11141 used_as_abstract_origin.
11142 * tree-inline.c (tree_function_versioning): Update
11143 used_as_abstract_origin; be ready for DECL_RESULT and
11144 DECL_ARGUMENTS to be NULL.
11146 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11147 for abstract functions.
11148 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11151 2013-08-01 Jan Hubicka <jh@suse.cz>
11153 * profile.c (compute_value_histograms): Fix thinko.
11155 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11157 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11158 aarch-common-protos.h to extra_headers.
11159 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11160 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11161 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11163 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11165 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11166 Include "../arm/types.md". Define "type" attribute for all patterns.
11167 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11170 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11172 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11173 to support power8 load fusion.
11174 (fusion_gpr_mem_load): Likewise.
11176 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11178 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11179 declarations for power8 load fusion.
11180 (emit_fusion_gpr_load): Likewise.
11182 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11183 tuning for power8, turn on fusion mode by default. Turn on sign
11184 extending fusion mode if normal fusion mode is on, and we are at
11186 (fusion_gpr_load_p): New function, return true if we can fuse an
11187 addis instruction with a dependent load to a GPR.
11188 (emit_fusion_gpr_load): Emit the instructions for power8 load
11191 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11192 (VSX load fusion peepholes): New peepholes to fuse together an
11193 addi instruction with a VSX load instruction.
11195 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11196 peepholes to fuse an addis instruction with a load to a GPR base
11197 register. If we are supporting sign extending fusions, convert
11198 sign extending loads to zero extending loads and add an explicit
11201 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11203 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11204 aarch-common-protos.h to extra_headers.
11205 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11206 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11207 (arm_early_store_addr_dep): Likewise.
11208 (arm_no_early_alu_shift_dep): Likewise.
11209 (arm_no_early_alu_shift_value_dep): Likewise.
11210 (arm_no_early_mul_dep): Likewise.
11211 (arm_no_early_store_addr_dep): Likewise.
11212 (arm_mac_accumulator_is_mul_result): Likewise.
11213 (arm_mac_accumulator_is_result): Likewise.
11214 * config/arm/aarch-common.c: ... here. New file.
11215 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11217 (arm_early_store_addr_dep): Likewise.
11218 (arm_no_early_alu_shift_dep): Likewise.
11219 (arm_no_early_alu_shift_value_dep): Likewise.
11220 (arm_no_early_mul_dep): Likewise.
11221 (arm_no_early_store_addr_dep): Likewise.
11222 (arm_mac_accumulator_is_mul_result): Likewise.
11223 (arm_mac_accumulator_is_result): Likewise.
11224 * config/arm/aarch-common-protos.h: ... here. New file.
11225 * config/arm/t-arm (aarch-common.o): Define.
11227 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11229 * config/arm/arm.md: Include new file "types.md".
11230 (define_attr "type"): Move from here to ...
11231 (define_attr "mul32"): Likewise.
11232 (define_attr "mul64"): Likewise.
11233 * config/arm/types.md: ... here. New file.
11235 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11237 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11238 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11240 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11242 * gen-pass-instances.awk: Fix offset of substr().
11244 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11246 * Makefile.in (pass-instances.def): New.
11247 (passes.o): Replace dependency on passes.def with one on
11250 * gen-pass-instances.awk: New.
11252 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11253 rather than passes.def, updating local definition of NEXT_PASS
11254 macro to add an extra NUM parameter (currently unused).
11256 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11258 * Makefile.in (PASS_MANAGER_H): New.
11259 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11260 (passes.o): Likewise.
11261 (statistics.o): Likewise.
11262 (cgraphunit.o): Likewise.
11263 (context.o): Depend on PASS_MANAGER_H.
11265 * pass_manager.h: New.
11267 * cgraphunit.c (cgraph_add_new_function): Update for moves
11268 of globals to fields of pass_manager.
11269 (analyze_function): Likewise.
11270 (expand_function): Likewise.
11271 (ipa_passes): Likewise.
11272 (compile): Likewise.
11274 * context.c (context::context): New.
11275 * context.h (context::context): New.
11276 (context::get_passes): New.
11277 (context::passes_): New.
11279 * lto-cgraph.c (input_node): Update for moves of globals to
11280 fields of pass_manager.
11282 * passes.c (all_passes): Remove, in favor of a field of the
11283 same name within the new class pass_manager.
11284 (all_small_ipa_passes): Likewise.
11285 (all_lowering_passes): Likewise.
11286 (all_regular_ipa_passes): Likewise.
11287 (all_late_ipa_passes): Likewise.
11288 (all_lto_gen_passes): Likewise.
11289 (passes_by_id): Likewise.
11290 (passes_by_id_size): Likewise.
11291 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11292 the new class pass_manager.
11293 (set_pass_for_id): Convert to...
11294 (pass_manager::set_pass_for_id): ...method.
11295 (get_pass_for_id): Convert to...
11296 (pass_manager::get_pass_for_id): ...method.
11297 (register_one_dump_file): Move body of implementation into...
11298 (pass_manager::register_one_dump_file): ...here.
11299 (register_dump_files_1): Convert to...
11300 (pass_manager::register_dump_files_1): ...method.
11301 (register_dump_files): Convert to...
11302 (pass_manager::register_dump_files): ...method.
11303 (create_pass_tab): Update for moves of globals to fields of
11305 (dump_passes): Move body of implementation into...
11306 (pass_manager::dump_passes): ...here.
11307 (register_pass): Move body of implementation into...
11308 (pass_manager::register_pass): ...here.
11309 (init_optimization_passes): Convert into...
11310 (pass_manager::pass_manager): ...constructor for new
11311 pass_manager class, and initialize the pass_lists array.
11312 (check_profile_consistency): Update for moves of globals to
11313 fields of pass_manager.
11314 (dump_profile_report): Move body of implementation into...
11315 (pass_manager::dump_profile_report): ...here.
11316 (ipa_write_summaries_1): Update for moves of pass lists from
11317 being globals to fields of pass_manager.
11318 (ipa_write_optimization_summaries): Likewise.
11319 (ipa_read_summaries): Likewise.
11320 (ipa_read_optimization_summaries): Likewise.
11321 (execute_all_ipa_stmt_fixups): Likewise.
11323 * statistics.c (statistics_fini): Update for moves of globals to
11324 fields of pass_manager.
11326 * toplev.c (general_init): Replace call to
11327 init_optimization_passes with construction of the pass_manager
11330 * tree-pass.h (all_passes): Remove, in favor of a field of the
11331 same name within the new class pass_manager.
11332 (all_small_ipa_passes): Likewise.
11333 (all_lowering_passes): Likewise.
11334 (all_regular_ipa_passes): Likewise.
11335 (all_lto_gen_passes): Likewise.
11336 (all_late_ipa_passes): Likewise.
11337 (passes_by_id): Likewise.
11338 (passes_by_id_size): Likewise.
11339 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11340 the new class pass_manager.
11341 (get_pass_for_id): Remove.
11343 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11345 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11348 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11350 * arm.md (mulhi3): New expand pattern.
11352 2013-07-30 Jan Hubicka <jh@suse.cz>
11353 Martin Liska <marxin.liska@gmail.com>
11355 * profile.c (compute_value_histograms): Do not ICE when
11356 there is mismatch only on some counters.
11358 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11360 PR rtl-optimization/57637
11361 * function.c (move_insn_for_shrink_wrap): Also check the
11362 GEN set of the LIVE problem for the liveness analysis
11363 if it exists, otherwise give up.
11365 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11367 PR tree-optimization/57993
11368 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11369 replaced statement in the candidate table.
11370 (phi_add_costs): Return infinite cost when the hidden basis does
11371 not dominate all phis on which the candidate is dependent.
11372 (replace_one_candidate): Record replaced statement in the
11375 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11377 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11378 (ashlv2si3): New expander.
11379 (*ashlv2si3_i): New define_insn_and_split.
11380 * predicates.md (float_operation): Allow patterns with three
11381 basic sub-patterns.
11383 PR rtl-optimization/58021
11384 * mode-switching.c (create_pre_exit): Always split off preceding
11385 insns if we are not at the basic block head.
11387 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11389 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11390 (UCLIBC_DYNAMIC_LINKER): New macro.
11391 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11393 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11394 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11396 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11397 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11398 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11399 for SF and DF modes. Use ieee_quad_format for TF mode.
11400 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11401 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11402 (mips_option_override): Handle `-mnan=legacy'.
11403 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11404 `-mabs=2008' and `-mnan=2008'.
11405 (OPTION_DEFAULT_SPECS): Add "nan" default.
11406 (ASM_SPEC): Handle `-mnan='.
11407 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11408 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11409 comment accordingly.
11410 (neg<mode>2): Likewise.
11411 * config/mips/mips.opt (mabs, mnan): New options.
11412 * doc/install.texi (Configuration): Document `--with-nan=' option.
11413 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11415 (MIPS Options): Document them.
11416 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11417 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11418 * configure: Regenerate.
11419 * config.in: Regenerate.
11421 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11423 * config/i386/i386.md (float post-reload splitters): Do not check
11424 for subregs of SSE registers.
11426 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11427 H.J. Lu <hongjiu.lu@intel.com>
11431 * config/i386/i386.md (post-reload splitter
11432 to avoid partial SSE reg dependency stalls): New pattern.
11434 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11436 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11438 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11440 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11441 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11443 2013-07-26 Cary Coutant <ccoutant@google.com>
11445 * dwarf2out.c (die_checksum_ordered): Don't include template
11446 instantiations in signature.
11447 (is_template_parameter): New function.
11448 (is_template_instantiation): New function.
11449 (generate_skeleton_bottom_up): Don't include template instantiations
11451 (generate_skeleton): Likewise.
11452 (break_out_comdat_types): Move recursive call to break out nested
11454 (prune_unused_types_mark_generic_parms_dies): Call
11455 is_template_parameter.
11457 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11459 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11460 uses vector registers.
11461 * config/aarch64/iterators.md: Add attributes rtn and vas.
11463 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11464 Richard Earnshaw <richard.earnshaw@arm.com>
11466 * combine.c (simplify_comparison): Re-canonicalize operands
11468 * config/arm/arm.md (movcond_addsi): New splitter.
11470 2013-07-25 Sterling Augustine <saugustine@google.com>
11472 * dwarf2out.c (size_of_pubnames): Move code to...
11473 (include_pubname_in_output): ...here. New.
11474 (want_pubnames): Rearrange.
11475 (output_pubnames): Call include_pubname_in_output. Move assertion.
11477 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11479 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11481 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11484 * doc/extend.texi: Remove obsolete builtins. Fix
11485 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11487 2013-07-25 Jan Hubicka <jh@suse.cz>
11489 * cgraph.c (release_function_body): Break out from ...
11490 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11491 and DECL_ARGUMENTS.
11492 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11493 old_tree in the map.
11494 (create_specialized_node): Update.
11495 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11497 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11498 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11499 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11500 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11503 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11505 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11506 addsi3_carryin_alt2_<optab>): Correct output template.
11508 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11510 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11511 Adjust for arm_restrict_it.
11512 Remove trailing whitespace.
11514 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11516 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11517 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11519 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11521 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11523 PR rtl-optimization/57960
11524 * lra-constraints.c (process_alt_operands): Use the right mode
11525 when checking strict_low.
11527 2013-07-25 Jan Hubicka <jh@suse.cz>
11529 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11530 * cgraph.c (cgraph_remove_node): Do not release function body
11531 when in cgraph streaming.
11532 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11533 in other partitions are not considered reachable; fix handling of
11536 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11538 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11540 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11545 * config/arm/arm.md ("*sibcall_insn): Replace use of
11546 Ss with US. Adjust output for v5 and v4t.
11547 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11549 * config/arm/constraints.md ("Ss"): Rename to US.
11551 2013-07-25 Terry Guo <terry.guo@arm.com>
11553 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11554 shift_add/shift_sub0/shift_sub1 RTXs.
11556 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11557 Anton Blanchard <anton@au1.ibm.com>
11559 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11560 (altivec_vpks<VI_char>ss): Likewise.
11561 (altivec_vpks<VI_char>us): Likewise.
11562 (altivec_vpku<VI_char>us): Likewise.
11563 (altivec_vpku<VI_char>um): Likewise.
11565 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11567 Introduce context class.
11569 * Makefile.in (CONTEXT_H): New.
11570 (OBJS): Add context.o.
11571 (toplev.o): Add CONTEXT_H to dependencies.
11574 * toplev.c (general_init): Create the singleton gcc::context instance.
11580 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11582 PR rtl-optimization/57968
11583 * mode-switching.c (create_pre_exit): Allow instructions that
11584 don't set a return register to need a non-exit mode.
11586 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11587 Anton Blanchard <anton@au1.ibm.com>
11589 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11590 operands to vperm for little endian.
11591 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11592 of lvsl to create the control mask for a vperm for little endian.
11594 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11595 Anton Blanchard <anton@au1.ibm.com>
11597 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11598 two operands for little-endian.
11600 2013-07-23 Steve Ellcey <sellcey@mips.com>
11602 * config/mips/mips.c (mips_case_values_threshold): New.
11603 (TARGET_CASE_VALUES_THRESHOLD): Define.
11605 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11606 Anton Blanchard <anton@au1.ibm.com>
11608 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11609 selection of field for vector splat in little endian mode.
11611 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11613 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11614 expanders to rs6000.md.
11615 (ior<mode>3): Likewise.
11616 (and<mode>3): Likewise.
11617 (one_cmpl<mode>2): Likewise.
11618 (nor<mode>3): Likewise.
11619 (andc<mode>3): Likewise.
11620 (eqv<mode>3): Likewise.
11621 (nand<mode>3): Likewise.
11622 (orc<mode>3): Likewise.
11624 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11627 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11628 to split multi-word logical operations.
11629 (rs6000_split_logical_di): Likewise.
11630 (rs6000_split_logical): Likewise.
11632 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11633 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11634 and allow TImode operations in 32-bit.
11635 (vsx_and<mode>3_64bit): Likewise.
11636 (vsx_ior<mode>3_32bit): Likewise.
11637 (vsx_ior<mode>3_64bit): Likewise.
11638 (vsx_xor<mode>3_32bit): Likewise.
11639 (vsx_xor<mode>3_64bit): Likewise.
11640 (vsx_one_cmpl<mode>2_32bit): Likewise.
11641 (vsx_one_cmpl<mode>2_64bit): Likewise.
11642 (vsx_nor<mode>3_32bit): Likewise.
11643 (vsx_nor<mode>3_64bit): Likewise.
11644 (vsx_andc<mode>3_32bit): Likewise.
11645 (vsx_andc<mode>3_64bit): Likewise.
11646 (vsx_eqv<mode>3_32bit): Likewise.
11647 (vsx_eqv<mode>3_64bit): Likewise.
11648 (vsx_nand<mode>3_32bit): Likewise.
11649 (vsx_nand<mode>3_64bit): Likewise.
11650 (vsx_orc<mode>3_32bit): Likewise.
11651 (vsx_orc<mode>3_64bit): Likewise.
11653 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11654 logical types in GPRs.
11656 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11657 logical insns to rs6000.md, and allow TImode operations in
11659 (altivec_ior<mode>3): Likewise.
11660 (altivec_xor<mode>3): Likewise.
11661 (altivec_one_cmpl<mode>2): Likewise.
11662 (altivec_nor<mode>3): Likewise.
11663 (altivec_andc<mode>3): Likewise.
11665 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11666 attributes for moving the 128-bit logical operations into
11668 (BOOL_REGS_OUTPUT): Likewise.
11669 (BOOL_REGS_OP1): Likewise.
11670 (BOOL_REGS_OP2): Likewise.
11671 (BOOL_REGS_UNARY): Likewise.
11672 (BOOL_REGS_AND_CR0): Likewise.
11673 (one_cmpl<mode>2): Add support for DI logical operations on
11674 32-bit, splitting the operations to 32-bit.
11675 (anddi3): Likewise.
11676 (iordi3): Likewise.
11677 (xordi3): Likewise.
11678 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11679 changes to combine the 32/64-bit code, allow logical operations on
11680 TI mode in 32-bit, and to use similar match_operator patterns like
11681 scalar mode uses. Combine the Altivec and VSX code for logical
11682 operations, and move it here.
11683 (ior<mode>3, 128-bit types): Likewise.
11684 (xor<mode>3, 128-bit types): Likewise.
11685 (one_cmpl<mode>3, 128-bit types): Likewise.
11686 (nor<mode>3, 128-bit types): Likewise.
11687 (andc<mode>3, 128-bit types): Likewise.
11688 (eqv<mode>3, 128-bit types): Likewise.
11689 (nand<mode>3, 128-bit types): Likewise.
11690 (orc<mode>3, 128-bit types): Likewise.
11691 (and<mode>3_internal): Likewise.
11692 (bool<mode>3_internal): Likewise.
11693 (boolc<mode>3_internal1): Likewise.
11694 (boolc<mode>3_internal2): Likewise.
11695 (boolcc<mode>3_internal1): Likewise.
11696 (boolcc<mode>3_internal2): Likewise.
11697 (eqv<mode>3_internal1): Likewise.
11698 (eqv<mode>3_internal2): Likewise.
11699 (one_cmpl1<mode>3_internal): Likewise.
11701 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11703 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11704 Rename flag_stack_usage to flag_stack_usage_info.
11706 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11708 * config/microblaze/sync.md: New file.
11709 * config/microblaze/microblaze.md: Include sync.md
11710 * config/microblaze/microblaze.c: Add print_operand 'y'.
11711 * config/microblaze/constraints.md: Add memory_contraint
11712 'Q' which is a single register.
11714 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11716 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11718 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11720 * reload.c (find_reloads): Exit loop once we find this operand
11721 cannot be reloaded somehow for this alternative.
11723 * reload.c (find_reloads): Exit loop once we find a hard register.
11725 * rtlanal.c (computed_jump_p): Exit loop once we find label
11728 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11730 * cfgloopmanip.c (remove_path): Exit loop after setting
11733 * gensupport.c (subst_dup): Avoid loop if code is not
11734 MATCH_DUP nor MATCH_OP_DUP.
11736 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11738 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11740 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11742 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11743 true for SP_REGNUM if mode == ptr_mode.
11744 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11745 with value R0_REGNUM + 31.
11747 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11749 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11750 pad pointer-typed argument downward.
11752 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11754 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11755 and __ILP32__ when the ILP32 model is in use.
11757 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11759 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11760 (aarch64_load_symref_appropriately): In the case of
11761 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11762 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11763 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11764 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11765 if 'mode' doesn't equal to 'ptr_mode'.
11766 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11767 'vcall_offset'; change to call aarch64_emit_move differently depending
11768 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11769 to calculate the upper bound of 'vcall_offset'.
11770 (aarch64_cannot_force_const_mem): Change to also return true if
11772 (aarch64_legitimize_reload_address): In the case of large
11773 displacements, add new local variable 'xmode' and an assertion
11774 based on it; change to use 'xmode' to generate the new rtx and
11776 (aarch64_asm_trampoline_template): Change to generate the template
11777 differently depending on TARGET_ILP32 or not; change to use
11778 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11779 (aarch64_trampoline_size): Removed.
11780 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11781 and replace immediate literals with it. Change to use 'ptr_mode'
11782 instead of 'DImode' and call convert_memory_address if the mode
11783 of 'fnaddr' doesn't equal to 'ptr_mode'.
11784 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11786 (aarch64_elf_asm_destructor): Likewise.
11787 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11788 on TARGET_ILP32 instead of aarch64_trampoline_size.
11789 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11790 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11791 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11792 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11793 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11794 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11795 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11796 depending on the value of 'mode'.
11797 (add_losym_<mode>): New.
11798 (ldr_got_small_<mode>): New, based on ldr_got_small.
11799 (ldr_got_small): Remove.
11800 (ldr_got_small_sidi): New.
11801 * config/aarch64/iterators.md (P): New.
11802 (PTR): Change to 'ptr_mode' in the condition.
11804 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11806 * config.gcc (aarch64*-*-*): Support --with-abi.
11807 (aarch64*-*-elf): Support --with-multilib-list.
11808 (aarch64*-*-linux*): Likewise.
11809 (supported_defaults): Add abi to aarch64*-*-*.
11810 * configure.ac: Mention AArch64 for --with-multilib-list.
11811 * configure: Re-generated.
11812 * config/aarch64/biarchilp32.h: New file.
11813 * config/aarch64/biarchlp64.h: New file.
11814 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11816 (MULTILIB_DEFAULTS): Ditto.
11817 (DRIVER_SELF_SPECS): Ditto.
11818 (ASM_SPEC): Update to also substitute -mabi.
11819 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11820 file whose name depends on -mabi= and -mbig-endian.
11821 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11823 (POINTER_SIZE): New define.
11824 (POINTERS_EXTEND_UNSIGNED): Ditto.
11825 (enum aarch64_abi_type): New enumeration tag.
11826 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11827 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11828 (TARGET_ILP32): New define.
11829 * config/aarch64/aarch64.opt (mabi): New.
11830 (aarch64_abi): New.
11831 (ilp32, lp64): New values for -mabi.
11832 * config/aarch64/t-aarch64 (comma): New define.
11833 (MULTILIB_OPTIONS): Ditto.
11834 (MULTILIB_DIRNAMES): Ditto.
11835 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11836 * doc/invoke.texi: Document -mabi for AArch64.
11838 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11840 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11842 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11843 Anton Blanchard <anton@au1.ibm.com>
11845 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11846 endianness when selecting field to splat.
11848 2013-07-22 Eric Christopher <echristo@gmail.com>
11850 * dwarf2out.c (die_odr_checksum): New function to use
11851 CHECKSUM_ macros and ULEB128 for DIE tag.
11852 (generate_type_signature): Use.
11854 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11856 * config.gcc (sparc*-*-*): Accept leon3 processor.
11857 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11858 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11859 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11860 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11861 (mfix-ut699): Adjust comment.
11862 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11863 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11864 (CPP_CPU_SPEC): Likewise.
11865 (ASM_CPU_SPEC): Likewise.
11866 * config/sparc/sparc.c (leon3_cost): New constant.
11867 (sparc_option_override): Add leon3 support.
11868 (mem_ref): New function.
11869 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11870 (sparc_do_work_around_errata): Look into the instruction in the delay
11871 slot and adjust accordingly. Add fix for the data cache nullify issues
11872 of the UT699. Change insertion position for the NOP.
11873 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11874 (leon3_load): New reservation.
11875 (leon_store): Bump latency to 2.
11876 (grfpu): New automaton.
11877 (grfpu_alu): New unit.
11878 (grfpu_ds): Likewise.
11879 (leon_fp_alu): Adjust.
11880 (leon_fp_mult): Delete.
11881 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11882 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11883 * config/sparc/sparc.md (cpu): Add leon3.
11884 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11885 (swapsi): Likewise.
11886 (atomic_test_and_set): Likewise.
11887 (ldstub): Likewise.
11889 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11891 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11892 default for R5900 targets.
11893 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11894 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11895 * config/mips/mips.c (mips_option_override): Report an error for
11896 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11897 for -march=r5900 -mhard-float.
11899 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11901 * df-problems.c (can_move_insns_across): Exit loop once we
11902 find a non-fixed, non-global register.
11904 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11907 * omega.c (omega_eliminate_red): Break after setting red_found.
11908 (omega_problem_has_red_equations): Similarly after setting found.
11909 (omega_query_variable): Similarly after setting coupled.
11911 2013-07-22 Marek Polacek <polacek@redhat.com>
11913 * gimplify.c: Don't include gimple.h twice.
11915 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11917 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11918 instead of TARGET_THUMB1.
11919 (Pz): New constraint.
11920 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11922 (compare_negsi_si): Likewise.
11923 (compare_addsi2_op0): Likewise.
11924 (compare_addsi2_op1): Likewise.
11925 (addsi3_carryin_<optab>): Likewise.
11926 (addsi3_carryin_alt2_<optab>): Likewise.
11927 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11928 for arm_restrict_it.
11929 (subsi3_carryin): Likewise.
11930 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11931 (minmax_arithsi): Disable for arm_restrict_it.
11932 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11933 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11934 (satsi_<SAT:code>_shift): Likewise.
11935 (arm_shiftsi3): Add alternative for 16-bit encoding.
11936 (arm32_movhf): Disable for arm_restrict_it.
11937 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11938 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11940 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11942 * config/arm/arm.md (attribute "insn"): Delete.
11943 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11944 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11945 (not_shiftsi): Update for attribute change.
11946 (not_shiftsi_compare0): Likewise.
11947 (not_shiftsi_compare0_scratch): Likewise.
11948 (arm_one_cmplsi2): Likewise.
11949 (thumb1_one_cmplsi2): Likewise.
11950 (notsi_compare0): Likewise.
11951 (notsi_compare0_scratch): Likewise.
11952 (thumb1_movdi_insn): Likewise.
11953 (arm_movsi_insn): Likewise.
11954 (movhi_insn_arch4): Likewise.
11955 (movhi_bytes): Likewise.
11956 (arm_movqi_insn): Likewise.
11957 (thumb1_movqi_insn): Likewise.
11958 (arm32_movhf): Likewise.
11959 (thumb1_movhf): Likewise.
11960 (arm_movsf_soft_insn): Likewise.
11961 (thumb1_movsf_insn): Likewise.
11962 (thumb_movdf_insn): Likewise.
11963 (movsicc_insn): Likewise.
11964 (movsfcc_soft_insn): Likewise.
11965 (and_scc): Likewise.
11966 (cond_move): Likewise.
11967 (if_move_not): Likewise.
11968 (if_not_move): Likewise.
11969 (if_shift_move): Likewise.
11970 (if_move_shift): Likewise.
11971 (if_shift_shift): Likewise.
11972 (if_not_arith): Likewise.
11973 (if_arith_not): Likewise.
11974 (cond_move_not): Likewise.
11975 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11976 (neon_mov<mode>): Likewise.
11977 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11978 (thumb2_movsi_vfp): Likewise.
11979 (movsf_vfp): Likewise.
11980 (thumb2_movsf_vfp): Likewise.
11981 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11983 (cortexa7_older_only): Likewise.
11984 (cortexa7_younger): Likewise.
11985 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11986 (1020alu_shift_op): Likewise.
11987 (1020alu_shift_reg_op): Likewise.
11988 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11989 (alu_shift_op): Likewise.
11990 (alu_shift_reg_op): Likewise.
11991 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11992 (11_alu_shift_op): Likewise.
11993 (11_alu_shift_reg_op): Likewise.
11994 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11995 (9_alu_shift_reg_op): Likewise.
11996 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11998 (cortex_a15_alu_shift): Likewise.
11999 (cortex_a15_alu_shift_reg): Likewise.
12000 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12001 (cortex_a5_alu_shift): Likewise.
12002 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12004 (cortex_a53_alu_shift): Likewise.
12005 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12007 (cortex_a7_alu_reg): Likewise.
12008 (cortex_a7_alu_shift): Likewise.
12009 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12010 (cortex_a8_alu_shift): Likewise.
12011 (cortex_a8_alu_shift_reg): Likewise.
12012 (cortex_a8_mov): Likewise.
12013 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12014 (cortex_a9_dp_shift): Likewise.
12015 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12016 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12017 (cortex_r4_mov): Likewise.
12018 (cortex_r4_alu_shift): Likewise.
12019 (cortex_r4_alu_shift_reg): Likewise.
12020 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12021 (526_alu_shift_op): Likewise.
12022 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12023 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12024 (626te_alu_shift_op): Likewise.
12025 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12026 (726te_alu_op): Likewise.
12027 (726te_alu_shift_op): Likewise.
12028 (726te_alu_shift_reg_op): Likewise.
12029 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12030 (mp626_alu_shift_op): Likewise.
12031 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12032 (pj4_alu_e1_conds): Likewise.
12033 (pj4_alu): Likewise.
12034 (pj4_alu_conds): Likewise.
12035 (pj4_shift): Likewise.
12036 (pj4_shift_conds): Likewise.
12037 (pj4_alu_shift): Likewise.
12038 (pj4_alu_shift_conds): Likewise.
12040 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12042 * config/arm/predicates.md (shiftable_operator_strict_it):
12044 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12045 Disable cond_exec version for arm_restrict_it.
12046 (thumb2_smaxsi3): Convert to generate cond_exec.
12047 (thumb2_sminsi3): Likewise.
12048 (thumb32_umaxsi3): Likewise.
12049 (thumb2_uminsi3): Likewise.
12050 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12051 (thumb2_neg_abssi2): Likewise.
12052 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12053 (thumb2_movsicc_insn): Adjust alternatives.
12054 (thumb2_mov_negscc): Disable for arm_restrict_it.
12055 (thumb2_mov_negscc_strict_it): New pattern.
12056 (thumb2_mov_notscc_strict_it): New pattern.
12057 (thumb2_mov_notscc): Disable for arm_restrict_it.
12058 (thumb2_ior_scc): Likewise.
12059 (thumb2_ior_scc_strict_it): New pattern.
12060 (thumb2_cond_move): Adjust for arm_restrict_it.
12061 (thumb2_cond_arith): Disable for arm_restrict_it.
12062 (thumb2_cond_arith_strict_it): New pattern.
12063 (thumb2_cond_sub): Adjust for arm_restrict_it.
12064 (thumb2_movcond): Likewise.
12065 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12066 (thumb2_zero_extendhisi2_v6): Likewise.
12067 (thumb2_zero_extendqisi2_v6): Likewise.
12068 (orsi_notsi_si): Likewise.
12069 (orsi_not_shiftsi_si): Likewise.
12071 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12073 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12074 instruction sequence is 1 byte shorter.
12076 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12078 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12079 it is not needed after split.
12081 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12084 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12085 second label for nonlocal goto receivers. Don't output pic base labels
12086 unless we're producing PIC; mark that action unreachable().
12087 (ix86_save_reg): If the function contains a nonlocal label, save the
12089 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12090 * config/darwin.c (emitted_pic_label_num): New GTY.
12091 (update_pic_label_number_if_needed): New.
12092 (machopic_output_function_base_name): Adjust for nonlocal receiver
12094 (machopic_should_output_picbase_label): New.
12095 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12096 (nonlocal_goto_receiver): New insn and split.
12098 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12100 * config/aarch64/aarch64-builtins.c
12101 (aarch64_fold_builtin): Fold abs in all modes.
12102 * config/aarch64/aarch64-simd-builtins.def
12103 (abs): Enable for all modes.
12104 * config/aarch64/arm_neon.h
12105 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12106 (vabs_f64): Add missing intrinsic.
12108 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12110 * config/aarch64/arm_neon.h (vabs_s64): New function
12112 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12115 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12116 * config/avr/avr.md (adjust_len): Add `round'.
12117 * config/avr/avr-protos.h (avr_out_round): New prototype.
12118 (avr_out_plus): Add `out_label' argument.
12119 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12120 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12121 Handle the case where `insn' is just a pattern.
12122 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12123 (avr_out_round): New function.
12124 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12126 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12128 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12129 Add check for flag_stack_usage to handle -fstack-usage support
12131 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12133 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12134 interaction for new Power8 flags and VSX.
12136 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12138 PR middle-end/57698
12139 * tree-inline.c (expand_call_inline): Emit errors during
12140 early_inlining only if optimization is not turned on.
12142 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12146 * passes.c (init_optimization_passes): Move the construction of
12147 the pass hierarchy into a new passes.def file.
12149 * Makefile.in (passes.o): Add dependency on passes.def.
12151 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12153 * passes.c (init_optimization_passes): Introduce macros for
12154 constructing the tree of passes (INSERT_PASSES_AFTER,
12155 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12156 TERMINATE_PASS_LIST).
12158 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12159 Wei Mi <wmi@google.com>
12161 PR rtl-optimization/57878
12162 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12164 (reload_pseudo_compare_func): Check nregs first for reload
12167 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12169 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12171 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12173 * read-rtl.c (validate_const_int): Once an invalid character is
12174 seen, quit the loop.
12176 * gengtype.c (write_roots): Similarly once we find the "deletable"
12177 or "if_marked" option.
12179 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12181 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12182 "xtab" and "sat". Move value "clz" from here to ...
12183 (attriubte "type"): ... here.
12184 (satsi_<SAT:code>): Delete "insn" attribute.
12185 (satsi_<SAT:code>_shift): Likewise.
12186 (arm_zero_extendqisi2addsi): Likewise.
12187 (arm_extendqisi2addsi): Likewise.
12188 (clzsi2): Update for attribute changes.
12189 (rbitsi2): Likewise.
12190 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12192 (arm_usatsihi): Likewise.
12193 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12195 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12197 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12198 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12199 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12200 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12201 in alphabetical order.
12202 (attribute "core_cycles"): Update for attribute changes.
12203 (arm_addsi3): Likewise.
12204 (addsi3_compare0): Likewise.
12205 (addsi3_compare0_scratch): Likewise.
12206 (addsi3_compare_op1): Likewise.
12207 (addsi3_compare_op2): Likewise.
12208 (compare_addsi2_op0): Likewise.
12209 (compare_addsi2_op1): Likewise.
12210 (addsi3_carryin_shift_<optab>): Likewise.
12211 (subsi3_carryin_shift): Likewise.
12212 (rsbsi3_carryin_shift): Likewise.
12213 (arm_subsi3_insn): Likewise.
12214 (subsi3_compare0): Likewise.
12215 (subsi3_compare): Likewise.
12216 (arm_andsi3_insn): Likewise.
12217 (thumb1_andsi3_insn): Likewise.
12218 (andsi3_compare0): Likewise.
12219 (andsi3_compare0_scratch): Likewise.
12220 (zeroextractsi_compare0_scratch
12221 (andsi_not_shiftsi_si): Likewise.
12222 (iorsi3_insn): Likewise.
12223 (iorsi3_compare0): Likewise.
12224 (iorsi3_compare0_scratch): Likewise.
12225 (arm_xorsi3): Likewise.
12226 (thumb1_xorsi3_insn): Likewise.
12227 (xorsi3_compare0): Likewise.
12228 (xorsi3_compare0_scratch): Likewise.
12229 (satsi_<SAT:code>_shift): Likewise.
12231 (arm_shiftsi3): Likewise.
12232 (shiftsi3_compare0): Likewise.
12233 (not_shiftsi): Likewise.
12234 (not_shiftsi_compare0): Likewise.
12235 (not_shiftsi_compare0_scratch): Likewise.
12236 (arm_one_cmplsi2): Likewise.
12237 (thumb_one_complsi2): Likewise.
12238 (notsi_compare0): Likewise.
12239 (notsi_compare0_scratch): Likewise.
12240 (thumb1_zero_extendhisi2): Likewise.
12241 (arm_zero_extendhisi2): Likewise.
12242 (arm_zero_extendhisi2_v6): Likewise.
12243 (arm_zero_extendhisi2addsi): Likewise.
12244 (thumb1_zero_extendqisi2): Likewise.
12245 (thumb1_zero_extendqisi2_v6): Likewise.
12246 (arm_zero_extendqisi2): Likewise.
12247 (arm_zero_extendqisi2_v6): Likewise.
12248 (arm_zero_extendqisi2addsi): Likewise.
12249 (thumb1_extendhisi2): Likewise.
12250 (arm_extendhisi2): Likewise.
12251 (arm_extendhisi2_v6): Likewise.
12252 (arm_extendqisi): Likewise.
12253 (arm_extendqisi_v6): Likewise.
12254 (arm_extendqisi2addsi): Likewise.
12255 (thumb1_extendqisi2): Likewise.
12256 (thumb1_movdi_insn): Likewise.
12257 (arm_movsi_insn): Likewise.
12258 (movsi_compare0): Likewise.
12259 (movhi_insn_arch4): Likewise.
12260 (movhi_bytes): Likewise.
12261 (arm_movqi_insn): Likewise.
12262 (thumb1_movqi_insn): Likewise.
12263 (arm32_movhf): Likewise.
12264 (thumb1_movhf): Likewise.
12265 (arm_movsf_soft_insn): Likewise.
12266 (thumb1_movsf_insn): Likewise.
12267 (movdf_soft_insn): Likewise.
12268 (thumb_movdf_insn): Likewise.
12269 (arm_cmpsi_insn): Likewise.
12270 (cmpsi_shiftsi): Likewise.
12271 (cmpsi_shiftsi_swp): Likewise.
12272 (arm_cmpsi_negshiftsi_si): Likewise.
12273 (movsicc_insn): Likewise.
12274 (movsfcc_soft_insn): Likewise.
12275 (arith_shiftsi): Likewise.
12276 (arith_shiftsi_compare0
12277 (arith_shiftsi_compare0_scratch
12278 (sub_shiftsi): Likewise.
12279 (sub_shiftsi_compare0
12280 (sub_shiftsi_compare0_scratch
12281 (and_scc): Likewise.
12282 (cond_move): Likewise.
12283 (if_plus_move): Likewise.
12284 (if_move_plus): Likewise.
12285 (if_move_not): Likewise.
12286 (if_not_move): Likewise.
12287 (if_shift_move): Likewise.
12288 (if_move_shift): Likewise.
12289 (if_shift_shift): Likewise.
12290 (if_not_arith): Likewise.
12291 (if_arith_not): Likewise.
12292 (cond_move_not): Likewise.
12293 (thumb1_ashlsi3): Set type attribute.
12294 (thumb1_ashrsi3): Likewise.
12295 (thumb1_lshrsi3): Likewise.
12296 (thumb1_rotrsi3): Likewise.
12297 (shiftsi3_compare0_scratch): Likewise.
12298 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12299 (neon_mov<mode>): Likewise.
12300 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12302 (thumb2_movsi_insn): Likewise.
12303 (thumb2_cmpsi_neg_shiftsi): Likewise.
12304 (thumb2_extendqisi_v6): Likewise.
12305 (thumb2_zero_extendhisi2_v6): Likewise.
12306 (thumb2_zero_extendqisi2_v6): Likewise.
12307 (thumb2_shiftsi3_short): Likewise.
12308 (thumb2_addsi3_compare0_scratch): Likewise.
12309 (orsi_not_shiftsi_si): Likewise.
12310 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12311 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12313 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12314 (1020alu_shift_op): Likewise.
12315 (1020alu_shift_reg_op): Likewise.
12316 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12317 (alu_shift_op): Likewise.
12318 (alu_shift_reg_op): Likewise.
12319 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12320 (11_alu_shift_op): Likewise.
12321 (11_alu_shift_reg_op): Likewise.
12322 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12323 (9_alu_shift_reg_op): Likewise.
12324 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12326 (cortex_a15_alu_shift): Likewise.
12327 (cortex_a15_alu_shift_reg): Likewise.
12328 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12330 (cortex_a5_alu_shift): Likewise.
12331 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12333 (cortex_a53_alu_shift): Likewise.
12334 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12336 (cortex_a7_alu_reg): Likewise.
12337 (cortex_a7_alu_shift): Likewise.
12338 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12340 (cortex_a8_alu_shift): Likewise.
12341 (cortex_a8_alu_shift_reg): Likewise.
12342 (cortex_a8_mov): Likewise.
12343 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12344 (cortex_a9_dp_shift): Likewise.
12345 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12347 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12349 (cortex_r4_mov): Likewise.
12350 (cortex_r4_alu_shift): Likewise.
12351 (cortex_r4_alu_shift_reg): Likewise.
12352 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12353 (526_alu_shift_op): Likewise.
12354 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12355 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12356 (626te_alu_shift_op): Likewise.
12357 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12358 (726te_alu_op): Likewise.
12359 (726te_alu_shift_op): Likewise.
12360 (726te_alu_shift_reg_op): Likewise.
12361 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12362 (mp626_alu_shift_op): Likewise.
12363 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12364 (pj4_alu_e1_conds): Likewise.
12365 (pj4_alu): Likewise.
12366 (pj4_alu_conds): Likewise.
12367 (pj4_shift): Likewise.
12368 (pj4_shift_conds): Likewise.
12369 (pj4_alu_shift): Likewise.
12370 (pj4_alu_shift_conds): Likewise.
12371 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12373 (cortexa7_older_only): Likewise.
12374 (cortexa7_younger): Likewise.
12376 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12378 * ipa-pure-const.c (generate_summary): Rename to...
12379 (pure_const_generate_summary): ... this.
12381 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12383 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12385 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12388 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12391 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12393 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12394 enabled without -march=zEC12.
12395 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12398 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12400 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12401 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12402 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12403 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12404 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12405 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12406 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12408 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12409 and ISA_HAS_NMADD3_NMSUB3.
12410 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12411 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12412 (nmsub4<mode>, nmsub3<mode>): Likewise.
12413 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12415 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12417 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12418 TARGET_MIPS5400 checking.
12420 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12421 Peter Bergner <bergner@vnet.ibm.com>
12423 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12424 registers in the comment.
12425 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12426 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12427 rather than FIRST_PSEUDO_REGISTERS.
12429 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12431 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12432 enable extra ISA flags with TARGET_HTM.
12434 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12436 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12439 2013-07-15 Cong Hou <congh@google.com>
12441 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12442 in compare function for sorting.
12444 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12446 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12447 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12448 * config/rs6000/rs6000.opt: Add -mhtm option.
12449 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12450 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12451 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12452 __HTM__ if the HTM instructions are available.
12453 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12454 htm_spr_reg_operand): New define_predicates.
12455 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12456 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12458 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12459 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12460 HTM builtin functions.
12461 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12462 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12463 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12464 (rs6000_builtin_mask_calculate): Likewise.
12465 (rs6000_option_override_internal): Likewise.
12466 (bdesc_htm): Add new HTM builtin support.
12467 (htm_spr_num): New function.
12468 (htm_spr_regno): Likewise.
12469 (rs6000_htm_spr_icode): Likewise.
12470 (htm_expand_builtin): Likewise.
12471 (htm_init_builtins): Likewise.
12472 (rs6000_expand_builtin): Add support for HTM builtin functions.
12473 (rs6000_init_builtins): Likewise.
12474 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12476 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12477 (TARGET_HTM, MASK_HTM): Define macros.
12478 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12479 (FIXED_REGISTERS): Likewise.
12480 (CALL_USED_REGISTERS): Likewise.
12481 (CALL_REALLY_USED_REGISTERS): Likewise.
12482 (REG_ALLOC_ORDER): Likewise.
12483 (enum reg_class): Likewise.
12484 (REG_CLASS_NAMES): Likewise.
12485 (REG_CLASS_CONTENTS): Likewise.
12486 (REGISTER_NAMES): Likewise.
12487 (ADDITIONAL_REGISTER_NAMES): Likewise.
12488 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12489 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12490 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12491 * config/rs6000/htm.md: New file.
12492 * config/rs6000/htmintrin.h: New file.
12493 * config/rs6000/htmxlintrin.h: New file.
12495 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12497 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12498 Define SYMBOL_TINY_GOT, update comment.
12499 * config/aarch64/aarch64.c
12500 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12501 (aarch64_expand_mov_immediate): Likewise.
12502 (aarch64_print_operand): Likewise.
12503 (aarch64_classify_symbol): Likewise.
12504 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12505 (ldr_got_tiny): Define.
12507 2013-07-13 Tobias Grosser <tobias@grosser.es>
12509 PR tree-optimization/54094
12510 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12511 scheduling dimension for the parallelism check from the polyhedral
12512 information in the AST.
12513 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12516 2013-07-13 Jason Merrill <jason@redhat.com>
12518 * print-tree.c (debug_vec_tree): Use debug_raw.
12519 (debug_raw (vec<tree, va_gc> &)): New.
12520 (debug_raw (vec<tree, va_gc> *)): New.
12521 * tree.h: Declare them.
12523 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12525 * ifcvt.c (ifcvt_after_combine): New static variable.
12526 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12528 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12529 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12530 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12532 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12534 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12536 2013-07-12 Michael Matz <matz@suse.de>
12538 PR middle-end/55771
12539 * convert.c (convert_to_real): Reject non-float inner types.
12541 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12543 * config/aarch64/aarch64-protos.h
12544 (aarch64_simd_immediate_valid_for_move): Remove.
12545 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12546 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12548 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12550 2013-07-11 Steve Ellcey <sellcey@mips.com>
12552 * config/mips/mips.c (mips_conditional_register_usage): Do not
12553 use t[0-7] registers in MIPS16 mode when optimizing for size.
12555 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12557 * config/i386/i386.c (dispatch_function_versions): Fix array
12558 indexing of function_version_info to match actual_versions.
12560 2013-07-11 Teresa Johnson <tejohnson@google.com>
12562 * vec.h (struct va_gc): Move release out-of-line.
12563 (va_gc::release): Call ggc_free on released vec.
12565 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12567 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12568 Require GOT register as additional operand in UNSPEC.
12569 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12570 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12571 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12572 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12573 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12574 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12575 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12577 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12580 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12581 name seen by assembler/linker rather if available.
12583 2013-07-11 Andreas Schwab <schwab@suse.de>
12585 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12587 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12589 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12591 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12593 * doc/tm.texi.in: Move hook documentation to ....
12594 * target.def: ... here.
12596 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12597 text on @hook line.
12598 * doc/tm.texi: Regenerate.
12600 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12603 * doc/invoke.texi: Document -Wconditionally-supported.
12605 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12608 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12611 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12614 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12615 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12616 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12617 Remove duplicate devices.
12618 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12619 * config/avr/t-multilib: Regenerate.
12620 * config/avr/avr-tables.opt: Regenerate.
12621 * doc/avr-mmcu.texi: Regenerate.
12623 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12626 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12628 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12630 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12631 the cost of MULT when optimizing for size.
12633 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12635 * config/cr16/cr16-protos.h: Don't include target.h.
12637 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12639 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12640 adjust register size for TDmode and TFmode for VSX registers.
12642 2013-07-08 Kai Tietz <ktietz@redhat.com>
12645 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12646 hook_bool_const_tree_true.
12648 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12650 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12651 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12652 * config/s390/s390.md: Define FPR*_REGNUM constants.
12653 Fix FPR2_REGNUM constant (18 -> 17).
12654 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12655 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12656 ("trunc<BFP:mode><DFP_ALL:mode>2")
12657 ("trunc<DFP_ALL:mode><BFP:mode>2")
12658 ("*extend<BFP:mode><DFP_ALL:mode>2")
12659 ("*extend<DFP_ALL:mode><BFP:mode>2")
12660 ("extend<BFP:mode><DFP_ALL:mode>2")
12661 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12664 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12666 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12668 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12670 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12671 and cfun_fpr_bit_p to cfun_fpr_save_p.
12672 (s390_frame_area, s390_register_info, s390_frame_info)
12673 (s390_emit_prologue, s390_emit_epilogue)
12674 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12676 * config/s390/s390.h: Define *_REGNUM macros for floating point
12679 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12681 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12683 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12685 PR rtl-optimization/57786
12686 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12687 and break out of the loop when it is set to false.
12689 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12692 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12693 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12694 (const_int 63)) 0)).
12695 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12696 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12697 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12699 PR rtl-optimization/57829
12700 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12701 mask bits outside of mode are just sign-extension from mode to HWI.
12703 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12705 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12706 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12707 adjust_address instead of change_address to keep info about alignment.
12708 (emit_strmov): Remove.
12709 (emit_memmov): New function.
12710 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12711 (expand_movmem_epilogue): Likewise and return updated rtx for
12713 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12714 destination and source.
12715 (decide_alignment): Refactor, handle vector_loop.
12716 (ix86_expand_movmem): Likewise.
12717 (ix86_expand_setmem): Likewise.
12718 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12720 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12722 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12723 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12725 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12727 * config/i386/sse.md (sse_movlhps): Change alternative 3
12728 of operand 2 to "m".
12730 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12733 * config/i386/sse.md (iptr): New mode attribute.
12734 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12735 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12736 (<sse>_vmmul<mode>3): Ditto.
12737 (<sse>_vmdiv<mode>3): Ditto.
12738 (sse_vmrcpv4sf2): Ditto.
12739 (<sse>_vmsqrt<mode>2): Ditto.
12740 (sse_vmrsqrtv4sf2): Ditto.
12741 (<sse>_vm<code><mode>3): Ditto.
12742 (avx_vmcmp<mode>3): Ditto.
12743 (<sse>_vmmaskcmp<mode>3): Ditto.
12744 (<sse>_comi): Ditto.
12745 (<sse>_ucomi): Ditto.
12746 (*xop_vmfrcz_<mode>): Ditto.
12747 (*fmai_fmadd_<mode>): Ditto.
12748 (*fmai_fmsub_<mode>): Ditto.
12749 (*fmai_fnmadd_<mode>): Ditto.
12750 (*fmai_fnmsub_<mode>): Ditto.
12751 (*fma4i_vmfmadd_<mode>): Ditto.
12752 (*fma4i_vmfmsub_<mode>): Ditto.
12753 (*fma4i_vmfnmadd_<mode>): Ditto.
12754 (*fma4i_vmfnmsub_<mode>): Ditto.
12755 (*xop_vmfrcz_<mode>): Ditto.
12756 (sse_cvtps2pi): Ditto.
12757 (sse_cvttps2pi): Ditto.
12758 (sse_cvtss2si): Ditto.
12759 (sse_cvtss2si_2): Ditto.
12760 (sse_cvtss2siq_2): Ditto.
12761 (sse_cvttss2si): Ditto.
12762 (sse_cvttss2siq): Ditto.
12763 (sse_cvtsd2si): Ditto.
12764 (sse_cvtsd2si_2): Ditto.
12765 (sse_cvtsd2siq_2): Ditto.
12766 (sse_cvttsd2si): Ditto.
12767 (sse_cvttsd2siq): Ditto.
12768 (sse_cvtsd2ss): Ditto.
12769 (sse_cvtss2sd): Ditto.
12770 (avx2_pbroadcast<mode>): Ditto.
12771 (avx2_pbroadcast<mode>_1): Ditto.
12772 (*avx_vperm_broadcast_v4sf): Ditto.
12774 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12775 (sse_movlhps): Ditto.
12776 (sse_storehps): Ditto.
12777 (sse_loadhps): Ditto.
12778 (sse_storelps): Ditto.
12779 (sse_loadlps): Ditto.
12780 (*vec_concatv4sf): Ditto.
12781 (*vec_interleave_highv2df): Ditto.
12782 (*vec_interleave_lowv2df): Ditto.
12783 (*vec_extractv2df_1_sse): Ditto.
12784 (*vec_extractv2df_0_sse): Ditto.
12785 (sse2_storelpd): Ditto.
12786 (sse2_loadlpd): Ditto.
12787 (sse2_movsd): Ditto.
12788 (*vec_concatv4si): Ditto.
12789 (vec_concatv2di): Ditto.
12791 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12792 for Intel asm dialect.
12793 (mmx_punpcklwd): Ditto.
12794 (mmx_punpckldq): Ditto.
12796 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12797 for intel assembler dialect.
12799 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12802 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12803 for BUILT_IN_C{LZ,LRSB}*.
12804 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12805 * tree-vrp.c (extract_range_basic): Handle
12806 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12807 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12808 fall thru to code calling set_value*.
12809 * builtins.c (expand_builtin): Remove *IMAX cases.
12810 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12811 if width is bigger than 2*HWI.
12813 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12815 PR rtl-optimization/55342
12816 * lra-int.h (lra_subreg_reload_pseudos): New.
12817 * lra.c: Add undoing optional reloads to the block diagram.
12818 (lra_subreg_reload_pseudos): New.
12819 (lra_optional_reload_pseudos): Change comments.
12820 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12821 lra_optional_reload_pseudos after undo transformations.
12822 * lra-assigns.c (pseudo_prefix_title): New.
12823 (lra_setup_reg_renumber): Use it.
12824 (spill_for): Ditto. Check subreg reload pseudos too.
12825 (assign_by_spills): Consider subreg reload pseudos too.
12826 * lra-constraints.c (simplify_operand_subreg): Use
12827 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12828 (curr_insn_transform): Recognize and do optional reloads.
12829 (undo_optional_reloads): New.
12830 (lra_undo_inheritance): Call undo_optional_reloads.
12832 2013-07-05 Thomas Quinot <quinot@adacore.com>
12834 * tree-complex.c (expand_complex_operations_1): Fix typo.
12836 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12838 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12839 (tune_params): New member 'const vec_costs'.
12840 * config/aarch64/aarch64.c (generic_vector_cost): New.
12841 (generic_tunings): New member 'generic_vector_cost'.
12842 (aarch64_builtin_vectorization_cost): New.
12843 (aarch64_add_stmt_cost): New.
12844 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12845 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12847 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12850 * config/i386/predicates.md (vsib_address_operand): Disallow
12851 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12853 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12855 PR middle-end/55030
12856 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12857 expand_builtin_setjmp_receiver.
12858 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12859 with NULL for the label parameter.
12860 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12861 the frame-pointer. Adjust comments.
12862 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12863 only if LABEL is non-NULL.
12865 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12867 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12868 (ARM_ABI_AAPCS64): Ditto.
12870 (ARM_DEFAULT_ABI): Ditto.
12872 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12874 * config/aarch64/aarch64-builtins.c
12875 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12876 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12878 * config/aarch64/aarch64-simd.md
12879 (aarch64_ld1<VALL:mode>): New.
12880 (aarch64_st1<VALL:mode>): Likewise.
12881 * config/aarch64/arm_neon.h
12882 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12884 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12886 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12888 (ix86_option_override_internal):Turn on all -mavx target flags by
12889 default as they are dependent on AVX anyway.
12891 2013-07-02 Cary Coutant <ccoutant@google.com>
12893 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12894 deterministic hash.
12895 (loc_checksum_ordered): Likewise.
12896 (hash_loc_operands): Remove inline keyword.
12898 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12900 PR tree-optimization/57741
12901 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12902 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12903 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12904 Allow REAL_CST step_exprs if flag_associative_math.
12905 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12907 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12909 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12911 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12913 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12915 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12917 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12919 (iorsi3_insn): Likewise.
12920 (arm_xorsi3): Likewise.
12922 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12924 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12926 (attribute "type"): ... here, and prefix with "wmmx_".
12927 (attribute "core_cycles"): Update for attribute changes.
12928 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12929 (tbcstv4hi): Likewise.
12930 (tbcstv2si): Likewise.
12931 (iwmmxt_iordi3): Likewise.
12932 (iwmmxt_xordi3): Likewise.
12933 (iwmmxt_anddi3): Likewise.
12934 (iwmmxt_nanddi3): Likewise.
12935 (iwmmxt_arm_movdi): Likewise.
12936 (iwmmxt_movsi_insn): Likewise.
12937 (mov<mode>_internal): Likewise.
12938 (and<mode>3_iwmmxt): Likewise.
12939 (ior<mode>3_iwmmxt): Likewise.
12940 (xor<mode>3_iwmmxt): Likewise.
12941 (add<mode>3_iwmmxt): Likewise.
12942 (ssaddv8qi3): Likewise.
12943 (ssaddv4hi3): Likewise.
12944 (ssaddv2si3): Likewise.
12945 (usaddv8qi3): Likewise.
12946 (usaddv4hi3): Likewise.
12947 (usaddv2si3): Likewise.
12948 (sub<mode>3_iwmmxt): Likewise.
12949 (sssubv8qi3): Likewise.
12950 (sssubv4hi3): Likewise.
12951 (sssubv2si3): Likewise.
12952 (ussubv8qi3): Likewise.
12953 (ussubv4hi3): Likewise.
12954 (ussubv2si3): Likewise.
12955 (mulv4hi3_iwmmxt): Likewise.
12956 (smulv4hi3_highpart): Likewise.
12957 (umulv4hi3_highpart): Likewise.
12958 (iwmmxt_wmacs): Likewise.
12959 (iwmmxt_wmacsz): Likewise.
12960 (iwmmxt_wmacu): Likewise.
12961 (iwmmxt_wmacuz): Likewise.
12962 (iwmmxt_clrdi): Likewise.
12963 (iwmmxt_clrv8qi): Likewise.
12964 (iwmmxt_clr4hi): Likewise.
12965 (iwmmxt_clr2si): Likewise.
12966 (iwmmxt_uavgrndv8qi3): Likewise.
12967 (iwmmxt_uavgrndv4hi3): Likewise.
12968 (iwmmxt_uavgv8qi3): Likewise.
12969 (iwmmxt_uavgv4hi3): Likewise.
12970 (iwmmxt_tinsrb): Likewise.
12971 (iwmmxt_tinsrh): Likewise.
12972 (iwmmxt_tinsrw): Likewise.
12973 (iwmmxt_textrmub): Likewise.
12974 (iwmmxt_textrmsb): Likewise.
12975 (iwmmxt_textrmuh): Likewise.
12976 (iwmmxt_textrmsh): Likewise.
12977 (iwmmxt_textrmw): Likewise.
12978 (iwmxxt_wshufh): Likewise.
12979 (eqv8qi3): Likewise.
12980 (eqv4hi3): Likewise.
12981 (eqv2si3): Likewise.
12982 (gtuv8qi3): Likewise.
12983 (gtuv4hi3): Likewise.
12984 (gtuv2si3): Likewise.
12985 (gtv8qi3): Likewise.
12986 (gtv4hi3): Likewise.
12987 (gtv2si3): Likewise.
12988 (smax<mode>3_iwmmxt): Likewise.
12989 (umax<mode>3_iwmmxt): Likewise.
12990 (smin<mode>3_iwmmxt): Likewise.
12991 (umin<mode>3_iwmmxt): Likewise.
12992 (iwmmxt_wpackhss): Likewise.
12993 (iwmmxt_wpackwss): Likewise.
12994 (iwmmxt_wpackdss): Likewise.
12995 (iwmmxt_wpackhus): Likewise.
12996 (iwmmxt_wpackwus): Likewise.
12997 (iwmmxt_wpackdus): Likewise.
12998 (iwmmxt_wunpckihb): Likewise.
12999 (iwmmxt_wunpckihh): Likewise.
13000 (iwmmxt_wunpckihw): Likewise.
13001 (iwmmxt_wunpckilb): Likewise.
13002 (iwmmxt_wunpckilh): Likewise.
13003 (iwmmxt_wunpckilw): Likewise.
13004 (iwmmxt_wunpckehub): Likewise.
13005 (iwmmxt_wunpckehuh): Likewise.
13006 (iwmmxt_wunpckehuw): Likewise.
13007 (iwmmxt_wunpckehsb): Likewise.
13008 (iwmmxt_wunpckehsh): Likewise.
13009 (iwmmxt_wunpckehsw): Likewise.
13010 (iwmmxt_wunpckelub): Likewise.
13011 (iwmmxt_wunpckeluh): Likewise.
13012 (iwmmxt_wunpckeluw): Likewise.
13013 (iwmmxt_wunpckelsb): Likewise.
13014 (iwmmxt_wunpckelsh): Likewise.
13015 (iwmmxt_wunpckelsw): Likewise.
13016 (ror<mode>3): Likewise.
13017 (ashr<mode>3_iwmmxt): Likewise.
13018 (lshr<mode>3_iwmmxt): Likewise.
13019 (ashl<mode>3_iwmmxt): Likewise.
13020 (ror<mode>3_di): Likewise.
13021 (ashr<mode>3_di): Likewise.
13022 (lshr<mode>3_di): Likewise.
13023 (ashl<mode>3_di): Likewise.
13024 (iwmmxt_wmadds): Likewise.
13025 (iwmmxt_wmaddu): Likewise.
13026 (iwmmxt_tmia): Likewise.
13027 (iwmmxt_tmiaph): Likewise.
13028 (iwmmxt_tmiabb): Likewise.
13029 (iwmmxt_tmiatb): Likewise.
13030 (iwmmxt_tmiabt): Likewise.
13031 (iwmmxt_tmiatt): Likewise.
13032 (iwmmxt_tmovmskb): Likewise.
13033 (iwmmxt_tmovmskh): Likewise.
13034 (iwmmxt_tmovmskw): Likewise.
13035 (iwmmxt_waccb): Likewise.
13036 (iwmmxt_wacch): Likewise.
13037 (iwmmxt_waccw): Likewise.
13038 (iwmmxt_waligni): Likewise.
13039 (iwmmxt_walignr): Likewise.
13040 (iwmmxt_walignr0): Likewise.
13041 (iwmmxt_walignr1): Likewise.
13042 (iwmmxt_walignr2): Likewise.
13043 (iwmmxt_walignr3): Likewise.
13044 (iwmmxt_wsadb): Likewise.
13045 (iwmmxt_wsadh): Likewise.
13046 (iwmmxt_wsadbz): Likewise.
13047 (iwmmxt_wsadhz): Likewise.
13048 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13049 (iwmmxt_wabsdiffb): Likewise.
13050 (iwmmxt_wabsdiffh): Likewise.
13051 (iwmmxt_wabsdiffw): Likewise.
13052 (iwmmxt_waddsubhx): Likewise
13053 (iwmmxt_wsubaddhx): Likewise.
13054 (addc<mode>3): Likewise.
13055 (iwmmxt_avg4): Likewise.
13056 (iwmmxt_avg4r): Likewise.
13057 (iwmmxt_wmaddsx): Likewise.
13058 (iwmmxt_wmaddux): Likewise.
13059 (iwmmxt_wmaddsn): Likewise.
13060 (iwmmxt_wmaddun): Likewise.
13061 (iwmmxt_wmulwsm): Likewise.
13062 (iwmmxt_wmulwum): Likewise.
13063 (iwmmxt_wmulsmr): Likewise.
13064 (iwmmxt_wmulumr): Likewise.
13065 (iwmmxt_wmulwsmr): Likewise.
13066 (iwmmxt_wmulwumr): Likewise.
13067 (iwmmxt_wmulwl): Likewise.
13068 (iwmmxt_wqmulm): Likewise.
13069 (iwmmxt_wqmulwm): Likewise.
13070 (iwmmxt_wqmulmr): Likewise.
13071 (iwmmxt_wqmulwmr): Likewise.
13072 (iwmmxt_waddbhusm): Likewise.
13073 (iwmmxt_waddbhusl): Likewise.
13074 (iwmmxt_wqmiabb): Likewise.
13075 (iwmmxt_wqmiabt): Likewise.
13076 (iwmmxt_wqmiatb): Likewise.
13077 (iwmmxt_wqmiatt): Likewise.
13078 (iwmmxt_wqmiabbn): Likewise.
13079 (iwmmxt_wqmiabtn): Likewise.
13080 (iwmmxt_wqmiatbn): Likewise.
13081 (iwmmxt_wqmiattn): Likewise.
13082 (iwmmxt_wmiabb): Likewise.
13083 (iwmmxt_wmiabt): Likewise.
13084 (iwmmxt_wmiatb): Likewise.
13085 (iwmmxt_wmiatt): Likewise.
13086 (iwmmxt_wmiabbn): Likewise.
13087 (iwmmxt_wmiabtn): Likewise.
13088 (iwmmxt_wmiatbn): Likewise.
13089 (iwmmxt_wmiattn): Likewise.
13090 (iwmmxt_wmiawbb): Likewise.
13091 (iwmmxt_wmiawbt): Likewise.
13092 (iwmmxt_wmiawtb): Likewise.
13093 (iwmmxt_wmiawtt): Likewise.
13094 (iwmmxt_wmiawbbn): Likewise.
13095 (iwmmxt_wmiawbtn): Likewise.
13096 (iwmmxt_wmiawtbn): Likewise.
13097 (iwmmxt_wmiawttn): Likewise.
13098 (iwmmxt_wmerge): Likewise.
13099 (iwmmxt_tandc<mode>3): Likewise.
13100 (iwmmxt_torc<mode>3): Likewise.
13101 (iwmmxt_torvsc<mode>3): Likewise.
13102 (iwmmxt_textrc<mode>3): Likewise.
13103 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13104 (wmmxt_pack): Likewise.
13105 (wmmxt_mult_c1): Likewise.
13106 (wmmxt_mult_c2): Likewise.
13107 (wmmxt_alu_c1): Likewise.
13108 (wmmxt_alu_c2): Likewise.
13109 (wmmxt_alu_c3): Likewise.
13110 (wmmxt_transfer_c1): Likewise.
13111 (wmmxt_transfer_c2): Likewise.
13112 (wmmxt_transfer_c3): Likewise.
13113 (marvell_f_iwmmxt_wstr): Likewise.
13114 (marvell_f_iwmmxt_wldr): Likewise.
13116 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13118 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13120 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13123 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13124 * lra-constraints.c (need_for_split_p): Check call used hard regs
13125 living through calls.
13127 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13128 call used regs for call insn.
13130 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13133 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13134 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13136 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13138 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13140 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13142 * lra-constraints.c (need_for_split_p): Check call used hard regs
13143 living through calls.
13145 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13148 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13149 to tie with any other modes. Eliminate Altivec vector mode tests,
13150 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13151 code, to return 0 if testing MODE2 for a condition, if we've
13152 already tested MODE1 for the same condition.
13154 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13156 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13159 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13161 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13162 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13164 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13166 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13168 (aarch64_symbolic_constant_p): Remove.
13169 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13170 static. Fix line length and white space.
13171 (aarch64_symbolic_constant_p): Remove.
13172 * config/aarch64/predicates.md (aarch64_valid_symref):
13173 Use aarch64_classify_symbol_expression.
13175 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13177 * config/arm/constraints.md (Ts): New constraint.
13178 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13180 (compare_scc): Use "Ts" constraint for operand 0.
13181 (ior_scc_scc): Likewise.
13182 (and_scc_scc): Likewise.
13183 (and_scc_scc_nodom): Likewise.
13184 (ior_scc_scc_cmp): Likewise for operand 7.
13185 (and_scc_scc_cmp): Likewise.
13186 * config/arm/thumb2.md (thumb2_movsi_insn):
13187 Add alternatives for 16-bit encodings.
13188 (thumb2_movhi_insn): Likewise.
13189 (thumb2_movsicc_insn): Likewise.
13190 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13191 (thumb2_negscc): Use "Ts" constraint.
13192 Move mvn instruction outside cond_exec block.
13193 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13194 for 16-bit encodings.
13196 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13198 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13200 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13201 (mulsi3subsi): Likewise.
13202 (mulsidi3adddi): Likewise.
13203 (mulsidi3_v6): Likewise.
13204 (umulsidi3_v6): Likewise.
13205 (umulsidi3adddi_v6): Likewise.
13206 (smulsi3_highpart_v6): Likewise.
13207 (umulsi3_highpart_v6): Likewise.
13208 (mulhisi3tb): Likewise.
13209 (mulhisi3bt): Likewise.
13210 (mulhisi3tt): Likewise.
13211 (maddhisi4): Likewise.
13212 (maddhisi4tb): Likewise.
13213 (maddhisi4tt): Likewise.
13214 (maddhidi4): Likewise.
13215 (maddhidi4tb): Likewise.
13216 (maddhidi4tt): Likewise.
13217 (zeroextractsi_compare0_scratch): Likewise.
13218 (insv_zero): Likewise.
13219 (insv_t2): Likewise.
13220 (anddi_notzesidi_di): Likewise.
13221 (anddi_notsesidi_di): Likewise.
13222 (andsi_notsi_si): Likewise.
13223 (iordi_zesidi_di): Likewise.
13224 (xordi_zesidi_di): Likewise.
13225 (andsi_iorsi3_notsi): Likewise.
13226 (smax_0): Likewise.
13227 (smax_m1): Likewise.
13228 (smin_0): Likewise.
13229 (not_shiftsi): Likewise.
13230 (unaligned_loadsi): Likewise.
13231 (unaligned_loadhis): Likewise.
13232 (unaligned_loadhiu): Likewise.
13233 (unaligned_storesi): Likewise.
13234 (unaligned_storehi): Likewise.
13235 (extv_reg): Likewise.
13236 (extzv_t2): Likewise.
13237 (divsi3): Likewise.
13238 (udivsi3): Likewise.
13239 (arm_zero_extendhisi2addsi): Likewise.
13240 (arm_zero_extendqisi2addsi): Likewise.
13241 (compareqi_eq0): Likewise.
13242 (arm_extendhisi2_v6): Likewise.
13243 (arm_extendqisi2addsi): Likewise.
13244 (arm_movt): Likewise.
13245 (thumb2_ldrd): Likewise.
13246 (thumb2_ldrd_base): Likewise.
13247 (thumb2_ldrd_base_neg): Likewise.
13248 (thumb2_strd): Likewise.
13249 (thumb2_strd_base): Likewise.
13250 (thumb2_strd_base_neg): Likewise.
13251 (arm_negsi2): Add alternative for 16-bit encoding.
13252 (arm_one_cmplsi2): Likewise.
13254 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13256 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13257 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13258 (movdfcc): Likewise.
13259 * config/arm/vfp.md (*thumb2_movsf_vfp):
13260 Disable predication for arm_restrict_it.
13261 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13262 (*thumb2_movdfcc_vfp): Likewise.
13263 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13264 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13265 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13266 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13267 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13268 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13269 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13270 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13271 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13272 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13273 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13274 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13275 Disable predication for arm_restrict_it.
13277 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13279 * config/i386/bmiintrin.h (_bextr_u32): New.
13280 (_bextr_u64): Ditto.
13282 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13284 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13285 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13286 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13287 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13288 * config/mips/n32-elf.h: ...this new file.
13290 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13293 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13294 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13296 2013-06-27 Catherine Moore <clm@codesourcery.com>
13298 * config/mips/mips-tables.opt: Regenerate.
13299 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13300 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13301 * doc/invoke.texi: Add -m14kc.
13303 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13306 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13307 constraints of operand 1 and 2.
13310 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13311 to match RTL canonicalization. Swap predicates and
13312 constraints of operand 1 and 2.
13314 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13316 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13317 Process OP_INOUT regs for splitting too.
13319 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13321 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13322 decl before the loop, initialize to NULL.
13323 (vectorizable_load): Initialize ptr_incr to NULL.
13325 2013-06-27 Martin Jambor <mjambor@suse.cz>
13328 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13329 * ipa-ref.c (ipa_maybe_record_reference): New function.
13330 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13331 * ipa-cp.c (create_specialized_node): Record potential references from
13333 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13335 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13337 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13338 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13339 'mode' to force_reg.
13340 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13341 (aarch64_expand_mov_immediate): Likewise.
13343 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13345 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13346 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13348 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13350 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13351 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13352 (struct machine_function): Add tbegin_p.
13353 (s390_canonicalize_comparison): Fold CC mode compares to
13354 conditional jump if possible.
13355 (s390_emit_jump): Return the emitted jump.
13356 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13357 Handle CCRAWmode compares.
13358 (s390_option_override): Default to -mhtm if available.
13359 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13360 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13361 FPRs instead of df_regs_ever_live_p.
13362 (s390_optimize_nonescaping_tx): New function.
13363 (s390_init_frame_layout): Extend clobbered_regs array to cover
13365 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13366 (s390_expand_tbegin): New function.
13367 (enum s390_builtin): New enum definition.
13368 (code_for_builtin): New array definition.
13369 (s390_init_builtins): New function.
13370 (s390_expand_builtin): New function.
13371 (TARGET_INIT_BUILTINS): Define.
13372 (TARGET_EXPAND_BUILTIN): Define.
13373 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13374 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13375 (s390_alc_comparison): Likewise.
13376 * config/s390/s390-modes.def: Add CCRAWmode.
13377 * config/s390/s390.h (processor_flags): Add PF_TX.
13378 (TARGET_CPU_HTM): Define macro.
13379 (TARGET_HTM): Define macro.
13380 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13381 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13382 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13383 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13384 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13385 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13386 ("*cc_to_int"): Move up.
13387 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13388 constants other than 0.
13389 ("*ccraw_to_int"): New insn and splitter definition.
13390 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13391 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13392 ("tx_assist"): New expander.
13393 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13394 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13395 * config/s390/s390.opt: Add -mhtm option.
13396 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13397 * config/s390/htmxlintrin.h: New file.
13398 * config/s390/htmintrin.h: New file.
13399 * config/s390/s390intrin.h: New file.
13400 * doc/extend.texi: Document htm builtins.
13401 * config.gcc: Add the new header files to extra_headers.
13403 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13405 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13406 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13408 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13409 Pat Haugen <pthaugen@us.ibm.com>
13410 Peter Bergner <bergner@vnet.ibm.com>
13412 * config/rs6000/power8.md: New.
13413 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13414 setting for power8 entry.
13415 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13416 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13417 test for Power4/Power5 only.
13418 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13420 (force_new_group): Adjust comment.
13421 * config/rs6000/rs6000.md: Include power8.md.
13423 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13425 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13426 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13428 (tune_params): Update comment.
13429 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13430 (arm_max_conditional_execute): New function.
13431 (thumb2_final_prescan_insn): Use max_insn_skipped and
13432 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13434 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13436 PR tree-optimization/57705
13437 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13438 SSA_NAME step, provided that it is not defined inside the loop.
13439 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13440 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13442 2013-06-25 Martin Jambor <mjambor@suse.cz>
13444 PR middle-end/57670
13445 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13446 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13448 (ipa_note_param_call): Initialize member_ptr flag.
13449 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13450 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13451 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13452 (ipa_read_indirect_edge_info): Likewise.
13454 2013-06-25 Richard Biener <rguenther@suse.de>
13456 PR middle-end/56977
13457 * passes.c (init_optimization_passes): Move pass_fold_builtins
13458 and pass_dce earlier with -Og.
13460 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13462 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13463 <BIT_FIELD_REF>: Remove trailing TAB.
13464 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13467 2013-06-24 Martin Jambor <mjambor@suse.cz>
13469 PR tree-optimization/57358
13470 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13471 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13472 (ipa_analyze_params_uses): Generate pessimistic info when true.
13474 2013-06-24 Martin Jambor <mjambor@suse.cz>
13476 PR tree-optimization/57539
13477 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13478 global.inlined_to of the new node to it. All callers changed.
13479 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13480 inlining_into, pass it to cgraph_clone_node.
13481 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13482 ipa_free_edge_args_substructures.
13483 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13484 rdesc linked list. Do not assert rdesc edges have inlined caller.
13485 Assert we have found an rdesc in the rdesc list.
13487 2013-06-24 Richard Biener <rguenther@suse.de>
13489 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13490 (pointer_set_lookup): Declare.
13491 (class pointer_map): New template class implementing a
13492 generic pointer to T map.
13493 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13494 pointer_map<T>::contains, pointer_map<T>::insert,
13495 pointer_map<T>::traverse): New functions.
13496 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13497 (pointer_set_lookup): New function.
13498 (pointer_set_contains): Use pointer_set_lookup.
13499 (pointer_set_insert): Likewise.
13500 (insert_aux): Remove.
13501 (struct pointer_map_t): Embed a pointer_set_t.
13502 (pointer_map_create): Adjust.
13503 (pointer_map_destroy): Likewise.
13504 (pointer_map_contains): Likewise.
13505 (pointer_map_insert): Likewise.
13506 (pointer_map_traverse): Likewise.
13507 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13508 pointer_map<unsigned> instead of a pointer_map_t.
13509 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13510 (streamer_tree_cache_lookup): Likewise.
13511 (streamer_tree_cache_create): Likewise.
13512 (streamer_tree_cache_delete): Likewise.
13513 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13514 pointer_map<unsigned> instead of a pointer_map_t.
13515 (lto_init_tree_ref_encoder): Adjust.
13516 (lto_destroy_tree_ref_encoder): Likewise.
13517 * lto-section-out.c (lto_output_decl_index): Likewise.
13518 (lto_record_function_out_decl_state): Likewise.
13519 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13521 2013-06-24 Richard Biener <rguenther@suse.de>
13523 PR tree-optimization/57488
13524 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13526 2013-06-24 Alan Modra <amodra@gmail.com>
13528 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13529 (gen_easy_altivec_constant): Likewise.
13530 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13531 easy_vector_constant_msb): Likewise.
13533 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13536 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13537 add missing return true.
13539 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13542 * config/sh/predicates.md (general_extend_operand): Invoke
13543 general_movsrc_operand for memory operands.
13544 (general_movsrc_operand): Allow reg+reg addressing, do not use
13545 general_operand for memory operands.
13547 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13549 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13550 when current target options does not apply.
13551 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13552 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13553 * config/i386/bmiintrin.h: Pass appropriate target
13554 attributes to header.
13555 * config/i386/mmintrin.h: Ditto.
13556 * config/i386/nmmintrin.h: Ditto.
13557 * config/i386/avx2intrin.h: Ditto.
13558 * config/i386/fxsrintrin.h: Ditto.
13559 * config/i386/tbmintrin.h: Ditto.
13560 * config/i386/xsaveintrin.h: Ditto.
13561 * config/i386/f16cintrin.h: Ditto.
13562 * config/i386/xtestintrin.h: Ditto.
13563 * config/i386/xsaveoptintrin.h: Ditto.
13564 * config/i386/bmi2intrin.h: Ditto.
13565 * config/i386/lzcntintrin.h: Ditto.
13566 * config/i386/smmintrin.h: Ditto.
13567 * config/i386/wmmintrin.h: Ditto.
13568 * config/i386/x86intrin.h: Remove all header include guards.
13569 * config/i386/prfchwintrin.h: Ditto.
13570 * config/i386/pmmintrin.h: Ditto.
13571 * config/i386/tmmintrin.h: Ditto.
13572 * config/i386/xmmintrin.h: Ditto.
13573 * config/i386/popcntintrin.h: Ditto.
13574 * config/i386/rdseedintrin.h: Ditto.
13575 * config/i386/ammintrin.h: Ditto.
13576 * config/i386/emmintrin.h: Ditto.
13577 * config/i386/immintrin.h: Remove all header include guards.
13578 * config/i386/fma4intrin.h: Ditto.
13579 * config/i386/lwpintrin.h: Ditto.
13580 * config/i386/xopintrin.h: Ditto.
13581 * config/i386/ia32intrin.h: Ditto.
13582 * config/i386/avxintrin.h: Ditto.
13583 * config/i386/rtmintrin.h: Ditto.
13584 * config/i386/fmaintrin.h: Ditto.
13585 * config/i386/mm3dnow.h: Ditto.
13587 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13589 * common/config/i386/i386-common.c: Handle LZCNT.
13591 2013-06-22 Andi Kleen <ak@linux.intel.com>
13593 * doc/extend.texi: Use __atomic_store_n instead of
13594 __atomic_store in HLE example.
13596 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13598 * config/sh/sh.c: Remove <cstdlib> workaround.
13600 2013-06-21 Andi Kleen <ak@linux.intel.com>
13602 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13604 2013-06-21 Andi Kleen <ak@linux.intel.com>
13606 * doc/extend.texi: Document that __atomic_clear and
13607 __atomic_test_and_set should only be used with bool.
13609 2013-06-20 Jan Hubicka <jh@suse.cz>
13611 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13612 types_same_for_odr.
13613 * tree.c (decls_same_for_odr): New function.
13614 (same_for_edr): New function.
13615 (types_same_for_odr): New function.
13616 (get_binfo_at_offset): Use it.
13617 * tree.h (types_same_for_odr): Declare.
13619 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13620 Jason Merrill <jason@redhat.com>
13622 * system.h: Include <cstdlib> as well as <stdlib.h>.
13624 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13627 * config/i386/i386.c (construct_container): Report error if
13628 long double is used with disabled x87 float returns.
13630 2013-06-20 Jan Hubicka <jh@suse.cz>
13632 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13634 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13636 PR rtl-optimization/57425
13637 PR rtl-optimization/57569
13638 * alias.c (write_dependence_p): Remove parameters mem_mode and
13639 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13640 Changed all callers.
13641 (canon_anti_dependence): Get comments and semantics in sync.
13642 Add parameter mem_canonicalized. Changed all callers.
13643 * rtl.h (canon_anti_dependence): Update prototype.
13645 2013-06-20 Richard Biener <rguenther@suse.de>
13647 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13648 byte case, inline streamer_read_uchar and defer section
13651 2013-06-20 Richard Biener <rguenther@suse.de>
13653 PR tree-optimization/57584
13654 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13655 SSA names into the expanded expression that take part in
13656 abnormal coalescing.
13658 2013-06-19 Sharad Singhai <singhai@google.com>
13660 * gcov.c (print_usage): Handle new option.
13661 (process_args): Ditto.
13662 (get_gcov_intermediate_filename): New function.
13663 (output_intermediate_file): New function.
13664 (output_gcov_file): New function
13665 (generate_results): Handle new option.
13666 (release_function): Relase demangled name.
13667 (read_graph_file): Handle demangled name.
13668 (output_lines): Ditto.
13669 * doc/gcov.texi: Document gcov intermediate format.
13671 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13674 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13675 (lra_emit_add): Use the functions. Add comment about Y as an
13678 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13681 * collect2.c (collect_atexit): New.
13682 (collect_exit): Delete.
13683 (main): Register collect_atexit with atexit.
13684 (collect_wait): Change collect_exit to exit.
13686 * collect2.h (collect_exit): Delete.
13687 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13689 2013-06-19 Wei Mi <wmi@google.com>
13691 PR rtl-optimization/57518
13692 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13693 if regno is used in paradoxical subreg.
13694 (update_equiv_regs): Check pdx_subregs[regno] before
13695 set a reg to be equivalent with a mem.
13697 2013-06-19 Matthias Klose <doko@ubuntu.com>
13700 * file-find.h (find_a_file): Add a mode parameter.
13701 * file-find.c (find_a_file): Likewise.
13702 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13703 with X_OK for the executables.
13704 * collect2.c (main): Call find_a_file with X_OK.
13706 2013-06-19 Steve Ellcey <sellcey@mips.com>
13709 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13710 Use NEXT_INSN instead of next_real_insn.
13712 2013-06-19 Jan Hubicka <jh@suse.cz>
13714 * cgraph.h (const_value_known_p): Replace by ...
13715 (ctor_for_folding): .. this one.
13716 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13717 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13718 * expr.c (expand_expr_real_1): Likewise.
13719 (string_constant): Likewise.
13720 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13721 * ipa.c (process_references): Likewise.
13722 (symtab_remove_unreachable_nodes): Likewise.
13723 * ipa-inline-analysis.c (param_change_prob): Likewise.
13724 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13725 (get_base_constructor): Likwise.
13726 * varpool.c (varpool_remove_node): Likewise.
13727 (varpool_remove_initializer): LIkewise.
13728 (dump_varpool_node): LIkwise.
13729 (const_value_known_p): Rewrite to ...
13730 (ctor_for_folding): ... this one.
13732 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13735 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13736 PERSONALITY in $PATH derived prefixes.
13738 2013-06-19 Jeff Law <law@redhat.com>
13740 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13743 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13744 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13747 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13749 * config/arm/vfp.md: Move VFP instruction classification documentation
13751 * config/arm/arm.md: ... here. Update instruction classification
13754 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13756 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13757 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13758 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13760 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13763 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13764 with NEXT_INSN. Use tablejump_p to check for jump table data
13767 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13770 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13771 that now in C++ the value is correct per the C++ standards.
13773 2013-06-19 Richard Biener <rguenther@suse.de>
13775 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13776 for global context.
13778 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13781 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13784 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13785 with next_active_insn.
13787 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13789 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13790 functions are inlined during failures to flag an error.
13791 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13792 in early inline pass.
13794 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13796 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13799 2013-06-18 Julian Brown <julian@codesourcery.com>
13801 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13802 Permit virtual register pre-reload if !strict.
13803 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13805 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13807 * config/arm/neon.md (movmisalign<mode>): Use
13808 neon_perm_struct_or_reg_operand instead of
13809 neon_struct_or_register_operand.
13810 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13811 neon_permissive_struct_operand instead of neon_struct_operand.
13812 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13813 neon_vector_mem_operand.
13814 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13815 neon_vector_mem_operand.
13816 (neon_permissive_struct_operand): New.
13817 (neon_struct_or_register_operand): Rename to...
13818 (neon_perm_struct_or_reg_operand): This. Adjust call to
13819 neon_vector_mem_operand.
13821 2013-06-18 Richard Biener <rguenther@suse.de>
13823 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13824 * lto-streamer.h: Include pointer-set.h.
13825 (struct lto_decl_slot): Remove.
13826 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13827 Remove next_index entry.
13828 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13829 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13830 (lto_init_tree_ref_encoder): Adjust.
13831 (lto_destroy_tree_ref_encoder): Likewise.
13832 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13833 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13834 (lto_output_decl_index): Adjust.
13835 (lto_new_out_decl_state): Likewise.
13836 (lto_record_function_out_decl_state): Likewise.
13837 * lto-streamer-out.c (copy_function): Likewise.
13839 2013-06-18 Richard Biener <rguenther@suse.de>
13841 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13842 * cgraphunit.c: Include cfgloop.h.
13843 (init_lowered_empty_function): Initialize the loop tree.
13844 (assemble_thunk): Insert new BBs into loops.
13846 2013-06-18 Richard Biener <rguenther@suse.de>
13848 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13849 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13850 the map from cache entry to cache index optional.
13851 (streamer_tree_cache_replace_tree): Adjust accordingly.
13852 (streamer_tree_cache_append): Likewise.
13853 (streamer_tree_cache_delete): Likewise.
13854 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13855 streamer cache map from cache entry to cache index.
13856 * lto-streamer-out.c (create_output_block): Adjust.
13858 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13860 * config/arm/arm.md (attribute "insn"): Move multiplication and
13861 division attributes to...
13862 (attribute "type"): ... here. Remove mult.
13863 (attribute "mul32"): New attribute.
13864 (attribute "mul64"): Add umaal.
13865 (*arm_mulsi3): Update attributes.
13866 (*arm_mulsi3_v6): Likewise.
13867 (*thumb_mulsi3): Likewise.
13868 (*thumb_mulsi3_v6): Likewise.
13869 (*mulsi3_compare0): Likewise.
13870 (*mulsi3_compare0_v6): Likewise.
13871 (*mulsi_compare0_scratch): Likewise.
13872 (*mulsi_compare0_scratch_v6): Likewise.
13873 (*mulsi3addsi): Likewise.
13874 (*mulsi3addsi_v6): Likewise.
13875 (*mulsi3addsi_compare0): Likewise.
13876 (*mulsi3addsi_compare0_v6): Likewise.
13877 (*mulsi3addsi_compare0_scratch): Likewise.
13878 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13879 (*mulsi3subsi): Likewise.
13880 (*mulsidi3adddi): Likewise.
13881 (*mulsi3addsi_v6): Likewise.
13882 (*mulsidi3adddi_v6): Likewise.
13883 (*mulsidi3_nov6): Likewise.
13884 (*mulsidi3_v6): Likewise.
13885 (*umulsidi3_nov6): Likewise.
13886 (*umulsidi3_v6): Likewise.
13887 (*umulsidi3adddi): Likewise.
13888 (*umulsidi3adddi_v6): Likewise.
13889 (*smulsi3_highpart_nov6): Likewise.
13890 (*smulsi3_highpart_v6): Likewise.
13891 (*umulsi3_highpart_nov6): Likewise.
13892 (*umulsi3_highpart_v6): Likewise.
13893 (mulhisi3): Likewise.
13894 (*mulhisi3tb): Likewise.
13895 (*mulhisi3bt): Likewise.
13896 (*mulhisi3tt): Likewise.
13897 (maddhisi4): Likewise.
13898 (*maddhisi4tb): Likewise.
13899 (*maddhisi4tt): Likewise.
13900 (maddhidi4): Likewise.
13901 (*maddhidi4tb): Likewise.
13902 (*maddhidi4tt): Likewise.
13903 (divsi3): Likewise.
13904 (udivsi3): Likewise.
13905 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13906 (thumb2_mulsi_short_compare0): Likewise.
13907 (thumb2_mulsi_short_compare0_scratch): Likewise.
13908 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13909 (1020mult2): Likewise.
13910 (1020mult3): Likewise.
13911 (1020mult4): Likewise.
13912 (1020mult5): Likewise.
13913 (1020mult6): Likewise.
13914 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13916 (cortex_a15_mult64): Likewise.
13917 (cortex_a15_sdiv): Likewise.
13918 (cortex_a15_udiv): Likewise.
13919 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13925 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13926 (pj4_ir_div): Likewise.
13927 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13928 (11_mult2): Likewise.
13929 (11_mult3): Likewise.
13930 (11_mult4): Likewise.
13931 (11_mult5): Likewise.
13932 (11_mult6): Likewise.
13933 (11_mult7): Likewise.
13934 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13935 (cortex_a8_mla): Likewise.
13936 (cortex_a8_mull): Likewise.
13937 (cortex_a8_smulwy): Likewise.
13938 (cortex_a8_smlald): Likewise.
13939 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13940 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13941 (cortex_r4_mul_3): Likewise.
13942 (cortex_r4_mla_4): Likewise.
13943 (cortex_r4_mla_3): Likewise.
13944 (cortex_r4_smlald): Likewise.
13945 (cortex_r4_mull): Likewise.
13946 (cortex_r4_sdiv): Likewise.
13947 (cortex_r4_udiv): Likewise.
13948 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13949 (cortex_a7_idiv): Likewise.
13950 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13951 (9_mult2): Likewise.
13952 (9_mult3): Likewise.
13953 (9_mult4): Likewise.
13954 (9_mult5): Likewise.
13955 (9_mult6): Likewise.
13956 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13957 (cortex_a53_sdiv): Likewise.
13958 (cortex_a53_udiv): Likewise.
13959 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13960 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13961 (mp626_mult2): Likewise.
13962 (mp626_mult3): Likewise.
13963 (mp626_mult4): Likewise.
13964 * config/arm/fa526.md (526_mult1): Update attribute change.
13965 (526_mult2): Likewise.
13966 * config/arm/arm-generic.md (mult): Update attribute change.
13967 (mult_ldsched_strongarm): Likewise.
13968 (mult_ldsched): Likewise.
13969 (multi_cycle): Likewise.
13970 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13971 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13972 (606te_mult2): Likewise.
13973 (606te_mult3): Likewise.
13974 (606te_mult4): Likewise.
13975 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13976 (cortex_a9_mac16): Likewise.
13977 (cortex_a9_multiply): Likewise.
13978 (cortex_a9_mac): Likewise.
13979 (cortex_a9_multiply_long): Likewise.
13980 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13981 (626te_mult2): Likewise.
13982 (626te_mult3): Likewise.
13983 (626te_mult4): Likewise.
13985 2013-06-18 Richard Biener <rguenther@suse.de>
13988 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13990 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13993 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13994 with next_active_insn.
13996 2013-06-18 Alan Modra <amodra@gmail.com>
13998 * config/rs6000/rs6000.h (enum data_align): New.
13999 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14000 (DATA_ABI_ALIGNMENT): Define.
14001 (CONSTANT_ALIGNMENT): Correct comment.
14002 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14003 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14005 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14007 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14008 ATTRIBUTE_UNUSED marking.
14010 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14012 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14013 alternative and update.
14014 (aarch64_dup_lanedi): Delete.
14015 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14016 * config/aarch64/aarch64-simd-builtins.def: Update.
14018 2013-06-17 Richard Biener <rguenther@suse.de>
14020 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14021 (lto_input_scc): Declare.
14022 (lto_input_tree_1): Likewise.
14023 (struct lto_stats_d): Add num_tree_bodies_output and
14024 num_pickle_refs_output.
14025 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14026 (lto_read_tree_1): Split out from ...
14027 (lto_read_tree): ... this.
14028 (lto_input_scc): New function.
14029 (lto_input_tree_1): Split out from ...
14030 (lto_input_tree): ... this. Handle LTO_tree_scc.
14031 (lto_data_in_create): Create the streamer cache without hashes.
14032 * lto-streamer-out.c (create_output_block): Create the streamer
14033 cache with hashes when not doing WPA.
14034 (lto_write_tree_1): Split out from ...
14035 (lto_write_tree): ... this.
14036 (get_symbol_initial_value): New function.
14037 (lto_output_tree_1): Split out from ...
14038 (lto_output_tree): ... this. Write trees as series of SCCs
14039 using a DFS walk via DFS_write_tree.
14040 (struct sccs, struct scc_entry): New types.
14041 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14042 (DFS_write_tree_body): New function.
14043 (DFS_write_tree): Likewise.
14044 (hash_tree): Likewise.
14045 (scc_entry_compare): Likewise.
14046 (hash_scc): Likewise.
14047 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14048 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14049 TREE_CHAIN as regular reference.
14050 (streamer_read_integer_cst): Remove.
14051 (streamer_get_pickled_tree): Adjust.
14052 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14053 of DECL_EXTERNALs in BLOCK_VARS for now.
14054 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14056 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14057 Add hash value argument and record that if hashes are recorded
14059 (streamer_tree_cache_insert_1): Adjust.
14060 (streamer_tree_cache_insert): Likewise.
14061 (streamer_tree_cache_insert_at): Rename to ...
14062 (streamer_tree_cache_replace_tree): ... this and adjust.
14063 (streamer_tree_cache_append): Adjust.
14064 (record_common_node): Likewise.
14065 (streamer_tree_cache_create): Add argument whether to
14066 record hash values together with trees.
14067 (streamer_tree_cache_delete): Adjust.
14068 * tree-streamer.h (struct streamer_tree_cache_d): Add
14070 (streamer_read_integer_cst): Remove.
14071 (streamer_tree_cache_insert): Adjust.
14072 (streamer_tree_cache_append): Likewise.
14073 (streamer_tree_cache_insert_at): Rename to ...
14074 (streamer_tree_cache_replace_tree): ... this and adjust.
14075 (streamer_tree_cache_create): Add argument whether to record hashes.
14076 (streamer_tree_cache_get): Rename to ...
14077 (streamer_tree_cache_get_tree): ... this.
14078 (streamer_tree_cache_get_hash): New function.
14079 * tree.c (cache_integer_cst): New function.
14080 * tree.h (cache_integer_cst): Declare.
14081 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14082 * lto-symtab.c (lto_varpool_replace_node): Only release
14083 DECL_INITIAL of non-prevailing decls.
14084 * varpool.c (varpool_remove_initializer): Do not release
14085 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14087 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14089 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14090 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14091 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14092 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14093 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14094 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14095 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14096 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14097 instead of TARGET_64BIT.
14098 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14099 Require ISA_HAS_<D>DIV.
14101 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14103 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14104 (mips*-*-linux*): Move default with_llsc setting to where other
14106 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14108 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14109 Likewise. Remove default with_tune setting. Move default float
14110 setting to its own block. Handle with_llsc in the same block as above.
14112 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14114 PR rtl-optimization/57425
14115 PR rtl-optimization/57569
14116 * alias.c (write_dependence_p): Add new parameters mem_mode,
14117 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14118 Changed all callers.
14119 (canon_anti_dependence): New function.
14120 * cse.c (check_dependence): Use canon_anti_dependence.
14121 * cselib.c (cselib_invalidate_mem): Likewise.
14122 * rtl.h (canon_anti_dependence): Declare.
14124 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14126 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14127 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14128 ".set mips3" for 64-bit targets.
14130 2013-06-15 Dehao Chen <dehao@google.com>
14132 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14133 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14134 (gimple_check_call_args): Likewise.
14135 * value-prof.c (check_ic_target): Likewise.
14136 * ipa-inline.c (early_inliner): Likewise.
14137 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14138 * cgraph.c (cgraph_create_edge_1): Likewise.
14139 (cgraph_make_edge_direct): Likewise.
14141 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14144 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14145 rs6000_output_move_128bit to handle emitting quad memory
14146 operations. Set attribute length to 8 bytes.
14148 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14150 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14152 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14153 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14154 (aarch64_<su>mlsl<mode>): Likewise.
14156 2013-06-14 Mike Stump <mikestump@comcast.net>
14158 * Makefile.in (TARGET_H): Add insn-codes.h.
14160 2013-06-14 Alan Modra <amodra@gmail.com>
14162 PR middle-end/57134
14163 PR middle-end/57586
14164 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14165 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14166 bitfield expansion when EXPAND_MEMORY.
14167 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14169 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14171 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14172 test for clearing quad memory on 32-bit later.
14174 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14176 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14177 (fold_negate_expr): Likewise.
14178 (fold_real_zero_addition_p): Handle vectors.
14179 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14181 2013-06-14 Alan Modra <amodra@gmail.com>
14183 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14185 2013-06-13 Jan Hubicka <jh@suse.cz>
14187 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14188 Local comdats are not externally visible.
14189 * symtab.c (dump_symtab_base): Dump externally visible.
14190 (verify_symtab_base): Verify back links in the symtab hash.
14192 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14194 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14195 CONVERT_EXPR as equal nodes.
14197 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14199 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14201 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14203 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14204 Generalize to complex and vector.
14205 * tree.c (build_all_ones_cst): New function.
14206 * tree.h (build_all_ones_cst): Declare it.
14208 2013-06-13 Alan Modra <amodra@gmail.com>
14210 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14211 * config/rs6000/rs6000.md (signbittf2): New insn.
14212 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14213 (abstf2_internal, cmptf_internal2): Likewise.
14214 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14216 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14217 Pat Haugen <pthaugen@us.ibm.com>
14218 Peter Bergner <bergner@vnet.ibm.com>
14220 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14221 power8 byte, half-word, and quad-word atomic instructions.
14222 (emit_store_conditional): Likewise.
14223 (rs6000_expand_atomic_compare_and_swap): Likewise.
14224 (rs6000_expand_atomic_op): Likewise.
14226 * config/rs6000/sync.md (larx): Add new modes for power8.
14228 (AINT): New mode iterator to include TImode as well as normal
14229 integer modes on power8.
14230 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14231 that VSX registers are not considered. Use AINT mode iterator
14232 instead of INT1 to allow inclusion of quad word atomic operations
14234 (load_locked<mode>): Likewise.
14235 (store_conditional<mode>): Likewise.
14236 (atomic_compare_and_swap<mode>): Likewise.
14237 (atomic_exchange<mode>): Likewise.
14238 (atomic_nand<mode>): Likewise.
14239 (atomic_fetch_<fetchop_name><mode>): Likewise.
14240 (atomic_nand_fetch<mode>): Likewise.
14241 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14243 (ATOMIC): On power8, add QImode, HImode modes.
14244 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14245 modes that promote to SImode.
14246 (load_lockedti): Convert TImode arguments to PTImode, so that we
14247 get a guaranteed even/odd register pair.
14248 (load_lockedpti): Likewise.
14249 (store_conditionalti): Likewise.
14250 (store_conditionalpti): Likewise.
14252 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14253 atomic load/store instructions.
14256 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14258 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14260 (insn_count): New attribute, with most cases extracted from...
14261 (length): ...here. Redefine most cases in terms of insn_count.
14262 (single_insn): Delete.
14263 (can_delay): Use insn_count to check for single instructions.
14264 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14265 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14266 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14267 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14268 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14269 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14270 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14271 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14272 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14273 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14274 rather than "length".
14275 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14276 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14277 Use "insn_count" rather than "length".
14278 * config/mips/mips-dsp.md
14279 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14280 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14283 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14285 PR tree-optimization/57361
14286 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14288 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14290 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14292 (aarch64_simd_combine<mode>): New instruction expansion.
14293 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14294 function prototype.
14295 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14296 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14298 2013-06-12 Jan Hubicka <jh@suse.cz>
14300 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14301 decl has when in streaming stage.
14302 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14303 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14305 2013-06-12 Roland Stigge <stigge@antcom.de>
14308 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14310 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14312 PR tree-optimization/57537
14313 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14314 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14316 2013-06-12 Richard Biener <rguenther@suse.de>
14318 * data-streamer.h (streamer_write_char_stream): CSE
14319 obs->current_pointer.
14320 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14321 streamer_write_char_stream manually and optimize the resulting loop.
14322 (streamer_write_hwi_stream): Likewise.
14324 2013-06-12 Jan Hubicka <jh@suse.cz>
14326 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14327 * cgraph.h (varpool_create_empty_node): Declare.
14328 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14330 * symtab.c (symtab_unregister_node): Be lax about missin entries
14332 (symtab_get_node): Update comment.
14333 * varpool.c (varpool_create_empty_node): Break out from ...
14334 (varpool_node_for_decl): ... here.
14335 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14337 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14339 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14340 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14341 part. Use straight-line flow at the end.
14342 <COMPONENT_REF>: Remove superfluous else.
14343 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14345 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14348 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14349 target hook even for !TREE_PUBLIC decls. If no resolution info
14350 is available, return false for common and external decls.
14352 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14354 * config/rl78/constraints.md (U): New constraint.
14355 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14358 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14361 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14362 to allow returning address to AT_PLATFORM name.
14364 2013-06-11 Jan Hubicka <jh@suse.cz>
14366 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14367 * cgraph.h (symtab_node_base): Add weakref flag.
14368 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14369 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14370 (output_weakrefs): Use weakref flag.
14371 * fold-const.c (simple_operand_p): Handle WEAK.
14372 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14373 * ipa.c (varpool_externally_visible_p): Drop weakref.
14374 (function_and_variable_visibility): Update comment; fix weakref
14375 sanity checks; do not clear DECL_WEAK on them.
14376 * lto-cgraph.c (lto_output_node): update.
14377 (lto_output_varpool_node): Update.
14378 (input_overwrite_node): Update.
14379 (input_node): Update.
14380 (input_varpool_node): Update.
14381 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14382 (lto_symtab_merge_symbols): Add sanity check.
14383 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14384 * passes.c (rest_of_decl_compilation): Set static flag, too.
14385 * symtab.c (dump_symtab_base): Dump weakref.
14386 (verify_symtab_base): Sanity check weakrefs.
14387 (symtab_make_decl_local): Remove duplicated code.
14388 (symtab_alias_ultimate_target): Simplify.
14389 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14391 2013-06-11 Tom de Vries <tom@codesourcery.com>
14393 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14394 sequence_vect == NULL.
14396 2013-06-11 DJ Delorie <dj@redhat.com>
14398 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14399 (rl78_unwind_word_mode): New.
14401 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14403 * final.c (debug_prefix_maps): Make static.
14405 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14407 * function.c (initial_trampoline): Remove stray copy.
14409 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14411 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14413 2013-06-11 Martin Jambor <mjambor@suse.cz>
14415 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14416 within bounds at the beginning of the function.
14418 2013-06-11 Alan Modra <amodra@gmail.com>
14420 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14422 (get_named_section): Don't NULL !DECL_P decl.
14424 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14426 * doc/invoke.texi (core-avx2): Document.
14428 (atom): Updated with MOVBE.
14430 2013-06-11 Richard Biener <rguenther@suse.de>
14432 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14434 2013-06-11 Anton Blanchard <anton@samba.org>
14436 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14437 correct shift value in little-endian mode.
14439 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14442 * varasm.c (get_variable_align): Move #endif to the right place.
14444 2013-06-10 Cary Coutant <ccoutant@google.com>
14446 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14447 for hash so that hash table traversal order is deterministic.
14449 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14450 Pat Haugen <pthaugen@us.ibm.com>
14451 Peter Bergner <bergner@vnet.ibm.com>
14453 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14454 of vectors in GPRS if they are direct moves or quad word load or
14457 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14459 (direct_move_p): Likewise.
14460 (quad_load_store_p): Likewise.
14462 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14463 classes into bins based on the physical register type.
14464 (reg_class_to_reg_type): Likewise.
14465 (IS_STD_REG_TYPE): Likewise.
14466 (IS_FP_VECT_REG_TYPE): Likewise.
14467 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14468 use direct move instructions.
14469 (reload_gpr_vsx): Likewise.
14470 (reload_vsx_gpr): Likewise.
14471 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14472 information that is a simplification of register classes. Also
14473 precalculate direct move reload helpers.
14474 (direct_move_p): New function to return true if the operation can
14475 be done as a direct move instruciton.
14476 (quad_load_store_p): New function to return true if the operation
14477 is a quad memory operation.
14478 (rs6000_legitimize_address): If quad memory, only allow register
14479 indirect for TImode addresses.
14480 (rs6000_legitimate_address_p): Likewise.
14481 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14482 (rs6000_reload_register_type): Likewise.
14483 (register_to_reg_type): Return register type.
14484 (rs6000_secondary_reload_simple_move): New helper function for
14485 secondary reload and secondary memory needed to identify anything
14486 that is a simple move, and does not need reloading.
14487 (rs6000_secondary_reload_direct_move): New helper function for
14488 secondary reload to identify cases that can be done with several
14489 instructions via the direct move instructions.
14490 (rs6000_secondary_reload_move): New helper function for secondary
14491 reload to identify moves between register types that can be done.
14492 (rs6000_secondary_reload): Add support for quad memory operations
14493 and for direct move.
14494 (rs6000_secondary_memory_needed): Likewise.
14495 (rs6000_debug_secondary_memory_needed): Change argument names.
14496 (rs6000_output_move_128bit): New function to return the move to
14497 use for 128-bit moves, including knowing about the various
14498 limitations of quad memory operations.
14500 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14501 memory operations. call rs6000_output_move_128bit for the actual
14502 instruciton(s) to generate.
14503 (vsx_movti_64bit): Likewise.
14505 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14506 (UNSPEC_P8V_MTVSRWZ): Likewise.
14507 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14508 (UNSPEC_P8V_MTVSRD): Likewise.
14509 (UNSPEC_P8V_XXPERMDI): Likewise.
14510 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14511 (UNSPEC_FUSION_GPR): Likewise.
14512 (FMOVE128_GPR): New iterator for direct move.
14513 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14514 (f32_sv): Likewise.
14515 (f32_dm): Likewise.
14516 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14517 loads and direct move instructions.
14518 (zero_extendsidi2_lfiwzx): Likewise.
14519 (extendsidi2_lfiwax): Likewise.
14520 (extendsidi2_nocell): Likewise.
14521 (floatsi<mode>2_lfiwax): Likewise.
14522 (lfiwax): Likewise.
14523 (floatunssi<mode>2_lfiwzx): Likewise.
14524 (lfiwzx): Likewise.
14525 (fix_trunc<mode>_stfiwx): Likewise.
14526 (fixuns_trunc<mode>_stfiwx): Likewise.
14527 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14528 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14529 (parity<mode>2_cmpb): Set length/type attr.
14530 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14531 for 'mr.' to fast_compare.
14532 (bpermd_<mode>): Change type attr to popcnt.
14533 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14534 (p8_mtvsrwz_1): Likewise.
14535 (p8_mtvsrwz_2): Likewise.
14536 (reload_fpr_from_gpr<mode>): Likewise.
14537 (p8_mtvsrd_1): Likewise.
14538 (p8_mtvsrd_2): Likewise.
14539 (p8_xxpermdi_<mode>): Likewise.
14540 (reload_vsx_from_gpr<mode>): Likewise.
14541 (reload_vsx_from_gprsf): Likewise.
14542 (p8_mfvsrd_3_<mode>): LIkewise.
14543 (reload_gpr_from_vsx<mode>): Likewise.
14544 (reload_gpr_from_vsxsf): Likewise.
14545 (p8_mfvsrd_4_disf): Likewise.
14546 (multi-word GPR splits): Do not split direct moves or quad memory
14549 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14551 * tree-into-ssa.c (interesting_blocks): Make static.
14553 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14556 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14557 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14558 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14559 (get_variable_align): New function.
14560 (get_variable_section, emit_bss, emit_common,
14561 assemble_variable_contents, place_block_symbol): Use
14562 get_variable_align instead of DECL_ALIGN.
14563 (assemble_noswitch_variable): Add align argument, use it
14564 instead of DECL_ALIGN.
14565 (assemble_variable): Adjust caller. Use get_variable_align
14566 instead of DECL_ALIGN.
14567 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14569 (DATA_ABI_ALIGNMENT): Define.
14570 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14571 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14572 opt is false, only return the psABI mandated alignment increase.
14573 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14574 (DATA_ABI_ALIGNMENT): ... this.
14575 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14576 (DATA_ABI_ALIGNMENT): ... this.
14577 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14578 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14579 (DATA_ABI_ALIGNMENT): ... this.
14580 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14581 * doc/tm.texi: Regenerated.
14583 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14585 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14586 cmp_code to construct REG_EQUAL note.
14588 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14591 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14592 that operands[2] doesn't overlap with operands[0].
14594 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14595 Jan Hubicka <jh@suse.cz>
14597 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14598 hack to mark symbols as used.
14600 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14602 PR rtl-optimization/57559
14603 * lra-constraints.c (process_alt_operands): Don't discourage
14604 memory with known offset for offsetable memory constraint.
14605 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14607 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14609 * varasm.c (struct oc_local_state): Reorder fields.
14610 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14611 and adjust accordingly.
14612 (output_constructor): Reorder initialization code and adjust call to
14613 output_constructor_bitfield.
14615 2013-06-07 Jan Hubicka <jh@suse.cz>
14617 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14619 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14621 * tree-object-size.c (unknown): Make const.
14623 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14625 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14626 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14627 for last alternative in the cpu_facility attribute.
14629 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14632 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14633 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14634 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14635 * config/arm/constraints.md (Dg): New constraint.
14636 * config/arm/neon.md (xordi3_neon): Remove.
14637 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14638 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14640 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14642 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14643 Clean up alternatives.
14645 2013-06-07 Alan Modra <amodra@gmail.com>
14647 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14650 2013-06-07 Alan Modra <amodra@gmail.com>
14652 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14654 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14656 * config/arm/constraints.md (Df): New constraint.
14657 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14658 Correct length attribute for last two alternatives.
14660 2013-06-07 Alan Modra <amodra@gmail.com>
14662 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14663 override user -mfp-in-toc.
14664 (offsettable_ok_by_alignment): Consider just the current access
14665 rather than the whole object, unless BLKmode. Handle
14666 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14667 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14668 for -mcmodel=medium.
14669 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14670 override user -mfp-in-toc or -msum-in-toc. Default to
14671 -mno-fp-in-toc for -mcmodel=medium.
14673 2013-06-06 DJ Delorie <dj@redhat.com>
14675 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14676 TARGET_VALID_POINTER_MODE.
14678 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14679 Pat Haugen <pthaugen@us.ibm.com>
14680 Peter Bergner <bergner@vnet.ibm.com>
14682 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14683 Document new power8 builtins.
14685 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14686 condition code register, to allow 128-bit logical operations to be
14687 done in the VSX or GPR registers.
14688 (nor<mode>3): Use the canonical form for nor.
14689 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14690 vclz*, and vpopcnt* vector instructions.
14691 (nand<mode>3): Likewise.
14692 (orc<mode>3): Likewise.
14693 (clz<mode>2): LIkewise.
14694 (popcount<mode>2): Likewise.
14696 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14697 that only the GPRs are recognized.
14699 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14700 support for new power8 builtins.
14702 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14704 (xscvdpspn): Likewise.
14710 (vpopcnt): Likewise.
14711 (vpopcntb): Likewise.
14712 (vpopcnth): Likewise.
14713 (vpopcntw): Likewise.
14714 (vpopcntd): Likewise.
14716 (vmrgew): Likewise.
14717 (vmrgow): Likewise.
14719 (eqv_v16qi3): Likewise.
14720 (eqv_v8hi3): Likewise.
14721 (eqv_v4si3): Likewise.
14722 (eqv_v2di3): Likewise.
14723 (eqv_v4sf3): Likewise.
14724 (eqv_v2df3): Likewise.
14726 (nand_v16qi3): Likewise.
14727 (nand_v8hi3): Likewise.
14728 (nand_v4si3): Likewise.
14729 (nand_v2di3): Likewise.
14730 (nand_v4sf3): Likewise.
14731 (nand_v2df3): Likewise.
14733 (orc_v16qi3): Likewise.
14734 (orc_v8hi3): Likewise.
14735 (orc_v4si3): Likewise.
14736 (orc_v2di3): Likewise.
14737 (orc_v4sf3): Likewise.
14738 (orc_v2df3): Likewise.
14740 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14741 allow power8 quad mode in 64-bit.
14742 (rs6000_builtin_vectorized_function): Add support to vectorize
14743 ISA 2.07 count leading zeros, population count builtins.
14744 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14745 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14746 (builtin_function_type): Add vgbbd builtin function which takes an
14748 (altivec_expand_vec_perm_const): Add support for new power8 merge
14751 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14752 that does not include TImdoe for use with 32-bit.
14753 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14755 (UNSPEC_VSX_CVDPSPN): Likewise.
14756 (vsx_xscvdpspn): Likewise.
14757 (vsx_xscvspdpn): Likewise.
14758 (vsx_xscvdpspn_scalar): Likewise.
14759 (vsx_xscvspdpn_directmove): Likewise.
14760 (vsx_and<mode>3): Split logical operations into 32-bit and
14761 64-bit. Add support to do logical operations on TImode as well as
14762 VSX vector types. Allow logical operations to be done in either
14763 VSX registers or in general purpose registers in 64-bit mode. Add
14764 splitters if GPRs were used. For AND, add clobber of CCmode to
14765 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14767 (vsx_and<mode>3_32bit): Likewise.
14768 (vsx_and<mode>3_64bit): Likewise.
14769 (vsx_ior<mode>3): Likewise.
14770 (vsx_ior<mode>3_32bit): Likewise.
14771 (vsx_ior<mode>3_64bit): Likewise.
14772 (vsx_xor<mode>3): Likewise.
14773 (vsx_xor<mode>3_32bit): Likewise.
14774 (vsx_xor<mode>3_64bit): Likewise.
14775 (vsx_one_cmpl<mode>2): Likewise.
14776 (vsx_one_cmpl<mode>2_32bit): Likewise.
14777 (vsx_one_cmpl<mode>2_64bit): Likewise.
14778 (vsx_nor<mode>3): Likewise.
14779 (vsx_nor<mode>3_32bit): Likewise.
14780 (vsx_nor<mode>3_64bit): Likewise.
14781 (vsx_andc<mode>3): Likewise.
14782 (vsx_andc<mode>3_32bit): Likewise.
14783 (vsx_andc<mode>3_64bit): Likewise.
14784 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14785 and xxlorc instructions.
14786 (vsx_eqv<mode>3_64bit): Likewise.
14787 (vsx_nand<mode>3_32bit): Likewise.
14788 (vsx_nand<mode>3_64bit): Likewise.
14789 (vsx_orc<mode>3_32bit): Likewise.
14790 (vsx_orc<mode>3_64bit): Likewise.
14792 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14794 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14796 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14797 (p8_vmrgow): Likewise.
14798 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14799 GPRs to be split under VSX.
14800 (p8v_clz<mode>2): Add power8 count leading zero support.
14801 (p8v_popcount<mode>2): Add power8 population count support.
14802 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14805 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14808 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14810 (vec_nand): Likewise.
14811 (vec_vclz): Likewise.
14812 (vec_vclzb): Likewise.
14813 (vec_vclzd): Likewise.
14814 (vec_vclzh): Likewise.
14815 (vec_vclzw): Likewise.
14816 (vec_vgbbd): Likewise.
14817 (vec_vmrgew): Likewise.
14818 (vec_vmrgow): Likewise.
14819 (vec_vpopcnt): Likewise.
14820 (vec_vpopcntb): Likewise.
14821 (vec_vpopcntd): Likewise.
14822 (vec_vpopcnth): Likewise.
14823 (vec_vpopcntw): Likewise.
14825 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14827 PR rtl-optimization/57468
14828 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14831 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14833 PR rtl-optimization/57459
14834 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14835 type when setting live regs.
14837 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14839 * config/s390/s390.opt (mlra): New option.
14840 * config/s390/s390.c (s390_decompose_address): Check displacement
14841 for all registers for LRA.
14842 (s390_secondary_reload): Don't used secondary reloads for LRA.
14843 (s390_lra_p): New function.
14844 (TARGET_LRA_P): Define.
14845 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14846 of attribute cpu_facility to zarch for the last alternative.
14847 (*cmpmem_short): Ditto.
14849 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14851 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14852 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14853 (arm_expand_prologue): Likewise.
14855 2013-06-06 Teresa Johnson <tejohnson@google.com>
14858 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14859 as this is now done by redirect_edge_and_branch_force.
14860 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14861 barriers, and fix interaction with splitting.
14862 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14863 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14864 reflect changes made in the routine.
14865 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14866 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14867 since this is called in cfglayout mode, and replace partition fixup
14868 with assert as that is now done by force_nonfallthru_and_redirect.
14869 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14870 already be marked with region crossing note.
14871 (insert_section_boundary_note): Make non-static, gate on flag
14872 has_bb_partition, rewrite to also check for multiple partitions.
14873 (rest_of_handle_reorder_blocks): Remove call to
14874 insert_section_boundary_note, now done later during free_cfg.
14875 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14876 * bb-reorder.h (insert_section_boundary_note): Declare.
14877 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14878 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14879 invoke insert_section_boundary_note.
14880 (try_redirect_by_replacing_jump): Remove unnecessary
14881 check for region crossing note.
14882 (fixup_partition_crossing): New function.
14883 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14884 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14885 in non-cfglayout mode.
14886 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14887 remove old code that tried to do this. Emit barrier correctly
14888 when we are in cfglayout mode.
14889 (last_bb_in_partition): New function.
14890 (rtl_split_edge): Correctly fixup partition boundaries.
14891 (commit_one_edge_insertion): Remove old code that tried to
14892 fixup region crossing edge since this is now handled in
14893 split_block, and set up insertion point correctly since
14894 block may now end in a jump.
14895 (verify_hot_cold_block_grouping): Guard against checking when not in
14896 linearized RTL mode.
14897 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14899 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14900 rtl_verify_flow_info, so not called in cfglayout mode.
14901 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14902 (fixup_reorder_chain): Remove old code that attempted to fixup region
14903 crossing note as this is now handled in force_nonfallthru_and_redirect.
14904 (duplicate_insn_chain): Don't duplicate switch section notes.
14905 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14907 * basic-block.h (emit_barrier_after_bb): Declare.
14909 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14911 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14912 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14913 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14915 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14917 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14919 * config/arm/ldmstm.md: Regenerate.
14921 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14923 * config/arm/sync.md (atomic_loaddi_1):
14924 Disable predication for arm_restrict_it.
14925 (arm_load_exclusive<mode>): Likewise.
14926 (arm_load_exclusivesi): Likewise.
14927 (arm_load_exclusivedi): Likewise.
14928 (arm_load_acquire_exclusive<mode>): Likewise.
14929 (arm_load_acquire_exclusivesi): Likewise.
14930 (arm_load_acquire_exclusivedi): Likewise.
14931 (arm_store_exclusive<mode>): Likewise.
14932 (arm_store_exclusive<mode>): Likewise.
14933 (arm_store_release_exclusivedi): Likewise.
14934 (arm_store_release_exclusive<mode>): Likewise.
14936 2013-06-06 Richard Biener <rguenther@suse.de>
14938 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14940 (lto_tag_is_tree_code_p): Adjust.
14941 (lto_tag_is_gimple_code_p): Likewise.
14942 (lto_gimple_code_to_tag): Likewise.
14943 (lto_tag_to_gimple_code): Likewise.
14944 (lto_tree_code_to_tag): Likewise.
14945 (lto_tag_to_tree_code): Likewise.
14946 * data-streamer.h (streamer_write_hwi_in_range): Use
14947 uhwi streaming to stream the normalized range.
14948 (streamer_read_hwi_in_range): Likewise.
14950 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14952 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14953 (predicable_short_it): Likewise.
14954 (predicated): Likewise.
14955 (enabled): Handle above.
14956 (define_cond_exec): Set predicated attribute to yes.
14958 2013-06-05 Mike Stump <mikestump@comcast.net>
14960 * gdbinit.in (__FUNCTION__): Add.
14962 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14964 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14965 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14967 2013-06-05 Jan Hubicka <jh@suse.cz>
14969 * varasm.c (mark_decl_referenced): Revert the removal until targets
14972 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14974 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14975 instead of mark_decl_referenced.
14977 2013-06-05 Jan Hubicka <jh@suse.cz>
14979 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14980 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14981 and symtab_used_from_object_file_p.
14982 (cgraph_make_node_local_1): Clear forced_by_abi.
14983 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14984 * cgraph.h (symtab_node_base): Add forced_by_abi.
14985 (decide_is_variable_needed): Remove.
14986 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14987 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14988 (decide_is_symbol_needed): ... this one; handle symbols in general;
14989 always analyze virtuals; honnor forced_by_abi.
14990 (cgraph_finalize_function): Update.
14991 (varpool_finalize_decl): Update.
14992 (symbol_defined_and_needed): Remove.
14993 (analyze_functions): Update.
14994 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14995 output_refs, input_overwrite_node): Handle forced_by_abi.
14996 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14997 (address_taken_from_non_vtable_p): ... this one.
14998 (comdat_can_be_unshared_p_1): New function.
14999 (cgraph_comdat_can_be_unshared_p): Rename to ...
15000 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15001 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15002 (function_and_variable_visibility): Clear forced_by_abi as needed.
15003 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15004 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15005 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15006 * varpool.c (decide_is_variable_needed): Remove.
15008 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15010 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15011 (arm_option_override): Override arm_restrict_it where appropriate.
15012 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15013 * config/arm/arm.opt (mrestrict-it): New command-line option.
15014 * doc/invoke.texi: Document -mrestrict-it.
15016 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15018 * tsan.c (tsan_atomic_table): Make const.
15020 2013-06-05 Richard Biener <rguenther@suse.de>
15022 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15023 index associated with the tree we are supposed to replace.
15024 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15025 TREE_ASM_WRITTEN as zero for everything but SSA names.
15027 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15029 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15031 2013-06-04 Jan Hubicka <jh@suse.cz>
15033 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15034 (input_node, input_varpool_node): Handle correctly external same
15036 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15037 nodes at ltrans stage.
15039 2013-06-04 Jan Hubicka <jh@suse.cz>
15041 * ipa-inline.c (update_caller_keys): Fix availability test.
15042 (update_callee_keys): Likewise.
15043 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15044 to follow ELF standard.
15046 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15048 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15049 (mips64r5900el-*-elf*): New configurations.
15050 * config/mips/mips-cpus.def (r5900): New processor.
15051 * config/mips/mips-tables.opt: Regenerate.
15052 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15053 (mips_issue_rate): Handle PROCESSOR_R5900.
15054 (mips_reorg_process_insns): Force reorder mode for the R5900.
15055 * config/mips/mips.h (TARGET_MIPS5900): Define.
15056 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15058 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15060 * config/mips/mips.md (processor): Add r5900.
15061 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15063 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15065 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15066 into function to generate MOVI instruction.
15067 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15068 (aarch64_preferred_simd_mode): Turn into wrapper.
15069 (aarch64_output_scalar_simd_mov_immediate): New function.
15070 * config/aarch64/aarch64-protos.h: Add prototype for above.
15072 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15074 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15075 element_char member.
15076 (sizetochar): Return signed char.
15077 (aarch64_simd_valid_immediate): Remove elchar and other
15078 unnecessary variables.
15079 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15080 Calculate element_char as required.
15081 * config/aarch64/aarch64-protos.h: Update and move prototype
15082 for aarch64_output_simd_mov_immediate.
15083 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15086 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15088 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15089 information completed by aarch64_simd_valid_immediate.
15090 (aarch64_legitimate_constant_p): Update arguments.
15091 (aarch64_simd_valid_immediate): Work with struct rather than many
15093 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15094 (aarch64_simd_make_constant): Update arguments.
15095 (aarch64_output_simd_mov_immediate): Work with struct rather than
15096 many pointers. Output immediate directly rather than as operand.
15097 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15099 * config/aarch64/constraints.md (Dn): Update arguments.
15101 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15103 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15105 (aarch64_simd_immediate_valid_for_move): Remove.
15106 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15107 (aarch64_simd_make_constant): Update call.
15108 (aarch64_output_simd_mov_immediate): Update call.
15109 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15111 * config/aarch64/constraints.md (Dn): Update call.
15113 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15115 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15116 return type to bool for prototype.
15117 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15118 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15119 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15121 2013-06-04 Catherine Moore <clm@codesourcery.com>
15123 * config/mips/mips.opt (meva): New.
15124 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15125 (ASM_SPEC): Handle -meva.
15126 * doc/invoke.texi (meva): Document.
15128 2013-06-04 Alan Modra <amodra@gmail.com>
15130 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15133 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15135 * rtl.def: Add extra fourth optional field to define_cond_exec.
15136 * gensupport.c (process_one_cond_exec): Process attributes from
15138 * doc/md.texi: Document fourth field in define_cond_exec.
15140 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15142 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15143 out the processing order as in store_bit_field_1.
15145 2013-06-04 Jan Hubicka <jh@suse.cz>
15147 PR middle-end/57500
15148 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15149 non-VAR_DECL node if it does not exist yet.
15151 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15153 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15154 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15155 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15156 target_cpu_default setting.
15158 2013-06-03 Teresa Johnson <tejohnson@google.com>
15160 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15161 default to -fopt-info=optimized instead of all.
15162 * doc/invoke.texi: Ditto.
15163 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15164 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15165 (execute_vect_slp): Emit BB vectorization success under
15166 MSG_OPTIMIZED_LOCATIONS.
15167 * tree-vect-slp.c (vect_slp_transform_bb): Change
15168 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15169 * tree-vect-loop.c (vect_transform_loop): Ditto.
15171 2013-06-03 Jason Merrill <jason@redhat.com>
15174 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15175 Use TARGET_EXPR for C++.
15177 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15179 PR rtl-optimization/57268
15180 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15181 if DEBUG_INSN_P (insn).
15184 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15186 PR rtl-optimization/57268
15187 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15188 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15190 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15192 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15193 (ix86_avoid_lea_for_addr): Likewise.
15194 (exact_dependency_1): Likewise.
15195 (ix86_adjust_cost): Likewise.
15196 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15198 (do_reorder_for_imul): Fix typo, formatting and
15199 !reload_completed check removed.
15200 (ix86_sched_reorder): Fix typo and formatting.
15201 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15204 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15206 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15208 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15210 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15211 <STRING_CST>: Likewise.
15212 <VECTOR_CST>: Likewise.
15214 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15215 Mikael Morin <mikael@gcc.gnu.org>
15217 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15218 * config.in: Regenerated.
15219 * configure: Regenerated.
15221 2013-06-01 Jan Hubicka <jh@suse.cz>
15223 PR middle-end/57366
15224 * cgraphunit.c (compile): When weakref is not supported,
15225 set up transparent aliases before final output pass.
15226 * varasm.c (assemble_alias): Do not try to do it here.
15228 2013-06-01 Jan Hubicka <jh@suse.cz>
15230 PR middle-end/57467
15231 * passes.c (for_per_function): Skip unanalyzed functions.
15233 2013-06-01 Jan Hubicka <jh@suse.cz>
15235 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15236 (lto_symtab_merge_symbols_1): ... this one.
15237 (lto_symtab_merge_cgraph_nodes): Rename to ...
15238 (lto_symtab_merge_symbols): ... this one; simplify.
15239 * cgraph.c (same_body_aliases_done): Rename to ...
15240 (cpp_implicit_aliases_done): ... this one.
15241 (cgraph_create_function_alias): Update.
15242 (cgraph_same_body_alias): Update.
15243 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15244 (verify_edge_corresponds_to_fndecl): Simplify.
15245 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15246 (cgraph_node): Remove same_body_alias.
15247 (varpool_node): Remove alias_of and extra_name_alias.
15248 (same_body_aliases_done): Rename to ..
15249 (cpp_implicit_aliases_done): ... this one.
15250 (symtab_alias_ultimate_target): Add default parameter.
15251 (symtab_resolve_alias): New function.
15252 (fixup_same_cpp_alias_visibility): Declare.
15253 (cgraph_function_node): Add default parameter.
15254 (cgraph_node_asm_name): Likewise.
15255 (cgraph_function_or_thunk_node): Add default parameter; do
15256 not ICE when it is NULL.
15257 (varpool_variable_node): Likewise.
15258 * tree-emutls.c (create_emultls_var): Update.
15259 (ipa_lower_emutls): Update.
15260 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15261 (cgraph_reset_node): Reset alias info.
15262 (cgraph_finalize_function): Update.
15263 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15264 (analyze_function): Simplify.
15265 (cgraph_process_same_body_aliases): Simplify.
15266 (analyze_functions): Fixup same body aliases.
15267 (handle_alias_pairs): Simplify.
15268 (assemble_thunk): Update.
15269 (assemble_thunks_and_aliases): Update.
15270 (output_weakrefs): Rewrite.
15271 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15272 (lto_output_varpool_node): Likewise.
15273 (compute_ltrans_boundary): Remve assert.
15274 (get_alias_symbol): New functoin.
15275 (input_node): Rewrite alias handling.
15276 (input_varpool_node): Likewise.
15277 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15278 * ipa.c (process_references): Handle weakrefs correctly.
15279 (symtab_remove_unreachable_nodes): Likewise.
15280 * trans-mem.c (get_cg_data): Update.
15281 (ipa_tm_create_version_alias): Update.
15282 (ipa_tm_execute): Update.
15283 * symtab.c (dump_symtab_base): Dump aliases.
15284 (verify_symtab_base): Verify aliases.
15285 (symtab_node_availability): New function.
15286 (symtab_alias_ultimate_target): Simplify.
15287 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15288 handle all the fixup cases.
15289 (symtab_resolve_alias): New function.
15290 * passes.c (ipa_write_summaries): Handle weakrefs.
15291 * varpool.c (varpool_analyze_node): Simplify.
15292 (assemble_aliases): Update.
15293 (varpool_create_variable_alias): Simplify.
15294 (varpool_extra_name_alias): Simplify.
15295 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15296 (lto_symtab_merge_symbols): ... this one.
15298 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15301 PR rtl-optimization/57268
15302 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15303 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15305 2013-06-01 Tobias Burnus <burnus@net-b.de>
15307 Partially reverted:
15308 2013-05-31 Tobias Burnus <burnus@net-b.de>
15310 PR middle-end/57073
15311 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15314 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15316 PR rtl-optimization/57268
15317 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15318 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15320 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15322 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15323 unordered comparison operators when -fno-trapping-math is in effect
15325 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15326 and implement unordered comparison operators properly on the e500.
15328 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15330 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15331 for constant scalar integers.
15332 (simplify_relational_operation_1): Likewise.
15334 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15336 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15337 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15340 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15341 Igor Zamyatin <igor.zamyatin@intel.com>
15343 Silvermont (SLM) architecture performance tuning.
15344 * config/i386/i386.h (enum ix86_tune_indices): Add
15345 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15346 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15348 * config/i386/i386.c (initial_ix86_tune_features)
15349 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15350 (ix86_lea_outperforms): Handle Silvermont tuning.
15351 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15353 (ix86_use_lea_for_mov): Likewise.
15354 (ix86_avoid_lea_for_addr): Likewise.
15355 (ix86_lea_for_add_ok): Likewise.
15356 (exact_dependency_1): New function.
15357 (exact_store_load_dependency): Likewise.
15358 (ix86_adjust_cost): Handle Silvermont tuning.
15359 (do_reoder_for_imul): Likewise.
15360 (swap_top_of_ready_list): New function.
15361 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15363 * config/i386/i386.md (peepholes that split memory operand in fp
15366 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15368 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15369 Remove un-necessary braces.
15371 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15373 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15374 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15376 2013-05-31 Tobias Burnus <burnus@net-b.de>
15378 PR middle-end/57073
15379 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15381 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15384 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15385 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15386 * config/arm/neon.md (iordi3_neon): Remove.
15387 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15388 * config/arm/predicates.md (imm_for_neon_logic_operand):
15389 Move to earlier in the file.
15390 (neon_logic_op2): Likewise.
15391 (arm_iordi_operand_neon): New predicate.
15393 2013-05-31 Richard Biener <rguenther@suse.de>
15395 PR tree-optimization/57478
15396 PR tree-optimization/57453
15397 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15400 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15402 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15403 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15405 2013-05-30 Tobias Burnus <burnus@net-b.de>
15406 Thomas Koenig <tkoenig@gcc.gnu.org>
15408 PR middle-end/57073
15409 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15410 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15412 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15414 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15416 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15418 * target.def (register_usage_leveling_p): New hook.
15419 * targhooks.c (default_register_usage_leveling_p): New.
15420 * targhooks.h (default_register_usage_leveling_p): New prototype.
15421 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15422 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15423 * doc/tm.texi: Update.
15424 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15426 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15428 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15429 (*insv_reg<mode>): New define_insn.
15431 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15433 PR rtl-optimization/57439
15434 * postreload.c (move2add_valid_value_p): Check that we have
15435 a zero subreg_regno_offset when accessing the register in
15436 the requested mode.
15438 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15439 Igor Zamyatin <igor.zamyatin@intel.com>
15441 Silvermont (SLM) architecture pipeline model, tuning and
15443 * config.gcc: Add slm config options and target.
15445 * config/i386/slm.md: New.
15447 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15449 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15451 (ix86_target_macros_internal): Likewise.
15453 * config/i386/i386.c (slm_cost): New cost.
15454 (m_SLM): New macro flag.
15455 (initial_ix86_tune_features): Set m_SLM.
15456 (x86_accumulate_outgoing_args): Likewise.
15457 (x86_arch_always_fancy_math_387): Likewise.
15458 (processor_target_table): Add slm cost.
15459 (cpu_names): Add slm cpu name.
15460 (x86_option_override_internal): Set SLM ISA.
15461 (ix86_issue_rate): New case PROCESSOR_SLM.
15462 (ia32_multipass_dfa_lookahead): Likewise.
15463 (fold_builtin_cpu): Add slm.
15465 * config/i386/i386.h (TARGET_SLM): New target macro.
15466 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15467 (processor_type): Add PROCESSOR_SLM.
15469 * config/i386/i386.md (cpu): Add new value "slm".
15470 (slm.md): Include slm.md.
15472 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15473 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15475 * config/arm/arm-protos.h: Add and update function protos.
15476 * config/arm/arm.c (use_simple_return_p): New added.
15477 (thumb2_expand_return): Check simple_return flag.
15478 * config/arm/arm.md: Add simple_return and conditional simple_return.
15479 * config/arm/iterators.md: Add iterator for return and simple_return.
15481 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15483 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15484 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15485 (arm_emit_vfp_multi_reg_pop): Likewise.
15486 (thumb2_emit_ldrd_pop): Likewise.
15487 (arm_expand_epilogue): Add misc REG_CFA notes.
15488 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15490 2013-05-29 Lawrence Crowl <crowl@google.com>
15492 * config/arm/t-arm: Update for below.
15494 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15495 Change type to hash_table. Update dependent calls and types.
15497 * config/i386/t-cygming: Update for below.
15499 * config/i386/t-interix: Update for below.
15501 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15502 Change type to hash_table. Update dependent calls and types.
15503 (i386_find_on_wrapper_list::wrappers): Likewise.
15505 * config/ia64/t-ia64: Update for below.
15507 * config/ia64/ia64.c (bundle_state_table):
15508 Change type to hash_table. Update dependent calls and types.
15510 * config/mips/mips.c (mips_reorg_process_insns::htab):
15511 Change type to hash_table. Update dependent calls and types.
15513 * config/sol2.c (solaris_comdat_htab):
15514 Change type to hash_table. Update dependent calls and types.
15516 * config/t-sol2: Update for above.
15518 2013-05-29 Teresa Johnson <tejohnson@google.com>
15520 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15521 functions are not yet marked as defined.
15523 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15524 Pat Haugen <pthaugen@us.ibm.com>
15525 Peter Bergner <bergner@vnet.ibm.com>
15527 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15531 (vrotl<mode>3): Likewise.
15532 (vashl<mode>3): Likewise.
15533 (vlshr<mode>3): Likewise.
15534 (vashr<mode>3): Likewise.
15536 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15537 support for power8 V2DI builtins.
15539 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15540 power8 V2DI builtins.
15541 (vupkhsw): Likewise.
15542 (vupklsw): Likewise.
15543 (vaddudm): Likewise.
15544 (vminsd): Likewise.
15545 (vmaxsd): Likewise.
15546 (vminud): Likewise.
15547 (vmaxud): Likewise.
15548 (vpkudum): Likewise.
15549 (vpksdss): Likewise.
15550 (vpkudus): Likewise.
15551 (vpksdus): Likewise.
15556 (vsubudm): Likewise.
15557 (vcmpequd): Likewise.
15558 (vcmpgtsd): Likewise.
15559 (vcmpgtud): Likewise.
15560 (vcmpequd_p): Likewise.
15561 (vcmpgtsd_p): Likewise.
15562 (vcmpgtud_p): Likewise.
15563 (vupkhsw): Likewise.
15564 (vupklsw): Likewise.
15565 (vaddudm): Likewise.
15566 (vmaxsd): Likewise.
15567 (vmaxud): Likewise.
15568 (vminsd): Likewise.
15569 (vminud): Likewise.
15570 (vpksdss): Likewise.
15571 (vpksdus): Likewise.
15572 (vpkudum): Likewise.
15573 (vpkudus): Likewise.
15578 (vsubudm): Likewise.
15580 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15581 support for power8 V2DI instructions.
15583 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15584 power8 V2DI instructions. Combine pack and unpack insns to use an
15585 iterator for each mode. Check whether a particular mode supports
15586 Altivec instructions instead of just checking TARGET_ALTIVEC.
15587 (UNSPEC_VPKUWUM): Likewise.
15588 (UNSPEC_VPKSHSS): Likewise.
15589 (UNSPEC_VPKSWSS): Likewise.
15590 (UNSPEC_VPKUHUS): Likewise.
15591 (UNSPEC_VPKSHUS): Likewise.
15592 (UNSPEC_VPKUWUS): Likewise.
15593 (UNSPEC_VPKSWUS): Likewise.
15594 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15595 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15596 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15597 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15598 (UNSPEC_VUPKHSB): Likewise.
15599 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15600 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15601 (UNSPEC_VUPKHSH): Likewise.
15602 (UNSPEC_VUPKLSB): Likewise.
15603 (UNSPEC_VUPKLSH): Likewise.
15605 (VI_char): Likewise.
15606 (VI_scalar): Likewise.
15607 (VI_unit): Likewise.
15609 (VP_small): Likewise.
15610 (VP_small_lc): Likewise.
15611 (VU_char): Likewise.
15612 (add<mode>3): Likewise.
15613 (altivec_vaddcuw): Likewise.
15614 (altivec_vaddu<VI_char>s): Likewise.
15615 (altivec_vadds<VI_char>s): Likewise.
15616 (sub<mode>3): Likewise.
15617 (altivec_vsubcuw): Likewise.
15618 (altivec_vsubu<VI_char>s): Likewise.
15619 (altivec_vsubs<VI_char>s): Likewise.
15620 (altivec_vavgs<VI_char>): Likewise.
15621 (altivec_vcmpbfp): Likewise.
15622 (altivec_eq<mode>): Likewise.
15623 (altivec_gt<mode>): Likewise.
15624 (altivec_gtu<mode>): Likewise.
15625 (umax<mode>3): Likewise.
15626 (smax<mode>3): Likewise.
15627 (umin<mode>3): Likewise.
15628 (smin<mode>3): Likewise.
15629 (altivec_vpkuhum): Likewise.
15630 (altivec_vpkuwum): Likewise.
15631 (altivec_vpkshss): Likewise.
15632 (altivec_vpkswss): Likewise.
15633 (altivec_vpkuhus): Likewise.
15634 (altivec_vpkshus): Likewise.
15635 (altivec_vpkuwus): Likewise.
15636 (altivec_vpkswus): Likewise.
15637 (altivec_vpks<VI_char>ss): Likewise.
15638 (altivec_vpks<VI_char>us): Likewise.
15639 (altivec_vpku<VI_char>us): Likewise.
15640 (altivec_vpku<VI_char>um): Likewise.
15641 (altivec_vrl<VI_char>): Likewise.
15642 (altivec_vsl<VI_char>): Likewise.
15643 (altivec_vsr<VI_char>): Likewise.
15644 (altivec_vsra<VI_char>): Likewise.
15645 (altivec_vsldoi_<mode>): Likewise.
15646 (altivec_vupkhsb): Likewise.
15647 (altivec_vupkhs<VU_char>): Likewise.
15648 (altivec_vupkls<VU_char>): Likewise.
15649 (altivec_vupkhsh): Likewise.
15650 (altivec_vupklsb): Likewise.
15651 (altivec_vupklsh): Likewise.
15652 (altivec_vcmpequ<VI_char>_p): Likewise.
15653 (altivec_vcmpgts<VI_char>_p): Likewise.
15654 (altivec_vcmpgtu<VI_char>_p): Likewise.
15655 (abs<mode>2): Likewise.
15656 (vec_unpacks_hi_v16qi): Likewise.
15657 (vec_unpacks_hi_v8hi): Likewise.
15658 (vec_unpacks_lo_v16qi): Likewise.
15659 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15660 (vec_unpacks_lo_v8hi): Likewise.
15661 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15662 (vec_pack_trunc_v8h): Likewise.
15663 (vec_pack_trunc_v4si): Likewise.
15664 (vec_pack_trunc_<mode>): Likewise.
15666 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15668 (vec_vmaxsd): Likewise.
15669 (vec_vmaxud): Likewise.
15670 (vec_vminsd): Likewise.
15671 (vec_vminud): Likewise.
15672 (vec_vpksdss): Likewise.
15673 (vec_vpksdus): Likewise.
15674 (vec_vpkudum): Likewise.
15675 (vec_vpkudus): Likewise.
15676 (vec_vrld): Likewise.
15677 (vec_vsld): Likewise.
15678 (vec_vsrad): Likewise.
15679 (vec_vsrd): Likewise.
15680 (vec_vsubudm): Likewise.
15681 (vec_vupkhsw): Likewise.
15682 (vec_vupklsw): Likewise.
15684 2013-05-29 Jan Hubicka <jh@suse.cz>
15686 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15687 flags; reorder rest of fields in more consistent way.
15688 (varpool_node): Remove analyzed, finalized and alias.
15689 (cgraph_ndoe): Likewise.
15690 (symtab_alias_ultimate_target): New function.
15691 (cgraph_function_node): Move offline.
15692 (cgraph_reset_node): Declare.
15693 (cgraph_comdat_can_be_unshared_p): Remove.
15694 (varpool_remove_initializer): Declare.
15695 (varpool_first_defined_variable, varpool_next_defined_variable
15696 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15697 (cgraph_function_with_gimple_body_p): Update.
15698 (varpool_all_refs_explicit_p): Update.
15699 (symtab_alias_target): New function.
15700 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15701 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15702 (cgraph_function_or_thunk_node): Simplify using
15703 symtab_alias_ultimate_target.
15704 (varpool_variable_node): Likewise.
15705 * cgraph.c (cgraph_create_function_alias): Update.
15706 (cgraph_add_thunk): Update.
15707 (cgraph_remove_node): Update.
15708 (dump_cgraph_node): Do not dump removed flags.
15709 (cgraph_function_body_availability): Update.
15710 (cgraph_propagate_frequency): Update.
15711 (verify_cgraph_node): Check sanity of local flag.
15712 (cgraph_function_node): Move here from cgraph.h; revamp for
15713 cgraph_function_or_thunk_node.
15714 * lto-symtab.c (lto_varpool_replace_node): Update.
15715 (lto_symtab_resolve_can_prevail_p): Update.
15716 (lto_symtab_merge_cgraph_nodes): Update.
15717 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15718 propagate_constants_accross_call, devirtualization_time_bonus,
15719 ipcp_propagate_stage): Update.
15720 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15721 * ipa-inline-transform.c (clone_inlined_nodes,
15722 preserve_function_body_p): Update.
15723 * ipa-reference.c (propagate): Update.
15724 (write_node_summary_p): Update.
15725 * toplev.c (wrapup_global_declaration_2): Update.
15726 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15727 (analyze_function) ... this one.
15728 (cgraph_process_new_functions): Update.
15729 (cgraph_reset_node): Export.
15730 (cgraph_finalize_function): Update.
15731 (cgraph_add_new_function): Update.
15732 (process_function_and_variable_attributes): Update.
15733 (varpool_finalize_decl): Update.
15734 (symbol_finalized): Remove.
15735 (symbol_finalized_and_needed): Rename to ...
15736 (symbol_defined_and_needed): ... update.
15737 (cgraph_analyze_functions): Update.
15738 (handle_alias_pairs): Update.
15739 (mark_functions_to_output): Update.
15740 (assemble_thunk): Update.
15741 (output_in_order): Update.
15742 (output_weakrefs): Update.
15743 (finalize_compilation_unit): Update.
15744 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15745 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15746 input_node, input_varpool_node): Update.
15747 * dbxout.c (dbxout_expand_expr): Update.
15748 * cgraphclones.c (cgraph_clone_node): Update.
15749 (cgraph_copy_node_for_versioning): Update.
15750 (cgraph_materialize_clone): Update.
15751 (cgraph_materialize_all_clones): Update.
15752 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15753 propagate_pure_const, propagate_nothrow): Update.
15754 * lto-streamer-out.c (lto_output, write_symbol): Update.
15755 * ipa-utils.c (ipa_reverse_postorder): Update.
15756 * ipa-inline.c (can_inline_edge_p): Update.
15757 (update_caller_keys, ipa_inline): Update.
15758 * dwarf2out.c (reference_to_unused,
15759 premark_types_used_by_global_vars_helper): Update.
15760 * tree-eh.c (tree_could_trap_p): Update.
15761 * ipa-split.c (consider_split, execute_split_functions): Update.
15762 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15763 has_addr_references_p): Update; move ahead in file for better
15765 (process_references): Simplify.
15766 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15767 bodies are removed.
15768 (cgraph_comdat_can_be_unshared_p): Make static.
15769 (cgraph_externally_visible_p): Update.
15770 (varpool_externally_visible_p): Update.
15771 (function_and_variable_visibility): Update.
15772 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15773 ipa_tm_mark_force_output_node): Update.
15774 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15775 estimate_edge_devirt_benefit, inline_generate_summary,
15776 inline_write_summary): Update.
15777 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15778 * ipa-prop.c (ipa_compute_jump_functions): Update.
15779 (ipa_print_node_params, ipa_prop_read_section,
15780 ipa_update_after_lto_read, read_replacements_section): Update.
15781 * varasm.c (mark_decl_referenced): Update.
15782 (assemble_alias, dump_tm_clone_pairs): Update.
15783 * tree-inline.c (copy_bb): Update.
15784 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15786 * symtab.c (dump_symtab_base): Print new flags.
15787 (verify_symtab_base): Verify new flags.
15788 (symtab_alias_ultimate_target): New function.
15789 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15790 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15792 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15794 * i386.c (ix86_get_function_versions_dispatcher,
15795 ix86_generate_version_dispatcher_body): Update.
15796 (fold_builtin_cpu): Use varpool_add_new_variable.
15797 * varpool.c (varpool_remove_initializer): Break out from ...
15798 (varpool_remove_node): ... this one.
15799 (dump_varpool_node, varpool_node_for_asm,
15800 cgraph_variable_initializer_availability, varpool_analyze_node,
15801 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15802 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15805 2013-05-29 Jan Hubicka <jh@suse.cz>
15807 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15809 2013-05-29 Easwaran Raman <eraman@google.com>
15811 PR tree-optimization/57442
15812 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15813 when control exits the main loop.
15815 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15817 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15819 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15820 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15821 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15824 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15826 PR tree-optimization/57441
15827 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15828 Don't limit size of incr_vec to number of candidates.
15830 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15832 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15833 and mips16 directories.
15834 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15835 (MULTILIB_DIRNAMES): Ditto.
15836 (MULTILIB_EXCEPTIONS): Add new exceptions.
15837 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15838 (MULTILIB_DIRNAMES): Ditto.
15839 (MULTILIB_EXCEPTIONS): Add new exceptions.
15841 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15842 Marcus Shawcroft <marcus.shawcroft@arm.com>
15844 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15845 SYMBOL_TINY_ABSOLUTE.
15846 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15847 SYMBOL_TINY_ABSOLUTE.
15848 (aarch64_expand_mov_immediate): Likewise.
15849 (aarch64_classify_symbol): Likewise.
15850 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15851 Permit SYMBOL_TINY_ABSOLUTE.
15852 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15854 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15855 Marcus Shawcroft <marcus.shawcroft@arm.com>
15857 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15858 Refactor if/switch. Replace gcc_assert with if.
15860 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15862 * config/i386/i386.c (initial_ix86_tune_features): Enable
15863 FP Reassociation for AMD bdver1 and bdver2.
15865 2013-05-29 Martin Jambor <mjambor@suse.cz>
15867 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15868 and IMAGPART_EXPR do not occur within other handled_components.
15870 2013-05-29 Richard Biener <rguenther@suse.de>
15872 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15873 access on whether the use is in the BB we currently try to
15875 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15876 vectorize to vect_bb_slp_scalar_cost.
15878 2013-05-29 Richard Biener <rguenther@suse.de>
15880 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15881 computing scalar cost offsetted by stmts that are kept live
15883 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15884 for computation of scalar cost.
15886 2013-05-28 Steve Ellcey <sellcey@mips.com>
15888 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15890 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15892 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15893 array notation built-in reduction functions.
15894 * doc/passes.texi (Passes): Added documentation about changes done
15896 * doc/invoke.texi (C Dialect Options): Added documentation about
15897 the -fcilkplus flag.
15898 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15899 (BUILTINS_DEF): Depend on cilkplus.def.
15900 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15901 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15902 * cilkplus.def: New file.
15904 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15906 PR rtl-optimization/57439
15907 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15909 2013-05-28 Easwaran Raman <eraman@google.com>
15911 PR tree-optimization/57337
15912 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15913 (find_insert_point): Correctly identify the insertion point
15914 when two statements with the same UID is compared.
15916 2013-05-28 Richard Biener <rguenther@suse.de>
15918 PR tree-optimization/56787
15919 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15920 from the list of data references.
15921 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15923 (vect_analyze_loop_operations): Likewise.
15924 (vect_transform_loop): Remove clobbers.
15926 2013-05-28 Martin Jambor <mjambor@suse.cz>
15928 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15929 and REALPART_EXPRs have scalar type.
15931 2013-05-28 Richard Biener <rguenther@suse.de>
15933 PR tree-optimization/57411
15934 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15936 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15937 virtual operand propagation.
15939 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15941 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15942 destination register for bmasksi_vis.
15943 (vector_init_bshuffle): Likewise.
15944 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15946 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15948 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15949 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15950 mode if the instruction isn't available in the original mode.
15951 * config/sparc/sparc.opt (mfix-ut699): New option.
15952 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15953 (divdf3): Turn into expander.
15954 (divdf3_nofix): New insn.
15955 (divdf3_fix): Likewise.
15956 (divsf3): Disable if -mfix-ut699.
15957 (sqrtdf2): Turn into expander.
15958 (sqrtdf2_nofix): New insn.
15959 (sqrtdf2_fix): Likewise.
15960 (sqrtsf2): Disable if -mfix-ut699.
15962 2013-05-27 Richard Biener <rguenther@suse.de>
15964 PR middle-end/57412
15965 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15966 block for the new loop.
15968 2013-05-27 Richard Biener <rguenther@suse.de>
15970 PR tree-optimization/57343
15971 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15972 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15973 (number_of_iterations_cond): Do not build the folded tree.
15975 2013-05-27 Richard Biener <rguenther@suse.de>
15978 PR middle-end/57381
15979 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15980 OEP_CONSTANT_ADDRESS_OF retained.
15982 PR tree-optimization/57417
15983 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15984 for unchanged base.
15985 (set_ssa_val_to): Compare addresses using
15986 get_addr_base_and_unit_offset.
15988 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15990 PR rtl-optimization/56833
15991 * postreload.c (move2add_record_mode): New function.
15992 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15993 (move2add_use_add2_insn): Use move2add_record_sym_value.
15994 (move2add_use_add3_insn): Likewise.
15995 (reload_cse_move2add): Use move2add_valid_value_p and
15996 move2add_record_mode. Invalidate call-clobbered and REG_INC
15997 affected regs by setting reg_mode to VOIDmode.
15998 (move2add_note_store): Don't pretend the inside of a SUBREG is
15999 the actual destination. Invalidate single/leading registers by
16000 setting reg_mode to VOIDmode.
16001 Use move2add_record_sym_value, move2add_valid_value_p and
16002 move2add_record_mode.
16004 2013-05-27 Richard Biener <rguenther@suse.de>
16006 PR tree-optimization/57396
16007 * tree-affine.c (double_int_constant_multiple_p): Properly
16008 return false for val == 0 and div != 0.
16010 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16012 * config/mips/mips.h: Use #elif in preprocessor conditions.
16014 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16017 * config/mips/constraints.md (kl): New constraint.
16018 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16019 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16020 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16021 from LO for MIPS16.
16022 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16024 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16027 * config/mips/mips.c (mips_can_inline_p): New function.
16028 (TARGET_CAN_INLINE_P): Define.
16030 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16032 * sched-int.h (ds_t, dw_t): Make unsigned int.
16033 Fix documentation that describes how all the ds_t bits are used.
16034 Reserve the last bit for delayed-branch scheduling.
16035 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16036 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16037 (gen_dep_weak_1): Remove prototype.
16038 * sched-deps.c (get_dep_weak_1): Make static.
16039 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16040 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16041 * doc/tm.texi: Regenerate.
16042 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16044 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16047 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16049 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16050 Sandra Loosemore <sandra@codesourcery.com>
16052 * config.gcc (powerpc-*): Allow native for with-cpu.
16054 2013-05-24 Jeff Law <law@redhat.com>
16056 PR tree-optimization/57124
16057 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16058 conversion feeding a condition if the range has an overflow
16059 if -fstrict-overflow. Add warnings for when we do make the
16062 2013-05-24 Dehao Chen <dehao@google.com>
16064 * tree-cfg.c (locus_discrim_map): Fix the typo.
16065 (locus_discrim_hasher): Likewise.
16066 (locus_discrim_hasher::hash): Likewise.
16067 (locus_discrim_hasher::equal): Likewise.
16069 2013-05-24 Martin Jambor <mjambor@suse.cz>
16071 PR tree-optimization/57294
16072 * cgraph.h (ipa_record_stmt_references): Declare.
16073 * cgraphbuild.c (ipa_record_stmt_references): New function.
16074 (build_cgraph_edges): Use ipa_record_stmt_references.
16075 (rebuild_cgraph_edges): Likewise.
16076 (cgraph_rebuild_references): Likewise.
16077 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16078 associated with the old statement and build references from the
16079 newly built statements.
16080 * ipa-ref.c (ipa_remove_stmt_references): New function.
16081 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16083 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16085 * lra-constraints.c (emit_spill_move): Use smaller mode for
16087 (check_and_process_move): Consider mem-reg moves for secondary
16089 (curr_insn_transform): Don't lose insns emitted before for
16090 secondary memory moves.
16091 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16092 reg set up in the current insn.
16094 2013-05-24 Dehao Chen <dehao@google.com>
16096 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16098 (locus_descrim_hasher::equal): Likewise.
16099 (build_gimple_cfg): New discriminator assignment algorithm.
16100 (make_edges): Likewise.
16101 (next_discriminator_for_locus): Likewise.
16102 (same_line_p): Likewise.
16103 (assign_discriminators): Likewise.
16104 (make_cond_expr_edges): Likewise.
16105 (make_gimple_switch_edges): Likewise.
16106 (make_goto_expr_edges): Likewise.
16107 (make_gimple_asm_edges): Likewise.
16109 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16111 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16112 X format specifier to only display bottom 16 bits.
16113 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16114 immediate to match for operand 2, since it will be masked.
16116 2013-05-24 Richard Biener <rguenther@suse.de>
16118 PR tree-optimization/57287
16119 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16120 all SSA names that occur in abnormal PHIs.
16122 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16125 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16126 that index is not negative.
16128 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16130 PR rtl-optimization/55177
16131 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16132 (simplify_byte_swapping_operation): New.
16133 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16134 (simplify_relational_operation_1): Deal with BSWAP.
16136 2013-05-23 Richard Henderson <rth@redhat.com>
16139 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16140 (ix86_reorg): Call it.
16142 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16145 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16146 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16147 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16149 2013-05-23 Christian Bruel <christian.bruel@st.com>
16152 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16154 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16155 Marcus Shawcroft <marcus.shawcroft@arm.com>
16157 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16158 * config/aarch64/constraints.md (Usa): Remove.
16159 * doc/md.texi (AArch64 Usa): Remove.
16161 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16162 Marcus Shawcroft <marcus.shawcroft@arm.com>
16164 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16165 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16166 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16167 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16169 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16171 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16172 instruction (AdvSIMD).
16173 * config/aarch64/aarch64-builtins.c
16174 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16175 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16177 2013-05-23 Martin Jambor <mjambor@suse.cz>
16179 PR middle-end/57347
16180 * tree.h (contains_bitfld_component_ref_p): Declare.
16181 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16182 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16184 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16185 not access a bit-field. Assert all final offsets are byte-aligned.
16187 2013-05-23 Richard Biener <rguenther@suse.de>
16189 PR tree-optimization/57380
16190 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16191 least one invariant or re-used load.
16192 * passes.c (init_optimization_passes): Move pass_phiprop before
16195 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16197 * config/aarch64/aarch64-simd.md
16198 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16200 2013-05-23 Richard Biener <rguenther@suse.de>
16202 PR middle-end/57381
16203 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16204 OEP_CONSTANT_ADDRESS_OF retained.
16206 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16208 PR middle-end/57344
16209 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16210 don't lower unit. Handle unit not being always BITS_PER_WORD.
16212 2013-05-23 Richard Biener <rguenther@suse.de>
16214 PR rtl-optimization/57341
16215 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16216 instead of true_dependence.
16218 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16220 * bb-reorder.c (branch_threshold): Make const.
16221 (exec_threshold): Ditto.
16223 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16224 Pat Haugen <pthaugen@us.ibm.com>
16225 Peter Bergner <bergner@vnet.ibm.com>
16227 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16228 documentation for the power8 crypto builtins.
16230 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16232 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16233 macros for defining power8 builtin functions.
16234 (BU_P8V_AV_2): Likewise.
16235 (BU_P8V_AV_P): Likewise.
16236 (BU_P8V_VSX_1): Likewise.
16237 (BU_P8V_OVERLOAD_1): Likewise.
16238 (BU_P8V_OVERLOAD_2): Likewise.
16239 (BU_CRYPTO_1): Likewise.
16240 (BU_CRYPTO_2): Likewise.
16241 (BU_CRYPTO_3): Likewise.
16242 (BU_CRYPTO_OVERLOAD_1): Likewise.
16243 (BU_CRYPTO_OVERLOAD_2): Likewise.
16244 (XSCVSPDP): Fix typo, point to the correct instruction.
16245 (VCIPHER): Add power8 crypto builtins.
16246 (VCIPHERLAST): Likewise.
16247 (VNCIPHER): Likewise.
16248 (VNCIPHERLAST): Likewise.
16249 (VPMSUMB): Likewise.
16250 (VPMSUMH): Likewise.
16251 (VPMSUMW): Likewise.
16252 (VPERMXOR_V2DI): Likewise.
16253 (VPERMXOR_V4SI: Likewise.
16254 (VPERMXOR_V8HI: Likewise.
16255 (VPERMXOR_V16QI: Likewise.
16256 (VSHASIGMAW): Likewise.
16257 (VSHASIGMAD): Likewise.
16258 (VPMSUM): Likewise.
16259 (VPERMXOR): Likewise.
16260 (VSHASIGMA): Likewise.
16262 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16263 __CRYPTO__ if the crypto instructions are available.
16264 (altivec_overloaded_builtins): Add support for overloaded power8
16267 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16268 support for power8 crypto builtins.
16269 (builtin_function_type): Likewise.
16270 (altivec_init_builtins): Add support for builtins that take vector
16271 long long (V2DI) arguments.
16273 * config/rs6000/crypto.md: New file, define power8 crypto
16276 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16277 Pat Haugen <pthaugen@us.ibm.com>
16278 Peter Bergner <bergner@vnet.ibm.com>
16280 * doc/invoke.texi (Option Summary): Add power8 options.
16281 (RS/6000 and PowerPC Options): Likewise.
16283 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16284 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16285 wm, wn, wr documentation.
16287 * config/rs6000/constraints.md (wm): New constraint for VSX
16288 registers if direct move instructions are enabled.
16289 (wn): New constraint for no registers.
16290 (wq): New constraint for quad word even GPR registers.
16291 (wr): New constraint if 64-bit instructions are enabled.
16292 (wv): New constraint if power8 vector instructions are enabled.
16293 (wQ): New constraint for quad word memory locations.
16295 * config/rs6000/predicates.md (const_0_to_15_operand): New
16296 constraint for 0..15 for crypto instructions.
16297 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16298 as GPR and floating point registers.
16299 (int_reg_operand): New predicate to match only GPR registers.
16300 (base_reg_operand): New predicate to match base registers.
16301 (quad_int_reg_operand): New predicate to match even GPR registers
16302 for quad memory operations.
16303 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16304 operations in both GPR and VSX registers.
16305 (quad_memory_operand): New predicate for quad memory operations.
16306 (reg_or_indexed_operand): New predicate for direct move support.
16308 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16309 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16310 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16311 (POWERPC_MASKS): Add power8 options.
16312 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16315 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16316 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16318 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16319 (-mpower8-fusion): New power8 options.
16320 (-mpower8-fusion-sign): Likewise.
16321 (-mpower8-vector): Likewise.
16322 (-mcrypto): Likewise.
16323 (-mdirect-move): Likewise.
16324 (-mquad-memory): Likewise.
16326 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16327 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16329 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16330 (rs6000_debug_vector_unit): Add p8_vector.
16331 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16332 definitions. Also print fusion state.
16333 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16334 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16335 (rs6000_option_override_internal): Add support for power8.
16336 (rs6000_common_init_builtins): Add debugging for skipped builtins
16337 if -mdebug=builtin.
16338 (rs6000_adjust_cost): Add power8 support.
16339 (rs6000_issue_rate): Likewise.
16340 (insn_must_be_first_in_group): Likewise.
16341 (insn_must_be_last_in_group): Likewise.
16342 (force_new_group): Likewise.
16343 (rs6000_register_move_cost): Likewise.
16344 (rs6000_opt_masks): Likewise.
16346 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16347 power8 capable assembler, default to power7 options.
16348 (TARGET_DIRECT_MOVE): Likewise.
16349 (TARGET_CRYPTO): Likewise.
16350 (TARGET_P8_VECTOR): Likewise.
16351 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16352 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16353 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16354 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16355 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16356 (TARGET_XSCVDPSPN): Likewise.
16357 (TARGET_XSCVSPDPN): Likewsie.
16358 (TARGET_SYNC_HI_QI): Likewise.
16359 (TARGET_SYNC_TI): Likewise.
16360 (MASK_CRYPTO): Likewise.
16361 (MASK_DIRECT_MOVE): Likewise.
16362 (MASK_P8_FUSION): Likewise.
16363 (MASK_P8_VECTOR): Likewise.
16364 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16365 temporary used by some of the direct move instructions to get two FP
16366 temporary registers does not force creation of a stack frame.
16367 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16368 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16369 that any VSX registers are tieable, even if they are also an
16370 Altivec vector mode.
16371 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16372 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16373 (RS6000_BTM_CRYPTO): Likewise.
16374 (RS6000_BTM_COMMON): Likewise.
16376 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16377 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16378 (enum rs6000_vector): Add power8 vector support.
16380 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16384 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16385 (any_sibcall_could_use_r3): this and handle indirect calls.
16386 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16388 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16390 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16392 2013-05-22 Richard Biener <rguenther@suse.de>
16394 PR middle-end/57349
16395 * profile.c (branch_prob): Do not split blocks that are
16396 abnormally receiving from ECF_RETURNS_TWICE functions.
16398 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16400 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16401 address mode. Move pointer mode initialization to the same place.
16403 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16405 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16406 while it has any effect.
16408 2013-05-21 Easwaran Raman <eraman@google.com>
16410 PR tree-optimization/57322
16411 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16412 UID of the statement added to the BB to be 1.
16414 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16416 PR tree-optimization/57331
16417 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16418 of conversion from pointer type to integral type with integer.
16420 2013-05-21 Martin Jambor <mjambor@suse.cz>
16423 * ipa-prop.c (ipa_read_node_info): Process param_used and
16424 controlled_uses in the same order as when writing.
16426 2013-05-21 Magnus Granberg <baldrick@free.fr>
16429 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16431 2013-05-21 Richard Biener <rguenther@suse.de>
16433 PR tree-optimization/57318
16434 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16435 estimate stmts with side-effects as likely eliminated.
16437 2013-05-21 Richard Biener <rguenther@suse.de>
16439 PR tree-optimization/57330
16440 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16441 preserve the call stmts fntype.
16443 2013-05-21 Richard Biener <rguenther@suse.de>
16445 PR tree-optimization/57303
16446 * tree-ssa-sink.c (statement_sink_location): Improve killing
16447 stmt detection and properly handle self-assignments.
16449 2013-05-21 Christian Bruel <christian.bruel@st.com>
16451 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16452 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16453 registers. Set register size out of the PARALLEL loop.
16455 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16458 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16459 (*fmasf4, *fmasf4_media): New insns.
16461 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16463 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16464 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16465 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16466 (mips_idiv_insns): Update the comments to say that the returned
16467 instruction counts are in units of BASE_INSN_LENGTH.
16468 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16469 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16470 using 2 rather than 4 as the length of indirect MIPS16 and
16471 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16472 length of a NOP. Don't divide MIPS16 lengths by 2.
16473 (mips16_split_long_branches): Assume a branch is long if the
16474 length is greater than 4 rather than 8.
16475 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16476 rather than multiplying them by 2. Multiply instruction counts
16477 by BASE_INSN_LENGTH rather than 4.
16478 (*jump_mips16, tls_get_tp_mips16_<mode>)
16479 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16481 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16483 * config/mips/mips.md (extended_mips16): Remove branch case.
16484 (length): Remove duplicated extended_mips16 test.
16486 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16488 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16490 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16492 * recog.h (Recog_data): Rename to...
16493 (recog_data_d): ...this.
16494 (recog_data): Update accordingly.
16495 * recog.c (recog_data): Likewise.
16496 * reload.c (save_recog_data): Likewise.
16497 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16498 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16500 2013-05-17 Julian Brown <julian@codesourcery.com>
16502 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16503 found in a REG_EQUAL note, invalidate it.
16505 2013-05-17 Easwaran Raman <eraman@google.com>
16507 * tree-ssa-reassoc.c (find_insert_point): New function.
16508 (insert_stmt_after): Likewise.
16509 (get_def_stmt): Likewise.
16510 (ensure_ops_are_available): Likewise.
16511 (not_dominated_by): Likewise.
16512 (rewrite_expr_tree): Do not move statements beyond what is
16513 necessary. Remove call to swap_ops_for_binary_stmt...
16514 (reassociate_bb): ... and move it here.
16515 (build_and_add_sum): Assign UIDs for new statements.
16516 (linearize_expr): Likewise.
16517 (do_reassoc): Renumber gimple statement UIDs.
16519 2013-05-17 Jan Hubicka <jh@suse.cz>
16521 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16523 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16524 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16526 (output_weakrefs): Update.
16528 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16529 Martin Jambor <mjambor@suse.cz>
16531 PR middle-end/57276
16532 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16533 value that corresponds to the given aggval is found in values vector.
16535 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16537 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16538 sse, sse2, sse3, ssse3 and sse4a flags to options.
16540 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16542 * gengtype-state.c: (s_expr_writer): New class, to handle
16543 prettifying of output layout of s-expressions.
16544 (state_writer): New class, to write out gtype.state.
16545 (state_written_type_count): Move this variable into member data of
16547 (s_expr_writer::s_expr_writer): New code: constructor for new class
16548 (state_writer::state_writer(): ditto
16549 (s_expr_writer::write_new_line): New function
16550 (s_expr_writer::write_any_indent): ditto
16551 (s_expr_writer::begin_s_expr): ditto
16552 (s_expr_writer::end_s_expr): ditto
16553 (write_state_fileloc): convert to method of state_writer...
16554 (state_writer:: write_state_fileloc): ...and use methods of
16555 s_expr_writer to write indentation into the gtype.state output file
16556 to visually represent the hierarchical structure of the list
16558 (write_state_fields): ditto, renaming to...
16559 (state_writer::write_state_fields)
16560 (write_state_a_string): ditto, renaming to...
16561 (state_writer::write_state_a_string)
16562 (write_state_string_option): ditto, renaming to...
16563 (state_writer::write_state_string_option)
16564 (write_state_type_option): ditto, renaming to...
16565 (state_writer::write_state_type_option)
16566 (write_state_nested_option): ditto, renaming to...
16567 (state_writer::write_state_nested_option)
16568 (write_state_option): ditto, renaming to...
16569 (state_writer::write_state_option)
16570 (write_state_options): ditto, renaming to...
16571 (state_writer::write_state_options)
16572 (write_state_lang_bitmap): ditto, renaming to...
16573 (state_writer::write_state_lang_bitmap)
16574 (write_state_version): ditto, renaming to...
16575 (state_writer::write_state_version)
16576 (write_state_scalar_type): ditto, renaming to...
16577 (state_writer::write_state_scalar_type)
16578 (write_state_string_type): ditto, renaming to...
16579 (state_writer::write_state_string_type)
16580 (write_state_undefined_type): ditto, renaming to...
16581 (state_writer::write_state_undefined_type)
16582 (write_state_struct_union_type): ditto, renaming to...
16583 (state_writer::write_state_struct_union_type)
16584 (write_state_struct_type): ditto, renaming to...
16585 (state_writer::write_state_struct_type)
16586 (write_state_user_struct_type): ditto, renaming to...
16587 (state_writer::write_state_user_struct_type)
16588 (write_state_lang_struct_type): ditto, renaming to...
16589 (state_writer::write_state_lang_struct_type)
16590 (write_state_param_struct_type): ditto, renaming to...
16591 (state_writer::write_state_param_struct_type)
16592 (write_state_pointer_type): ditto, renaming to...
16593 (state_writer::write_state_pointer_type)
16594 (write_state_array_type): ditto, renaming to...
16595 (state_writer::write_state_array_type)
16596 (write_state_gc_used): ditto, renaming to...
16597 (state_writer::write_state_gc_used)
16598 (write_state_common_type_content): ditto, renaming to...
16599 (state_writer::write_state_common_type_content)
16600 (write_state_type): ditto, renaming to...
16601 (state_writer::write_state_type)
16602 (write_state_pair_list): ditto, renaming to...
16603 (state_writer::write_state_pair_list)
16604 (write_state_pair): ditto, renaming to...
16605 (state_writer::write_state_pair)
16606 (write_state_typedefs): ditto, renaming to...
16607 (state_writer::write_state_typedefs)
16608 (write_state_structures): ditto, renaming to...
16609 (state_writer::write_state_structures)
16610 (write_state_param_structs): ditto, renaming to...
16611 (state_writer::write_state_param_structs)
16612 (write_state_variables): ditto, renaming to...
16613 (state_writer::write_state_variables)
16614 (write_state_srcdir): ditto, renaming to...
16615 (state_writer::write_state_srcdir)
16616 (write_state_files_list): ditto, renaming to...
16617 (state_writer::write_state_files_list)
16618 (write_state_languages): ditto, renaming to...
16619 (state_writer::write_state_languages)
16620 (write_state): create a state_writer instance and use it when
16621 writing out the state file
16623 2013-05-17 Mike Stump <mikestump@comcast.net>
16625 PR rtl-optimization/57304
16626 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16627 accessing DF_REF_REAL_LOC.
16629 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16631 PR rtl-optimization/57281
16632 PR rtl-optimization/57300
16633 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16634 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16635 what the other splitter did if the registers are dead.
16637 2013-05-17 Richard Biener <rguenther@suse.de>
16639 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16642 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16644 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16647 2013-05-17 Marek Polacek <polacek@redhat.com>
16649 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16650 length when doing non-zero store of storing '\0' to '\0'.
16652 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16654 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16655 vect_external_def oprnd1 with loop_vinfo, try to emit
16656 optional cast, negation and and stmts on the loop preheader
16657 edge instead of into the pattern def seq.
16659 PR tree-optimization/57051
16660 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16661 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16663 2013-05-16 Nick Clifton <nickc@redhat.com>
16665 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16666 (rl78_is_naked_func): New function.
16667 (rl78_expand_prologue): Skip prologue generation for naked functions.
16668 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16669 * doc/extend.texi (naked): Add RL78 to the list of processors
16670 that supports this attribute.
16672 2013-05-16 Jeff Law <law@redhat.com>
16674 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16676 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16678 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16679 cache parameters using detect_caches_amd also for CYRIX,
16680 NSC and TM2 signatures.
16682 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16683 Dzianis Kahanovich <mahatma@eu.by>
16687 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16688 VIA/Centaur processors and determine their cache parameters
16689 using detect_caches_amd.
16691 2013-05-16 Teresa Johnson <tejohnson@google.com>
16693 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16694 (rtl_verify_edges): New function.
16695 (rtl_verify_bb_insns): Ditto.
16696 (rtl_verify_bb_pointers): Ditto.
16697 (rtl_verify_bb_insn_chain): Ditto.
16698 (rtl_verify_fallthru): Ditto.
16699 (rtl_verify_bb_layout): Ditto.
16700 (rtl_verify_flow_info_1): Outline checks into new functions.
16701 (rtl_verify_flow_info): Ditto.
16703 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16705 * cfghooks.c (copy_bbs): Add update_dominance argument.
16706 * cfghooks.h (copy_bbs): Update prototype.
16707 * tree-cfg.c (gimple_duplicate_sese_region):
16708 Add update_dominance argument.
16709 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16710 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16711 gimple_duplicate_sese_region call.
16712 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16713 Update copy_bbs call.
16714 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16715 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16717 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16719 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16720 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16721 vect_recog_rotate_pattern.
16722 (vect_recog_rotate_pattern): New function.
16724 2013-05-16 Jason Merrill <jason@redhat.com>
16726 * Makefile.in (LLINKER): New variable.
16727 (mostlyclean): Remove link mutex.
16728 * configure.ac: Handle --enable-link-mutex.
16729 * lock-and-run.sh: New script.
16731 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16734 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16737 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16739 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16741 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16743 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16744 * config/arm/arm.c (next_consecutive_mem): New function.
16745 (gen_movmem_ldrd_strd): Likewise.
16746 * config/arm/arm.md (movmemqi): Update condition and code.
16747 (unaligned_loaddi, unaligned_storedi): New patterns.
16749 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16751 * config.gcc: Obsolete *-*-solaris2.9*.
16752 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16754 2013-05-16 Richard Biener <rguenther@suse.de>
16756 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16757 earlier, after GRAPHITE transforms and IV canonicalization.
16759 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16761 * omp-low.c (extract_omp_for_data): For collapsed loops,
16762 if at least one of the loops is known at compile time to
16763 iterate zero times, set count to 0.
16764 (expand_omp_regimplify_p): New function.
16765 (expand_omp_for_generic): For collapsed loops, if at least
16766 one of the loops isn't known to iterate at least once,
16767 add runtime check with setting count to 0.
16768 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16769 For unsigned types if it isn't known at compile time that
16770 the loop will iterate at least once, add runtime check to bypass
16771 the whole loop if initial condition isn't true.
16773 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16775 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16777 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16779 PR middle-end/57286
16780 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16781 transformations to avoid an infinite loop.
16783 2013-05-16 Marek Polacek <polacek@redhat.com>
16785 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16787 2013-05-15 Leif Ekblad <leif@rdos.net>
16789 * config/i386/i386.c (ix86_decompose_address): Use
16790 DEFAULT_TLS_SEG_REG to access TLS segment register.
16791 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16792 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16793 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16795 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16798 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16799 sibling calls to functions that would normally be lazily bound,
16800 unless $gp is call-clobbered.
16802 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16804 * config/i386/i386.c (ix86_option_override_internal): Update
16805 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16806 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16807 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16808 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16810 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16812 2013-05-15 Andreas Schwab <schwab@suse.de>
16814 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16815 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16818 2013-05-15 Teresa Johnson <tejohnson@google.com>
16820 * loop-unroll.c (report_unroll_peel): Check decision before
16821 emitting unroll/peel message.
16823 2013-05-15 Teresa Johnson <tejohnson@google.com>
16825 * function.h (has_bb_partition): New rtl_data flag.
16826 (bb_reorder_complete): Ditto.
16827 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16828 instead of flag_reorder_blocks_and_partition.
16829 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16830 with some enhancements.
16831 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16832 * bb-reorder.c (connect_traces): Check for has_bb_partition
16833 instead of flag_reorder_blocks_and_partition.
16834 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16835 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16836 verify_hot_cold_block_grouping.
16837 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16839 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16842 * config/arm/predicates.md (call_insn_operand): New predicate.
16843 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16844 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16845 if insn is not a tail call.
16846 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16848 * config/arm/arm.h (enum reg_class): New caller save register class.
16849 (REG_CLASS_NAMES): Likewise.
16850 (REG_CLASS_CONTENTS): Likewise.
16851 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16854 2013-05-15 Richard Biener <rguenther@suse.de>
16856 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16857 of MSG_OPTIMIZED_LOCATIONS.
16858 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16859 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16861 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16862 of MSG_OPTIMIZED_LOCATIONS.
16863 (execute_vect_slp): Likewise.
16864 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16865 (vect_create_cond_for_alias_checks): Likewise.
16866 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16867 (vect_recog_widen_mult_pattern): Likewise.
16868 (vect_recog_widen_sum_pattern): Likewise.
16869 (vect_recog_over_widening_pattern): Likewise.
16870 (vect_recog_widen_shift_pattern): Likewise.
16871 (vect_recog_vector_vector_shift_pattern): Likewise.
16872 (vect_recog_divmod_pattern): Likewise.
16873 (vect_recog_mixed_size_cond_pattern): Likewise.
16874 (vect_recog_bool_pattern): Likewise.
16875 (vect_pattern_recog_1): Likewise.
16877 2013-05-15 Martin Jambor <mjambor@suse.cz>
16879 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16880 non-functions to builtin_unreachable.
16881 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16882 correct when new direct edges were discovered.
16884 2013-05-15 Martin Jambor <mjambor@suse.cz>
16886 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16887 header, print symbol order instead of node uid, print more information
16888 about indirect edge targets.
16889 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16891 (ipa_make_edge_direct_to_target): Likewise.
16892 (remove_described_reference): Likewise.
16893 (propagate_controlled_uses): Likewise.
16894 (ipa_print_node_params): Also print symbol order.
16895 (ipcp_transform_function): Print symbol order instead of node uids.
16896 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16897 (cgraph_get_create_real_symbol_node): Likewise.
16898 * ipa-cp.c (print_lattice): Likewise.
16899 (print_all_lattices): Likewise.
16900 (determine_versionability): Likewise.
16901 (initialize_node_lattices): Likewise.
16902 (estimate_local_effects): Likewise.
16903 (update_profiling_info): Likewise.
16904 (create_specialized_node): Likewise.
16905 (perhaps_add_new_callers): Likewise.
16906 (decide_about_value): Likewise.
16907 (decide_whether_version_node): Likewise.
16908 (identify_dead_nodes): Likewise.
16909 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16910 (dump_inline_summary): Likewise.
16911 (estimate_node_size_and_time): Likewise.
16912 (inline_analyze_function): Likewise.
16913 * ipa-inline.c (report_inline_failed_reason): Likewise.
16914 (want_early_inline_function_p): Likewise.
16915 (edge_badness): Likewise.
16916 (update_edge_key): Likewise.
16917 (inline_small_functions): Likewise. Add dumping of order to two other
16919 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16920 instead of node uids.
16921 (propagate_pure_const): Likewise.
16922 (propagate_pure_const): Likewise.
16923 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16924 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16925 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16927 * tree-pretty-print.c (dump_function_header): Likewise.
16928 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16929 Print symbol order instead of node uids.
16931 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16933 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16934 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16936 2013-05-15 Richard Biener <rguenther@suse.de>
16938 PR tree-optimization/57275
16939 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16940 return value for fail to do runtime alias checks for gather loads.
16942 2013-05-15 Jan Hubicka <jh@suse.cz>
16946 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16947 weakrefs are not external.
16948 (lto_symtab_merge_decls): Fix thinko when dealing with
16949 non-lto_symtab decls.
16950 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16951 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16952 * varpool.c (dump_varpool_node): Dump more flags.
16954 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16956 * config/i386/i386.c (processor_alias_table): Add instruction
16957 FSGSBASE for AMD bdver3 architecture.
16959 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16961 * tree.c (warn_deprecated_use): Print file:line using locus color.
16962 * diagnostic.c (diagnostic_report_current_module): Print file:line
16963 and file:line:column using locus color.
16965 2013-05-14 Mike Stump <mikestump@comcast.net>
16967 * gdbinit.in: Add __null.
16969 2013-05-14 Mike Stump <mikestump@comcast.net>
16971 * recog.h: Rename struct recog_data to Recog_data.
16972 * recog.c: Likewise.
16973 * reload.c (can_reload_into): Likewise.
16974 * config/picochip/picochip.c: Likewise.
16976 2013-05-14 Mike Stump <mikestump@comcast.net>
16978 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16980 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16982 * resource.h (struct resources): Remove unch_memory member.
16983 (CLEAR_RESOURCE): Don't clear unch_memory.
16984 * resource.c (mark_referenced_resources): Don't set it.
16985 (mark_set_resources): Likewise.
16986 (mark_target_live_regs): Don't clear it.
16987 (init_resource_info): Likewise.
16988 * reorg.c (resource_conflicts_p): Don't compare it.
16989 (redundant_insn): Don't set it.
16991 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16993 * emit-rtl.c (next_label): Remove unused function.
16994 (skip_consecutive_labels, link_cc0_insns): Move to ...
16995 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16996 only place where these functions are used, and make them static.
16998 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17000 * fold-const.c (fold_negate_expr): Handle vectors.
17001 (fold_truth_not_expr): Make it static.
17002 (fold_invert_truthvalue): New static function.
17003 (invert_truthvalue_loc): Handle vectors. Do not call
17004 fold_truth_not_expr directly.
17005 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17006 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17007 (fold_comparison): Handle vector constants.
17008 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17009 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17010 * tree.h (fold_truth_not_expr): Remove declaration.
17012 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17014 * config/aarch64/aarch64-simd.md
17015 (aarch64_vcond_internal<mode>): Rename to...
17016 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17017 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17018 float modes. Clarify all iterator modes.
17019 (vcond<mode><mode>): Use new name for vcond expanders.
17020 (vcond<v_cmp_result><mode>): Likewise.
17021 (vcondu<mode><mode>: Likewise.
17022 * config/aarch64/iterators.md (VDQF_COND): New.
17024 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17027 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17028 variable for the shift amount. Check that we shift by non-negative
17031 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17034 * config/arm/arm.h (EPILOGUE_USES): Only return true
17035 for LR_REGNUM after epilogue_completed.
17037 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17039 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17040 is error_mark_node.
17042 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17045 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17046 and Solaris 11+/x86 with gld.
17047 * configure: Regenerate.
17049 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17051 * expmed.c (expand_shift_1): Canonicalize rotates by
17052 constant bitsize / 2 to bitsize - 1.
17053 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17054 case ROTATERT>: Likewise.
17057 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17059 * config/i386/i386.md (rotateinv): New code attr.
17060 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17061 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17062 roll $31, %eax, etc.
17064 2013-05-14 Richard Biener <rguenther@suse.de>
17066 PR middle-end/57235
17067 * tree-eh.c (sink_clobbers): Give up for successors with
17068 multiple predecessors and no virtual uses.
17070 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17072 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17073 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17075 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17077 PR middle-end/57251
17078 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17079 the case when both op0 and op1 have VOIDmode.
17081 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17083 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17084 in multiply-accumulate mode.
17086 2013-05-13 Guozhi Wei <carrot@google.com>
17088 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17090 2013-05-13 Kai Tietz <ktietz@redhat.com>
17093 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17094 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17095 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17096 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17097 (ix86_expand_prologue): Likewise.
17098 (ix86_expand_split_stack_prologue): Likewise.
17099 (legitimate_pic_address_disp_p): Likewise.
17100 (legitimize_pic_address): Likewise.
17101 (legitimize_tls_address): Likewise.
17102 (legitimize_pe_coff_symbol): Likewise.
17103 (output_pic_addr_const): Likewise.
17104 (construct_plt_address): Likewise.
17105 (ix86_expand_call): Likewise.
17106 (x86_output_mi_thunk): Likewise.
17107 (x86_function_profiler): Likewise.
17109 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17111 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17112 similar switch cases.
17113 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17114 (aarch64_simd_mov_to_<mode>low): Delete.
17115 (aarch64_simd_mov_to_<mode>high): Delete.
17116 (move_lo_quad_<mode>): Add w<-r alternative.
17117 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17118 (aarch64_simd_mov_from_*): Update type attribute.
17119 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17122 2013-05-13 Jan Hubicka <jh@suse.cz>
17124 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17125 * config/i386/i386.c (ix86_compute_frame_layout,
17126 ix86_expand_epilogue, emit_i387_cw_initialization,
17127 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17128 ix86_local_alignment): Fix use of size/speed predicates.
17130 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17132 PR tree-optimization/45216
17133 PR tree-optimization/57157
17134 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17135 the (-Y) & (B - 1) variant if OP is |.
17136 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17137 return shifted. Use (-op1) & (prec - 1) as other_amount
17138 instead of prec - op1.
17140 2013-05-13 Martin Jambor <mjambor@suse.cz>
17142 PR middle-end/42371
17143 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17144 (ipa_constant_data): New type.
17145 (ipa_jump_func): Use ipa_constant_data to hold information about
17146 constant jump functions.
17147 (ipa_get_jf_constant): Adjust to jump function type changes.
17148 (ipa_get_jf_constant_rdesc): New function.
17149 (ipa_param_descriptor): New field controlled_uses.
17150 (ipa_get_controlled_uses): New function.
17151 (ipa_set_controlled_uses): Likewise.
17152 * ipa-ref.h (ipa_find_reference): Declare.
17153 * ipa-prop.c (ipa_cst_ref_desc): New type.
17154 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17156 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17157 New parameter cs. Adjust all callers.
17158 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17159 (remove_described_reference): New function.
17160 (jfunc_rdesc_usable): Likewise.
17161 (try_make_edge_direct_simple_call): Decrement controlled use count,
17162 attempt to remove reference if it hits zero.
17163 (combine_controlled_uses_counters): New function.
17164 (propagate_controlled_uses): Likewise.
17165 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17166 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17167 (ipa_print_node_params): Print described use counter.
17168 (ipa_write_jump_function): Adjust to jump function type changes.
17169 (ipa_read_jump_function): New parameter CS, pass it to
17170 ipa_set_jf_constant. Adjust caller.
17171 (ipa_write_node_info): Stream controlled use count
17172 (ipa_read_node_info): Likewise.
17173 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17175 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17176 count. Remove cloning-added reference if it reaches zero.
17177 * ipa-ref.c (ipa_find_reference): New function.
17179 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17181 * config/i386/i386.c (processor_target_table): Modified default
17182 alignment values for AMD BD and BT architectures.
17184 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17186 * tree-vect-generic.c (uniform_vector_p): Move ...
17187 * tree.c (uniform_vector_p): ... here.
17188 * tree.h (uniform_vector_p): Declare it.
17189 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17192 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17194 PR tree-optimization/57230
17195 * tree-ssa-strlen.c (handle_char_store): Record length for
17196 array store from STRING_CST.
17198 PR tree-optimization/57230
17199 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17202 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17204 * config/epiphany/epiphany.c (epiphany_init): Check size of
17205 NUM_MODES_FOR_MODE_SWITCHING.
17206 (epiphany_expand_prologue):
17207 Remove CONFIG_REGNUM initial value handling code.
17208 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17209 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17210 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17211 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17212 Don't return 1 for FP_MODE_NONE.
17213 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17214 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17215 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17216 * config/epiphany/epiphany.md (save_config): New pattern.
17218 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17220 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17222 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17224 * config/i386/i386.md (memory): Handle sseishft1.
17225 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17226 (*vec_extractv2di_1): Ditto.
17228 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17230 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17233 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17235 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17236 Add mthumb/march=armv7-a multilib.
17237 Add mthumb/march=armv7-r multilib.
17238 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17240 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17242 * config/v850/t-rtems: Add more multilibs.
17244 2013-05-10 Richard Biener <rguenther@suse.de>
17246 PR tree-optimization/57214
17247 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17248 not propagate from SSA names that occur in abnormal PHI nodes.
17250 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17252 * stor-layout.c (element_precision): New function.
17253 * machmode.h (element_precision): Declare it.
17254 * tree.c (build_minus_one_cst): New function.
17255 (element_precision): Likewise.
17256 * tree.h (build_minus_one_cst): Declare new function.
17257 (element_precision): Likewise.
17258 * fold-const.c (operand_equal_p): Use element_precision.
17259 (fold_binary_loc): Handle vector types.
17260 * convert.c (convert_to_integer): Use element_precision.
17261 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17264 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17266 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17267 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17268 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17269 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17270 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17271 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17272 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17273 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17274 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17275 (Uuw8): New constraints.
17276 (Usb4): Move into alphabetical order.
17277 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17278 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17279 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17280 previously unnamed patterns.
17281 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17282 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17283 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17284 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17285 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17286 of specific lengths.
17288 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17290 * config/i386/i386.md (rotateinv): New code attr.
17291 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17292 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17293 roll $31, %eax, etc.
17295 PR tree-optimization/45216
17296 PR tree-optimization/57157
17297 * tree-ssa-forwprop.c (simplify_rotate): New function.
17298 (ssa_forward_propagate_and_combine): Call it.
17300 2013-05-10 Richard Biener <rguenther@suse.de>
17302 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17303 disable peeling when we version for aliasing.
17304 (vector_alignment_reachable_p): Honor explicit user alignment.
17305 (vect_supportable_dr_alignment): Likewise.
17306 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17307 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17308 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17309 then peeling to arrange for the cost-model check to come first.
17311 2013-05-10 Alan Modra <amodra@gmail.com>
17313 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17314 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17315 * configure: Regenerate.
17317 2013-05-10 Alan Modra <amodra@gmail.com>
17320 * varasm.c (default_elf_select_section): Move !DECL_P check..
17321 (get_named_section): ..to here before calling get_section_name.
17323 (default_section_type_flags): Add DECL_P check.
17324 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17325 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17327 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17329 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17330 When using gen_stack_adjust_str with a register offset, add a
17331 REG_FRAME_RELATED_EXPR note.
17333 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17335 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17336 (*vec_extractv4si_zext_mem): Ditto.
17337 (*vec_extractv2di): Add 0->x and x->x alternatives.
17338 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17339 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17341 2013-05-09 Jason Merrill <jason@redhat.com>
17343 N3639 C++1y VLA support
17344 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17347 * tree.c (build_constructor_va): New.
17348 * tree.h: Declare it.
17350 2013-05-09 Martin Jambor <mjambor@suse.cz>
17353 * gimple-fold.c (canonicalize_constructor_val): Call
17354 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17356 2013-05-09 Jan Hubicka <jh@suse.cz>
17357 Richard Biener <rguenther@suse.de>
17360 * symtab.c (symtab_make_decl_local): Do not add private names.
17362 2013-05-09 Jan Hubicka <jh@suse.cz>
17365 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17366 (unlink_from_assembler_name_hash): Likewise.
17367 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17368 symtab_unregister_node, symtab_initialize_asm_name_hash,
17369 change_decl_assembler_name): Update.
17371 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17373 * config/aarch64/aarch64.md: New movtf split.
17374 (*movtf_aarch64): Update.
17375 (aarch64_movdi_tilow): Handle TF modes and rename to
17376 aarch64_movdi_<mode>low.
17377 (aarch64_movdi_tihigh): Handle TF modes and rename to
17378 aarch64_movdi_<mode>high
17379 (aarch64_movtihigh_di): Handle TF modes and rename to
17380 aarch64_mov<mode>high_di
17381 (aarch64_movtilow_di): Handle TF modes and rename to
17382 aarch64_mov<mode>low_di
17383 (aarch64_movtilow_tilow): Remove spurious whitespace.
17384 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17386 (aarch64_print_operand): Update.
17388 2013-05-09 Alan Modra <amodra@gmail.com>
17390 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17392 * configure: Regenerate.
17394 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17396 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17397 splitter preparation statements.
17398 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17399 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17402 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17404 * gimple-ssa-strength-reduction.c (count_candidates): Change
17405 return value to int.
17406 (analyze_candidates_and_replace): Change type of length to int.
17408 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17410 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17411 (*vec_extract<mode>): Use VI12_128 mode iterator.
17412 (*vec_extract<mode>_mem): Ditto.
17413 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17416 2013-05-08 Diego Novillo <dnovillo@google.com>
17421 2012-08-17 Diego Novillo <dnovillo@google.com>
17424 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17425 * config.in: Regenerate.
17426 * configure: Regenerate.
17427 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17429 2013-05-08 Jan Hubicka <jh@suse.cz>
17432 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17433 * cgraph.h (symtab_node_base): Add unique_name.
17434 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17435 input_overwrite_node, input_varpool_node): Stream unique_name.
17436 * cgraphclones.c (cgraph_create_virtual_clone,
17437 cgraph_function_versioning): Set unique_name.
17438 * ipa.c (function_and_variable_visibility): Set unique_name.
17440 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17442 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17443 (alloc_cand_and_find_basis): Restrict conditional candidate
17444 processing to CAND_MULTs.
17446 2013-05-08 Jan Hubicka <jh@suse.cz>
17449 lto-symtab.c (lto_symtab_symbol_p): New function.
17450 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17451 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17452 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17453 Skip static symbols.
17455 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17457 PR tree-optimization/57200
17458 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17459 Only call inform if the preceding warning_at returns true.
17461 2013-05-07 Han Shen <shenhan@google.com>
17463 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17464 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17465 * common.opt (fstack-protector-strong): New option.
17466 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17467 * doc/invoke.texi (Optimization Options): Document
17468 "-fstack-protector-strong".
17469 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17471 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17473 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17475 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17477 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17478 it is not lhs of the insn.
17479 (setup_reg_equiv): Remove insn having equiv memory which it is not
17481 * lra-constraints.c (process_address): Try to improve generation
17482 code for address base + disp.
17483 (lra_constraints): Make correct the code for checking insn setting
17484 up backward equivalence. Remove insn only if it is in the init
17486 * lra-eliminations.c (update_reg_eliminate): Change return value.
17487 (lra_eliminate): Use the result.
17489 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17491 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17492 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17493 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17494 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17495 register target operands.
17496 (*vec_extractv8hi_sse2): New pattern.
17497 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17498 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17499 (*vec_extract<mode>_mem): New insn and split pattern.
17501 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17503 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17504 (TARGET_ASAN_SHADOW_OFFSET): Define.
17505 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17506 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17508 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17510 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17511 (incr_vec_index): Return -1 if increment not found.
17512 (create_add_on_incoming_edge): Assert if increment not found.
17513 (record_increment): Limit number of increments recorded.
17514 (all_phi_incrs_profitable): Return false if an increment not found.
17515 (replace_profitable_candidates): Don't process increments that were
17517 (analyze_candidates_and_replace): Limit size of incr_vec.
17519 2013-05-07 Richard Biener <rguenther@suse.de>
17521 * calls.c (special_function_p): setjmp-like functions are leaf.
17522 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17523 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17525 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17527 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17528 (aarch64_simd_mov<mode>): New expander.
17529 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17530 (aarch64_simd_mov_to_<mode>high): Likewise.
17531 (aarch64_simd_mov_from_<mode>low): Likewise.
17532 (aarch64_simd_mov_from_<mode>high): Likewise.
17533 (aarch64_dup_lane<mode>): Update.
17534 (aarch64_dup_lanedi): New instruction pattern.
17535 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17536 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17538 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17540 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17541 (replace_mult_candidate): Remove unnecessary argument; remove
17542 unnecessary parameter from call to introduce_cast_before_cand.
17543 (replace_unconditional_candidate): Remove unnecessary parameter
17544 from call to replace_mult_candidate.
17545 (replace_conditional_candidate): Likewise.
17546 (insert_initializers): Use make_temp_ssa_name.
17547 (introduce_cast_before_cand): Remove unnecessary argument; use
17548 make_temp_ssa_name.
17549 (replace_one_candidate): Remove unnecessary argument; remove
17550 unnecessary parameter from calls to introduce_cast_before_cand.
17551 (replace_profitable_candidates): Remove unnecessary parameters
17552 from calls to replace_one_candidate.
17554 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17556 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17557 phi def as possibly hiding a basis for a CAND_ADD whose operands
17558 have been commuted in the analysis.
17559 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17561 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17563 * config/aarch64/aarch64.md
17564 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17565 shift value between 0-4.
17567 2013-05-07 Richard Biener <rguenther@suse.de>
17569 * double-int.h (rshift): New overload.
17570 * double-int.c (rshift): New function.
17571 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17572 (create_reference_ops_from_ref): Remove.
17573 (vn_reference_insert): Use shared ops for constructing the
17574 reference and copy it.
17576 2013-05-07 Richard Biener <rguenther@suse.de>
17578 PR middle-end/57190
17579 * tree-eh.c (sink_clobbers): Properly propagate
17580 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17582 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17584 PR tree-optimization/57149
17585 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17586 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17587 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17588 uninit_undefined_value_p instead of ssa_undefined_value_p.
17591 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17592 for modifier == EXPAND_INITIALIZER.
17594 2013-05-07 Anton Blanchard <anton@samba.org>
17596 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17597 for powerpc64 little endian.
17598 * configure: Regenerate.
17600 2013-05-06 Graham Stott <grahams@btinternet.com>
17602 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17603 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17604 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17605 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17607 2013-05-06 Graham Stott <grahams@btinternet.com>
17609 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17610 codes which allow non-lvalues.
17612 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17614 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17615 components are all 1s.
17616 (integer_minus_onep): New function.
17617 * tree.h (integer_minus_onep): Declare it.
17618 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17619 integer_minus_onep instead of integer_all_onesp.
17621 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17624 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17625 variations of these patterns.
17627 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17629 * config/i386/i386.md (isa): Add x64_sse4 member.
17630 (enabled): Handle x64_sse4.
17631 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17632 instruction for 64bit SSE4_1 targets. Update insn attributes.
17633 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17634 instruction for SSE4_1 targets. Update insn attributes.
17635 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17636 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17637 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17639 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17640 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17641 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17643 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17646 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17648 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17650 * final.c (do_assembler_dialects): Don't handle curly braces and
17651 vertical bar escaped by % as dialect delimiters.
17652 (output_asm_insn): Print curly braces and vertical bar if escaped
17653 by % and ASSEMBLER_DIALECT defined.
17654 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17655 * doc/tm.texi: Regenerated.
17657 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17659 * config/mips/mips.c: Include tree-pass.h.
17660 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17661 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17662 (pass_mips_machine_reorg2): New machine specific pass.
17663 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17664 (mips_option_override): Register the new pass.
17665 * rtl.h (cleanup_barriers): Remove prototype.
17666 (dbr_schedule): Likewise.
17667 * jump.c (cleanup_barriers): Make static.
17668 * reorg.c (dbr_schedule): Likewise.
17670 2013-05-06 Richard Biener <rguenther@suse.de>
17672 PR tree-optimization/57185
17673 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17674 SSA names properly.
17676 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17679 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17680 between "first_arg" and "insn", not "last" and "insn".
17682 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17684 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17685 (find_candidates_in_block): Re-enable slsr_process_phi.
17686 (create_phi_basis): Fix double counting of candidate adjustment.
17688 2013-05-06 Richard Biener <rguenther@suse.de>
17690 PR middle-end/57147
17691 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17692 the edge is also fallthru, preserve it and just clear the
17694 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17695 also complex, preserve that and just clear the fallthru flag.
17696 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17697 update virtual operands.
17699 2013-05-06 Alan Modra <amodra@gmail.com>
17701 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17702 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17703 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17704 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17705 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17706 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17708 2013-05-06 Alan Modra <amodra@gmail.com>
17710 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17711 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17712 (DEFAULT_ASM_ENDIAN): Define.
17713 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17714 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17715 Update -K PIC clause from sysv4.h.
17716 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17717 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17719 2013-05-06 Alan Modra <amodra@gmail.com>
17721 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17722 twice for little-endian.
17723 (ashrdi3_no_power, ashrdi3): Support little-endian.
17725 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17728 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17729 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17731 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17732 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17733 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17736 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17737 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17739 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17740 * config/pa/pa.opt: Make mbig-switch a no-op.
17741 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17742 (CASE_VECTOR_MODE): Always return SImode.
17743 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17744 for the !TARGET_BIG_SWITCH case.
17745 * config/pa/pa-linux.h: Likewise.
17746 * config/pa/pa-openbsd.h: Likewise.
17747 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17748 * config/pa/pa.md (short_jump): Remove define_insn.
17749 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17750 (casesi0): Remove define_insn.
17751 (type): Remove btable_branch.
17752 (pa_combine_type): Likewise.
17753 (in_nullified_branch_delay): Likewise.
17754 (in_call_delay): Likewise.
17755 (define_delay): Likewise.
17756 (define_insn_reservation "Z3"): Likewise.
17757 (define_insn_reservation "Z4"): Likewise.
17758 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17759 (pa_adjust_insn_length): Remove adjustment for btable branches.
17760 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17763 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17765 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17766 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17767 Add m->r,x alternatives.
17768 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17769 splitters using SWI48x mode iterator.
17770 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17771 TARGET_64BIT. Add m->x alternative.
17772 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17773 Add o->x alternative. Enable for TARGET_SSE.
17774 (sse_storeq): Remove expander.
17775 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17776 with memory input operand.
17777 (*vec_extractv2di_1 splitter): New.
17778 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17779 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17781 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17783 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17784 (INT_LOWPART): Delete.
17785 (extract_MB): Adjust.
17786 (extract_ME): Adjust.
17787 (print_operand): Adjust.
17789 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17791 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17792 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17793 (reg_or_logical_cint_operand, easy_fp_constant,
17794 logical_const_operand): Delete "CONST_DOUBLE" case.
17795 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17796 "HOST_BITS_PER_WIDE_INT == 64" test.
17797 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17798 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17799 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17800 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17802 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17803 CONST_DOUBLE DImode/VOIDmode case.
17804 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17805 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17806 CONST_DOUBLE VOIDmode case.
17807 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17808 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17809 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17810 Delete CONST_DOUBLE case.
17811 (splitters for mov FMOVE64 const_double): Delete
17812 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17813 "HOST_BITS_PER_WIDE_INT >= 64" test.
17814 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17816 (mov DI const_double): Delete.
17818 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17820 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17821 on op shows all bits zero in mode of a lowpart subreg, return zero.
17823 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17826 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17827 to save TFmode registers and DImode to save TImode registers for
17828 caller save operations.
17829 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17830 mark being partially clobbered since they only use the first
17833 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17834 and TDmode only use the upper 64-bits of each VSX register.
17836 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17838 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17839 (find_candidates_in_block): Disable slsr_process_phi.
17841 2013-05-03 Guozhi Wei <carrot@google.com>
17843 * coverage.c (coverage_obj_init): Move the construction of gcov
17845 (build_init_ctor): ... here.
17847 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17849 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17850 (slsr_cand_d): Redefine def_phi.
17851 (stride_status, phi_adjust_status, count_phis_status): New enums.
17852 (find_phi_def): New.
17853 (find_basis_for_base_expr): New.
17854 (find_basis_for_candidate): Handle hidden bases.
17855 (alloc_cand_and_find_basis): Handle phi candidates.
17856 (slsr_process_phi): New.
17857 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17858 (create_mul_imm_cand): Likewise.
17859 (create_add_ssa_cand): Exclude phi base candidates.
17860 (create_add_imm_cand): Likewise.
17861 (slsr_process_cast): Likewise.
17862 (slsr_process_copy): Likewise.
17863 (find_candidates_in_block): Handle phi candidates.
17864 (dump_candidate): Likewise.
17865 (unconditional_cands): Delete.
17866 (unconditional_cands_with_known_stride_p): Delete.
17867 (phi_dependent_cand_p): New.
17868 (cand_increment): Handle phi-dependent candidates.
17869 (replace_dependent): Delete.
17870 (replace_mult_candidate): New.
17871 (replace_unconditional_candidate): New.
17872 (incr_vec_index): Move to avoid forward reference.
17873 (create_add_on_incoming_edge): New.
17874 (create_phi_basis): New.
17875 (replace_dependents): Delete.
17876 (replace_conditional_candidate): New.
17877 (phi_add_costs): New.
17878 (replace_uncond_cands_and_profitable_phis): New.
17879 (record_increment): Handle phi adjustments.
17880 (record_phi_increments): New.
17881 (record_increments): Handle phi adjustments.
17882 (phi_incr_cost): New.
17883 (lowest_cost_path): Handle phis.
17884 (total_savings): Likewise.
17885 (analyze_increments): Likewise.
17886 (ncd_with_phi): New.
17887 (ncd_of_cand_and_phis): New.
17888 (nearest_common_dominator_for_cands): Handle phi increments.
17889 (all_phi_incrs_profitable): New.
17890 (replace_profitable_candidates): Handle phi-dependent candidates.
17891 (analyze_candidates_and_replace): Likewise.
17893 2013-05-03 Teresa Johnson <tejohnson@google.com>
17896 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17897 do not exceed REG_BR_PROB_BASE.
17899 2013-05-03 Jeff Law <law@redhat.com>
17901 PR tree-optimization/57144
17902 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17903 operand of the condition will bit into the new type when eliminating
17904 a cast feeding a condition.
17906 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17908 PR rtl-optimization/57130
17909 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17910 of COMPARE as in_code to the recursive call if needed.
17912 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17914 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17915 (enabled): Handle new members.
17916 * config/i386/sse.md (*vec_concatv2si): Merge from
17917 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17918 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17920 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17922 PR tree-optimization/57027
17923 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17924 for fnms opportunity, check we got the prerequisite kind
17925 of tree / gimple before using accessor functions.
17927 2013-05-03 Richard Biener <rguenther@suse.de>
17929 * double-int.h (lshift): New overload without precision
17930 and arith argument.
17931 (operator *=, operator +=, operator -=): Move ...
17932 * double-int.c (operator *=, operator +=, operator -=): ... here
17933 and implement more efficiently.
17934 (mul_double_with_sign): Remove.
17935 (lshift_double): Adjust to take unsinged shift argument, push
17936 dispatching code to callers.
17937 (mul_double_wide_with_sign): Add early out for callers that
17938 are not interested in high parts or overflow.
17939 (lshift): New function.
17940 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17941 dispatch code here.
17942 (lrotate, rrotate): Use logical shifts.
17943 * expr.c (get_inner_reference): Use lshift.
17944 * fixed-value.c (do_fixed_divide): Likewise.
17945 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17946 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17947 (indirect_refs_may_alias_p): Likewise.
17948 (stmt_kills_ref_p_1): Likewise.
17950 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17952 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17954 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17956 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17957 scalar form of FABD instruction.
17959 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17961 * lra-constraints.c (process_alt_operands): Add checking alt
17962 number to choose the best alternative.
17964 2013-05-02 Richard Biener <rguenther@suse.de>
17966 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17967 bitmap and its handling.
17968 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17970 2013-05-02 Richard Biener <rguenther@suse.de>
17972 PR middle-end/57140
17973 * tree-inline.c (copy_loops): Properly handle removed loops.
17974 (copy_cfg_body): Mark destination loops for fixup if source
17975 loops needed fixup.
17977 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17980 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17981 generating simple_return for naked functions.
17983 2013-05-02 Martin Jambor <mjambor@suse.cz>
17985 PR middle-end/56988
17986 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17987 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17989 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17990 ipa_agg_replacement_value structures.
17991 (known_aggs_to_agg_replacement_list): Likewise.
17992 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17993 (read_agg_replacement_chain): Likewise.
17994 (ipcp_transform_function): Also check that by_ref flags match.
17996 2013-05-02 Richard Biener <rguenther@suse.de>
17998 * graphds.h (struct graph): Add obstack member.
17999 * graphds.c (new_graph): Initialize obstack and allocate
18001 (add_edge): Allocate edge from the obstack.
18002 (free_graph): Free the obstack instead of all edges and vertices.
18004 2013-05-02 Teresa Johnson <tejohnson@google.com>
18006 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18008 * cfg.c (update_bb_profile_for_threading): Ditto.
18009 * tree-inline.c (copy_bb): Ditto.
18010 (copy_edges_for_bb): Ditto.
18011 (initialize_cfun): Ditto.
18012 (copy_cfg_body): Ditto.
18013 (expand_call_inline): Ditto.
18014 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18015 (estimate_node_size_and_time): Ditto.
18016 (inline_merge_summary): Ditto.
18017 * cgraphclones.c (cgraph_clone_edge): Ditto.
18018 (cgraph_clone_node): Ditto.
18019 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18020 (compute_trg_info): Ditto.
18022 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18024 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18025 S reg when fp attribute set.
18026 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18028 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18030 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18032 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18033 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18034 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18036 2013-05-02 Richard Biener <rguenther@suse.de>
18038 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18039 (struct instantiate_cache_entry): New type.
18040 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18041 (struct instantiate_cache_type): New type.
18042 (set_instantiated_value, get_instantiated_value): Remove.
18043 (get_instantiated_value_entry): New function.
18044 (instantiate_scev_name): Use the new cache and adjust.
18045 (instantiate_scev_poly): Adjust.
18046 (instantiate_scev_binary): Likewise.
18047 (instantiate_array_ref): Likewise.
18048 (instantiate_scev_convert): Likewise.
18049 (instantiate_scev_not): Likewise.
18050 (instantiate_scev_3): Likewise.
18051 (instantiate_scev_2): Likewise.
18052 (instantiate_scev_r): Likewise.
18053 (instantiate_scev): Likewise.
18054 (resolve_mixers): Likewise.
18056 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18059 * lra-constraints.c (best_small_class_operands_num): Remove.
18060 (process_alt_operands): Remove small_class_operands_num. Take
18061 small classes operands into losers and only if the operand is not
18062 matched. Modify debugging output.
18063 (curr_insn_transform): Remove best_small_class_operands_num.
18066 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18068 * config/aarch64/aarch64-builtins.c
18069 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18070 * config/aarch64/aarch64-simd-builtins.def
18071 (reduc_splus_): Add new modes.
18072 (reduc_uplus_): New.
18073 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18074 (reduc_uplus_v4sf): Likewise.
18075 (reduc_splus_v4sf): Likewise.
18076 (aarch64_addv<mode>): Likewise.
18077 (reduc_uplus_<mode>): Likewise.
18078 (reduc_splus_<mode>): Likewise.
18079 (aarch64_addvv2di): Likewise.
18080 (reduc_uplus_v2di): Likewise.
18081 (reduc_splus_v2di): Likewise.
18082 (aarch64_addvv2si): Likewise.
18083 (reduc_uplus_v2si): Likewise.
18084 (reduc_splus_v2si): Likewise.
18085 (reduc_<sur>plus_<mode>): New.
18086 (reduc_<sur>plus_v2di): Likewise.
18087 (reduc_<sur>plus_v2si): Likewise.
18088 (reduc_<sur>plus_v4sf): Likewise.
18089 (aarch64_addpv4sf): Likewise.
18090 * config/aarch64/arm_neon.h
18091 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18092 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18093 add UNSPEC_SADDV, UNSPEC_UADDV.
18095 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18097 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18099 * config/aarch64/arm_neon.h
18100 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18102 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18104 * config/aarch64/aarch64-builtins
18105 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18107 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18109 * config/aarch64/aarch64-simd-builtins.def
18110 (reduc_smax_): New.
18111 (reduc_smin_): Likewise.
18112 (reduc_umax_): Likewise.
18113 (reduc_umin_): Likewise.
18114 (reduc_smax_nan_): Likewise.
18115 (reduc_smin_nan_): Likewise.
18118 (smax): Update for V2SF, V4SF and V2DF modes.
18121 (smin_nan): Likewise.
18122 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18123 (<su><maxmin><mode>3): ...This, refactor.
18124 (s<maxmin><mode>3): New.
18125 (<maxmin_uns><mode>3): Likewise.
18126 (reduc_<maxmin_uns>_<mode>): Refactor.
18127 (reduc_<maxmin_uns>_v4sf): Likewise.
18128 (reduc_<maxmin_uns>_v2si): Likewise.
18129 (aarch64_<fmaxmin><mode>: Remove.
18130 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18132 (vmin<q>_f<32,64>): Likewise.
18133 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18135 (su): Add mappings for smax, smin, umax, umin.
18137 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18138 (FMAXMIN): Rename as...
18139 (FMAXMIN_UNS): ...This.
18141 (fmaxminv): Likewise.
18142 (fmaxmin): Likewise.
18144 (maxmin_uns_op): Likewise.
18146 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18148 * config/aarch64/arm_neon.h
18149 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18150 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18151 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18153 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18155 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18156 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18158 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18160 * config/aarch64/aarch64-simd.md
18161 (vcond<mode>_internal): Handle special cases for constant masks.
18162 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18163 (vcondu<mode><mode>): Likewise.
18164 (vcond<v_cmp_result><mode>): New.
18166 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18168 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18169 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18170 * config/aarch64/aarch64-simd-builtins.def
18171 (cmeq): Update to BUILTIN_VALLDI.
18176 * config/aarch64/arm_neon.h
18177 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18178 to builtins or C as appropriate.
18180 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18182 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18184 (cmhi): Rename to...
18186 * config/aarch64/aarch64-simd.md
18187 (simd_mode): Add SF.
18188 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18189 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18190 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18191 (cstore<mode>_neg): ...This.
18192 * config/aarch64/iterators.md
18194 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18195 (COMPARISONS): New.
18196 (UCOMPARISONS): Likewise.
18197 (optab): Add missing comparisons.
18203 (VCMP_S): Likewise.
18204 (VCMP_U): Likewise.
18205 (V_cmp_result): Add DF, SF modes.
18206 (v_cmp_result): Likewise.
18208 (vmtype): Likewise.
18209 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18211 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18213 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18214 define_insn to define_insn_and_split.
18215 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18216 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18217 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18218 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18219 (thumb2_negscc): Likewise.
18221 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18223 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18225 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18227 * config/arm/thumb2.md: Remove trailing whitespaces.
18229 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18231 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18232 Use gen_int_mode rather than GEN_INT.
18234 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18236 * value-prof.c (stream_in_histogram_value): Remove the strayed
18239 2013-04-30 Richard Biener <rguenther@suse.de>
18241 PR middle-end/57122
18242 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18244 2013-04-30 Richard Biener <rguenther@suse.de>
18246 PR middle-end/57107
18247 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18249 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18251 PR rtl-optimization/56957
18252 PR rtl-optimization/57105
18253 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18254 variable. Use just INSN_UID for determining whether an insn
18255 should be only disconnected from the insn stream.
18256 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18258 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18260 PR tree-optimization/57104
18261 * tsan.c (instrument_expr): Don't instrument accesses to
18262 DECL_HARD_REGISTER VAR_DECLs.
18264 2013-04-30 Richard Biener <rguenther@suse.de>
18266 * function.h (loops_for_fn): New inline function.
18267 (set_loops_for_fn): Likewise.
18268 * cfgloop.h (place_new_loop): Add struct function parameter.
18269 (get_loop): Likewise.
18270 (get_loops): Likewise.
18271 (number_of_loops): Likewise.
18272 (fel_next): Adjust.
18273 (fel_init): Likewise.
18274 * cfg.c (get_loop_copy): Adjust.
18275 * cfgloop.c (flow_loops_dump): Likewise.
18276 (record_loop_exits): Likewise.
18277 (verify_loop_structure): Likewise.
18278 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18279 (estimate_reg_pressure_cost): Likewise.
18280 (mark_loop_exit_edges): Likewise.
18281 * cfgloopmanip.c (place_new_loop): Likewise.
18282 (add_loop): Likewise.
18283 (duplicate_loop): Likewise.
18284 * graph.c (draw_cfg_nodes): Likewise.
18285 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18286 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18287 (extract_affine_chrec): Likewise.
18288 (build_scop_iteration_domain): Likewise.
18289 * graphite.c (graphite_initialize): Likewise.
18290 * ira-build.c (create_loop_tree_nodes): Likewise.
18291 (more_one_region_p): Likewise.
18292 (rebuild_regno_allocno_maps): Likewise.
18293 (mark_loops_for_removal): Likewise.
18294 (mark_all_loops_for_removal): Likewise.
18295 (remove_unnecessary_regions): Likewise.
18296 (ira_build): Likewise.
18297 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18298 * loop-init.c (fix_loop_structure): Likewise.
18299 (gate_rtl_move_loop_invariants): Likewise.
18300 (gate_rtl_unswitch): Likewise.
18301 (gate_rtl_unroll_and_peel_loops): Likewise.
18302 (rtl_doloop): Likewise.
18303 * lto-streamer-in.c (input_cfg): Likewise.
18304 * lto-streamer-out.c (output_cfg): Likewise.
18305 * modulo-sched.c (sms_schedule): Likewise.
18306 * predict.c (tree_estimate_probability): Likewise.
18307 (tree_estimate_probability_driver): Likewise.
18308 (estimate_loops): Likewise.
18309 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18310 (move_sese_region_to_fn): Likewise.
18311 (debug_loop_num): Likewise.
18312 * tree-chrec.c (chrec_evaluate): Likewise.
18313 (hide_evolution_in_other_loops_than_loop): Likewise.
18314 (chrec_component_in_loop_num): Likewise.
18315 (reset_evolution_in_loop): Likewise.
18316 (evolution_function_is_invariant_rec_p): Likewise.
18317 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18318 * tree-inline.c (copy_loops): Likewise.
18319 (copy_cfg_body): Likewise.
18320 (tree_function_versioning): Likewise.
18321 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18322 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18324 (add_to_evolution_1): Likewise.
18325 (scev_const_prop): Likewise.
18326 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18327 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18328 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18329 (tree_ssa_lim_initialize): Likewise.
18330 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18331 (verify_loop_closed_ssa): Likewise.
18332 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18333 (tree_ssa_loop_im): Likewise.
18334 (tree_ssa_loop_unswitch): Likewise.
18335 (tree_vectorize): Likewise.
18336 (check_data_deps): Likewise.
18337 (tree_ssa_loop_ivcanon): Likewise.
18338 (tree_ssa_loop_bounds): Likewise.
18339 (tree_complete_unroll): Likewise.
18340 (tree_complete_unroll_inner): Likewise.
18341 (tree_parallelize_loops): Likewise.
18342 (tree_ssa_loop_prefetch): Likewise.
18343 (tree_ssa_loop_ivopts): Likewise.
18344 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18345 * tree-vectorizer.c (vectorize_loops): Likewise.
18347 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18349 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18350 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18351 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18352 with EABI_LINK_SPEC.
18354 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18357 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18360 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18363 * lra-constraints.c (process_alt_operands): Discourage a bit more
18364 using memory for pseudos. Print cost dump for alternatives.
18365 Modify cost values for conflicts with early clobbers.
18366 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18368 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18371 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18373 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18375 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18376 from/to S register.
18377 (movdi_aarch64): Support LDR/STR from/to D register.
18379 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18381 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18382 or higher by default.
18384 2013-04-29 Richard Biener <rguenther@suse.de>
18386 PR middle-end/57075
18387 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18388 even if not adding abnormal edges for calls that can make
18391 2013-04-29 Richard Biener <rguenther@suse.de>
18393 PR middle-end/57103
18394 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18396 (move_stmt_r): Remove redundant checking.
18398 2013-04-29 Teresa Johnson <tejohnson@google.com>
18401 * basic-block.h (apply_scale): New function.
18402 (apply_probability): Use apply_scale.
18403 * gimple-streamer-in.c (input_bb): Ditto.
18404 * lto-streamer-in.c (input_cfg): Ditto.
18405 * lto-cgraph.c (merge_profile_summaries): Ditto.
18406 * tree-optimize.c (execute_fixup_cfg): Ditto.
18407 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18408 (copy_edges_for_bb): Ditto.
18409 (copy_cfg_body): Ditto.
18411 2013-04-29 Tom de Vries <tom@codesourcery.com>
18413 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18414 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18415 (tail_merge_optimize): Handle current_loops == NULL.
18417 2013-04-26 Jeff Law <law@redhat.com>
18419 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18420 (simplify_cond_using_ranges): Generalize code to simplify
18421 COND_EXPRs where one argument is a constant and the other
18422 is an SSA_NAME created by an integral type conversion.
18424 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18426 * config/arm/arm.md (store_minmaxsi): Use only when
18427 optimize_insn_for_size_p.
18429 2013-04-29 Christian Bruel <christian.bruel@st.com>
18432 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18434 2013-04-29 Richard Biener <rguenther@suse.de>
18436 PR middle-end/57089
18437 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18438 loop tree make sure to schedule a fixup for the child as well.
18439 (expand_omp_for_generic): Properly add loops.
18440 (expand_omp_for_static_nochunk): Likewise.
18441 (expand_omp_for_static_chunk): Likewise.
18442 (expand_omp_for): For the degenerate case fixup loops.
18443 (expand_omp_sections): Fix default bb placement in loops.
18444 (expand_omp_atomic_pipeline): Properly add loops.
18446 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18448 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18450 2013-04-29 Tom de Vries <tom@codesourcery.com>
18452 * tree-ssa-tail-merge.c: Update header comment.
18454 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18456 * config/aarch64/arm_neon.h
18457 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18458 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18459 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18460 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18461 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18462 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18463 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18464 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18466 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18468 * config/aarch64/aarch64-simd.md
18469 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18470 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18471 fix_trunc, fixuns_trunc.
18472 (ftrunc<VDQF:mode>2): New.
18473 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18474 (fix_trunc_optab): New.
18476 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18478 * config/aarch64/aarch64-builtins.c
18479 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18480 iceilf, lround, iroundf.
18482 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18485 * config/i386/i386.h (enum ix86_tune_indices)
18486 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18487 New, split from X86_TUNE_INTER_UNIT_MOVES.
18488 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18489 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18490 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18491 (TARGET_INTER_UNIT_MOVES): Remove.
18492 * config/i386/i386.c (initial_ix86_tune_features): Update.
18493 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18494 (ix86_expand_convert_uns_didf_sse): Use
18495 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18496 (ix86_expand_vector_init_one_nonzero): Ditto.
18497 (ix86_expand_vector_init_interleave): Ditto.
18498 (inline_secondary_memory_needed): Return true for moves from SSE class
18499 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18500 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18501 * config/i386/constraints.md (Yi, Ym): Depend on
18502 TARGET_INTER_UNIT_MOVES_TO_VEC.
18503 (Yj, Yn): New constraints.
18504 * config/i386/i386.md (*movdi_internal): Change constraints of
18505 operand 1 from Yi to Yj and from Ym to Yn.
18506 (*movsi_internal): Ditto.
18507 (*movdf_internal): Ditto.
18508 (*movsf_internal): Ditto.
18509 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18510 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18511 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18512 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18513 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18514 * config/i386/sse.md (movdi_to_sse): Ditto.
18515 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18516 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18517 TARGET_INTER_UNIT_MOVES.
18518 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18519 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18520 instead of TARGET_INTER_UNIT_MOVES.
18521 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18522 operand 1 from Yi to Yj and from Ym to Yn.
18524 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18526 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18527 (float_truncate_hi_): Likewise.
18528 (float_extend_lo_): Likewise.
18529 (float_truncate_lo_): Likewise.
18530 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18531 (aarch64_float_extend_lo_v2df): Likewise.
18532 (vec_unpacks_hi_v4sf): Likewise.
18533 (aarch64_float_truncate_lo_v2sf): Likewise.
18534 (aarch64_float_truncate_hi_v4sf): Likewise.
18535 (vec_pack_trunc_v2df): Likewise.
18536 (vec_pack_trunc_df): Likewise.
18538 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18540 * config/aarch64/aarch64-builtins.c
18541 (aarch64_fold_builtin): Fold float conversions.
18542 * config/aarch64/aarch64-simd-builtins.def
18543 (floatv2si, floatv4si, floatv2di): New.
18544 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18545 * config/aarch64/aarch64-simd.md
18546 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18547 * config/aarch64/iterators.md (FLOATUORS): New.
18548 (optab): Add float, floatuns.
18549 (su_optab): Likewise.
18551 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18553 * config/aarch64/aarch64-builtins.c
18554 (aarch64_builtin_vectorized_function): Use new names for
18556 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18557 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18558 (fcvtzu): Split as...
18559 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18560 (fcvtas): Split as...
18561 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18562 (fcvtau): Split as...
18563 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18564 (fcvtps): Split as...
18565 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18566 (fcvtpu): Split as...
18567 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18568 (fcvtms): Split as...
18569 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18570 (fcvtmu): Split as...
18571 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18572 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18573 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18574 (lfrintnusf, lfrintnudf): Likewise.
18575 * config/aarch64/aarch64-simd.md
18576 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18578 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18579 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18580 (fcvt_pattern): Likewise.
18582 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18584 * config/aarch64/aarch64-simd.md
18585 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18586 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18588 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18590 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18591 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18592 (vrnd<a,m,n,p>_f32): Implement using builtins.
18593 (vrnd<i,x><q>_f<32, 64>): New.
18595 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18597 * config/aarch64/aarch64-builtins.c
18598 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18599 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18600 (frintz): Rename to...
18602 (frintp): Rename to...
18604 (frintm): Rename to...
18606 (frinti): Rename to...
18607 (nearbyint): ...this.
18608 (frintx): Rename to...
18610 (frinta): Rename to...
18612 * config/aarch64/aarch64-simd.md
18613 (aarch64_frint<frint_suffix><mode>): Delete.
18614 (<frint_pattern><mode>2): Convert to insn.
18615 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18616 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18617 (frint_pattern): Likewise.
18618 (frint_suffix): Likewise.
18620 2013-04-29 Richard Biener <rguenther@suse.de>
18622 PR tree-optimization/57081
18623 * loop-init.c: Include tree-flow.h.
18624 (loop_optimizer_finalize): Free number of iteration estimates.
18625 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18627 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18629 PR tree-optimization/57083
18630 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18631 non-singleton shift count range, zero extend low_bound for uns case.
18633 * config/i386/predicates.md (general_vector_operand): New predicate.
18634 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18635 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18636 if they aren't nonimmediate operands. If their original values
18637 satisfy const_vector_equal_evenodd_p, don't shift them.
18638 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18639 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18640 if not nonimmediate_operand.
18641 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18642 instead of register_operand.
18643 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18645 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18647 * stor-layout.c (finalize_size_functions): Allocate a structure and
18648 reset cfun before dumping the functions.
18650 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18652 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18655 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18656 use xop_pmacsdqh if uns_p.
18657 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18658 the immediate rotate count.
18660 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18662 * rtl.h (struct rtx_def): Add comment for field jump.
18663 (LRA_SUBREG_P): New macro.
18664 * recog.c (register_operand): Check LRA_SUBREG_P.
18665 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18667 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18668 (lra_final_code_change): Skip subreg change for operators.
18669 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18670 if there are no operand changes.
18671 * lra-constraints.c (curr_insn_set): New.
18672 (match_reload): Set LRA_SUBREG_P.
18673 (emit_spill_move): Ditto.
18674 (check_and_process_move): Use curr_insn_set. Process only single
18675 set insns. Don't initialize sec_mem_p and change_p.
18676 (simplify_operand_subreg): Use LRA_SUBREG_P.
18677 (reg_in_class_p): New function.
18678 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18679 of #ifdef. Add code to remove cycling.
18680 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18681 non-null disp. Reload inner instead of disp when base and index
18682 are null. Try to put lo_sum into register.
18683 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18684 (check_and_process_move): Move code for move cost check to
18685 simple_move_p. Remove equiv_substitution.
18686 (simple_move_p): New function.
18687 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18688 curr_insn_set. Call check_and_process_move only for single set
18689 insns. Use the new function. Move call of check_and_process_move
18690 after operand equiv substitution and address process.
18692 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18695 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18696 with nonlocal goto receivers or returns twice calls, ignore
18697 unininitialized values from abnormal edges to nl goto receiver
18698 or returns twice call.
18700 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18702 PR tree-optimization/57051
18703 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18704 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18707 2013-04-26 Richard Biener <rguenther@suse.de>
18709 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18710 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18711 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18712 (fixup_loop_arrays_after_move): New function.
18713 (move_sese_region_to_fn): Properly outline the loop tree parts
18714 of the SESE region.
18716 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18718 * config/i386/i386.md (type, unit): Fix long lines.
18720 2013-04-26 Richard Biener <rguenther@suse.de>
18722 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18723 (lto-streamer-out.o): Likewise.
18724 * cfgloop.c (init_loops_structure): Export, add struct function
18725 argument and adjust.
18726 (flow_loops_find): Adjust.
18727 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18728 (init_loops_structure): Declare.
18729 * lto-streamer-in.c: Include cfgloop.h.
18730 (input_cfg): Input the loop tree.
18731 * lto-streamer-out.c: Include cfgloop.h.
18732 (output_cfg): Output the loop tree.
18733 (output_struct_function_base): Do not drop PROP_loops.
18735 2013-03-26 Richard Biener <rguenther@suse.de>
18737 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18738 (pass_build_cfg): Provide PROP_loops.
18739 (move_sese_region_to_fn): Remove loops that are outlined into fn
18741 * tree-inline.c: Include cfgloop.h.
18742 (initialize_cfun): Do not drop PROP_loops.
18743 (copy_loops): New function.
18744 (copy_cfg_body): Copy loop structure.
18745 (tree_function_versioning): Initialize destination loop tree.
18746 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18747 (pass_parallelize_loops): Do IL verification.
18748 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18749 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18750 the CFG make sure we fixup loops as well.
18751 * tree-ssa-tail-merge.c: Include cfgloop.h.
18752 (replace_block_by): When merging loop latches mark loops for fixup.
18753 * lto-streamer-out.c (output_struct_function_base): Drop
18754 PROP_loops for now.
18755 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18756 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18757 * ipa-split.c: Include cfgloop.h.
18758 (split_function): Add the new return block to the loop tree root.
18759 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18760 whether we have removed the forwarder block.
18761 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18762 * cfgloop.h (place_new_loop): Declare.
18763 * cfgloopmanip.c (place_new_loop): Export.
18764 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18765 (tree-switch-conversion.o): Likewise.
18766 (tree-complex.o): Likewise.
18767 (tree-inline.o): Likewise.
18768 (tree-ssa-tailmerge.o): Likewise.
18769 (ipa-split.o): Likewise.
18770 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18771 (tree-ssa-copy.o): Likewise.
18772 * tree-switch-conversion.c: Include cfgloop.h
18773 (process_switch): If we emit a bit-test cascade, schedule loops
18775 * tree-complex.c: Include cfgloop.h.
18776 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18777 * asan.c: Include cfgloop.h.
18778 (create_cond_insert_point): Properly add new basic-blocks to
18779 loops, schedule loop fixup.
18780 * cfgloop.c (verify_loop_structure): Check that looks are not
18782 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18784 (expand_omp_for_generic): Likewise.
18785 (expand_omp_sections): Likewise.
18786 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18787 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18788 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18789 is initialized, not when loops are present.
18790 * tree-parloops.c (parallelize_loops): Remove checking here.
18791 * passes.c (init_optimization_passes): Schedule a copy-propagation
18792 pass before complete unrolling of inner loops.
18794 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18796 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18797 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18798 (colorize_init): Add argument to _WIN32 version.
18799 * toplev.c: Include diagnostic-color.h.
18800 (process_options): Default to -fdiagnostics-color=auto if
18801 GCC_COLORS env var is in the environment.
18802 * common.opt (fdiagnostics-color=): Add Var and Init.
18803 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18804 env var is in the environment, the default is auto rather than never.
18806 * diagnostic.h (file_name_as_prefix): Add context argument.
18807 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18808 the string as locus.
18809 * langhooks.c (lhd_print_error_function): Adjust caller.
18811 2013-04-25 Lawrence Crowl <crowl@google.com>
18813 * var-tracking.c (shared_hash_def::htab):
18814 Change type to hash_table. Update dependent calls and types.
18816 2013-04-25 Lawrence Crowl <crowl@google.com>
18818 * Makefile.in: Update as needed below.
18820 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18821 Move declaration to after the type's method definitons.
18823 * attribs.c (htab_t scoped_attributes::attribute_hash):
18824 Change type to hash_table. Update dependent calls and types.
18826 * bitmap.c (htab_t bitmap_desc_hash):
18827 Change type to hash_table. Update dependent calls and types.
18829 * cselib.c (htab_t cselib_hash_table):
18830 Change type to hash_table. Update dependent calls and types.
18832 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18833 (hash_string_slot_node): Move implementation into lto-streamer.h
18834 struct string_slot_hasher.
18835 (eq_string_slot_node): Likewise.
18837 * data-streamer-out.c: Update output_block::string_hash_table
18838 dependent calls and types.
18840 * dwarf2cfi.c (htab_t trace_index):
18841 Change type to hash_table. Update dependent calls and types.
18843 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18844 Change type to hash_table. Update dependent calls and types.
18845 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18846 (htab_t optimize_external_refs::map): Likewise.
18847 (htab_t output_comp_unit::extern_map): Likewise.
18848 (htab_t output_comdat_type_unit::extern_map): Likewise.
18849 (htab_t output_macinfo::macinfo_htab): Likewise.
18850 (htab_t optimize_location_lists::htab): Likewise.
18851 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18853 * except.c (htab_t ehspec_hash_type):
18854 Change type to hash_table. Update dependent calls and types.
18855 (assign_filter_values::ttypes): Likewise.
18856 (assign_filter_values::ehspec): Likewise.
18857 (sjlj_assign_call_site_values::ar_hash): Likewise.
18858 (convert_to_eh_region_ranges::ar_hash): Likewise.
18860 * gcse.c (htab_t pre_ldst_table):
18861 Change type to hash_table. Update dependent calls and types.
18863 * ggc-common.c (htab_t saving_htab):
18864 Change type to hash_table. Update dependent calls and types.
18865 (htab_t loc_hash): Likewise.
18866 (htab_t ptr_hash): Likewise.
18867 (call_count): Rename ggc_call_count.
18868 (call_alloc): Rename ggc_call_alloc.
18869 (loc_descriptor): Rename make_loc_descriptor.
18870 (add_statistics): Rename ggc_add_statistics.
18872 * ggc-common.c (saving_htab):
18873 Change type to hash_table. Update dependent calls and types.
18875 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18876 (push_gimplify_context): Likewise.
18877 (pop_gimplify_context): Likewise.
18878 (struct gimple_temp_hash_elt): Added.
18879 (struct gimplify_hasher): Likewise.
18880 (struct gimplify_ctx.temp_htab):
18881 Change type to hash_table. Update dependent calls and types.
18883 * gimple-fold.c: Include gimplify-ctx.h.
18885 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18886 Change type to hash_table. Update dependent calls and types.
18887 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18888 avoid potential global name collision.
18890 * gimplify.c: Include gimplify-ctx.h.
18891 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18892 (htab_t gimplify_ctx::temp_htab):
18893 Update dependent calls and types for new type hash_table.
18894 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18895 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18897 * gimplify-ctx.h: New.
18898 (struct gimple_temp_hash_elt): Move from gimplify.c.
18899 (class gimplify_hasher): New.
18900 (struct gimplify_ctx): Move from gimple.h.
18901 (htab_t gimplify_ctx::temp_htab):
18902 Change type to hash_table. Update dependent calls and types.
18904 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18905 (htab_t ivs_params::newivs_index):
18906 Change type to hash_table. Update dependent calls and types.
18907 (htab_t ivs_params::params_index): Likewise.
18908 (htab_t print_generated_program::params_index): Likewise.
18909 (htab_t gloog::newivs_index): Likewise.
18910 (htab_t gloog::params_index): Likewise.
18912 * graphite.c: Include graphite-htab.h.
18913 4htab_t graphite_transform_loops::bb_pbb_mapping):
18914 Change type to hash_table. Update dependent calls and types.
18916 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18917 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18918 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18920 * graphite-dependences.c: Include graphite-htab.h.
18921 (loop_is_parallel_p): Change hash table type of parameter.
18923 * graphite-htab.h: New.
18924 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18925 (extern find_pbb_via_hash): Move from graphite-poly.h.
18926 (extern loop_is_parallel_p): Move from graphite-poly.h.
18927 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18929 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18930 (extern loop_is_parallel_p): Move to graphite-htab.h.
18931 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18933 * haifa-sched.c (htab_t delay_htab):
18934 Change type to hash_table. Update dependent calls and types.
18935 (htab_t delay_htab_i2): Likewise.
18937 * ira-color.c (htab_t allocno_hard_regs_htab):
18938 Change type to hash_table. Update dependent calls and types.
18940 * ira-costs.c (htab_t cost_classes_htab):
18941 Change type to hash_table. Update dependent calls and types.
18943 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18944 Change type to hash_table. Update dependent calls and types.
18946 * loop-iv.c (htab_t bivs):
18947 Change type to hash_table. Update dependent calls and types.
18949 * loop-unroll.c (htab_t opt_info::insns_to_split):
18950 Change type to hash_table. Update dependent calls and types.
18951 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18953 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18954 (struct string_slot_hasher): New.
18955 (htab_t output_block::string_hash_table):
18956 Change type to hash_table. Update dependent calls and types.
18958 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18959 (htab_t file_name_hash_table):
18960 Change type to hash_table. Update dependent calls and types.
18962 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18965 * lto-streamer.c (htab_t tree_htab):
18966 Change type to hash_table. Update dependent calls and types.
18968 * omp-low.c: Include gimplify-ctx.h.
18970 * passes.c (htab_t name_to_pass_map):
18971 Change type to hash_table. Update dependent calls and types.
18972 (pass_traverse): Rename to passes_pass_traverse.
18974 * plugin.c (htab_t event_tab):
18975 Change type to hash_table. Update dependent calls and types.
18977 * postreload-gcse.c (htab_t expr_table):
18978 Change type to hash_table. Update dependent calls and types.
18979 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18981 * sese.c (debug_rename_map_1): Make extern.
18982 (htab_t copy_bb_and_scalar_dependences::rename_map):
18983 Change type to hash_table. Update dependent calls and types.
18985 * sese.h (extern debug_rename_map): Move to .c file.
18987 * store-motion.c (htab_t store_motion_mems_table):
18988 Change type to hash_table. Update dependent calls and types.
18990 * trans-mem.c (htab_t tm_new_mem_hash):
18991 Change type to hash_table. Update dependent calls and types.
18993 * tree-browser.c (htab_t TB_up_ht):
18994 Change type to hash_table. Update dependent calls and types.
18996 * tree-cfg.c (htab_t discriminator_per_locus):
18997 Change type to hash_table. Update dependent calls and types.
18999 * tree-complex.c: Include tree-hasher.h
19000 (htab_t complex_variable_components):
19001 Change type to hash_table. Update dependent calls and types.
19003 * tree-eh.c (htab_t finally_tree):
19004 Change type to hash_table. Update dependent calls and types.
19006 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19007 struct int_tree_hasher.
19008 (extern int_tree_map_eq): Likewise.
19009 (uid_decl_map_hash): Removed.
19010 (extern decl_tree_map_eq): Likewise.
19012 * tree-hasher.h: New.
19013 (struct int_tree_hasher): New.
19014 (typedef int_tree_htab_type): New.
19016 * tree-inline.c: Include gimplify-ctx.h.
19018 * tree-mudflap.c: Include gimplify-ctx.h.
19020 * tree-parloops.c: Include tree-hasher.h.
19021 (htab_t eliminate_local_variables_stmt::decl_address):
19022 Change type to hash_table. Update dependent calls and types.
19023 (htab_t separate_decls_in_region::decl_copies): Likewise.
19025 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19026 Change type to hash_table. Update dependent calls and types.
19028 * tree-sra.c (candidates):
19029 Change type to hash_table. Update dependent calls and types.
19031 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19033 (int_tree_map_hash): Likewise.
19035 * tree-ssa-dom.c (htab_t avail_exprs):
19036 Change type to hash_table. Update dependent calls and types.
19038 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19039 Change type to hash_table. Update dependent calls and types.
19041 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19042 Change type to hash_table. Update dependent calls and types.
19044 * tree-ssa-phiopt.c (seen_ssa_names):
19045 Change type to hash_table. Update dependent calls and types.
19047 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19048 Change type to hash_table. Update dependent calls and types.
19050 * tree-ssa-uncprop.c (equiv):
19051 Change type to hash_table. Update dependent calls and types.
19053 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19055 PR rtl-optimization/57003
19056 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19057 call note_stores with kill_clobbered_value callback again after
19058 killing regs_invalidated_by_call.
19060 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19062 * config/aarch64/aarch64-simd.md
19063 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19064 (aarch64_simd_bsl<mode>): Likewise.
19065 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19067 2013-04-25 Marek Polacek <polacek@redhat.com>
19069 PR tree-optimization/57066
19070 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19072 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19074 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19076 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19078 * config/aarch64/aarch64-builtins.c
19079 (aarch64_fold_builtin): New.
19080 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19081 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19082 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19083 * config/aarch64/arm_neon.h
19084 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19086 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19087 Tejas Belagod <tejas.belagod@arm.com>
19089 * config/aarch64/aarch64-builtins.c
19090 (aarch64_gimple_fold_builtin): New.
19091 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19092 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19093 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19094 (addvv4sf): Update.
19095 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19097 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19099 * config/aarch64/aarch64.md
19100 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19102 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19104 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19105 (*ngcsi_uxtw): New pattern.
19107 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19108 Julian Brown <julian@codesourcery.com>
19110 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19111 (TB_DREG): Add T_V4HF.
19112 (v4hf_UP): New macro.
19113 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19114 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19115 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19117 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19118 (arm_vector_mode_supported_p): Handle V4HF.
19119 (arm_mangle_map): Handle V4HFmode.
19120 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19121 * config/arm/arm_neon_builtins.def: Add entries for
19122 vcvtv4hfv4sf, vcvtv4sfv4hf.
19123 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19124 (neon_vcvtv4hfv4sf): Likewise.
19125 * config/arm/neon-gen.ml: Handle half-precision floating point
19127 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19128 * config/arm/arm_neon.h: Regenerate.
19129 * config/arm/neon.ml (type elts): Add F16.
19130 (type vectype): Add T_float16x4, T_floatHF.
19131 (type vecmode): Add V4HF.
19132 (type features): Add Requires_FP_bit feature.
19133 (elt_width): Handle F16.
19134 (elt_class): Likewise.
19135 (elt_of_class_width): Likewise.
19136 (mode_of_elt): Refactor.
19137 (type_for_elt): Handle F16, fix error messages.
19138 (vectype_size): Handle T_float16x4.
19139 (vcvt_sh): New function.
19140 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19141 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19142 (string_of_mode): Handle V4HF.
19143 * doc/arm-neon-intrinsics.texi: Regenerate.
19145 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19147 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19148 format specifier in 'X' case.
19150 2013-04-25 Alan Modra <amodra@gmail.com>
19153 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19154 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19155 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19156 Repeat for many other rotate/shift and mask patterns using subregs.
19157 Name lshiftrt insns.
19158 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19159 on WORDS_BIG_ENDIAN.
19161 2013-04-25 Alan Modra <amodra@gmail.com>
19163 * config.gcc: Support little-endian powerpc-linux targets.
19164 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19165 (LINK_OS_LINUX_SPEC): Define.
19166 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19167 Preserve MASK_LITTLE_ENDIAN.
19168 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19169 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19170 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19171 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19172 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19173 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19174 Correct fp word order for little-endian. Don't shift toc entries
19175 smaller than a word for little-endian.
19176 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19177 (bswapdi2 splits): Correct low-part subreg for little-endian.
19178 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19179 low/high where such is correct only for be.
19180 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19181 little-endian for -mcall-aixdesc.
19183 2013-04-25 Alan Modra <amodra@gmail.com>
19185 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19186 replace_equiv_address_nv.
19188 2013-04-25 Alan Modra <amodra@gmail.com>
19190 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19192 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19195 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19196 * rtl.h (struct rtx_def): ...
19198 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19200 PR rtl-optimizations/57046
19201 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19202 for multi-reg splits.
19204 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19206 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19208 2013-04-24 Sterling Augustine <saugustine@google.com>
19210 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19211 (comp_dir_string, debug_str_dwo_section): New.
19212 (DEBUG_STR_DWO_SECTION): Rename to ...
19213 (DEBUG_DWO_STR_SECTION): ... this.
19214 (DEBUG_NORM_STR_SECTION): Delete.
19215 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19216 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19217 (find_AT_string): Move most logic to ...
19218 (find_AT_string_in_table): ... here. New.
19219 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19220 add_skeleton_AT_string. Delete logic.
19221 (output_skeleton_debug_sections): Remove call to
19222 add_top_level_skeleton_die_attrs.
19223 (add_comp_dir_attribute): Move logic to comp_dir_string.
19224 (dwarf2out_init): Initialize debug_str_dwo_section.
19225 (output_indirect_string): Call find_string_form.
19226 (output_indirect_strings): Rewrite.
19227 (prune_unused_types): Empty skeleton_debug_str_hash.
19228 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19229 (dwarf2out_finish): Call output_indirect_strings.
19231 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19233 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19235 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19237 * rtl.h (struct rtx_def): Add comment for field jump.
19238 (LRA_SUBREG_P): New macro.
19239 * recog.c (register_operand): Check LRA_SUBREG_P.
19240 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19242 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19243 (lra_final_code_change): Skip subreg change for operators.
19244 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19245 if there are no operand changes.
19246 * lra-constraints.c (curr_insn_set): New.
19247 (match_reload): Set LRA_SUBREG_P.
19248 (emit_spill_move): Ditto.
19249 (check_and_process_move): Use curr_insn_set. Process only single
19250 set insns. Don't initialize sec_mem_p and change_p.
19251 (simplify_operand_subreg): Use LRA_SUBREG_P.
19252 (reg_in_class_p): New function.
19253 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19254 of #ifdef. Add code to remove cycling.
19255 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19256 non-null disp. Reload inner instead of disp when base and index
19257 are null. Try to put lo_sum into register.
19258 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19259 (check_and_process_move): Move code for move cost check to
19260 simple_move_p. Remove equiv_substitution.
19261 (simple_move_p): New function.
19262 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19263 curr_insn_set. Call check_and_process_move only for single set
19264 insns. Use the new function. Move call of check_and_process_move
19265 after operand equiv substitution and address process.
19267 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19269 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19270 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19271 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19273 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19275 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19277 2013-04-24 Marek Polacek <polacek@redhat.com>
19279 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19280 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19281 (select_loops_exit_conditions): Likewise.
19282 (number_of_iterations_for_all_loops): Likewise.
19283 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19284 (scev_analysis): Likewise.
19286 2013-04-02 Catherine Moore <clm@codesourcery.com>
19287 Chao-ying Fu <fu@mips.com>
19289 * config/mips/micromips.md (jraddiusp): New pattern.
19290 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19291 instruction if possible.
19293 2013-04-24 Alan Modra <amodra@gmail.com>
19295 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19297 2013-04-24 Julian Brown <julian@codesourcery.com>
19298 Chung-Lin Tang <cltang@codesourcery.com>
19300 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19301 dependency behavior in enumeration type DIE generation. Add TODO note
19302 to comments about future DW_FORM_sdata/udata re-work of related code.
19304 2013-04-23 Lawrence Crowl <crowl@google.com>
19306 * Makefile.in: Update as needed below.
19308 * hash-table.h (class hash_table):
19309 Correct many methods with parameter types compare_type to the correct
19310 value_type. (Correct code was unlikely to notice the change.)
19311 (hash_table::elements_with_deleted) New.
19312 (class hashtable::iterator): New.
19313 (hashtable::begin()): New.
19314 (hashtable::end()): New.
19315 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19317 * statistics.c (statistics_hashes):
19318 Change type to hash_table. Update dependent calls and types.
19320 * tree-into-ssa.c (var_infos):
19321 Change type to hash_table. Update dependent calls and types.
19323 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19324 Change type to hash_table. Update dependent calls and types.
19326 * tree-ssa-loop-im.c (struct mem_ref.refs):
19327 Change type to hash_table. Update dependent calls and types.
19329 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19330 Change type to hash_table. Update dependent calls and types.
19332 * tree-ssa-sccvn.c (vn_tables_s::nary):
19333 Change type to hash_table. Update dependent calls and types.
19334 (vn_tables_s::phis): Likewise.
19335 (vn_tables_s::references): Likewise.
19337 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19338 (vn_reference_eq): Update parameter and return types.
19340 * tree-ssa-structalias.c (pointer_equiv_class_table):
19341 Change type to hash_table. Update dependent calls and types.
19342 (location_equiv_class_table): Likewise.
19344 * tree-vect-data-refs.c: Consequential changes for making
19345 peeling a hash_table.
19347 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19348 (destroy_loop_vec_info): Dependent hash_table update.
19350 * tree-vectorizer.h (peeling_htab):
19351 Change type to hash_table. Update dependent calls and types.
19353 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19355 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19356 to check the register content is equal or not.
19357 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19358 to assign register content record.
19359 * lra-eliminations.c (update_reg_eliminate): Use
19360 lra_update_reg_val_offset to update register content offset.
19361 * lra-int.h (struct lra_reg): Add offset member.
19362 (lra_reg_val_equal_p): New static inline function.
19363 (lra_update_reg_val_offset): New static inline function.
19364 (lra_assign_reg_val): New static inline function.
19365 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19366 to assign register content record.
19367 (initialize_lra_reg_info_element): Initial offset to zero.
19369 2013-04-23 Catherine Moore <clm@codesourcery.com>
19371 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19372 operands. Record compression.
19374 2013-04-23 Xinliang David Li <davidxl@google.com>
19376 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19378 2013-04-23 Richard Biener <rguenther@suse.de>
19380 PR middle-end/57036
19381 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19382 parameter, only add abnormal goto edges from the copied body
19383 if the call could perform abnormal gotos.
19384 (copy_cfg_body): Adjust.
19386 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19388 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19390 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19392 * coretypes.h (gimple_stmt_iterator): Add struct to make
19395 2013-04-23 Richard Biener <rguenther@suse.de>
19397 PR tree-optimization/57026
19398 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19399 from SSA names occuring in abnormal PHI nodes.
19401 2013-04-22 Andi Kleen <ak@linux.intel.com>
19403 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19405 2013-04-22 Andi Kleen <ak@linux.intel.com>
19407 * lto/lto.c (print_lto_report_1): Declare early.
19408 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19410 2013-04-22 Andi Kleen <ak@linux.intel.com>
19412 * common.opt (-flto-report-wpa): Add.
19413 * doc/invoke.texi (-flto-report-wpa): Add.
19414 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19417 2013-04-22 Xinliang David Li <davidxl@google.com>
19419 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19420 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19421 * Makefile.in: New dependency
19423 David Daney <ddaney.cavm@gmail.com>
19425 * configure.ac (gcc_cv_as_micromips_support): Use the
19426 --fatal-warnings option.
19427 * configure: Regenerate.
19429 2013-04-22 Marek Polacek <polacek@redhat.com>
19432 * tsan.c (instrument_expr): Don't instrument expression
19433 in case its size is zero.
19435 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19439 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19441 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19443 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19445 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19446 (gimple_stmt_iterator): New typedef.
19447 * gimple.h (gimple_stmt_iterator): Rename to...
19448 (gimple_stmt_iterator_d): ... This.
19449 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19450 trees be valid for GIMPLE and GENERIC.
19451 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19452 * gimple-fold.c (gimple_fold_call): Call target hook
19453 gimple_fold_builtin.
19454 * hooks.c (hook_bool_gsiptr_false): New.
19455 * hooks.h (hook_bool_gsiptr_false): New.
19456 * target.def (fold_stmt): New.
19457 * doc/tm.texi: Regenerate.
19459 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19462 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19463 a set sp if no stack realignment.
19465 2013-04-22 Nick Clifton <nickc@redhat.com>
19467 * config.gcc (tilegx-linux): Extend extra_objs rather than
19469 (tilepro-linux): Likewise.
19471 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19473 * config/aarch64/aarch64-builtins.c
19475 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19476 (VAR<1-12>): Add MAP parameter.
19477 (BUILTIN_*): Likewise.
19478 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19479 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19480 (aarch64_ushl_n<mode>): Likewise.
19481 (aarch64_sshr_n<mode>): Likewise.
19482 (aarch64_ushr_n<mode>): Likewise.
19483 (aarch64_<maxmin><mode>): Likewise.
19484 (aarch64_sqrt<mode>): Likewise.
19485 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19486 (vshr<q>_n_*): Likewise.
19488 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19490 * config/aarch64/aarch64-builtins.c
19491 (aarch64_simd_builtin_type_mode): Handle SF types.
19493 (BUILTIN_GPF): Define.
19494 (aarch64_init_simd_builtins): Handle SF types.
19495 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19496 (frecps): Likewise.
19497 (frecpx): Likewise.
19498 * config/aarch64/aarch64-simd.md
19499 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19500 (aarch64_frecpe<mode>): New.
19501 (aarch64_frecps<mode>): Likewise.
19502 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19503 (v8type): Add frecp<esx>.
19504 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19505 (aarch64_frecps<mode>): Likewise.
19506 * config/aarch64/iterators.md (FRECP): New.
19507 (frecp_suffix): Likewise.
19508 * config/aarch64/arm_neon.h
19509 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19511 2013-04-22 Christian Bruel <christian.bruel@st.com>
19514 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19515 (REG_CLASS_NAMES): Idem.
19516 (REG_CLASS_CONTENTS): Idem.
19517 (REGCLASS_HAS_FP_REG): Idem.
19518 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19519 (sh_conditional_register_usage): Idem.
19521 2013-04-21 Jeff Law <law@redhat.com>
19523 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19524 (ssa_forward_propagate_and_combine): Use it.
19526 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19528 * lra.c: Update the flow chart diagram.
19530 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19532 PR rtl-optimization/56847
19533 * lra-constraints.c (process_alt_operands): Discourage alternative
19534 with non-matche doffsettable memory constraint fro memory with
19537 2013-04-19 Richard Biener <rguenther@suse.de>
19539 PR tree-optimization/56982
19540 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19542 * gimplify.c (gimplify_call_expr): Notice special calls.
19543 (gimplify_modify_expr): Likewise.
19544 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19545 abnormal control flow receivers.
19546 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19547 in the same way as cfun->has_nonlocal_labels.
19548 (gimple_purge_dead_abnormal_call_edges): Likewise.
19549 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19550 receivers start a basic-block.
19552 2013-04-19 Richard Biener <rguenther@suse.de>
19554 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19556 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19557 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19558 (SLP_TREE_LOAD_PERMUTATION): Add.
19559 (vect_transform_slp_perm_load): Adjust prototype.
19560 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19561 (vect_free_slp_instance): Likewise.
19562 (vect_create_new_slp_node): Likewise.
19563 (vect_supported_slp_permutation_p): Remove.
19564 (vect_slp_rearrange_stmts): Adjust.
19565 (vect_supported_load_permutation_p): Likewise. Inline
19566 vect_supported_slp_permutation_p here.
19567 (vect_analyze_slp_instance): Compute load permutations per
19568 slp node instead of per instance.
19569 (vect_get_slp_defs): Adjust.
19570 (vect_transform_slp_perm_load): Likewise.
19571 (vect_schedule_slp_instance): Remove redundant code.
19572 (vect_schedule_slp): Remove hack for PR56270, add it ...
19573 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19574 CSE loads for SLP. Adjust.
19576 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19578 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19579 spelling in two comments.
19581 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19584 * config/arm/arm.c (load_multiple_sequence): Require SP
19585 as base register for loads if SP is in the register list.
19587 2013-04-19 Martin Jambor <mjambor@suse.cz>
19589 PR tree-optimization/56718
19590 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19591 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19592 and made public. Adjusted all callers.
19593 (ipa_intraprocedural_devirtualization): New function.
19594 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19595 (ipa_intraprocedural_devirtualization): Likewise.
19596 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19598 2013-04-19 Richard Biener <rguenther@suse.de>
19600 PR tree-optimization/57000
19601 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19603 2013-04-19 Terry Guo <terry.guo@arm.com>
19605 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19607 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19608 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19609 (cortex_m4_fmacs): Use new reservations.
19610 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19612 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19614 PR rtl-optimization/56999
19615 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19617 (lra_coalesce): Remove split_origin_bitmap and related code.
19618 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19619 ranges if necessary.
19621 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19623 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19625 (ix86_expand_call): Remove clobbered_registers array and use
19626 x86_64_ms_sysv_extra_clobbered_registers instead.
19627 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19629 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19631 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19632 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19633 (*call_value_rex64_ms_sysv): Ditto.
19635 2013-04-18 Cary Coutant <ccoutant@google.com>
19637 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19638 parent instead of die_mark.
19640 2013-04-18 Diego Novillo <dnovillo@google.com>
19642 * gimple.c (create_gimple_tmp): New.
19643 (get_expr_type): New.
19644 (build_assign): New.
19645 (build_type_cast): New.
19646 * gimple.h (enum ssa_mode): Define.
19647 (gimple_seq_set_location): New.
19648 * asan.c (build_check_stmt): Change some gimple_build_* calls
19649 to use build_assign and build_type_cast.
19651 2013-04-18 Richard Biener <rguenther@suse.de>
19653 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19654 handle negative step. Remove redundant checks.
19655 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19656 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19657 for negative step and grouped loads fail to vectorize.
19659 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19661 * emit-rtl.c (reset_insn_used_flags): New function.
19662 (reset_all_used_flags): Use it.
19663 (verify_insn_sharing): New function.
19664 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19666 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19668 PR tree-optimization/56984
19669 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19670 and (x >> M) >= N don't register any assertion if N << M is the
19673 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19675 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19676 do not delete_insn insns that have not yet been emitted, only
19677 unlink them with remove_insn.
19678 * df-scan.c (df_insn_delete): Revert r197492.
19680 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19682 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19683 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19685 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19687 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19688 define_insn_and_split.
19689 (and_scc,ior_scc,negscc): Likewise.
19690 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19692 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19694 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19695 can benefit from using a sequence of LDRD instructions in epilogue
19696 instead of a single LDM instruction.
19698 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19701 * doc/extend.texi: Fix typo.
19703 2013-04-17 Richard Biener <rguenther@suse.de>
19705 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19706 (vect_build_slp_tree): ... here.
19707 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19708 match. Remove special-casing of mismatched complex loads.
19709 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19710 re-try the match with swapped commutative operands.
19711 (vect_supported_load_permutation_p): Remove special-casing of
19712 mismatched complex loads.
19713 (vect_analyze_slp_instance): Adjust.
19715 2013-04-17 Richard Biener <rguenther@suse.de>
19717 PR rtl-optimization/56921
19718 * cfgloop.h (struct loop): Add simple_loop_desc member.
19719 (struct niter_desc): Mark with GTY(()).
19720 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19721 * loop-iv.c (get_simple_loop_desc): Likewise.
19722 (free_simple_loop_desc): Likewise.
19725 2013-04-16 Richard Biener <rguenther@suse.de>
19727 PR rtl-optimization/56921
19728 * loop-init.c (pass_rtl_move_loop_invariants): Add
19729 TODO_do_not_ggc_collect to todo_flags_finish.
19730 (pass_rtl_unswitch): Same.
19731 (pass_rtl_unroll_and_peel_loops): Same.
19732 (pass_rtl_doloop): Same.
19734 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19736 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19737 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19738 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19740 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19741 * tree-streamer.c (record_common_node): Adjust reference in comment.
19743 2013-04-17 Terry Guo <terry.guo@arm.com>
19745 * config/arm/cortex-m4.md: Add a new bypass.
19747 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19749 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19751 (*subs_<optab><mode>_multp2): New pattern.
19752 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19753 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19755 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19757 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19758 (*subs_mul_imm_<mode>): New pattern.
19760 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19763 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19764 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19765 (vsx_movti_32bit): Change j->wa to O->wa.
19767 2013-04-16 Richard Biener <rguenther@suse.de>
19769 PR rtl-optimization/56921
19770 * loop-init.c (pass_rtl_move_loop_invariants): Add
19771 TODO_do_not_ggc_collect to todo_flags_finish.
19772 (pass_rtl_unswitch): Same.
19773 (pass_rtl_unroll_and_peel_loops): Same.
19774 (pass_rtl_doloop): Same.
19776 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19778 * config/arm/arm.c (emit_multi_reg_push): New declaration
19779 for an existing function.
19780 (arm_emit_strd_push): New function.
19781 (arm_expand_prologue): Used here.
19782 (arm_emit_ldrd_pop): New function.
19783 (arm_expand_epilogue): Used here.
19784 (arm_get_frame_offsets): Update condition.
19785 (arm_emit_multi_reg_pop): Add a special case for load of a single
19786 register with writeback.
19788 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19790 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19793 2013-04-16 Richard Biener <rguenther@suse.de>
19795 PR tree-optimization/56756
19796 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19797 (first_mem_ref_loc): New.
19798 (execute_sm): Place the load temporarily before a previous
19799 access instead of in the latch edge to ensure its SSA dependencies
19800 are defined at points dominating the load.
19802 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19804 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19805 correct fix by moving header and footer insn to the footer of
19806 the merged basic block. Clear BB_END of the merged-away block.
19808 PR middle-end/43631
19809 * emit-rtl.c (make_note_raw): New function.
19810 (link_insn_into_chain): New static inline function.
19811 (add_insn): Use it.
19812 (add_insn_before, add_insn_after): Factor insn chain linking code...
19813 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19814 using link_insn_into_chain.
19815 (note_outside_basic_block_p): New helper function for emit_note_after
19816 and emit_note_before.
19817 (emit_note_after): Use nobb variant of add_insn_after if the note
19818 should not be contained in a basic block.
19819 (emit_note_before): Use nobb variant of add_insn_before if the note
19820 should not be contained in a basic block.
19821 (emit_note_copy): Use make_note_raw.
19822 (emit_note): Likewise.
19823 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19824 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19825 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19826 the moved barrier the tail of the basic block it follows.
19827 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19829 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19831 PR tree-optimization/56962
19832 * gimple-ssa-strength-reduction.c (record_increment): Only set
19833 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19834 either rhs1 or rhs2 is equal to c->base_expr.
19836 2013-04-15 Richard Biener <rguenther@suse.de>
19838 PR tree-optimization/56933
19839 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19841 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19842 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19843 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19844 dependence check ...
19845 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19847 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19848 GROUP_READ_WRITE_DEPENDENCE.
19850 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19852 * emit-rtl.c (reset_all_used_flags): New function.
19853 (verify_rtl_sharing): Call reset_all_used_flags before and after
19854 performing the checks.
19856 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19858 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19859 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19860 * config/arm/constraints.md (De): New constraint.
19861 * config/arm/neon.md (anddi3_neon): Delete.
19862 (neon_vand<mode>): Expand to standard anddi3 pattern.
19863 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19864 Move earlier in the file.
19865 (neon_inv_logic_op2): Likewise.
19866 (arm_anddi_operand_neon): New predicate.
19868 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19870 * configure.ac (gcc_cv_ld_as_needed): Set
19871 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19872 Use -z ignore, -z record on *-*-solaris2*.
19873 (HAVE_LD_AS_NEEDED): Update comment.
19874 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19875 * configure: Regenerate.
19876 * config.in: Regenerate.
19877 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19878 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19879 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19880 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19881 equivalents. Fix markup.
19882 * doc/tm.texi: Regenerate.
19884 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19886 * config/i386/i386.opt: New option mstack-protector-guard=.
19887 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19888 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19889 TARGET_SSP_TLS_GUARD.
19890 * config/i386/i386.c (ix86_option_override_internal): Set
19891 ix86_stack_protector_guard.
19892 * config/i386/i386.md (stack_protect_set): Enable for
19893 TARGET_SSP_TLS_GUARD only.
19894 (stack_protect_set_<mode>): Ditto.
19895 (stack_protect_test): Ditto.
19896 (stack_protect_test_<mode>): Ditto.
19897 * doc/invoke.texi (i386 Option): Document.
19899 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19902 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19903 (S_MODES): Set H_MODE bit.
19904 (SF_MODES): Set only S_MODE and SF_MODE bits.
19905 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19906 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19907 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19908 <MODE_FLOAT>: Likewise.
19910 2013-04-15 Joey Ye <joey.ye@arm.com>
19912 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19914 2013-04-15 Joey Ye <joey.ye@arm.com>
19916 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19918 (thumb_far_jump_used_p): Count instruction size and set
19921 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19923 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19924 * resource.c (mark_target_live_regs): Reformat conditional block.
19926 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19928 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19929 notes, they are emitted only just before final.
19930 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19932 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19934 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19935 * cfgrtl.c (delete_insn): Call it here instead.
19936 * lra-spills.c (lra_final_code_change): Use delete_insn.
19937 * haifa-sched.c (sched_remove_insn): Likewise.
19938 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19939 returning to the nop pool.
19940 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19941 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19943 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19945 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19946 * doc/tm.texi: Regenerated.
19948 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19950 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19953 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19955 * df-core.c (df_find_def): Compare register numbers.
19956 (df_find_use): Likewise.
19958 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19961 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19962 lra_in_progress for return.
19964 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19966 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19967 define_insn into define_insn_and_split and emit movsicc patterns.
19969 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19971 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19973 2013-04-12 Richard Biener <rguenther@suse.de>
19975 * tree-pass.h (TODO_do_not_ggc_collect): New.
19976 * passes.c (execute_one_ipa_transform_pass): Honor
19977 TODO_do_not_ggc_collect.
19978 (execute_one_pass): Likewise.
19981 2013-04-10 Richard Biener <rguenther@suse.de>
19983 * passes.c (init_optimization_passes): Remove reload pass.
19984 * ira.c (do_reload): Merge into ...
19986 (rest_of_handle_reload): Remove.
19987 (pass_reload): Likewise.
19988 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19989 of reload for vzeroupper pass placement.
19991 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19993 PR tree-optimization/56918
19994 PR tree-optimization/56920
19995 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19996 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19997 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19998 use wide_mul_with_sign method.
20000 2013-04-12 Richard Biener <rguenther@suse.de>
20002 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20003 not be considered a gimple constant.
20005 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20007 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20008 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20011 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20012 Jakub Jelinek <jakub@redhat.com>
20014 * opts.c: Include diagnostic-color.h.
20015 (common_handle_option): Handle OPT_fdiagnostics_color_.
20016 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20017 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20018 (diagnostic-color.o): New.
20019 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20020 (diagnostic_color_rule): New enum.
20021 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20022 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20023 the location string.
20024 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20025 either NULL, or color kind.
20026 * diagnostic-color.c: New file.
20027 * diagnostic-color.h: New file.
20028 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20030 * doc/invoke.texi (-fdiagnostics-color): Document.
20031 * pretty-print.h (pp_show_color): Define.
20032 (struct pretty_print_info): Add show_color field.
20033 * diagnostic.c: Include diagnostic-color.h.
20034 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20035 macros. Colorize error:, warning: etc. strings and also the location
20037 (diagnostic_show_locus): Colorize the caret line.
20038 * pretty-print.c: Include diagnostic-color.h.
20039 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20040 inside of %< %> quotes or quoted through q format modifier.
20042 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20044 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20046 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20048 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20049 code in CC_NZ mode.
20050 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20053 2013-04-11 Marek Polacek <polacek@redhat.com>
20055 PR tree-optimization/48184
20056 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20058 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20060 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20061 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20062 (skip_simple_arithmetic): Tidy up.
20063 * tree.h (skip_simple_constant_arithmetic): Declare.
20065 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20067 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20069 2013-04-11 Richard Biener <rguenther@suse.de>
20071 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20072 generate vector constants.
20074 2013-04-11 Richard Biener <rguenther@suse.de>
20076 PR tree-optimization/56878
20077 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20078 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20080 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20081 Prefer to align the DR with the most invariant base address.
20083 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20085 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20088 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20090 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20091 floating-point vector comparisons against 0.
20093 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20095 PR tree-optimization/56899
20096 * fold-const.c (extract_muldiv_1): Apply distributive law
20097 only if TYPE_OVERFLOW_WRAPS (ctype).
20099 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20102 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20103 loading has side effect.
20105 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20107 * config/sparc/sparc.c: Include tree-pass.h.
20108 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20109 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20110 head of file. Change return type. Split off gate function.
20111 (sparc_gate_work_around_errata): New function.
20112 (pass_work_around_errata): New pass definition.
20113 (insert_pass_work_around_errata) New pass insert definition to
20114 insert pass_work_around_errata just after delayed-branch scheduling.
20115 (sparc_option_override): Insert the pass.
20116 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20118 2013-04-10 David S. Miller <davem@davemloft.net>
20120 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20121 or -mcpu=hypersparc.
20123 * target.def (cstore_mode): New hook.
20124 * target.h: Include insn-codes.h
20125 * targhooks.c: Likewise.
20126 (default_cstore_mode): New function.
20127 * targhooks.h: Declare it.
20128 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20129 * doc/tm.texi: Rebuild.
20130 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20131 target hook, rather than inspecting the insn_data.
20132 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20133 (TARGET_CSTORE_MODE): Redefine.
20134 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20136 * config/sparc/predicates.md (cstore_result_operand): New special
20138 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20139 Use it for operand 0.
20140 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20141 (*snesi_special): Likewise.
20142 (*snesi_zero): Likewise.
20143 (*seqsi_zero): Likewise.
20144 (*sltu_insn): Likewise.
20145 (*sgeu_insn): Likewise.
20146 (*seqdi_special): Make operand 0 and comparison operation be of
20148 (*snedi_special): Likewise.
20149 (*snedi_special_vis3): Likewise.
20150 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20151 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20152 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20153 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20154 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20155 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20156 (*sltu_extend_sp64): Likewise.
20157 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20158 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20159 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20160 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20161 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20163 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20165 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20166 (aarch64_start_file): Use the new function.
20168 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20169 Jason Merrill <jason@redhat.com>
20171 * common.opt: Add -gdwarf.
20172 * opts.c (common_handle_option): Handle it.
20173 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20175 2013-04-10 Richard Biener <rguenther@suse.de>
20177 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20178 (execute_one_ipa_transform_pass): But unconditionally here.
20179 (execute_one_pass): And here.
20180 (init_optimization_passes): Remove reload pass.
20181 * tree-pass.h (TODO_ggc_collect): Remove.
20182 (pass_reload): Likewise.
20183 * ira.c (do_reload): Merge into ...
20185 (rest_of_handle_reload): Remove.
20186 (pass_reload): Likewise.
20187 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20188 of reload for vzeroupper pass placement.
20189 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20190 and todo_flags_finish of all passes.
20192 2013-04-10 Richard Biener <rguenther@suse.de>
20194 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20195 first_const_oprnd field, rename first_def_type to first_op_type.
20196 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20197 (vect_get_and_check_slp_defs): Always use the type of the
20198 operand. Allow mixed vect_external_def, vect_constant_def types.
20199 (vect_get_constant_vectors): Handle mixed vect_external_def,
20200 vect_constant_def types.
20202 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20204 PR tree-optimization/55524
20205 * tree-ssa-math-opts.c
20206 (convert_mult_to_fma): Don't use an fms construct
20207 when we don't have an fms operation, but fnma, and it looks
20208 likely that we'll be able to use the latter.
20210 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20212 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20214 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20215 inline fail caused by overwritable functions.
20217 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20219 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20220 unnecessary bits in the constant power of two case.
20222 2013-04-10 Richard Biener <rguenther@suse.de>
20224 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20225 broken code swapping operands.
20226 (vect_build_slp_tree): Do not compute load permutations here.
20227 (vect_analyze_slp_instance): Compute load permutations here,
20228 after building the SLP tree.
20230 2013-04-09 Christian Bruel <christian.bruel@st.com>
20232 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20233 of next/prev_real_insn.
20235 2013-04-09 Jan Hubicka <jh@suse.cz>
20237 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20238 Drop aliased parameter.
20239 (function_and_variable_visibility): Do not handle alias pairs.
20240 * cgraph.c (varpool_externally_visible_p): Update prototype.
20241 * varpool.c (varpool_add_new_variable): Update.
20243 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20245 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20247 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20249 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20251 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20253 2013-04-09 Marek Polacek <polacek@redhat.com>
20255 PR tree-optimization/48762
20256 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20258 2013-04-09 Richard Biener <rguenther@suse.de>
20260 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20262 (vect_build_slp_tree): Likewise.
20263 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20264 calculating the cost of a SLP instance.
20265 (vect_analyze_slp_instance): Use it from here, after building
20268 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20270 PR middle-end/56883
20271 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20272 expand_omp_for_static_chunk): Use simple_p = true in
20273 force_gimple_operand_gsi calls when assigning to addressable decls.
20275 2013-04-09 Jeff Law <law@redhat.com>
20277 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20278 when the boolean was created by converting a wider object which
20279 had a boolean range.
20281 2013-04-09 Richard Biener <rguenther@suse.de>
20283 * tree-vectorizer.h (slp_void_p): Remove.
20284 (slp_tree): Typedef before _slp_tree declaration.
20285 (struct _slp_tree): Use a vector of slp_tree as children.
20286 (vect_get_place_in_interleaving_chain): Remove.
20287 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20289 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20291 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20292 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20293 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20294 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20295 Use slp_node instead of slp_void_p and adjust.
20297 2013-04-09 Richard Biener <rguenther@suse.de>
20299 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20300 work that is not necessary.
20302 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20304 PR tree-optimization/56854
20305 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20306 forward into clobber stmts if it would change MEM_REF lhs into
20309 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20311 * tree.c (type_hash_lookup, type_hash_add): Make static.
20312 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20314 2013-04-09 Richard Biener <rguenther@suse.de>
20316 * tree.h (unsave_expr_now): Remove.
20317 * tree-inline.c (mark_local_for_remap_r): Remove.
20318 (unsave_expr_1): Likewise.
20319 (unsave_r): Likewise.
20320 (unsave_expr_now): Likewise.
20321 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20322 (propagate_tree_value): Likewise.
20324 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20326 * doc/rtl.texi (sequence): Rewrite documentation to match the
20327 current use of SEQUENCE rtl objects.
20328 * rtl.def (SEQUENCE): Likewise.
20330 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20331 Update documentation.
20332 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20333 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20335 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20337 2013-04-08 Teresa Johnson <tejohnson@google.com>
20339 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20340 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20342 (estimate_edge_size_and_time): Add comment to suggest using rounding
20344 (estimate_node_size_and_time): Ditto.
20345 (remap_edge_change_prob): Use helper rounding divide methods.
20346 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20347 (gimple_mod_pow2_value_transform): Ditto.
20348 (gimple_mod_subtract_transform): Ditto.
20349 (gimple_ic_transform): Ditto.
20350 (gimple_stringops_transform): Ditto.
20351 * stmt.c (conditional_probability): Ditto.
20352 (emit_case_dispatch_table): Ditto.
20353 * lto-cgraph.c (merge_profile_summaries): Ditto.
20354 * tree-optimize.c (execute_fixup_cfg): Ditto.
20355 * cfgcleanup.c (try_forward_edges): Ditto.
20356 * cfgloopmanip.c (scale_loop_profile): Ditto.
20358 (duplicate_loop_to_header_edge): Ditto.
20359 (lv_adjust_loop_entry_edge): Ditto.
20360 * tree-vect-loop.c (vect_transform_loop): Ditto.
20361 * profile.c (compute_branch_probabilities): Ditto.
20362 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20363 * lto-streamer-in.c (input_cfg): Ditto.
20364 * gimple-streamer-in.c (input_bb): Ditto.
20365 * ipa-cp.c (update_profiling_info): Ditto.
20366 (update_specialized_profile): Ditto.
20367 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20368 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20370 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20371 (compute_trg_info): Ditto.
20372 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20373 (purge_dead_edges): Ditto.
20374 * loop-unswitch.c (unswitch_loop): Ditto.
20375 * cgraphclones.c (cgraph_clone_edge): Ditto.
20376 (cgraph_clone_node): Ditto.
20377 * tree-inline.c (copy_bb): Ditto.
20378 (copy_edges_for_bb): Ditto.
20379 (initialize_cfun): Ditto.
20380 (copy_cfg_body): Ditto.
20381 (expand_call_inline): Ditto.
20383 2013-04-08 Kai Tietz <ktietz@redhat.com>
20385 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20386 TARGET_CYGWIN64 by TARGET_64BIT.
20388 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20390 * config/epiphany/epiphany.md (GPR_1): New constant.
20391 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20392 * config/epiphany/epiphany.c (gen_compare_reg):
20393 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20394 is already in place.
20395 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20396 Don't require being called during rtl expansion; If y operlaps r0,
20398 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20399 (epiphany_expand_epilogue): Likewise.
20401 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20402 Don't use CC_FPmode for ORDERED / UNORDERED.
20403 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20405 * config/epiphany/constraints.md (CnL): New constraint.
20406 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20407 * config/epiphany/predicates.md (add_operand): Allow 1024.
20409 * config/epiphany/epiphany.md (logical_op): New code iterator.
20410 (op_mnc): New code attribute.
20411 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20412 (mov_f+1, mov_f+2): New peephole2 patterns.
20414 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20415 (cstoresi4): Also allow re-use of zero result when doing a NE
20416 comparison to a non-zero operand.
20417 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20419 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20420 Use gen_addsi3_i / gen_subsi3_i.
20422 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20426 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20427 contain anything but clobbers, at most one __builtin_stack_restore,
20428 optionally debug stmts and final resx, and if it has at least one
20429 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20430 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20431 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20432 which isn't defaut definition, remove them.
20433 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20434 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20435 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20436 with MEM_REF LHS with SSA_NAME address.
20438 2013-04-08 Jeff Law <law@redhat.com>
20440 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20442 2013-04-08 Richard Biener <rguenther@suse.de>
20444 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20446 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20447 determined vector type.
20448 (vect_analyze_data_refs): Likewise.
20449 (vect_get_new_vect_var): Adjust.
20450 (vect_create_destination_var): Preserve SSA name versions.
20451 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20452 not dump anything here.
20454 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20456 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20457 Add member lr_slot_known.
20458 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20460 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20461 Remove code that sets lr_slot_offset according to what a previous
20462 version of epiphany_emit_save_restore used to do.
20463 (epiphany_emit_save_restore): When doing an lr save or restore,
20464 set/verify lr_slot_known and lr_slot_offset.
20466 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20469 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20472 2013-04-08 Richard Biener <rguenther@suse.de>
20474 * alias.c (find_base_term): Fix thinko in previous change.
20476 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20478 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20479 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20480 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20481 if possible to compute val.
20482 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20483 For QImode integers don't require anything about precision. Use
20484 const_with_all_bytes_same to find out if the constant doesn't have
20485 repeated bytes in it.
20487 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20489 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20492 2013-04-08 Marek Polacek <polacek@redhat.com>
20494 PR rtl-optimization/48182
20495 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20498 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20501 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20502 nuses, make sure we have a label.
20504 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20507 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20508 (rs6000_emit_swdiv_low_precision): Remove.
20509 (rs6000_emit_swdiv): Rewrite to handle between one and four
20510 iterations of Newton-Raphson generally; modify required number of
20511 iterations for some cases.
20512 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20514 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20516 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20517 set-but-unused variable.
20519 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20520 basic blocks of released function bodies garbage-collectable.
20522 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20523 (struct rtl_opt_pass): Add TODO_df_finish.
20525 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20527 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20529 * config/arm/constraints.md (q): New constraint.
20530 * config/arm/ldrdstrd.md: New file.
20531 * config/arm/arm.md (ldrdstrd.md) New include.
20532 (arm_movdi): Use "q" instead of "r" constraint
20533 for double-word memory access.
20534 (movdf_soft_insn): Likewise.
20535 * config/arm/vfp.md (movdi_vfp): Likewise.
20536 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20537 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20538 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20539 (mem_ok_for_ldrd_strd): Likewise.
20540 (output_move_double): Update assertion.
20542 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20544 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20546 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20548 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20549 define_insn_and_split.
20550 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20552 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20554 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20555 define_insn_and_split.
20556 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20557 (shiftsi3_compare): New pattern.
20558 (rrx): New pattern.
20559 * config/arm/unspecs.md (UNSPEC_RRX): New.
20561 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20563 * config/arm/arm.md (negdi_extendsidi): New pattern.
20564 (negdi_zero_extendsidi): Likewise.
20566 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20568 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20569 define_insn_and_split.
20570 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20571 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20573 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20575 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20576 define_insn_and_split.
20577 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20578 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20582 * config/arm/arm.md (subsi3_carryin): New pattern.
20583 (subsi3_carryin_const): Likewise.
20584 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20585 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20587 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20589 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20591 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20593 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20594 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20596 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20598 * config/arm/arm.c (arm_expand_builtin): Change fcode
20599 type to unsigned int.
20601 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20603 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20605 2013-04-04 Ian Lance Taylor <iant@google.com>
20607 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20610 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20612 PR middle-end/56729
20613 * df-scan.c (df_insn_delete): Disable failing assert.
20615 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20617 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20618 New function prototype.
20619 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20620 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20621 (arm_builtin_vectorized_function): New function.
20623 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20625 * config/arm/arm_neon_builtins.def: New file.
20626 * config/arm/arm.c (neon_builtin_data): Move contents to
20627 arm_neon_builtins.def.
20628 (enum arm_builtins): Include neon builtin definitions.
20629 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20630 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20632 2013-04-04 Marek Polacek <polacek@redhat.com>
20634 PR tree-optimization/48186
20635 * predict.c (maybe_hot_frequency_p): Return false if
20636 HOT_BB_FREQUENCY_FRACTION is 0.
20637 (cgraph_maybe_hot_edge_p): Likewise.
20639 2013-04-04 Richard Biener <rguenther@suse.de>
20641 PR tree-optimization/56826
20642 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20645 2013-04-04 Richard Biener <rguenther@suse.de>
20647 PR tree-optimization/56213
20648 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20649 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20651 2013-04-04 Richard Biener <rguenther@suse.de>
20653 PR tree-optimization/56837
20654 * tree-loop-distribution.c (classify_partition): For non-zero
20655 values require that the value has the same precision as its
20656 mode to be useful as memset value.
20658 2013-04-03 Nick Clifton <nickc@redhat.com>
20660 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20661 (fmssf4): Use fmsf.s on E3V5 architectures.
20662 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20663 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20665 2013-04-03 Jeff Law <law@redhat.com>
20667 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20668 (lra-eliminations.o): Likewise.
20670 2013-04-03 Teresa Johnson <tejohnson@google.com>
20672 * gcov-io.c (compute_working_sets): Moved most of body of old
20673 compute_working_sets here from profile.c.
20674 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20675 (gcov_working_set_t): Moved typedef here from basic-block.h
20676 (compute_working_set): Declare.
20677 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20678 (get_working_sets): Renamed from compute_working_set,
20679 replace most of body with call to new compute_working_sets.
20680 (get_exec_counts): Replace call to compute_working_sets
20681 to get_working_sets.
20682 * profile.h (get_working_sets): Renamed from compute_working_set.
20683 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20684 to get_working_sets.
20685 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20686 * gcov-dump.c (dump_working_sets): New function.
20688 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20690 * hwint.c (sext_hwi, zext_hwi): New functions.
20691 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20692 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20693 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20694 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20695 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20696 (sext_hwi, zext_hwi): New functions.
20698 2013-04-03 Jeff Law <law@redhat.com>
20700 PR tree-optimization/56799
20701 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20702 back test for widening conversion erroneously dropped in prior change.
20704 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20707 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20708 instead of next_real_insn.
20710 2013-04-03 Marek Polacek <polacek@redhat.com>
20713 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20715 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20718 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20720 (thumb1_output_casesi): Likewise.
20721 (thumb2_output_casesi): Likewise.
20723 2013-04-03 Richard Biener <rguenther@suse.de>
20725 PR tree-optimization/56817
20726 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20728 (tree_unroll_loops_completely_1): ... new function to manually
20729 walk the loop tree, properly defering outer loops of unrolled
20730 loops to later iterations.
20732 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20734 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20735 (vectorizable_load): Likewise.
20736 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20737 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20739 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20741 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20744 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20746 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20748 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20750 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20752 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20754 PR tree-optimization/56790
20755 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20758 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20760 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20762 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20763 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20766 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20769 * tree.h (force_folding_builtin_constant_p): New decl.
20770 * builtins.c (force_folding_builtin_constant_p): New variable.
20771 (fold_builtin_constant_p): Fold immediately also if
20772 force_folding_builtin_constant_p.
20774 2013-04-03 Richard Biener <rguenther@suse.de>
20776 PR tree-optimization/56812
20777 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20778 DRs of the same interleaving chain are independent.
20780 2013-04-02 Jason Merrill <jason@redhat.com>
20782 * gdbinit.in (pbb): Use debug fn.
20784 2013-04-02 Lawrence Crowl <crowl@google.com>
20786 * sese.h (struct ivtype_map_elt_s): Remove unused.
20787 (extern debug_ivtype_map): Remove unused.
20788 (extern eq_ivtype_map_elts): Remove unused.
20789 * sese.c (debug_ivtype_map): Removed unused.
20790 (debug_ivtype_map_1): Removed unused.
20791 (debug_ivtype_elt): Remove unused.
20792 (eq_ivtype_map_elts): Remove unused.
20795 2013-04-02 Kai Tietz <ktietz@redhat.com>
20798 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20799 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20800 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20802 (legitimize_pe_coff_symbol): Likewise.
20803 (is_imported_p): New helper-function.
20804 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20805 for Windows x64 targets.
20806 (ix86_expand_prologue): Optimize for pe-coff targets.
20807 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20808 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20809 medium/large code-model.
20810 (legitimize_pic_address): Likewise.
20811 (legitimize_tls_address): Likewise.
20812 (ix86_expand_call): Likewise.
20813 (x86_output_mi_thunk): Likewise.
20814 (get_dllimport_decl): Add new beimport argument.
20815 (construct_plt_address): Don't assert for x64 pe-coff targets.
20816 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20818 (SYMBOL_FLAG_STUBVAR): New macro.
20819 (SYMBOL_REF_STUBVAR_P): Likewise.
20820 * config/i386/winnt.c (stub_list): New structure.
20821 (stub_head): New local variable.
20822 (i386_pe_record_stub): New function.
20823 (i386_pe_file_end): Emit refptr-stubs.
20825 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20827 PR rtl-optimization/56745
20828 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20829 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20832 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20833 and both of them are MEM_REFs, just compare first argument for
20834 equality and attempt to deal even with differing offsets.
20837 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20838 of gimple_clobber_p to be MEM_REF.
20839 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20840 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20841 after gimplification.
20842 * asan.c (get_mem_ref_of_assignment): Don't instrument
20843 gimple_clobber_p stmts.
20844 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20845 gimple_clobber_p stmt if they have MEM_REF lhs and
20846 are dead because of another gimple_clobber_p stmt.
20847 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20848 gimple_clobber_p stmts like debug stmts.
20849 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20850 that refer to unused VAR_DECLs or uninitialized values.
20851 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20852 gimple_clobber_p stmts if they refer to removed parameters.
20853 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20856 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20858 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20859 using SWI48 mode attribute.
20861 2013-04-02 Wei Mi <wmi@google.com>
20863 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20864 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20865 *<rotate_insn><mode>3_mask in i386.md.
20867 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20869 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20871 2013-04-02 Richard Biener <rguenther@suse.de>
20873 PR tree-optimization/56778
20874 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20875 Runtime alias tests are not supported for gather loads.
20876 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20877 stmts referenced from SSA operands before updating SSA form.
20879 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20880 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20882 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20883 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20884 * config/arm/cortex-a53.md: New file.
20885 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20886 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20887 * config/arm/arm.c (arm_issue_rate): Likewise.
20888 * config/arm/arm-tune.md: Regenerate
20889 * config/arm/arm-tables.opt: Regenerate.
20890 * config/arm/arm-cores.def: Add cortex-a53.
20892 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20894 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20897 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20899 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20900 scalar load/store operations using B/H registers.
20901 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20903 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20905 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20907 * config/aarch64/aarch64.c
20908 (aarch64_simd_scalar_immediate_valid_for_move): New.
20909 * config/aarch64/aarch64-protos.h
20910 (aarch64_simd_scalar_immediate_valid_for_move): New.
20911 * config/aarch64/constraints.md (Dh, Dq): New.
20912 * config/aarch64/iterators.md (hq): New.
20914 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20916 * reorg.c (get_branch_condition): Deal with conditional returns.
20917 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20919 2013-04-01 Wei Mi <wmi@google.com>
20921 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20922 Truncate operand 2 using %b asm operand modifier.
20923 (*<shift_insn><mode>3_mask): Ditto.
20924 (*<rotate_insn><mode>3_mask): Ditto.
20926 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20928 PR middle-end/56798
20929 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20931 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20933 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20935 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20937 2013-03-30 Lawrence Crowl <crowl@google.com>
20939 * dse.c (clear_alias_sets): Remove never set.
20940 (disqualified_clear_alias_sets): Remove never set.
20941 (clear_alias_mode_pool): Remove never set.
20942 (dse_step0): Remove condition that is never true.
20943 (canon_address): Remove condition that is never true.
20944 (dse_step7): Remove condition that is never true.
20945 (rest_of_handle_dse): Remove condition that is never true.
20946 (rest_of_handle_dse::did_global): Remove never read from above.
20947 (dse_step2_spill): Remove never called from above.
20948 (dse_step5_spill): Remove never called from above.
20950 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20952 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20953 JUMP_TABLE_DATA changes.
20954 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20955 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20956 (Insns) <jump_table_data>: New entry.
20957 * doc/tm.texi: Regenerate.
20959 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20961 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20962 for table jump at the end of a basic block using tablejump_p.
20963 * targhooks.c (default_invalid_within_doloop): Likewise.
20964 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20965 target hook implementation that is identical to the default hook.
20966 (rs6000_invalid_within_doloop): Remove.
20968 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20969 unused variable from tablejump_p call.
20971 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20972 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20973 (INSN_DELETED_P): Likewise.
20974 (emit_jump_table_data): New prototype.
20975 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20976 after 4th as unused.
20977 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20978 * sched-vis.c (print_insn): Likewise.
20979 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20980 insn for compatibility with back ends that use next_active_insn to
20981 identify jump table data.
20982 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20983 (remove_insn): Likewise.
20984 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20986 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20987 (emit_jump_table_data): New function.
20989 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20990 basic block, a JUMP_TABLE_DATA never is.
20991 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20992 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20993 off from code handling real insns.
20994 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20995 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20996 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20997 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20998 is not a NONDEBUG_INSN_P.
20999 * ira-costs.c (scan_one_insn): Likewise.
21000 * jump.c (mark_all_labels): Likewise.
21001 (mark_jump_label_1): Likewise.
21002 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21003 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21005 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21006 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21007 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21008 JUMP_TABLE_DATA_P insns.
21009 (calculate_elim_costs_all_insns): Likewise.
21010 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21011 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21012 (delete_output_reload): Code style fixups.
21013 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21014 insn flags on this non-insn.
21015 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21016 as scheduling barriers, for pre-change compatibility.
21017 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21018 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21020 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21021 redundant JUMP_TABLE_DATA_P test.
21022 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21023 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21024 (frv_for_each_packet): Likewise.
21025 * config/i386/i386.c (min_insn_size): Likewise.
21026 (ix86_avoid_jump_mispredicts): Likewise.
21027 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21028 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21029 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21030 (mips16_insn_length): Robustify.
21031 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21032 (mips16_split_long_branches): Likewise.
21033 * config/pa/pa.c (pa_combine_instructions): Likewise.
21034 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21035 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21036 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21037 as contributing to pool range lengths.
21038 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21039 Remove redundant JUMP_TABLE_DATA_P test.
21040 (sh_loop_align): Likewise.
21041 (split_branches): Likewise.
21042 (sh_insn_length_adjustment): Likewise.
21043 * config/spu/spu.c (get_branch_target): Likewise.
21045 2013-03-29 Jan Hubicka <jh@suse.cz>
21047 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21048 gcov streaming; stream hot bb threshold to ltrans.
21049 * predict.c (get_hot_bb_threshold): Break out from ....
21050 (maybe_hot_count_p): ... here.
21051 (set_hot_bb_threshold): New function.
21052 * lto-section-in.c (lto_section_name): Add profile.
21053 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21054 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21055 and data-streamer.h
21056 (histogram_entry): New structure.
21057 (histogram, histogram_pool): New global vars.
21058 (histogram_hash): New structure.
21059 (histogram_hash::hash): New method.
21060 (histogram_hash::equal): Likewise.
21061 (account_time_size): New function.
21062 (cmp_counts): New function.
21063 (dump_histogram): New function.
21064 (ipa_profile_generate_summary): New function.
21065 (ipa_profile_write_summary): New function.
21066 (ipa_profile_read_summary): New function.
21067 (ipa_profile): Decide on threshold.
21068 (pass_ipa_profile): Add ipa_profile_write_summary and
21069 ipa_profile_read_summary.
21070 * Makefile.in (ipa.o): Update dependencies.
21071 * lto-streamer.h (LTO_section_ipa_profile): New section.
21073 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21075 * tree.h (VAR_P): New.
21077 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21080 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21082 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21084 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21085 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21086 (control_flow_insn_p): Likewise.
21087 * cfgrtl.c (duplicate_insn_chain): Likewise.
21088 * final.c (get_attr_length_1): Likewise.
21089 (shorten_branches): Likewise.
21090 (final_scan_insn): Likewise.
21091 * function.c (instantiate_virtual_regs): Likewise.
21092 * gcse.c (insert_insn_end_basic_block): Likewise.
21093 * ira-costs.c (scan_one_insn): Likewise.
21094 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21095 * lra.c (check_rtl): Likewise.
21096 * reload1.c (elimination_costs_in_insn): Likewise.
21097 * reorg.c (follow_jumps): Likewise.
21099 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21100 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21101 (thumb_far_jump_used_p): Likewise.
21102 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21103 (workaround_speculation): Likewise.
21104 (add_sched_insns_for_speculation): Likewise.
21105 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21106 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21107 (frv_for_each_packet): Likewise.
21108 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21109 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21110 (final_emit_insn_group_barriers): Likewise.
21111 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21112 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21113 (mips16_insn_length): Likewise.
21114 * config/pa/pa.c (pa_reorg): Likewise.
21115 (pa_combine_instructions): Likewise.
21116 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21117 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21118 (sh_reorg): Likewise.
21119 (split_branches): Likewise.
21120 * config/spu/spu.c (get_branch_target): Likewise.
21122 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21125 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21127 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21128 Fix declaration name.
21130 2013-03-28 Lawrence Crowl <crowl@google.com>
21132 * graphds.h (struct graph.indicies): Remove unused.
21133 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21134 (SCOP_ORIGINAL_PDDRS): Remove unused.
21135 * sese.h (extern insert_loop_close_phis): Removed unused.
21136 (extern insert_guard_phis): Removed unused.
21137 (extern ivtype_map_elt_info): Removed unused.
21138 (new_ivtype_map_elt): Removed unused.
21139 * sese.c (ivtype_map_elt_info): Removed unused.
21141 2013-03-28 Lawrence Crowl <crowl@google.com>
21143 * Makefile.in: Add several missing include dependences.
21145 (test-dump.o): New. This object is not added to any executable,
21146 but is present for ad-hoc testing.
21148 (debug (const bitmap_head_def &)): New.
21149 (debug (const bitmap_head_def *)): New.
21151 (extern debug (const bitmap_head_def &)): New.
21152 (extern debug (const bitmap_head_def *)): New.
21154 (debug (edge_def &)): New.
21155 (debug (edge_def *)): New.
21157 (debug (basic_block_def &)): New.
21158 (debug (basic_block_def *)): New.
21160 (dump_node (const_tree, int, FILE *)): Correct source file.
21162 (debug (die_struct &)): New.
21163 (debug (die_struct *)): New.
21165 (extern debug (die_struct &)): New.
21166 (extern debug (die_struct *)): New.
21167 * gimple-pretty-print.c
21168 (debug (gimple_statement_d &)): New.
21169 (debug (gimple_statement_d *)): New.
21170 * gimple-pretty-print.h
21171 (extern debug (gimple_statement_d &)): New.
21172 (extern debug (gimple_statement_d *)): New.
21174 (debug (ira_allocno_copy &)): New.
21175 (debug (ira_allocno_copy *)): New.
21176 (debug (ira_allocno &)): New.
21177 (debug (ira_allocno *)): New.
21179 (extern debug (ira_allocno_copy &)): New.
21180 (extern debug (ira_allocno_copy *)): New.
21181 (extern debug (ira_allocno &)): New.
21182 (extern debug (ira_allocno *)): New.
21184 (debug (live_range &)): New.
21185 (debug (live_range *)): New.
21187 (debug (lra_live_range &)): New.
21188 (debug (lra_live_range *)): New.
21190 (debug (lra_live_range &)): New.
21191 (debug (lra_live_range *)): New.
21193 (debug (omega_pb_d &)): New.
21194 (debug (omega_pb_d *)): New.
21196 (extern debug (omega_pb_d &)): New.
21197 (extern debug (omega_pb_d *)): New.
21199 (debug (const rtx_def &)): New.
21200 (debug (const rtx_def *)): New.
21202 (debug_tree (tree): Move within file.
21203 (debug_raw (const tree_node &)): New.
21204 (debug_raw (const tree_node *)): New.
21205 (dump_tree_via_hooks (const tree_node *, int)): New.
21206 (debug (const tree_node &)): New.
21207 (debug (const tree_node *)): New.
21208 (debug_verbose (const tree_node &)): New.
21209 (debug_verbose (const tree_node *)): New.
21210 (debug_head (const tree_node &)): New.
21211 (debug_head (const tree_node *)): New.
21212 (debug_body (const tree_node &)): New.
21213 (debug_body (const tree_node *)): New.
21214 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21215 (debug (vec<tree, va_gc> &)): New.
21216 (debug (vec<tree, va_gc> *)): New.
21218 (extern debug (const rtx_def &)): New.
21219 (extern debug (const rtx_def *)): New.
21221 (debug_raw (simple_bitmap_def &)): New.
21222 (debug_raw (simple_bitmap_def *)): New.
21223 (debug (simple_bitmap_def &)): New.
21224 (debug (simple_bitmap_def *)): New.
21226 (extern debug (simple_bitmap_def &)): New.
21227 (extern debug (simple_bitmap_def *)): New.
21228 (extern debug_raw (simple_bitmap_def &)): New.
21229 (extern debug_raw (simple_bitmap_def *)): New.
21231 (debug (vinsn_def &)): New.
21232 (debug (vinsn_def *)): New.
21233 (debug_verbose (vinsn_def &)): New.
21234 (debug_verbose (vinsn_def *)): New.
21235 (debug (expr_def &)): New.
21236 (debug (expr_def *)): New.
21237 (debug_verbose (expr_def &)): New.
21238 (debug_verbose (expr_def *)): New.
21239 (debug (vec<rtx> &)): New.
21240 (debug (vec<rtx> *)): New.
21242 (extern debug (vinsn_def &)): New.
21243 (extern debug (vinsn_def *)): New.
21244 (extern debug_verbose (vinsn_def &)): New.
21245 (extern debug_verbose (vinsn_def *)): New.
21246 (extern debug (expr_def &)): New.
21247 (extern debug (expr_def *)): New.
21248 (extern debug_verbose (expr_def &)): New.
21249 (extern debug_verbose (expr_def *)): New.
21250 (extern debug (vec<rtx> &)): New.
21251 (extern debug (vec<rtx> *)): New.
21253 (_list_iter_cond_expr): Make inline instead of static.
21255 (debug (sreal &)): New.
21256 (debug (sreal *)): New.
21258 (extern debug (sreal &)): New.
21259 (extern debug (sreal *)): New.
21261 (extern debug_raw (const tree_node &)): New.
21262 (extern debug_raw (const tree_node *)): New.
21263 (extern debug (const tree_node &)): New.
21264 (extern debug (const tree_node *)): New.
21265 (extern debug_verbose (const tree_node &)): New.
21266 (extern debug_verbose (const tree_node *)): New.
21267 (extern debug_head (const tree_node &)): New.
21268 (extern debug_head (const tree_node *)): New.
21269 (extern debug_body (const tree_node &)): New.
21270 (extern debug_body (const tree_node *)): New.
21271 (extern debug (vec<tree, va_gc> &)): New.
21272 (extern debug (vec<tree, va_gc> *)): New.
21274 (debug (struct loop &)): New.
21275 (debug (struct loop *)): New.
21276 (debug_verbose (struct loop &)): New.
21277 (debug_verbose (struct loop *)): New.
21278 * tree-dump.c: Add header dependence.
21280 (extern debug (struct loop &)): New.
21281 (extern debug (struct loop *)): New.
21282 (extern debug_verbose (struct loop &)): New.
21283 (extern debug_verbose (struct loop *)): New.
21285 (debug (data_reference &)): New.
21286 (debug (data_reference *)): New.
21287 (debug (vec<data_reference_p> &)): New.
21288 (debug (vec<data_reference_p> *)): New.
21289 (debug (vec<ddr_p> &)): New.
21290 (debug (vec<ddr_p> *)): New.
21292 (extern debug (data_reference &)): New.
21293 (extern debug (data_reference *)): New.
21294 (extern debug (vec<data_reference_p> &)): New.
21295 (extern debug (vec<data_reference_p> *)): New.
21296 (extern debug (vec<ddr_p> &)): New.
21297 (extern debug (vec<ddr_p> *)): New.
21299 (debug (pt_solution &)): New.
21300 (debug (pt_solution *)): New.
21302 (extern debug (pt_solution &)): New.
21303 (extern debug (pt_solution *)): New.
21305 (debug (_var_map &)): New.
21306 (debug (_var_map *)): New.
21307 (debug (tree_live_info_d &)): New.
21308 (debug (tree_live_info_d *)): New.
21310 (extern debug (_var_map &)): New.
21311 (extern debug (_var_map *)): New.
21312 (extern debug (tree_live_info_d &)): New.
21313 (extern debug (tree_live_info_d *)): New.
21315 2013-03-28 Jan Hubicka <jh@suse.cz>
21317 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21319 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21321 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21322 record only when desired or required.
21324 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21326 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21327 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21329 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21331 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21332 (*andsi3_compare0_uxtw): New pattern.
21333 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21334 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21336 2013-03-28 Jan Hubicka <jh@suse.cz>
21338 * data-streamer-in.c (streamer_read_gcov_count): New function.
21339 * gimple-streamer-out.c: Include value-prof.h.
21340 (output_gimple_stmt): Output histogram.
21341 (output_bb): Use streamer_write_gcov_count.
21342 * value-prof.c: Include data-streamer.h
21343 (dump_histogram_value): Add HIST_TYPE_MAX.
21344 (stream_out_histogram_value): New function.
21345 (stream_in_histogram_value): New function.
21346 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21347 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21348 * data-streamer-out.c (streamer_write_gcov_count): New function.
21349 (streamer_write_gcov_count_stream): New function.
21350 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21351 (lto_output_node): Likewise.
21352 (input_node, input_edge): Likewise.
21353 * lto-streamer-out.c (output_cfg): Update streaming.
21354 * lto-streamer-in.c (input_cfg): Likewise.
21355 * data-streamer.h (streamer_write_gcov_count,
21356 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21357 * gimple-streamer-in.c: Include value-prof.h
21358 (input_gimple_stmt): Input histograms.
21359 (input_bb): Update profile streaming.
21361 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21363 * genmodes.c (emit_max_int): New function.
21364 (emit_insn_modes_h): Added call to emit_max_function.
21365 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21367 * machmode.def: Fixed comment.
21369 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21371 * combine.c (try_combine): Removed useless assert.
21372 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21374 2013-03-28 Marek Polacek <polacek@redhat.com>
21375 Richard Biener <rguenther@suse.de>
21377 PR tree-optimization/56695
21378 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21379 build signed result of a vector comparison.
21380 * tree-cfg.c (verify_gimple_comparison): Check that a result
21381 of a vector comparison has signed type.
21383 2013-03-28 Richard Biener <rguenther@suse.de>
21385 PR tree-optimization/37021
21386 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21387 do not restrict gaps between groups.
21388 * tree-vect-stmts.c (vectorizable_load): Properly account for
21389 a gap between groups.
21391 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21393 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21394 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21397 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21399 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21400 * doc/extend.texi (Named Address Spaces): Ditto.
21401 (Variable Attributes): Ditto.
21403 2013-03-27 Kai Tietz <ktietz@redhat.com>
21405 * config.build: Add support for cygwin x64 target.
21406 * config.gcc: Likewise.
21407 * config.host: Likewise.
21408 * configure.ac: Likewise
21409 * configure: Regenerated.
21411 2013-03-27 Kai Tietz <ktietz@redhat.com>
21413 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21414 * config/i386/t-cygwin-w64: New file.
21415 * config/i386/cygwin-w64.h: New file.
21416 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21417 and add support for x64-cygwin target.
21418 (CPP_SPEC): Likewise.
21419 (CXX_WRAP_SPEC_LIST): Undefine before define.
21420 (LIBGCJ_SONAME): Use 15 as version.
21422 2013-03-27 Richard Biener <rguenther@suse.de>
21424 PR tree-optimization/56716
21425 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21426 dumping for ref nodes.
21428 2013-03-27 Martin Jambor <mjambor@suse.cz>
21430 PR tree-optimization/55334
21431 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21432 restricted pointers to arrays.
21434 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21436 * Makefile.in (.SUFFIXES): Add .cc.
21437 (.c.o): Apply same recipe for implicit rule .cc.o.
21439 2013-03-27 Richard Biener <rguenther@suse.de>
21441 PR tree-optimization/37021
21442 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21443 REALPART/IMAGPART_EXPRs around the supported refs.
21444 * tree-ssa-structalias.c (find_func_aliases): Assume that
21445 floating-point values are not used to transfer pointers.
21447 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21449 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21450 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21451 * doc/tm.texi: Regenerate.
21452 * targhooks.h (default_has_ifunc_p): New.
21453 * targhooks.c (default_has_ifunc_p): Ditto.
21454 * config/linux-protos.h: New file.
21455 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21456 hook for linux which disables support of indirect functions in android.
21457 * config/linux-android.c: New file.
21458 * config/t-linux-android.c: Ditto.
21459 * config.gcc: Added new object file linux-android.o.
21460 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21461 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21462 * varasm.c (do_assemble_alias): Likewise.
21463 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21464 doesn't support indirect functions.
21465 * configure: Regenerate.
21467 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21470 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21471 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21474 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21476 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21478 2013-03-27 Terry Guo <terry.guo@arm.com>
21480 * config/arm/arm-cores.def: Added core cortex-r7.
21481 * config/arm/arm-tune.md: Regenerated.
21482 * config/arm/arm-tables.opt: Regenerated.
21483 * doc/invoke.texi: Added entry for core cortex-r7.
21485 2013-03-27 Walter Lee <walt@tilera.com>
21487 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21488 double-decrement of next_scratch_regno.
21490 2013-03-27 Walter Lee <walt@tilera.com>
21492 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21494 (insn_v1mulus): Ditto.
21495 (insn_v2muls): Ditto.
21497 2013-03-27 Walter Lee <walt@tilera.com>
21499 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21500 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21502 2013-03-27 Walter Lee <walt@tilera.com>
21504 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21505 (*sibcall_value): Ditto.
21507 2013-03-27 Walter Lee <walt@tilera.com>
21509 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21510 (insn_mnz_v8qi): ... this ...
21511 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21513 (insn_v<n>mnz): Replaced by ...
21514 (insn_v1mnz): ... this ...
21515 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21517 (insn_mz_<mode>): Replaced by ...
21518 (insn_mz_v8qi): ... this ...
21519 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21521 (insn_v<n>mz): Replaced by ...
21522 (insn_v1mz): ... this ...
21523 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21526 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21528 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21530 2013-03-26 Roland McGrath <mcgrathr@google.com>
21532 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21533 than fprintf with a non-constant, non-format string.
21535 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21537 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21538 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21539 operand 0 predicate.
21540 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21541 attribute. Use general_x64nomem_operand as operand 1 predicate.
21542 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21543 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21544 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21545 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21546 (mov<mode>_insv_1): Remove expander. Merge insn with
21547 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21548 Use general_x64nomem_operand as operand 1 predicate.
21549 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21550 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21551 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21552 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21553 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21554 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21555 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21556 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21557 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21558 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21559 (general_x64nomem_operand): Ditto.
21561 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21563 * config/rtems.opt: Add -pthread option.
21565 2013-03-26 Richard Biener <rguenther@suse.de>
21567 * alias.c (find_base_term): Avoid redundant and not used recursion.
21568 (base_alias_check): Get the initial base term from the caller.
21569 (true_dependence_1): Compute and pass base terms to base_alias_check.
21570 (write_dependence_p): Likewise.
21571 (may_alias_p): Likewise.
21573 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21575 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21576 PC-relative load in SI modes and above only.
21578 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21580 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21581 * config/arm/iwmmxt.md (WCGR0): Update.
21582 (WCGR1, WCGR2, WCGR3): Likewise.
21584 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21586 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21587 Use x64 and nox64 isa attributes.
21589 2013-03-26 Richard Biener <rguenther@suse.de>
21591 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21592 alignment computations and rely on get_object_alignment_1
21593 for the !TYPE_P case.
21594 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21596 2013-03-26 Walter Lee <walt@tilera.com>
21598 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21599 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21601 2013-03-25 Jeff Law <law@redhat.com>
21603 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21604 check for INTEGRAL_TYPE_P that was missing due to checking in
21605 wrong version of prior patch.
21607 2013-03-25 Walter Lee <walt@tilera.com>
21609 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21610 TILEGX_INSN_SHUFFLEBYTES1.
21611 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21613 (tilegx_builtins): Ditto.
21614 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21616 2013-03-25 Walter Lee <walt@tilera.com>
21618 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21619 (floatunssisf2): New pattern.
21620 (floatsidf2): New pattern.
21621 (floatunssidf2): New pattern.
21623 2013-03-25 Walter Lee <walt@tilera.com>
21625 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21626 tests for constraint J, K, N, P.
21628 2013-03-25 Walter Lee <walt@tilera.com>
21630 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21631 Use indirect/pcrel encoding.
21632 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21635 2013-03-25 Steve Ellcey <sellcey@mips.com>
21637 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21638 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21639 * config/mips/mips.c (mips_option_override): Set IMADD default.
21640 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21641 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21642 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21643 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21644 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21646 2013-03-25 Jeff Law <law@redhat.com>
21648 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21649 slightly to avoid creating and folding useless trees. Simplify
21650 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21652 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21654 * config/i386/i386.md (*zero_extendsidi2): Merge with
21655 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21656 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21657 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21658 targets. Clarify comment.
21660 2013-03-25 Martin Jambor <mjambor@suse.cz>
21662 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21663 pass-through jump functions differently.
21664 (ipa_read_jump_function): Likewise. Also use setter functions to set
21667 2013-03-25 Martin Jambor <mjambor@suse.cz>
21669 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21670 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21672 (ipa_get_indirect_edge_target): New function.
21673 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21674 ipa_get_indirect_edge_target. Update all callers.
21675 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21676 ipa_get_indirect_edge_target_1 instead of calling
21677 ipa_get_indirect_edge_target.
21678 (create_specialized_node): Pass aggvlas to
21679 ipcp_discover_new_direct_edges.
21681 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21683 * config/arm/arm.md (f_sels, f_seld): New types.
21684 (*cmov<mode>): New pattern.
21685 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21688 2013-03-25 Kai Tietz <ktietz@redhat.com>
21690 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21691 POSIX-printf for mingw-hosted builds.
21693 2013-03-25 Richard Biener <rguenther@suse.de>
21695 PR middle-end/56694
21696 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21697 must-not-throw stmt location.
21699 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21701 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21702 Emit load-acquire versions when acq is true.
21703 (arm_emit_store_exclusive): Add rel parameter.
21704 Emit store-release versions when rel is true.
21705 (arm_split_compare_and_swap): Use acquire-release instructions
21707 of barriers when appropriate.
21708 (arm_split_atomic_op): Likewise.
21709 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21710 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21711 (VUNSPEC_SLX): Likewise.
21712 (VUNSPEC_LDA): Likewise.
21713 (VUNSPEC_STL): Likewise.
21714 * config/arm/sync.md (atomic_load<mode>): New pattern.
21715 (atomic_store<mode>): Likewise.
21716 (arm_load_acquire_exclusive<mode>): Likewise.
21717 (arm_load_acquire_exclusivesi): Likewise.
21718 (arm_load_acquire_exclusivedi): Likewise.
21719 (arm_store_release_exclusive<mode>): Likewise.
21721 2013-03-25 Catherine Moore <clm@codesourcery.com>
21723 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21724 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21725 * config/mip/predicates.md (lwsp_swsp_operand,
21726 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21727 sb16_operand, db4_operand, db7_operand, ib3_operand,
21728 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21729 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21730 andi16_operand): New predicates.
21731 * config/mips/mips.md (compression): New attribute.
21732 (enabled): New attribute.
21733 (length): Consider compression in computing length.
21734 (shift_compression): New code attribute.
21735 (*add<mode>3): New operands. Record compression.
21736 (sub<mode>3): Likewise.
21737 (one_cmpl<mode>2): Likewise.
21738 (*and<mode>3): Likewise.
21739 (*ior<mode>3): Likewise.
21740 (unnamed pattern for xor): Likewise.
21741 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21742 (*<optab><mode>3): Likewise.
21743 (*mov<mode>_internal: Likewise.
21744 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21745 (mips_unsigned_immediate_p): New.
21746 (umips_lwsp_swsp_address_p): New.
21747 (m16_based_address_p): New.
21748 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21749 (mips_unsigned_immediate_p): New prototype.
21750 (lwsp_swsp_address_p): New prototype.
21751 (m16_based_address_p): New prototype.
21752 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21753 (mips_signed_immediate_p): New function.
21754 (m16_based_address_p): New function.
21755 (lwsp_swsp_address_p): New function.
21756 (mips_print_operand_punctuation): Recognize short delay slot insns
21757 for microMIPS.add<mode>3"
21759 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21762 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21763 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21765 2013-03-25 Richard Biener <rguenther@suse.de>
21767 PR tree-optimization/56689
21768 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21771 2013-03-25 Richard Biener <rguenther@suse.de>
21773 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21775 (memory_references): Likewise.
21776 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21777 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21778 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21779 (gather_mem_refs_in_loops): Fold into ...
21780 (analyze_memory_references): ... this. Move initialization
21781 to tree_ssa_lim_initialize.
21782 (fill_always_executed_in): Rename to ...
21783 (fill_always_executed_in_1): ... this.
21784 (fill_always_executed_in): Move contains_call computation to
21785 this new function from ...
21786 (tree_ssa_lim_initialize): ... here.
21787 (tree_ssa_lim): Call fill_always_executed_in.
21789 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21791 * postreload.c (reload_combine): Fix code detecting returns.
21793 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21795 * function.c (emit_use_return_register_into_block): On cc0 targets,
21796 do not emit the sequence between cc0 setter and user.
21798 2013-03-25 Kai Tietz <ktietz@redhat.com>
21800 * config/i386/predicates.md (local_symbolic_operand): Interpret
21801 dll-imported symbols as none-local.
21803 2013-03-25 Richard Biener <rguenther@suse.de>
21805 * tree-ssa-loop-im.c (struct depend): Remove.
21806 (struct lim_aux_data): Make depends a vec of gimples.
21807 (free_lim_aux_data): Adjust.
21808 (add_dependency): Likewise.
21809 (set_level): Likewise.
21811 2013-03-25 Richard Biener <rguenther@suse.de>
21813 PR middle-end/56434
21814 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21815 the pointer returned by calls with ECF_MALLOC set.
21817 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21819 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21821 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21823 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21824 using MMXMODE mode iterator.
21825 (*move<mode>_internal): Merge with *movv2sf_internal and
21826 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21828 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21830 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21831 (record_last_mem_set_info): Likewise.
21833 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21834 of XNEWVEC followed by memset.
21835 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21837 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21839 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21840 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21841 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21842 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21843 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21844 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21845 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21846 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21847 BARRIER_P instead of GET_CODE.
21849 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21851 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21852 inaccuracy in the probing code.
21854 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21855 (ctrapdi4): Likewise.
21857 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21859 * calls.c (expand_call): Add missing guard to code handling return
21860 of non-BLKmode structures in MSB.
21861 * function.c (expand_function_end): Likewise.
21863 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21865 * combine.c (try_combine): Adjust comment. Do not add the set of
21866 insn #0 if the destination indirectly is set or dies in insn #2.
21867 Tidy up code to distribute a new note.
21869 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21871 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21872 also for alternatives 16 and 17.
21874 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21876 * config/i386/sse.md (*mov<mode>_internal): Merge with
21877 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21878 Emit insn template depending on type attribute. Use
21879 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21880 movd instead of movq mnemonic for interunit moves. Rewrite mode
21881 attribute calculation. Remove unit attribute calculation.
21882 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21883 Set prefix_data16 attribute for DImode ssemov types.
21884 Use Ym instead of y for SSE-MMX conversion alternatives.
21885 Reorder operand constraints.
21887 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21889 * df.h (df_insn_delete): Adjust prototype.
21890 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21891 and let it decide whether mark the basic block dirty.
21892 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21893 * df-scan.c (df_insn_info_delete): New helper function, split
21894 off from df_insn_delete.
21895 (df_scan_free_bb_info): Use it.
21896 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21898 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21899 that the insn is actually an insn and it has a non-NULL basic block.
21900 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21902 2013-03-22 Richard Biener <rguenther@suse.de>
21904 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21906 (mem_ref_alloc): Do not allocate them.
21907 (refs_independent_p): Do not query or maintain a cache.
21909 2013-03-22 Richard Biener <rguenther@suse.de>
21911 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21912 (gather_mem_refs_in_loops): Do not compute it.
21913 (analyze_memory_references): Do not allocate it.
21914 (tree_ssa_lim_finalize): Do not free it.
21915 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21917 2013-03-22 Richard Biener <rguenther@suse.de>
21919 * is-a.h (as_a): Use gcc_checking_assert.
21921 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21923 * config/aarch64/aarch64.c (aarch64_print_operand): New
21924 format specifier for printing a constant in hex.
21925 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21926 format specifier for printing second operand.
21928 2013-03-22 Richard Biener <rguenther@suse.de>
21930 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21932 (gather_mem_refs_in_loops): Perform store accumulation here.
21933 (create_vop_ref_mapping_loop): Remove.
21934 (create_vop_ref_mapping): Likewise.
21935 (analyze_memory_references): Initialize refs_stored_in_loop.
21936 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21937 (record_indep_loop): Remove.
21938 (record_dep_loop): New function.
21939 (ref_indep_loop_p_1): Adjust to only walk over references
21940 in the loop, not its subloops.
21941 (ref_indep_loop_p): Rename to ...
21942 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21943 maintaining a more fine-grained cache.
21944 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21945 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21947 2013-03-22 Richard Biener <rguenther@suse.de>
21949 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21950 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21951 aggregate mem_ref_loc.
21952 (free_mem_ref_locs): Inline into ...
21953 (memref_free): ... this and adjust.
21954 (mem_ref_alloc): Adjust.
21955 (mem_ref_locs_alloc): Remove.
21956 (record_mem_ref_loc): Adjust.
21957 (get_all_locs_in_loop): Rewrite into ...
21958 (for_all_locs_in_loop): ... this iterator.
21959 (rewrite_mem_ref_loc): New functor.
21960 (rewrite_mem_refs): Use for_all_locs_in_loop.
21961 (sm_set_flag_if_changed): New functor.
21962 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21963 (ref_always_accessed): New functor.
21964 (ref_always_accessed_p): Use for_all_locs_in_loop.
21966 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21968 * tree-pass.h (PROP_gimple_lvec): New.
21969 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21970 (init_optimization_passes): Move pass_lower_vector.
21971 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21973 (pass_lower_vector): Provide PROP_gimple_lvec.
21974 (pass_lower_vector_ssa): Likewise.
21975 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21977 2013-03-21 Mark Wielaard <mjw@redhat.com>
21979 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21981 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21983 * config/i386/i386.md (*movdi_internal): Disparage slightly
21984 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21985 conversion alternatives.
21987 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21989 PR middle-end/48087
21990 * diagnostic.def (DK_WERROR): New kind.
21991 * diagnostic.h (werrorcount): Define.
21992 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21993 promoted to DK_ERROR, increment DK_WERROR counter instead of
21995 * toplev.c (toplev_main): Call print_ignored_options even if
21996 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21997 even if just werrorcount is non-zero.
22000 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22002 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22003 (string_cst_pool_decl): New function.
22004 (optimize_one_addr_into_implicit_ptr): New function.
22005 (resolve_addr_in_expr): Optimize DWARF location expression
22006 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22007 which doesn't live in memory, but has DW_AT_location or
22008 DW_AT_const_value, or refers to a string literal, into
22009 DW_OP_GNU_implicit_pointer.
22010 (optimize_location_into_implicit_ptr): New function.
22011 (resolve_addr): If removing DW_AT_location of a variable because
22012 it was DW_OP_addr of address of the variable, but the variable doesn't
22013 live in memory, try to emit const value attribute for the initializer.
22015 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22017 * tree.h (VECTOR_TYPE_P): New macro.
22018 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22019 TYPE_MODE): Use it.
22020 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22021 VEC_COND_EXPR cannot be lvalues.
22022 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22024 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22026 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22027 Restrict the transformation to equal modes.
22029 2013-03-21 Richard Biener <rguenther@suse.de>
22031 PR tree-optimization/39326
22032 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22033 (MEM_ANALYZABLE): Adjust.
22034 (record_mem_ref_loc): Move bitmap ops ...
22035 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22036 unanalyzable refs, do not record locations for it.
22037 (analyze_memory_references): Allocate ref zero as shared
22039 (refs_independent_p): Do not test for unanalyzed mems here.
22040 (ref_indep_loop_p_1): Special-case disambiguation against
22041 the unanalyzed ref.
22042 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22044 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22046 * config/arm/arm-protos.h (tune_params): Add
22047 prefer_neon_for_64bits field.
22048 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22049 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22050 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22051 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22052 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22053 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22054 (arm_option_override): Handle -mneon-for-64bits new option.
22055 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22056 (prefer_neon_for_64bits): Declare new variable.
22057 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22058 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22059 (arch_enabled): Handle new arch types. Remove support for onlya8
22061 (one_cmpldi2): Use new arch names.
22062 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22063 * config/arm/arm.opt (mneon-for-64bits): Add option.
22064 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22065 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22066 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22068 * doc/invoke.texi (-mneon-for-64bits): Document.
22070 2013-03-21 Richard Biener <rguenther@suse.de>
22072 PR tree-optimization/39326
22073 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22074 (sort_bbs_in_loop_postorder_cmp): New function.
22075 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22077 2013-03-21 Richard Biener <rguenther@suse.de>
22079 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22080 (vect_insert_into_interleaving_chain): Likewise.
22081 (vect_drs_dependent_in_basic_block): Inline ...
22082 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22084 (vect_analyze_data_ref_dependence): ... here. Simplify.
22085 (vect_check_interleaving): Simplify.
22086 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22087 (vect_slp_analyze_data_ref_dependences): ... this new function.
22088 (dr_group_sort_cmp): New function.
22089 (vect_analyze_data_ref_accesses): Compute data-reference groups
22090 here instead of in vect_analyze_data_ref_dependence. Use
22091 a more efficient algorithm.
22092 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22093 vect_slp_analyze_data_ref_dependences. Call
22094 vect_analyze_data_ref_accesses earlier.
22095 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22096 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22097 (vect_slp_analyze_data_ref_dependences): New prototype.
22099 2013-03-21 Richard Biener <rguenther@suse.de>
22101 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22102 ref is stored in the loop.
22103 (find_refs_for_sm): Walk only over all stores.
22104 (store_motion_loop): Allocate from lim_bitmap_obstack.
22105 (store_motion): Likewise.
22107 2013-03-21 Richard Biener <rguenther@suse.de>
22109 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22110 Update virtual SSA form.
22112 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22114 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22115 * configure: Regenerate.
22116 * config.in: Regenerate.
22117 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22118 if !HAVE_LD_EH_FRAME_CIEV3.
22120 2013-03-21 Richard Biener <rguenther@suse.de>
22122 * tree-cfg.c (verify_expr_no_block): New function.
22123 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22124 nor DECL_VALUE_EXPR have locations with associated blocks.
22125 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22126 (clear_unused_block_pointer): Remove code dealing with
22127 blocks in DECL_DEBUG_EXPR locations.
22129 2013-03-21 Richard Biener <rguenther@suse.de>
22131 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22132 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22133 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22134 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22135 instead of DECL_DEBUG_EXPR_IS_FROM.
22136 * gimplify.c (gimplify_modify_expr): Likewise.
22137 * tree-cfg.c (verify_expr_location_1): Likewise.
22138 * tree-complex.c (create_one_component_var): Likewise.
22139 * tree-sra.c (create_access_replacement): Likewise.
22140 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22141 (clear_unused_block_pointer): Likewise.
22142 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22143 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22144 * var-tracking.c (var_debug_decl): Likewise.
22145 (track_expr_p): Likewise.
22146 * tree-inline.c (add_local_variables): Likewise. Set
22147 DECL_HAS_DEBUG_EXPR_P after copying it.
22148 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22149 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22151 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22154 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22155 * configure: Regenerate.
22156 * config.in: Regenerate.
22157 * config/i386/i386.md (*movdf_internal): Use
22158 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22159 movd instead of movq mnemonic for interunit moves.
22160 (*movdi_internal): Ditto.
22162 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22164 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22165 (abd<mode>_3): New pattern.
22166 (aba<mode>_3): New pattern.
22167 (fabd<mode>_3): New pattern.
22169 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22171 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22172 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22173 occurrence of REGISTER_PREFIX as its empty string.
22175 2013-03-20 Jeff Law <law@redhat.com>
22177 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22178 addititional equivalences for equality comparisons between an SSA_NAME
22179 and a constant where the SSA_NAME was set from a widening conversion.
22181 2013-03-20 Walter Lee <walt@tilera.com>
22183 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22185 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22187 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22188 depending on type attribute.
22189 (*movti_internal): Ditto.
22190 (*movtf_internal): Ditto.
22191 (*movxf_internal): Ditto.
22192 (*movdf_internal): Ditto.
22193 (*movsf_internal): Ditto.
22195 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22197 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22198 maybe_vex for sselog1 and ssemov types.
22199 (*movdi_internal): Reorder operand constraints.
22200 (*movsi_internal): Ditto. Set prefix attribute to
22201 maybe_vex for sselog1 and ssemov types.
22202 (*movtf_internal): Set prefix attribute to maybe_vex
22203 for sselog1 and ssemov types.
22204 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22205 DImode ssemov types. Reorder operand constraints.
22206 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22207 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22208 attribute for SImode ssemov types. Reorder operand constraints.
22210 2013-03-20 Martin Jambor <mjambor@suse.cz>
22212 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22213 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22215 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22217 * config/rs6000/predicates.md (indexed_address, update_address_mem
22218 update_indexed_address_mem): New predicates.
22219 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22220 attribute for load/store instructions.
22221 * config/rs6000/dfp.md (movsd_store): Likewise.
22222 (movsd_load): Likewise.
22223 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22224 (unnamed HI->DI extend define_insn): Likewise.
22225 (unnamed SI->DI extend define_insn): Likewise.
22226 (unnamed QI->SI extend define_insn): Likewise.
22227 (unnamed QI->HI extend define_insn): Likewise.
22228 (unnamed HI->SI extend define_insn): Likewise.
22229 (unnamed HI->SI extend define_insn): Likewise.
22230 (extendsfdf2_fpr): Likewise.
22231 (movsi_internal1): Likewise.
22232 (movsi_internal1_single): Likewise.
22233 (movhi_internal): Likewise.
22234 (movqi_internal): Likewise.
22235 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22236 attribute for load/store instructions.
22237 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22239 (mov<mode>_softfloat): Likewise.
22240 (mov<mode>_hardfloat32): Likewise.
22241 (mov<mode>_hardfloat64): Likewise.
22242 (mov<mode>_softfloat64): Likewise.
22243 (movdi_internal32): Likewise.
22244 (movdi_internal64): Likewise.
22245 (probe_stack_<mode>): Likewise.
22247 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22249 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22250 floating point, and decimal floating point to reload iterator.
22252 * config/rs6000/constraints.md (wl constraint): New constraints to
22253 return FLOAT_REGS if certain options are used to reduce the number
22254 of separate patterns that exist in the file.
22255 (wx constraint): Likewise.
22256 (wz constraint): Likewise.
22258 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22259 -mdebug=reg, print wg, wl, wx, and wz constraints.
22260 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22261 Initialize the reload functions for 64-bit binary/decimal floating
22263 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22264 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22265 create the buffer on the stack to overcome not having a 32-bit
22267 (rs6000_emit_move): Likewise.
22268 (rs6000_secondary_memory_needed_rtx): Likewise.
22269 (rs6000_alloc_sdmode_stack_slot): Likewise.
22270 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22271 via xxlxor, just like DFmode 0.0.
22273 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22274 define as 1 if we are running on a power7 or newer.
22275 (enum r6000_reg_class_enum): Add new constraints.
22277 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22278 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22279 with other moves by using conditional constraits (wg). Use LFIWZX
22280 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22281 (movsd splitter): Likewise.
22282 (movsd_hardfloat): Likewise.
22283 (movsd_softfloat): Likewise.
22285 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22286 binary and decimal floating point moves.
22287 (fmove_ok): New attributes to combine binary and decimal floating
22288 point moves, and to combine power6x (mfpgpr) moves along normal
22290 (real_value_to_target): Likewise.
22291 (f32_lr): Likewise.
22292 (f32_lm): Likewise.
22293 (f32_li): Likewise.
22294 (f32_sr): Likewise.
22295 (f32_sm): Likewise.
22296 (f32_si): Likewise.
22297 (movsf): Combine binary and decimal floating point moves. Combine
22298 power6x (mfpgpr) moves with other moves by using conditional
22299 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22300 (mov<mode> for SFmode/SDmode); Likewise.
22301 (SFmode/SDmode splitters): Likewise.
22302 (movsf_hardfloat): Likewise.
22303 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22304 (movsf_softfloat): Likewise.
22305 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22307 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22308 wx and wz constraints.
22310 * config/rs6000/constraints.md (wg constraint): New constraint to
22311 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22313 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22316 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22317 -mdebug=reg, print wg, wl, wx, and wz constraints.
22318 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22319 Initialize the reload functions for 64-bit binary/decimal floating
22321 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22322 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22323 create the buffer on the stack to overcome not having a 32-bit
22325 (rs6000_emit_move): Likewise.
22326 (rs6000_secondary_memory_needed_rtx): Likewise.
22327 (rs6000_alloc_sdmode_stack_slot): Likewise.
22328 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22329 via xxlxor, just like DFmode 0.0.
22331 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22332 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22333 with other moves by using conditional constraits (wg). Use LFIWZX
22334 and STFIWX for loading SDmode on power7.
22335 (movdd splitters): Likewise.
22336 (movdd_hardfloat32): Likewise.
22337 (movdd_softfloat32): Likewise.
22338 (movdd_hardfloat64_mfpgpr): Likewise.
22339 (movdd_hardfloat64): Likewise.
22340 (movdd_softfloat64): Likewise.
22342 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22343 64-bit binary and decimal floating point moves.
22344 (FMOVE64X): Likewise.
22345 (movdf): Combine 64-bit binary and decimal floating point moves.
22346 Combine power6x (mfpgpr) moves with other moves by using
22347 conditional constraits (wg).
22348 (mov<mode> for DFmode/DDmode): Likewise.
22349 (DFmode/DDmode splitters): Likewise.
22350 (movdf_hardfloat32): Likewise.
22351 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22352 (movdf_softfloat32): Likewise.
22353 (movdf_hardfloat64_mfpgpr): Likewise.
22354 (movdf_hardfloat64): Likewise.
22355 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22356 (movdf_softfloat64): Likewise.
22357 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22358 (reload_<mode>_load): Move to later in the file so they aren't in
22359 the middle of the floating point move insns.
22360 (reload_<mode>_store): Likewise.
22362 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22365 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22366 constraint if -mdebug=reg.
22367 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22368 Enable using dd reload support if needed.
22370 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22371 binary and decimal floating point moves in rs6000.md.
22372 (movtd_internal): Likewise.
22374 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22375 decimal floating point moves.
22377 (movtf_internal): Likewise.
22378 (mov<mode>_internal, TDmode/TFmode): Likewise.
22379 (movtf_softfloat): Likewise.
22380 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22382 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22383 movdi_internal64, using wg constraint for move direct operations.
22384 (movdi_internal64): Likewise.
22386 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22387 MODES_TIEABLE_P for selected modes. Print the numerical value of
22388 the various virtual registers. Use GPR/FPR first/last values,
22389 instead of hard coding the register numbers. Print which modes
22390 have reload functions registered.
22391 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22392 settings before/after setting cpu, target and subtarget settings.
22393 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22394 and for secondary reload failures in rs6000_secondary_reload_inner.
22395 (rs6000_secondary_reload_fail): Likewise.
22396 (rs6000_secondary_reload_inner): Likewise.
22398 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22399 macros for first/last GPR and FPR registers.
22400 (LAST_GPR_REGNO): Likewise.
22401 (FIRST_FPR_REGNO): Likewise.
22402 (LAST_FPR_REGNO): Likewise.
22404 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22405 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22406 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22407 (vcond<mode><mode>): Likewise.
22408 (vcondu<mode><mode>): Likewise.
22409 (vector_gtu<mode>): Likewise.
22410 (vector_gte<mode>): Likewise.
22411 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22412 to prevent the compiler from converting DImode operations to TImode.
22413 (ior<mode>3): Likewise.
22414 (and<mode>3): Likewise.
22415 (one_cmpl<mode>2): Likewise.
22416 (nor<mode>3): Likewise.
22417 (andc<mode>3): Likewise.
22419 * config/rs6000/constraints.md (wt constraint): New constraint
22420 that returns VSX_REGS if TImode is allowed in VSX registers.
22422 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22423 constant under VSX.
22425 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22426 similar to TImode, but it is restricted to being in the GPRs.
22428 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22429 TImode to occupy a single VSX register.
22431 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22432 -mvsx-timode for power7/power8.
22433 (power7 cpu): Likewise.
22434 (power8 cpu): Likewise.
22436 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22437 sure that TFmode/TDmode take up two registers if they are ever
22438 allowed in the upper VSX registers.
22439 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22441 (rs6000_init_hard_regno_mode_ok): Likewise.
22442 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22443 constraint. Print if LRA is turned on.
22444 (rs6000_option_override_internal): Give an error if -mvsx-timode
22445 and VSX is not enabled.
22446 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22447 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22448 to reg+offset addressing. Use PTImode when checking offset
22449 addresses for validity.
22450 (reg_offset_addressing_ok_p): Likewise.
22451 (rs6000_legitimate_offset_address_p): Likewise.
22452 (rs6000_legitimize_address): Likewise.
22453 (rs6000_legitimize_reload_address): Likewise.
22454 (rs6000_legitimate_address_p): Likewise.
22455 (rs6000_eliminate_indexed_memrefs): Likewise.
22456 (rs6000_emit_move): Likewise.
22457 (rs6000_secondary_reload): Likewise.
22458 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22459 reloads to fpr registers to continue to use reg+offset addressing,
22460 but 64-bit reloads to altivec registers need reg+reg addressing.
22461 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22462 it. Treat LO_SUM like a PLUS operation.
22463 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22464 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22465 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22466 registers to share a register with a smaller sized type, since VSX
22467 puts scalars in the upper 64-bits.
22468 (print_operand): Add support for PTImode.
22469 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22470 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22471 registers, but don't have arithmetic support.
22472 (rs6000_memory_move_cost): Add test for VSX.
22473 (rs6000_opt_masks): Add -mvsx-timode.
22475 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22478 (VSr): Use wt constraint for TImode.
22479 (VSv): Drop TImode support.
22480 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22481 (vsx_movti_64bit): Likewise.
22482 (vsx_movti_32bit): Likewise.
22483 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22484 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22485 one '?' on the appropriate output constraint. Do not allow TImode
22486 logical operations on 32-bit systems.
22487 (vsx_ior<mode>3): Likewise.
22488 (vsx_xor<mode>3): Likewise.
22489 (vsx_one_cmpl<mode>2): Likewise.
22490 (vsx_nor<mode>3): Likewise.
22491 (vsx_andc<mode>3): Likewise.
22492 (vsx_concat_<mode>): Likewise.
22493 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22495 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22496 OPTION_MASK_VSX_TIMODE.
22497 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22498 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22500 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22501 (TI2 iterator): New iterator for TImode, PTImode.
22502 (wd mode attribute): Add values for vector types.
22503 (movti_string): Replace TI move operations with operations for TImode
22504 and PTImode. Add support for TImode being allowed in VSX registers.
22505 (mov<mode>_string, TImode/PTImode): Likewise.
22506 (movti_ppc64): Likewise.
22507 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22508 (TI mode splitters): Likewise.
22510 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22513 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22515 PR tree-optimization/56355
22516 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22517 Also handle integers with undefined overflow.
22519 2013-03-20 Catherine Moore <clm@codesourcery.com>
22520 Maciej W. Rozycki <macro@codesourcery.com>
22521 Tom de Vries <tom@codesourcery.com>
22522 Nathan Sidwell <nathan@codesourcery.com>
22523 Iain Sandoe <iain@codesourcery.com>
22524 Nathan Froyd <froydnj@codesourcery.com>
22525 Chao-ying Fu <fu@mips.com>
22527 * doc/extend.texi: (micromips, nomicromips, nocompression):
22528 Document new function attributes.
22529 * doc/invoke.texi (minterlink-compressed, mmicromips,
22530 m14k, m14ke, m14kec): Document new options.
22531 (minterlink-mips16): Update documentation.
22532 * doc/md.texi (ZC, ZD): Document new constraints.
22533 * configure.ac (gcc_cv_as_micromips): Check if linker
22534 supports the .set micromips directive.
22535 * configure: Regenerate.
22536 * config.in: Regenerate.
22537 * config/mips/mips-tables.opt: Regenerate.
22538 * config/mips/micromips.md: New file.
22539 * constraints.md (ZC, ZD): New constraints.
22540 * config/mips/predicates.md (movep_src_register): New predicate.
22541 (movep_src_operand): New predicate.
22542 (non_volatile_mem_operand): New predicate.
22543 * config/mips/mips.md (multimem): New type.
22544 (length): Differentiate between 17-bit and 18-bit branch offsets.
22545 (MOVEP1, MOVEP2): New mode iterator.
22546 (mov_<load>l): Use ZC constraint.
22547 (mov_<load>r): Likewise.
22548 (mov_<store>l): Likewise.
22549 (mov_<store>r): Likewise.
22550 (*branch_equality<mode>_inverted): Add microMIPS support.
22551 (*branch_equality<mode>): Likewise.
22552 (*jump_absolute): Likewise.
22553 (indirect_jump_<mode>): Likewise.
22554 (tablejump_<mode>): Likewise.
22555 (<optab>_internal): Likewise.
22556 (sibcall_internal): Likewise.
22557 (sibcall_value_internal): Likewise.
22558 (prefetch): Use constraint ZD.
22559 * config/mips/mips.opt (minterlink-compressed): New option.
22560 (minterlink-mips16): Now an alias for minterlink-compressed.
22561 (mmicromips): New option.
22562 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22563 (compare_and_swap_12): Likewise.
22564 (sync_add<mode>): Likewise.
22565 (sync_<optab>_12): Likewise.
22566 (sync_old_<optab>_12): Likewise.
22567 (sync_new_<optab>_12): Likewise.
22568 (sync_nand_12): Likewise.
22569 (sync_old_nand_12): Likewise.
22570 (sync_new_nand_12): Likewise.
22571 (sync_sub<mode>): Likewise.
22572 (sync_old_add<mode>): Likewise.
22573 (sync_old_sub<mode>): Likewise.
22574 (sync_new_add<mode>): Likewise.
22575 (sync_new_sub<mode>): Likewise.
22576 (sync_<optab><mode>): Likewise.
22577 (sync_old_<optab><mode>): Likewise.
22578 (sync_new_<optab><mode>): Likewise.
22579 (sync_nand<mode>): Likewise.
22580 (sync_old_nand<mode>): Likewise.
22581 (sync_new_nand<mode>): Likewise.
22582 (sync_lock_test_and_set<mode>): Likewise.
22583 (test_and_set_12): Likewise.
22584 (atomic_compare_and_swap<mode>): Likewise.
22585 (atomic_exchange<mode>_llsc): Likewise.
22586 (atomic_fetch_add<mode>_llsc): Likewise.
22587 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22588 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22589 (umips_save_restore_pattern_p): Likewise.
22590 (umips_load_store_pair_p): Likewise.
22591 (umips_output_load_store_pair): Likewise.
22592 (umips_movep_target_p): Likewise.
22593 (umips_12bit_offset_address_p): Likewise.
22594 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22595 (mips_base_mips16): Rename this...
22596 (mips_base_compression_flags): ...to this. Update all uses.
22597 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22598 (mips_mips16_decl_p): Delete.
22599 (mips_nomips16_decl_p): Delete.
22600 (mips_get_compress_on_flags): New function.
22601 (mips_get_compress_off_flags): New function.
22602 (mips_get_compress_mode): New function.
22603 (mips_get_compress_on_name): New function.
22604 (mips_get_compress_off_name): New function.
22605 (mips_insert_attributes): Support multiple compression types.
22606 (mips_merge_decl_attributes): Likewise.
22607 (umips_12bit_offset_address_p): New function.
22608 (mips_start_function_definition): Emit .set micromips directive.
22609 (mips_call_may_need_jalx_p): New function.
22610 (mips_function_ok_for_sibcall): Add microMIPS support.
22611 (mips_print_operand_punctuation): Support short delay slots and
22613 (umips_swm_mask, umips_swm_encoding): New.
22614 (umips_build_save_restore): New function.
22615 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22616 (was_mips16_p): Remove.
22617 (old_compression_mode): New.
22618 (mips_set_compression_mode): New function.
22619 (mips_set_current_function): Add microMIPS support.
22620 (mips_option_override): Likewise.
22621 (umips_save_restore_pattern_p): New function.
22622 (umips_output_save_restore): New function.
22623 (umips_load_store_pair_p_1): New function.
22624 (umips_load_store_pair_p): New function.
22625 (umips_output_load_store_pair_1): New function.
22626 (umips_output_load_store_pair): New function.
22627 (umips_movep_target_p) New function.
22628 (mips_prepare_pch_save): Add microMIPS support.
22629 * config/mips/mips.h (TARGET_COMPRESSION): New.
22630 (TARGET_CPU_CPP_BUILTINS): Update macro
22631 to use new compression flags and to support microMIPS.
22632 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22633 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22634 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22635 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22636 (ASM_SPEC): Support mmicromips and mno-micromips.
22637 (M16STORE_REG_P): New macro.
22638 (MIPS_CALL): Support TARGET_MICROMIPS.
22639 (MICROMIPS_J): New macro.
22640 (mips_base_mips16): Rename this...
22641 (mips_base_compression_flags): ...to this.
22642 (UMIPS_12BIT_OFFSET_P): New macro.
22643 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22644 (MULTILIB_DIRNAMES): Likewise.
22645 2013-03-20 Richard Biener <rguenther@suse.de>
22647 PR tree-optimization/56661
22648 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22649 the result does not have to be distinct.
22651 2013-03-20 Richard Biener <rguenther@suse.de>
22653 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22656 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22657 Steven Bosscher <steven@gcc.gnu.org>
22659 PR rtl-optimization/56605
22660 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22662 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22665 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22666 that require movd instead of movq.
22668 2013-03-20 Richard Biener <rguenther@suse.de>
22670 * tree-ssa-structalias.c (struct variable_info): Add pointer
22671 to the first field of an aggregate with sub-vars. Make
22672 this and the pointer to the next subfield its ID.
22673 (vi_next): New function.
22674 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22675 storedanything_id, integer_id): Increment by one.
22676 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22677 get_call_clobber_vi): Adjust.
22678 (solution_set_expand): Simplify and speedup.
22679 (solution_set_add): Inline into ...
22680 (set_union_with_increment): ... this. Adjust accordingly.
22681 (do_sd_constraint): Likewise.
22682 (do_ds_constraint): Likewise.
22683 (do_complex_constraint): Simplify.
22684 (build_pred_graph): Adjust.
22685 (solve_graph): Likewise. Simplify and speedup.
22686 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22687 get_constraint_for_component_ref, get_constraint_for_1,
22688 first_vi_for_offset, first_or_preceding_vi_for_offset,
22689 create_function_info_for, create_variable_info_for_1,
22690 create_variable_info_for, intra_create_variable_infos): Adjust.
22691 (init_base_vars): Push NULL for ID zero.
22692 (compute_points_to_sets): Adjust.
22694 2013-03-20 Richard Biener <rguenther@suse.de>
22696 * cfgloop.c (verify_loop_structure): Streamline and avoid
22697 ICEing on corrupt loop tree.
22698 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22701 2013-03-20 Richard Biener <rguenther@suse.de>
22703 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22704 check whether an SSA update is needed.
22706 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22708 * config/mips/constraints.md (T): Rename to...
22712 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22713 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22715 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22717 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22718 (*subsi3_carryin_uxtw): Likewise.
22720 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22722 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22723 (*rorsi3_insn_uxtw): Likewise.
22725 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22727 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22728 (*extrsi5_insn_uxtw): Likewise.
22730 2013-03-19 Richard Biener <rguenther@suse.de>
22732 PR tree-optimization/56273
22733 * passes.c (init_optimization_passes): Move second VRP after DOM.
22735 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22737 * config/i386/i386.md (*movti_internal): Merge from
22738 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22739 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22740 nox64 isa attributes.
22742 2013-03-18 Richard Biener <rguenther@suse.de>
22744 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22746 (merge_node_constraints): Likewise.
22747 (build_succ_graph): Likewise.
22748 (valid_graph_edge): Inline into single caller.
22749 (unify_nodes): Likewise. Use bitmap_set_bit return value
22751 (scc_visit): Fix formatting and variable use.
22752 (do_sd_constraint): Use gcc_checking_assert.
22753 (do_ds_constraint): Likewise.
22754 (do_complex_constraint): Likewise.
22755 (condense_visit): Likewise. Cleanup.
22756 (dump_pred_graph): New function.
22757 (perform_var_substitution): Dump the pred-graph before
22758 variable substitution.
22759 (find_equivalent_node): Use gcc_checking_assert.
22760 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22762 2013-03-18 Richard Biener <rguenther@suse.de>
22764 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22765 Remove cond_expr_stmt_list argument and do not gimplify the
22767 (vect_loop_versioning): Adjust.
22768 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22769 Cleanup to use less temporaries.
22770 (vect_create_data_ref_ptr): Cleanup.
22772 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22774 PR tree-optimization/56635
22775 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22776 require types_compatible_p types.
22778 2013-03-18 Nick Clifton <nickc@redhat.com>
22780 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22781 spurious backslash.
22783 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22784 Add missing line to comment describing function.
22786 2013-03-18 Richard Biener <rguenther@suse.de>
22788 PR tree-optimization/56210
22789 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22790 Handle string / character search functions.
22791 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22793 2013-03-18 Richard Biener <rguenther@suse.de>
22795 PR middle-end/56483
22796 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22797 and implement properly.
22798 * gimple.h (gimple_cond_single_var_p): Remove.
22800 2013-03-18 Richard Biener <rguenther@suse.de>
22802 * tree-data-ref.h (find_data_references_in_loop): Declare.
22803 * tree-data-ref.c (get_references_in_stmt): Use a stack
22804 vector pre-allocated in the callers.
22805 (find_data_references_in_stmt): Adjust.
22806 (graphite_find_data_references_in_stmt): Likewise.
22807 (create_rdg_vertices): Likewise.
22808 (find_data_references_in_loop): Export.
22809 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22810 Compute dependences here...
22811 (vect_analyze_data_refs): ...not here. When we encounter
22812 a non-vectorizable data reference in basic-block vectorization
22813 truncate the data reference vector. Do not bother to
22814 fixup data-dependence information for gather loads.
22815 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22816 of data references, as reported.
22818 2013-03-18 Richard Biener <rguenther@suse.de>
22820 PR tree-optimization/3713
22821 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22822 has_constants and expr.
22823 (stmt_has_constants): Properly valueize SSA names when deciding
22824 whether the stmt has constants.
22826 2013-03-18 Richard Biener <rguenther@suse.de>
22828 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22829 whole function when there is nothing to do.
22830 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22831 * tree-vectorizer.c (vectorize_loops): Update virtual and
22832 loop-closed SSA once.
22833 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22835 2013-03-18 Richard Biener <rguenther@suse.de>
22837 PR middle-end/56113
22838 * domwalk.c (bb_postorder): New global static.
22839 (cmp_bb_postorder): New function.
22840 (walk_dominator_tree): Replace scheme imposing an order for
22841 visiting dominator sons by one sorting them at the time they
22842 are pushed on the stack.
22844 2013-03-18 Richard Biener <rguenther@suse.de>
22846 PR tree-optimization/39326
22847 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22848 (struct mem_ref): Replace mem member with ao_ref typed member.
22849 (MEM_ANALYZABLE): Adjust.
22850 (memref_eq): Likewise.
22851 (mem_ref_alloc): Likewise.
22852 (gather_mem_refs_stmt): Likewise.
22853 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22854 (execute_sm_if_changed_flag_set): Adjust.
22855 (execute_sm): Likewise.
22856 (ref_always_accessed_p): Likewise.
22857 (refs_independent_p): Likewise.
22858 (can_sm_ref_p): Likewise.
22860 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22863 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22864 return 1 even for !unsignedp.
22866 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22868 * config/i386/i386.md (isa): Add x64 and nox64.
22869 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22870 (*pushtf): Enable *roF alternative for x64 isa only.
22871 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22872 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22873 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22874 (*movtf_internal): Merge from *movtf_internal_rex64 and
22875 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22876 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22877 nox64 isa attributes.
22878 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22879 nox64 isa attributes.
22880 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22882 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22884 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22886 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22889 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22890 class names. Remove trailing comma after "ALL_REGS".
22892 2013-03-16 Jan Hubicka <jh@suse.cz>
22894 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22895 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22896 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22897 of cgraph_get_create_node.
22898 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22900 2013-03-16 Jason Merrill <jason@redhat.com>
22903 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22904 with DW_AT_default_value.
22906 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22908 * BASE-VER: Set to 4.9.0.
22910 2013-03-14 Andi Kleen <ak@linux.intel.com>
22913 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22914 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22915 Document _x* TSX intrinsics.
22917 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22918 David Holsgrove <david.holsgrove@xilinx.com>
22920 * configure.ac: Add MicroBlaze TLS support detection.
22921 * configure: Regenerate.
22922 * config/microblaze/microblaze-protos.h
22923 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22924 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22925 * config/microblaze/microblaze.c (microblaze_address_type): Add
22926 ADDRESS_TLS and tls_reloc address types.
22927 (microblaze_address_info): Add tls_reloc.
22928 (TARGET_HAVE_TLS): Define.
22929 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22930 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22931 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22932 load_tls_operand, microblaze_call_tls_get_addr,
22933 microblaze_legitimize_tls_address): New functions.
22934 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22935 (get_base_reg): Use microblaze_tls_symbol_p.
22936 (microblaze_classify_address): Handle TLS.
22937 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22938 label_mentioned_p and microblaze_tls_referenced_p.
22939 (microblaze_legitimize_address): Handle TLS.
22940 (microblaze_address_insns): Handle ADDRESS_TLS.
22941 (pic_address_needs_scratch): Handle TLS.
22942 (print_operand_address): Handle TLS.
22943 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22944 (microblaze_expand_move): Handle TLS.
22945 (microblaze_legitimate_constant_p): Check
22946 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22947 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22948 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22949 (PIC_OFFSET_TABLE_REGNUM): Set.
22950 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22951 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22952 (addsi3, movsi_internal2, movdf_internal): Update constraints
22953 * config/microblaze/predicates.md (arith_plus_operand): Define
22954 (move_operand): Redefine as move_src_operand,
22955 check microblaze_tls_referenced_p.
22957 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22959 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22960 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22962 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22964 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22967 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22969 PR tree-optimization/53265
22970 * common.opt (Waggressive-loop-optimizations): New option.
22971 * tree-ssa-loop-niter.c: Include tree-pass.h.
22972 (do_warn_aggressive_loop_optimizations): New function.
22973 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22974 if number_of_latch_executions returned constant.
22975 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22976 early. If number_of_latch_executions returned constant, set
22977 nb_iterations_upper_bound back to it.
22978 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22980 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22981 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22983 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22984 (MULTILIB_OSDIRNAMES): Set.
22985 * genmultilib: If defaultosdirname doesn't start with :: , set
22986 defaultosdirname2 instead, clear it and emit two . multilib_raw
22987 entries instead of just one.
22989 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22991 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22992 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22993 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22994 (SUBTARGET_OVERRIDE_OPTIONS): New.
22996 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22999 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23000 (musermode): Convert to Var(TARGET_USERMODE).
23001 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23002 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23003 * config/sh/sh.c (sh_option_override): Use
23004 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23005 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23007 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23009 (udivsi3_i4_single, divsi3_i4_single): Use
23010 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23012 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23014 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23017 2013-03-13 Richard Biener <rguenther@suse.de>
23019 PR tree-optimization/56608
23020 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23021 calls when vectorizing basic-blocks.
23023 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23026 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23029 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23031 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23033 2013-03-11 Jan Hubicka <jh@suse.cz>
23036 * lto-streamer-out.c (output_symbol_p): Skip references from
23037 constructors of external variables.
23039 2013-03-11 Jan Hubicka <jh@suse.cz>
23041 PR middle-end/56571
23042 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23044 * emit-rtl.c (verify_rtx_sharing): Likewise.
23045 (copy_insn_1): Likewise.
23046 * rtl.c (copy_rtx): Likewise.
23048 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23051 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23052 output_operand_lossage message.
23054 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23057 * arm.c (shift_op): Validate RTL pattern on the fly.
23058 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23061 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23064 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23065 new call patterns that clobber %fr12.
23066 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23067 split and postreload patterns.
23068 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23069 registers %fr12 and %fr12R as call used.
23071 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23073 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23074 (canon_address, record_store, replace_read, check_mem_read_rtx,
23075 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23076 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23077 rest_of_handle_dse): Likewise.
23079 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23081 PR middle-end/56524
23082 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23084 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23085 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23086 (save_optabs_if_changed): Replace with...
23087 (init_tree_optimization_optabs): ...this.
23088 * optabs.c (save_optabs_if_changed): Rename to...
23089 (init_tree_optimization_optabs): ...this. Take the optimization node
23090 as argument. Do nothing if the base optabs are already correct.
23091 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23092 to recompute optabs.
23093 * function.h (function): Remove optabs field.
23094 * function.c (invoke_set_current_function_hook): Call
23095 init_tree_optimization_optabs. Use the result to initialize
23098 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23100 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23101 if GTMA_HAS_NO_INSTRUMENTATION.
23102 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23103 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23104 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23105 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23106 GTMA_HAS_NO_INSTRUMENTATION.
23108 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23110 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23113 2013-03-08 Marek Polacek <polacek@redhat.com>
23114 Jakub Jelinek <jakub@redhat.com>
23116 PR tree-optimization/56478
23117 * predict.c (is_comparison_with_loop_invariant_p): Change the
23118 type of loop_step to tree.
23119 (predict_loops): Adjust.
23120 (predict_iv_comparison): Perform the computations on double_ints.
23122 2013-03-08 Richard Biener <rguenther@suse.de>
23124 PR tree-optimization/56570
23125 * tree-cfg.c (verify_expr_location_1): Verify locations for
23127 * tree-sra.c (create_access_replacement): Strip locations
23128 from DECL_DEBUG_EXPRs.
23130 2013-03-08 Richard Biener <rguenther@suse.de>
23132 * tree-inline.c (expand_call_inline): Do not associate
23133 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23134 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23136 2013-03-08 Richard Biener <rguenther@suse.de>
23138 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23139 or block changes with -Og. Fix for location / block encoding
23140 changes and PHI arguments with locations.
23142 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23144 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23146 (struct output_info): Likewise.
23147 (register_overhead): Remove bad gcc_assert.
23148 (bitmap_find_bit): If there is only a single bitmap element, do not
23149 count a miss as a search.
23150 (print_statistics): Update for counter type changes.
23151 (dump_bitmap_statistics): Likewise. Print headers such that they
23152 are properly lined up with the printed counters.
23154 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23156 PR tree-optimization/56559
23157 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23158 check that it has only a single use.
23160 2013-03-07 Richard Biener <rguenther@suse.de>
23162 * doc/invoke.texi (fwhole-program): Discourage use in combination
23165 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23167 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23169 PR tree-optimization/56539
23170 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23171 instead of GSI_CONTINUE_LINKING as last argument to
23172 force_gimple_operand_gsi. Adjust function comment.
23174 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23177 PR middle-end/56548
23178 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23179 promoted mode, convert the result back to the original mode.
23181 2013-03-06 Richard Biener <rguenther@suse.de>
23183 PR middle-end/56294
23184 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23185 (insert_updated_phi_nodes_compare_uids): New function.
23186 (update_ssa): Sort symbols_to_rename after UID before
23187 traversing it to insert PHI nodes.
23189 2013-03-06 Richard Biener <rguenther@suse.de>
23191 PR middle-end/50494
23192 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23193 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23196 2013-02-13 Richard Biener <rguenther@suse.de>
23199 * varasm.c (output_constant_def_1): Get the decl representing
23200 the constant as argument.
23201 (output_constant_def): Wrap output_constant_def_1.
23202 (make_decl_rtl): Use output_constant_def_1 with the decl
23203 representing the constant.
23204 (build_constant_desc): Optionally re-use a decl already
23205 representing the constant.
23206 (tree_output_constant_def): Adjust.
23208 2013-03-06 Joey Ye <joey.ye@arm.com>
23211 * gcc.c (convert_white_space): New function.
23212 (main): Handles white space in function name.
23214 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23217 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23218 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23219 to SH_DIV_CALL_TABLE for TARGET_SH2.
23220 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23222 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23223 call-table options.
23225 2013-03-05 Sterling Augustine <saugustine@google.com>
23226 Cary Coutant <ccoutant@google.com>
23229 * dwarf2out.c (resolve_addr): Don't call
23230 remove_loc_list_addr_table_entries a second time for the same
23233 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23236 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23237 (avoid_complex_debug_insns): New function.
23238 (expand_debug_locations): Call it.
23240 PR rtl-optimization/56484
23241 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23242 lifetimes of hard registers on small register class machines.
23244 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23246 * config/microblaze/microblaze-protos.h: Rename
23247 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23248 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23250 (microblaze_fast_interrupt_function_p): New function.
23251 (microblaze_is_interrupt_handler): Rename to
23252 microblaze_is_interrupt_variant and add fast_interrupt check.
23253 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23254 (save_restore_insns): Likewise.
23255 (compute_frame_size): Likewise.
23256 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23257 (microblaze_globalize_label): Likewise.
23258 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23259 * config/microblaze/microblaze.md: Use wrapper
23260 microblaze_is_interrupt_variant.
23262 2013-03-05 Kai Tietz <ktietz@redhat.com>
23264 * sdbout.c (sdbout_one_type): Switch to current function's section
23265 supporting cold/hot.
23267 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23269 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23272 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23274 PR middle-end/56461
23275 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23276 if VALGRIND_GET_VBITS is defined, temporarily make object
23277 memory all defined, and restore previous valgrind addressability
23278 and definability afterwards. Free this_object at the end.
23280 PR middle-end/56461
23281 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23282 right before calling lra_create_live_ranges, also call it
23283 when clearing live_p. Only call lra_clear_live_ranges
23284 at the end if live_p.
23286 PR middle-end/56461
23287 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23289 2013-03-05 Richard Biener <rguenther@suse.de>
23291 PR tree-optimization/56521
23292 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23295 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23298 * except.h (remove_unreachable_eh_regions): New prototype.
23299 * except.c (remove_eh_handler_splicer): New function, split out
23300 of remove_eh_handler.
23301 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23302 warning about running it on many EH regions one at a time.
23303 (remove_unreachable_eh_regions_worker): New function, walk the
23304 EH tree in depth-first order and remove non-marked regions.
23305 (remove_unreachable_eh_regions): New function.
23306 * tree-eh.c (mark_reachable_handlers): New function, split out
23307 from remove_unreachable_handlers.
23308 (remove_unreachable_handlers): Use mark_reachable_handlers and
23309 remove_unreachable_eh_regions.
23310 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23311 and remove_unreachable_eh_regions.
23313 2013-03-05 Richard Biener <rguenther@suse.de>
23315 PR middle-end/56525
23316 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23317 not freeing them until the end.
23319 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23321 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23323 2013-03-05 Richard Biener <rguenther@suse.de>
23325 PR tree-optimization/56270
23326 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23327 of loads after scheduling an SLP instance.
23329 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23331 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23333 (check_gcc_parallelize): Run guality.exp as a separate job from
23334 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23335 struct-layout-1.exp with stackalign.exp.
23337 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23339 PR middle-end/56461
23340 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23341 load_index sbitmap even if some bit in it isn't set.
23343 PR middle-end/56461
23344 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23345 (discover_iteration_bound_by_body_walk): Change queues to
23346 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23347 spelling in comment. Call safe_push on queues[bound_index] directly.
23348 Release queues[queue_index] in every iteration unconditionally.
23349 Release bounds vector.
23351 PR middle-end/56461
23352 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23353 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23354 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23357 2013-03-05 Richard Biener <rguenther@suse.de>
23360 * tree-inline.c (remap_blocks_to_null): New function.
23361 (expand_call_inline): When expanding a call stmt without
23362 an associated block inline remap all callee blocks to NULL.
23364 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23366 PR rtl-optimization/56494
23367 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23368 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23369 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23371 PR middle-end/56461
23372 * sel-sched-ir.c (free_sched_pools): Release
23373 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23374 if succs_info_pool.max_top isn't -1.
23377 * opts.c (opts_obstack, opts_concat): Moved to...
23378 * opts-common.c (opts_obstack, opts_concat): ... here.
23380 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23382 PR middle-end/56461
23383 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23385 2013-03-04 Martin Jambor <mjambor@suse.cz>
23387 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23388 all appropriate places.
23390 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23392 PR tree-optimization/56424
23393 * ipa-split.c (split_function): Do not set the RSO flag if result is
23394 not by reference and its type is a register type.
23396 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23398 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23399 (microblaze_legitimate_pic_operand): Likewise
23400 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23401 new function microblaze_legitimate_pic_operand
23402 * config/microblaze/microblaze-protos.h
23403 (microblaze_legitimate_pic_operand): Declare.
23405 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23407 * config/microblaze/predicates.md (call_insn_simple_operand):
23408 New predicate for supported rtx code types.
23409 * config/microblaze/microblaze.md (call_internal1): Use
23410 call_insn_simple_operand predicate.
23412 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23414 PR middle-end/56461
23415 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23416 partitions.ordered_remove.
23418 PR middle-end/56461
23419 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23420 vec_oprnds0.create (1) for modifier == NONE.
23422 PR middle-end/56461
23423 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23424 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23425 vec_oprnds1 right before pushing anything to it for
23428 PR middle-end/56461
23429 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23430 set nbbs to 0 instead of having separate code path.
23431 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23432 instead of false as last argument if returning NULL.
23434 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23436 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23437 the attribute is now called "target" instead of "option".
23438 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23439 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23440 attribute/pragma name for TARGET_OPTION_VALID_P and
23441 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23442 * doc/tm.texi: Regenerated.
23444 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23446 * config/microblaze/microblaze.c:
23447 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23448 * config/microblaze/microblaze.h: Add -mxl-reorder to
23450 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23451 instructions emitted if TARGET_REORDER.
23452 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23453 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23456 2013-03-01 Xinliang David Li <davidxl@google.com>
23458 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23461 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23463 PR middle-end/56461
23464 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23465 vector even when returning true. Fix up function comment formatting.
23467 PR middle-end/56461
23468 * ira-build.c (ira_loop_nodes_count): New variable.
23469 (create_loop_tree_nodes): Initialize it.
23470 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23472 PR middle-end/56461
23473 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23474 method on dr_chain and result_chain.
23475 * tree-vect-stmts.c (vectorizable_store): Only call
23476 result_chain.create if j == 0.
23478 PR middle-end/56461
23479 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23480 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23481 before overwriting it.
23483 2013-03-01 Tobias Burnus <burnus@net-b.de>
23485 * doc/extended.texi (C Extensions): Change order in @menu
23487 (Other MIPS Built-in Functions): Move last MIPS entry before
23488 "picoChip Built-in Functions".
23489 (SH Built-in Functions): Move after RX Built-in Functions.
23490 * doc/gcc.texi (Introduction): Change order in @menu
23492 * doc/md.texi (Constraints): Ditto.
23493 * gty.texi (Type Information): Ditto.
23494 (User-provided marking routines for template types): Make
23496 * doc/invoke.texi (AArch64 Options): Move before
23497 "Adapteva Epiphany Options".
23499 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23500 Jakub Jelinek <jakub@redhat.com>
23503 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23504 no_address_safety_analysis attribute.
23505 * doc/extend.texi (no_address_safety_attribute): Rename to
23506 no_sanitize_address attribute, mention no_address_safety_analysis
23507 attribute as deprecated alias.
23509 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23511 PR middle-end/56461
23512 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23513 type to vec<vec<tree> > *.
23514 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23515 to be vec<tree> instead of vec<tree> *, set vec_defs
23516 to vNULL and call vec_defs.create (number_of_vects), adjust other
23518 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23519 vectorizable_condition): Adjust vect_get_slp_defs callers.
23521 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23523 * config/aarch64/aarch64.c
23524 (aarch64_float_const_representable): Remove unused variable.
23526 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23528 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23530 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23532 * config/aarch64/aarch64-builtins.c
23533 (aarch64_init_simd_builtins): Make static.
23535 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23537 * config/aarch64/aarch64.c
23538 (aarch64_simd_make_constant): Make static.
23540 2013-02-28 Martin Jambor <mjambor@suse.cz>
23542 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23543 with no initialization to the RHS of debug statements.
23545 2013-02-28 Martin Jambor <mjambor@suse.cz>
23547 PR tree-optimization/56294
23548 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23550 (get_access_replacement): Do not call create_access_replacement.
23551 Assert a replacement exists.
23552 (get_repl_default_def_ssa_name): Create the replacement declaration
23555 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23557 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23558 final_end_function.
23560 2013-02-28 Marek Polacek <polacek@redhat.com>
23562 PR rtl-optimization/56466
23563 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23564 if we're changing a loop.
23565 (peel_loops_completely): Likewise.
23567 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23570 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23572 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23575 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23576 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23577 INTX_FTYPE_FX, FX_FTYPE_INTX.
23578 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23580 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23582 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23583 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23584 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23585 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23586 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23587 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23588 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23589 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23590 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23591 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23592 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23593 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23594 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23595 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23596 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23597 (avrxmega6): Increase max flash segments from 5 to 6.
23598 * config/avr/t-multilib: Regenerate.
23599 * config/avr/avr-tables.opt: Regenerate.
23600 * doc/avr-mmcu.texi: Regenerate.
23602 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23604 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23605 (avr_device_to_arch): Rename to avr_device_to_ld.
23606 (avr_device_to_as): New prototype.
23607 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23608 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23609 * config/avr/driver-avr.c (avr_device_to_as): New.
23610 (avr_device_to_arch): Rename to avr_device_to_ld.
23612 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23614 PR middle-end/56461
23615 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23616 method on dr_chain and result_chain.
23618 PR middle-end/56461
23619 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23620 pointer_set_destroy on not_executed_last_iteration.
23622 PR middle-end/56461
23623 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23625 PR middle-end/56461
23626 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23627 FOR_EACH_DEFINED_FUNCTION when freeing state.
23629 PR middle-end/56461
23630 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23632 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23635 PR middle-end/56461
23636 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23637 known_aggs[i].items and release known_aggs vector.
23639 PR middle-end/56461
23640 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23642 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23644 * config/microblaze/microblaze.c (microblaze_emit_compare):
23645 Use xor for EQ/NE comparisions.
23646 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23647 (cbranchsf4): Adjust operator to comparison_operator.
23649 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23651 PR middle-end/56461
23652 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23654 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23655 vec_safe_push, always update *slot.
23656 (redirect_edge_var_map_clear): Use vec_free.
23657 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23658 (free_var_map_entry): Use vec_free.
23659 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23660 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23662 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23664 PR middle-end/45472
23665 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23666 when the may_trap_p bit of the exprs being merged differs.
23667 Reorder tests for speculativeness in the logical and operator.
23669 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23671 * incpath.c (add_standard_paths): Use reconcat instead of concat
23672 where appropriate and avoid leaking memory.
23674 * opts.h: Include obstack.h.
23675 (opts_concat): New prototype.
23676 (opts_obstack): New declaration.
23677 * opts.c (opts_concat): New function.
23678 (opts_obstack): New variable.
23679 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23680 (finish_options): Use opts_concat instead of concat
23681 and XOBNEWVEC instead of XNEWVEC.
23682 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23683 generate_option): Likewise.
23684 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23685 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23688 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23689 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23691 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23693 PR middle-end/56461
23694 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23696 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23698 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23699 (arm_block_move_unaligned_straight): Likewise.
23700 (arm_adjust_block_mem): Likewise.
23702 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23705 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23706 temp, cond and label.
23707 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23710 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23711 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23714 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23715 of prologue/epilogue functions.
23718 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23719 Remove unused variable cfa_offset.
23720 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23723 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23724 type promotion to unsigned.
23727 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23728 for HOST_WIDE_INT of 32 bit / same size as int.
23729 (arm_block_move_unaligned_straight): Likewise.
23730 (arm_adjust_block_mem): Likewise.
23733 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23736 2013-02-26 Marek Polacek <polacek@redhat.com>
23738 PR tree-optimization/56426
23739 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23741 2013-02-26 Richard Biener <rguenther@suse.de>
23744 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23745 unused variable loops.
23747 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23749 PR tree-optimization/56448
23750 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23751 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23752 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23753 later operands of the references, or even first operand for
23754 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23756 PR tree-optimization/56443
23757 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23758 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23759 to type_for_mode langhook.
23761 2013-02-25 Matt Turner <mattst88@gmail.com>
23763 * doc/invoke.texi: Document r4700.
23765 2013-02-25 Richard Biener <rguenther@suse.de>
23767 PR tree-optimization/56175
23768 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23770 (simplify_bitwise_binary): ... here. Also guard the conversion
23771 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23773 2013-02-25 Catherine Moore <clm@codesourcery.com>
23776 2013-02-24 Catherine Moore <clm@codesourcery.com>
23777 Maciej W. Rozycki <macro@codesourcery.com>
23778 Tom de Vries <tom@codesourcery.com>
23779 Nathan Sidwell <nathan@codesourcery.com>
23780 Iain Sandoe <iain@codesourcery.com>
23781 Nathan Froyd <froydnj@codesourcery.com>
23782 Chao-ying Fu <fu@mips.com>
23784 * doc/extend.texi: (micromips, nomicromips, nocompression):
23785 Document new function attributes.
23786 * doc/invoke.texi (minterlink-compressed, mmicromips,
23787 m14k, m14ke, m14kec): Document new options.
23788 (minterlink-mips16): Update documentation.
23789 * doc/md.texi (ZC, ZD): Document new constraints.
23790 * configure.ac (gcc_cv_as_micromips): Check if linker
23791 supports the .set micromips directive.
23792 * configure: Regenerate.
23793 * config.in: Regenerate.
23794 * config/mips/mips-tables.opt: Regenerate.
23795 * config/mips/micromips.md: New file.
23796 * constraints.md (ZC, AD): New constraints.
23797 * config/mips/predicates.md (movep_src_register): New predicate.
23798 (movep_src_operand): New predicate.
23799 (non_volatile_mem_operand): New predicate.
23800 * config/mips/mips.md (multimem): New type.
23801 (length): Differentiate between 17-bit and 18-bit branch offsets.
23802 (MOVEP1, MOVEP2): New mode iterator.
23803 (mov_<load>l): Use ZC constraint.
23804 (mov_<load>r): Likewise.
23805 (mov_<store>l): Likewise.
23806 (mov_<store>r): Likewise.
23807 (*branch_equality<mode>_inverted): Add microMIPS support.
23808 (*branch_equality<mode>): Likewise.
23809 (*jump_absolute): Likewise.
23810 (indirect_jump_<mode>): Likewise.
23811 (tablejump_<mode>): Likewise.
23812 (<optab>_internal): Likewise.
23813 (sibcall_internal): Likewise.
23814 (sibcall_value_internal): Likewise.
23815 (prefetch): Use constraint ZD.
23816 * config/mips/mips.opt (minterlink-compressed): New option.
23817 (minterlink-mips16): Now an alias for minterlink-compressed.
23818 (mmicromips): New option.
23819 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23820 (compare_and_swap_12): Likewise.
23821 (sync_add<mode>): Likewise.
23822 (sync_<optab>_12): Likewise.
23823 (sync_old_<optab>_12): Likewise.
23824 (sync_new_<optab>_12): Likewise.
23825 (sync_nand_12): Likewise.
23826 (sync_old_nand_12): Likewise.
23827 (sync_new_nand_12): Likewise.
23828 (sync_sub<mode>): Likewise.
23829 (sync_old_add<mode>): Likewise.
23830 (sync_old_sub<mode>): Likewise.
23831 (sync_new_add<mode>): Likewise.
23832 (sync_new_sub<mode>): Likewise.
23833 (sync_<optab><mode>): Likewise.
23834 (sync_old_<optab><mode>): Likewise.
23835 (sync_new_<optab><mode>): Likewise.
23836 (sync_nand<mode>): Likewise.
23837 (sync_old_nand<mode>): Likewise.
23838 (sync_new_nand<mode>): Likewise.
23839 (sync_lock_test_and_set<mode>): Likewise.
23840 (test_and_set_12): Likewise.
23841 (atomic_compare_and_swap<mode>): Likewise.
23842 (atomic_exchange<mode>_llsc): Likewise.
23843 (atomic_fetch_add<mode>_llsc): Likewise.
23844 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23845 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23846 (umips_save_restore_pattern_p): Likewise.
23847 (umips_load_store_pair_p): Likewise.
23848 (umips_output_load_store_pair): Likewise.
23849 (umips_movep_target_p): Likewise.
23850 (umips_12bit_offset_address_p): Likewise.
23851 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23852 (mips_base_mips16): Rename this...
23853 (mips_base_compression_flags): ...to this. Update all uses.
23854 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23855 (mips_mips16_decl_p): Delete.
23856 (mips_nomips16_decl_p): Delete.
23857 (mips_get_compress_on_flags): New function.
23858 (mips_get_compress_off_flags): New function.
23859 (mips_get_compress_mode): New function.
23860 (mips_get_compress_on_name): New function.
23861 (mips_get_compress_off_name): New function.
23862 (mips_insert_attributes): Support multiple compression types.
23863 (mips_merge_decl_attributes): Likewise.
23864 (umips_12bit_offset_address_p): New function.
23865 (mips_start_function_definition): Emit .set micromips directive.
23866 (mips_call_may_need_jalx_p): New function.
23867 (mips_function_ok_for_sibcall): Add microMIPS support.
23868 (mips_print_operand_punctuation): Support short delay slots and
23870 (umips_swm_mask, umips_swm_encoding): New.
23871 (umips_build_save_restore): New function.
23872 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23873 (was_mips16_p): Remove.
23874 (old_compression_mode): New.
23875 (mips_set_compression_mode): New function.
23876 (mips_set_current_function): Add microMIPS support.
23877 (mips_option_override): Likewise.
23878 (umips_save_restore_pattern_p): New function.
23879 (umips_output_save_restore): New function.
23880 (umips_load_store_pair_p_1): New function.
23881 (umips_load_store_pair_p): New function.
23882 (umips_output_load_store_pair_1): New function.
23883 (umips_output_load_store_pair): New function.
23884 (umips_movep_target_p) New function.
23885 (mips_prepare_pch_save): Add microMIPS support.
23886 * config/mips/mips.h (TARGET_COMPRESSION): New.
23887 (TARGET_CPU_CPP_BUILTINS): Update macro
23888 to use new compression flags and to support microMIPS.
23889 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23890 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23891 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23892 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23893 (ASM_SPEC): Support mmicromips and mno-micromips.
23894 (M16STORE_REG_P): New macro.
23895 (MIPS_CALL): Support TARGET_MICROMIPS.
23896 (MICROMIPS_J): New macro.
23897 (mips_base_mips16): Rename this...
23898 (mips_base_compression_flags): ...to this.
23899 (UMIPS_12BIT_OFFSET_P): New macro.
23900 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23901 (MULTILIB_DIRNAMES): Likewise.
23903 2013-02-25 Tom de Vries <tom@codesourcery.com>
23905 PR rtl-optimization/56131
23906 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23907 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23908 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23910 2013-02-25 Tobias Burnus <burnus@net-b.de>
23912 * doc/invoke.texi (-fsanitize=): Move from optimization
23913 to debugging options.
23915 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23917 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23919 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23920 Alexander Monakov <amonakov@ispras.ru>
23922 PR middle-end/56077
23923 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23924 flush pending lists also on non-jumps. Adjust comment.
23926 2013-02-24 Catherine Moore <clm@codesourcery.com>
23927 Maciej W. Rozycki <macro@codesourcery.com>
23928 Tom de Vries <tom@codesourcery.com>
23929 Nathan Sidwell <nathan@codesourcery.com>
23930 Iain Sandoe <iain@codesourcery.com>
23931 Nathan Froyd <froydnj@codesourcery.com>
23932 Chao-ying Fu <fu@mips.com>
23934 * doc/extend.texi: (micromips, nomicromips, nocompression):
23935 Document new function attributes.
23936 * doc/invoke.texi (minterlink-compressed, mmicromips,
23937 m14k, m14ke, m14kec): Document new options.
23938 (minterlink-mips16): Update documentation.
23939 * doc/md.texi (ZC, ZD): Document new constraints.
23940 * configure.ac (gcc_cv_as_micromips): Check if linker
23941 supports the .set micromips directive.
23942 * configure: Regenerate.
23943 * config.in: Regenerate.
23944 * config/mips/mips-tables.opt: Regenerate.
23945 * config/mips/micromips.md: New file.
23946 * constraints.md (ZC, AD): New constraints.
23947 * config/mips/predicates.md (movep_src_register): New predicate.
23948 (movep_src_operand): New predicate.
23949 (non_volatile_mem_operand): New predicate.
23950 * config/mips/mips.md (multimem): New type.
23951 (length): Differentiate between 17-bit and 18-bit branch offsets.
23952 (MOVEP1, MOVEP2): New mode iterator.
23953 (mov_<load>l): Use ZC constraint.
23954 (mov_<load>r): Likewise.
23955 (mov_<store>l): Likewise.
23956 (mov_<store>r): Likewise.
23957 (*branch_equality<mode>_inverted): Add microMIPS support.
23958 (*branch_equality<mode>): Likewise.
23959 (*jump_absolute): Likewise.
23960 (indirect_jump_<mode>): Likewise.
23961 (tablejump_<mode>): Likewise.
23962 (<optab>_internal): Likewise.
23963 (sibcall_internal): Likewise.
23964 (sibcall_value_internal): Likewise.
23965 (prefetch): Use constraint ZD.
23966 * config/mips/mips.opt (minterlink-compressed): New option.
23967 (minterlink-mips16): Now an alias for minterlink-compressed.
23968 (mmicromips): New option.
23969 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23970 (compare_and_swap_12): Likewise.
23971 (sync_add<mode>): Likewise.
23972 (sync_<optab>_12): Likewise.
23973 (sync_old_<optab>_12): Likewise.
23974 (sync_new_<optab>_12): Likewise.
23975 (sync_nand_12): Likewise.
23976 (sync_old_nand_12): Likewise.
23977 (sync_new_nand_12): Likewise.
23978 (sync_sub<mode>): Likewise.
23979 (sync_old_add<mode>): Likewise.
23980 (sync_old_sub<mode>): Likewise.
23981 (sync_new_add<mode>): Likewise.
23982 (sync_new_sub<mode>): Likewise.
23983 (sync_<optab><mode>): Likewise.
23984 (sync_old_<optab><mode>): Likewise.
23985 (sync_new_<optab><mode>): Likewise.
23986 (sync_nand<mode>): Likewise.
23987 (sync_old_nand<mode>): Likewise.
23988 (sync_new_nand<mode>): Likewise.
23989 (sync_lock_test_and_set<mode>): Likewise.
23990 (test_and_set_12): Likewise.
23991 (atomic_compare_and_swap<mode>): Likewise.
23992 (atomic_exchange<mode>_llsc): Likewise.
23993 (atomic_fetch_add<mode>_llsc): Likewise.
23994 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23995 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23996 (umips_save_restore_pattern_p): Likewise.
23997 (umips_load_store_pair_p): Likewise.
23998 (umips_output_load_store_pair): Likewise.
23999 (umips_movep_target_p): Likewise.
24000 (umips_12bit_offset_address_p): Likewise.
24001 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24002 (mips_base_mips16): Rename this...
24003 (mips_base_compression_flags): ...to this. Update all uses.
24004 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24005 (mips_mips16_decl_p): Delete.
24006 (mips_nomips16_decl_p): Delete.
24007 (mips_get_compress_on_flags): New function.
24008 (mips_get_compress_off_flags): New function.
24009 (mips_get_compress_mode): New function.
24010 (mips_get_compress_on_name): New function.
24011 (mips_get_compress_off_name): New function.
24012 (mips_insert_attributes): Support multiple compression types.
24013 (mips_merge_decl_attributes): Likewise.
24014 (umips_12bit_offset_address_p): New function.
24015 (mips_start_function_definition): Emit .set micromips directive.
24016 (mips_call_may_need_jalx_p): New function.
24017 (mips_function_ok_for_sibcall): Add microMIPS support.
24018 (mips_print_operand_punctuation): Support short delay slots and
24020 (umips_swm_mask, umips_swm_encoding): New.
24021 (umips_build_save_restore): New function.
24022 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24023 (was_mips16_p): Remove.
24024 (old_compression_mode): New.
24025 (mips_set_compression_mode): New function.
24026 (mips_set_current_function): Add microMIPS support.
24027 (mips_option_override): Likewise.
24028 (umips_save_restore_pattern_p): New function.
24029 (umips_output_save_restore): New function.
24030 (umips_load_store_pair_p_1): New function.
24031 (umips_load_store_pair_p): New function.
24032 (umips_output_load_store_pair_1): New function.
24033 (umips_output_load_store_pair): New function.
24034 (umips_movep_target_p) New function.
24035 (mips_prepare_pch_save): Add microMIPS support.
24036 * config/mips/mips.h (TARGET_COMPRESSION): New.
24037 (TARGET_CPU_CPP_BUILTINS): Update macro
24038 to use new compression flags and to support microMIPS.
24039 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24040 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24041 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24042 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24043 (ASM_SPEC): Support mmicromips and mno-micromips.
24044 (M16STORE_REG_P): New macro.
24045 (MIPS_CALL): Support TARGET_MICROMIPS.
24046 (MICROMIPS_J): New macro.
24047 (mips_base_mips16): Rename this...
24048 (mips_base_compression_flags): ...to this.
24049 (UMIPS_12BIT_OFFSET_P): New macro.
24050 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24051 (MULTILIB_DIRNAMES): Likewise.
24053 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24056 * target-globals.c (save_target_globals): For init_reg_sets and
24057 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24059 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24061 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24062 * config/aarch64/t-aarch64
24063 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24065 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24067 PR inline-asm/56148
24068 * lra-constraints.c (process_alt_operands): Reload operand
24069 conflicting with earlier clobber only if no more other conflicting
24072 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24075 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24076 if not linking a shared library.
24078 2013-02-22 Seth LaForge <sethml@google.com>
24080 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24082 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24084 * config/arm/arm.md (split for extendsidi): Update condition.
24085 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24086 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24087 (qhs_zextenddi_cstr): Likewise.
24089 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24091 PR middle-end/56420
24092 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24093 avoid signed wrapping.
24094 (expand_mult): Handle properly multiplication by
24095 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24096 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24097 in the compiler if coeff is HOST_WIDE_INT_MIN.
24098 (expand_divmod): Don't make ext_op1 static, change it's type to
24099 uhwi. Avoid undefined behavior in -INTVAL (op1).
24101 PR rtl-optimization/50339
24102 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24104 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24105 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24106 into splitting_ashiftrt field.
24107 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24109 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24112 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24114 PR middle-end/56108
24115 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24116 are sure to go irrevocable.
24118 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24120 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24121 scalars are valid operands.
24123 2013-02-21 Martin Jambor <mjambor@suse.cz>
24125 PR tree-optimization/56310
24126 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24127 only matching indices and non-negative final offsets.
24128 (intersect_aggregates_with_edge): Pass src_idx to
24129 agg_replacements_to_vector. Pass src_idx insstead of index to
24130 intersect_with_agg_replacements.
24132 2013-02-21 Martin Jambor <mjambor@suse.cz>
24134 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24135 instead of hard-wired defaults.
24137 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24139 * doc/invoke.texi (MIPS Options): Update documentation of the
24140 floating-point multiply-accumulate instruction restrictions.
24142 2013-02-21 Kostya Serebryany <kcc@google.com>
24144 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24145 asan_shadow_offset on x86_64 linux.
24147 2013-02-21 Richard Biener <rguenther@suse.de>
24149 PR tree-optimization/56415
24151 2013-02-11 Richard Biener <rguenther@suse.de>
24153 PR tree-optimization/56273
24154 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24157 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24160 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24163 PR inline-asm/56405
24164 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24165 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24167 2013-02-20 Jan Hubicka <jh@suse.cz>
24169 PR tree-optimization/56265
24170 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24171 when target is referenced for first time.
24173 2013-02-20 Richard Biener <rguenther@suse.de>
24175 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24176 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24177 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24178 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24179 not return anything.
24180 (rename_ssa_copies): Do not remove unused locals.
24181 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24182 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24183 * passes.c (execute_function_todo): Do not schedule unused locals
24184 removal if cleanup_tree_cfg did something.
24185 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24186 about the number of removed locals.
24188 2013-02-20 Richard Biener <rguenther@suse.de>
24190 PR tree-optimization/56398
24191 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24193 2013-02-20 Martin Jambor <mjambor@suse.cz>
24195 PR tree-optimization/55334
24196 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24197 restricted pointers to arrays.
24199 2013-02-20 Richard Biener <rguenther@suse.de>
24200 Jakub Jelinek <jakub@redhat.com>
24202 PR tree-optimization/56396
24203 * tree-ssa-ccp.c (n_const_val): New static variable.
24204 (get_value): Return NULL for SSA names we don't have a lattice
24206 (ccp_initialize): Initialize n_const_val.
24207 * tree-ssa-copy.c (n_copy_of): New static variable.
24208 (init_copy_prop): Initialize n_copy_of.
24209 (get_value): Return NULL_TREE for SSA names we don't have a
24212 2013-02-20 Martin Jambor <mjambor@suse.cz>
24214 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24216 2013-02-20 Richard Biener <rguenther@suse.de>
24218 * genpreds.c (write_lookup_constraint): Do not compare first
24219 letter of the constraint again.
24221 2013-02-20 Richard Biener <rguenther@suse.de>
24223 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24225 (get_use_iv_cost): Terminate hashtable walk when coming across
24228 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24230 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24231 reassociation for avx2 targets.
24233 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24235 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24236 Add version check for v8.10.a to enable microblaze_has_clz
24237 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24238 version and TARGET_PATTERN_COMPARE check
24239 * config/microblaze/microblaze.md: New clzsi2 instruction
24241 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24243 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24244 function before branching.
24246 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24248 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24250 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24252 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24254 PR middle-end/55889
24255 * sel-sched.c: Include ira.h.
24256 (implicit_clobber_conflict_p): New function.
24257 (moveup_expr): Use it.
24258 * Makefile.in (sel-sched.o): Depend on ira.h.
24260 2013-02-19 Richard Biener <rguenther@suse.de>
24262 PR tree-optimization/56384
24263 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24264 (vn_hash_type): Split out from ...
24265 (vn_hash_constant_with_type): ... here.
24266 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24267 (vn_phi_eq): Compare types from vn_phi_s structure.
24268 (vn_phi_lookup): Populate vn_phi_s type.
24269 (vn_phi_insert): Likewise.
24271 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24273 PR tree-optimization/56350
24274 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24275 if haven't found reduction or nested cycle operand, rather than
24276 asserting we must find it.
24278 PR tree-optimization/56381
24279 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24282 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24283 Jakub Jelinek <jakub@redhat.com>
24286 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24287 (swap_optab_enable): Same.
24288 (init_all_optabs): Use argument instead of global.
24289 * tree.h (struct tree_optimization_option): New field target_optabs.
24290 * expr.h (init_all_optabs): Add argument to prototype.
24291 (TREE_OPTIMIZATION_OPTABS): New.
24292 (save_optabs_if_changed): Protoize.
24293 * optabs.h: Declare this_fn_optabs.
24294 * optabs.c (save_optabs_if_changed): New.
24295 Declare this_fn_optabs.
24296 (init_optabs): Add argument to init_all_optabs() call.
24297 * function.c (invoke_set_current_function_hook): Handle per
24299 * function.h (struct function): New field optabs.
24300 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24301 optimization_current_node has changed.
24302 * target-globals.h (save_target_globals_default_opts): Protoize.
24303 * target-globals.c (save_target_globals_default_opts): New.
24305 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24308 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24309 registers %fr12 and %fr12R as call used.
24312 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24313 and HImode, require all displacements to be an integer multiple of
24315 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24316 only allow QImode and HImode when reload is in progress and strict is
24317 true. Likewise for symbolic addresses. Use base14_operand to check
24318 displacements in REG+BASE addresses.
24320 2013-02-18 Richard Biener <rguenther@suse.de>
24322 PR tree-optimization/56366
24323 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24324 handle sign-conversion of outer-loop initial induction value.
24326 2013-02-18 Richard Biener <rguenther@suse.de>
24328 PR middle-end/56349
24329 * cfghooks.c (merge_blocks): If we merge a latch into another
24330 block adjust references to it.
24331 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24332 (verify_loop_structure): Verify that a recorded latch is in fact
24335 2013-02-18 Richard Biener <rguenther@suse.de>
24337 PR tree-optimization/56321
24338 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24339 order SSA name release and virtual operand unlinking.
24341 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24343 * config/microblaze/microblaze.md (save_stack_block): Define.
24344 (restore_stack_block): Likewise.
24346 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24348 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24349 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24350 * config/microblaze/microblaze.c (microblaze_option_override):
24351 Bail out early for PIC modes when target does not support PIC.
24353 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24355 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24356 Replace with a microblaze version.
24357 (microblaze_trampoline_init): Adapt for microblaze.
24358 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24361 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24362 Dodji Seketeli <dodji@redhat.com>
24365 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24366 (instrument_mem_region_access): Do not forget to always put
24367 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24368 0) statement, even for cases where either 'base' or 'base + len'
24369 are not instrumented -- because they have been previously
24370 instrumented. Simplify the logic by putting all the statements
24371 instrument 'base + len' inside a sequence, and then insert that
24372 sequence right before the current insertion point. Then, to
24373 instrument 'base + len', just get an iterator on that statement.
24374 And do not forget to update the pointer to iterator the function
24375 received as argument.
24377 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24379 PR rtl-optimization/56348
24380 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24382 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24384 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24385 (clean_graph_dump_file): Pass base to start_graph_dump.
24387 2013-02-14 Richard Henderson <rth@redhat.com>
24390 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24392 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24394 * collect2-aix.h: Define F_LOADONLY.
24396 2013-02-14 Richard Biener <rguenther@suse.de>
24399 * varasm.c (output_constant_def_1): Get the decl representing
24400 the constant as argument.
24401 (output_constant_def): Wrap output_constant_def_1.
24402 (make_decl_rtl): Use output_constant_def_1 with the decl
24403 representing the constant.
24404 (build_constant_desc): Optionally re-use a decl already
24405 representing the constant.
24406 (tree_output_constant_def): Adjust.
24408 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24411 * asan.c (instrument_builtin_call): Really put the length of the
24412 second source argument into src1_len.
24414 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24416 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24417 argument. If it is false, don't create edge from then_bb to
24419 (insert_if_then_before_iter): Pass true to it.
24420 (build_check_stmt): Pass false to it.
24421 (transform_statements): Flush hash table only on extended basic
24422 block boundaries, rather than at the beginning of every bb.
24423 Don't flush hash table on nonfreeing_call_p calls.
24424 * tree-flow.h (nonfreeing_call_p): New prototype.
24425 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24427 2013-02-13 David S. Miller <davem@davemloft.net>
24429 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24431 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24434 * ira.c (max_regno_before_ira): Move from ...
24436 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24439 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24441 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24443 2013-02-13 Richard Biener <rguenther@suse.de>
24446 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24447 globals in MEM_REFs.
24449 2013-02-13 Richard Biener <rguenther@suse.de>
24451 * loop-init.c (loop_optimizer_init): Clear loop state when
24452 re-initializing preserved loops.
24453 * loop-unswitch.c (unswitch_single_loop): Return whether
24454 we unswitched the loop. Do not verify loop state here.
24455 (unswitch_loops): When we unswitched a loop discover new loops.
24457 2013-02-13 Kostya Serebryany <kcc@google.com>
24459 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24461 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24463 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24465 Avoid instrumenting duplicated memory access in the same basic block
24466 * Makefile.in (asan.o): Add new dependency on hash-table.h
24467 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24468 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24469 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24470 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24471 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24472 (get_mem_ref_of_assignment): New functions.
24473 (get_mem_refs_of_builtin_call): Extract from
24474 instrument_builtin_call and tweak a little bit to make it fit with
24476 (instrument_builtin_call): Use the new
24477 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24478 of is_gimple_builtin_call.
24479 (instrument_derefs, instrument_mem_region_access): Insert the
24480 instrumented memory reference into the hash table.
24481 (maybe_instrument_assignment): Renamed instrument_assignment into
24482 this, and change it to advance the iterator when instrumentation
24483 actually happened and return true in that case. This makes it
24484 homogeneous with maybe_instrument_assignment, and thus give a
24485 chance to callers to be more 'regular'.
24486 (transform_statements): Clear the memory reference hash table
24487 whenever we enter a new BB, when we cross a function call, or when
24488 we are done transforming statements. Use
24489 maybe_instrument_assignment instead of instrumentation. No more
24490 need to special case maybe_instrument_assignment and advance the
24491 iterator after calling it; it's now handled just like
24492 maybe_instrument_call. Update comment.
24494 2013-02-13 Richard Biener <rguenther@suse.de>
24496 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24497 Fix loop discovery code.
24499 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24501 PR inline-asm/56148
24502 * lra-constraints.c (process_alt_operands): Match early clobber
24503 operand with itself. Check conflicts with earlyclobber only if
24504 the operand is not reloaded. Prefer to reload conflicting operand
24505 if earlyclobber and matching operands are the same.
24507 2013-02-12 Richard Biener <rguenther@suse.de>
24510 * lto-streamer-out.c (write_symbol): Do not output symbols
24511 for hard register variables.
24513 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24516 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24517 (umulsidi3_insn, mulsidi3_insn): New insns.
24519 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24521 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24522 (struct tune_params): Add vec_costs field.
24523 * config/arm/arm.c (arm_builtin_vectorization_cost)
24524 (arm_add_stmt_cost): New functions.
24525 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24526 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24527 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24528 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24529 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24530 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24531 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24533 2013-02-12 Richard Biener <rguenther@suse.de>
24536 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24537 decls again if possible.
24539 2013-02-12 Richard Biener <rguenther@suse.de>
24541 PR middle-end/56288
24542 * tree-ssa.c (verify_ssa_name): Fix check, move
24543 SSA_NAME_IN_FREE_LIST check up.
24545 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24546 Steven Bosscher <steven@gcc.gnu.org>
24548 PR rtl-optimization/56151
24549 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24550 equal to op0 or op1, and last_insn pattern is CODE operation
24551 with MEM dest and one of the operands matches that MEM.
24553 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24555 * doc/extend.texi: Document Function Multiversioning and "default"
24556 parameter string to target attribute.
24557 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24558 target attribute parameter is "default".
24559 (ix86_compare_version_priority): Remove checks for target attribute.
24560 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24561 Remove check for target attribute equal to NULL. Add assert.
24562 (ix86_generate_version_dispatcher_body): Change error to sorry.
24564 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24565 Jack Howarth <howarth@bromo.med.uc.edu>
24566 Patrick Marlier <patrick.marlier@gmail.com>
24569 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24570 define ENDFILE_SPEC as TM_DESTRUCTOR.
24571 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24573 2013-02-11 Alexander Potapenko <glider@google.com>
24574 Jack Howarth <howarth@bromo.med.uc.edu>
24575 Jakub Jelinek <jakub@redhat.com>
24578 * config/darwin.c (cdtor_record): Rename ctor_record.
24579 (sort_cdtor_records): Rename sort_ctor_records.
24580 (finalize_dtors): New routine to sort destructors by
24581 priority before use in assemble_integer.
24582 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24584 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24586 PR rtl-optimization/56275
24587 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24588 offset is non-negative and less than cmode size before
24589 calling simplify_subreg.
24591 2013-02-11 Richard Biener <rguenther@suse.de>
24593 PR tree-optimization/56264
24594 * cfgloop.h (fix_loop_structure): Adjust prototype.
24595 * loop-init.c (fix_loop_structure): Return the number of
24596 newly discovered loops.
24597 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24598 are discovered, do a full loop-closed SSA rewrite.
24600 2013-02-11 Richard Biener <rguenther@suse.de>
24602 PR tree-optimization/56273
24603 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24605 (check_array_ref): Fix missing newline in dumps.
24606 (search_for_addr_array): Likewise.
24608 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24610 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24612 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24615 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24617 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24619 PR rtl-optimization/56246
24620 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24622 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24623 constraints are satisfied.
24625 2013-02-08 Jeff Law <law@redhat.com>
24628 * emit-rtl.c (reg_is_parm_p): New function.
24629 * regs.h (reg_is_parm_p): New prototype.
24630 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24631 callee-clobbered registers.
24633 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24636 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24637 If there is no implicit builtin declaration, just return NULL.
24639 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24641 * config/i386/sse.md (FMAMODEM): New mode iterator.
24642 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24643 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24645 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24647 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24648 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24649 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24651 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24653 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24654 (microblaze*-*-elf): Likewise.
24655 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24657 * config/microblaze/microblaze-c.c: Add builtin defines for
24658 _LITTLE_ENDIAN and _BIG_ENDIAN.
24659 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24660 add to TARGET_DEFAULT flags.
24661 Expand ASM_SPEC and LINK_SPEC.
24662 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24663 * config/microblaze/microblaze.md: Update extendsidi2 and
24664 movdi_internal instructions to use low-order / high-order reg
24666 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24667 options and inversemask / mask of LITTLE_ENDIAN.
24668 * config/microblaze/t-microblaze: Expand multilib options to
24669 include mlittle-endian (le) and update exceptions patterns.
24671 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24673 PR rtl-optimization/56195
24674 * lra-constraints.c (get_reload_reg): Don't reuse regs
24675 if they have smaller mode than requested, if they have
24676 wider mode than requested, try to return a SUBREG.
24678 PR tree-optimization/56250
24679 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24680 if type is unsigned and code isn't MULT_EXPR.
24682 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24684 PR tree-optimization/56064
24685 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24686 bits according to mode.
24687 * fixed-value.h (fixed_from_double_int)
24688 (const_fixed_from_double_int): Adjust comments.
24690 2013-02-08 Richard Biener <rguenther@suse.de>
24693 * lto-streamer.h (struct data_in): Remove current_file, current_line
24694 and current_col members.
24695 * lto-streamer-out.c (lto_output_location): Stream changed bits
24696 en-block for efficiency.
24697 * lto-streamer-in.c (clear_line_info): Remove.
24698 (lto_input_location): Cache current file, line and column
24699 globally via local statics. Read changed bits en-block.
24700 (input_function): Do not call clear_line_info.
24701 (lto_read_body): Likewise.
24702 (lto_input_toplevel_asms): Likewise.
24704 2013-02-08 Michael Matz <matz@suse.de>
24706 PR tree-optimization/52448
24707 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24708 (nt_call_phase): New static.
24709 (add_or_mark_expr): Only mark accesses with newer phase than any
24711 (nonfreeing_call_p): New.
24712 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24713 (nt_fini_block): Keep blocks marked as visited.
24714 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24716 2013-02-08 Richard Biener <rguenther@suse.de>
24718 * ira.c (ira): Free broken dominator information.
24720 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24722 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24724 2013-02-08 Marek Polacek <polacek@redhat.com>
24726 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24728 2013-02-08 Richard Biener <rguenther@suse.de>
24730 PR middle-end/56181
24731 * cfgloop.h (flow_loops_find): Adjust.
24732 (bb_loop_header_p): Declare.
24733 * cfgloop.c (bb_loop_header_p): New function split out from ...
24734 (flow_loops_find): ... here. Adjust function signature,
24735 support incremental loop structure update.
24736 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24737 * cfgloopmanip.c (fix_loop_structure): Move ...
24738 * loop-init.c (fix_loop_structure): ... here.
24739 (apply_loop_flags): Split out from ...
24740 (loop_optimizer_init): ... here.
24741 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24742 in incremental mode, only remove dead loops here.
24744 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24747 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24748 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24749 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24750 (*round<mode>3.libgcc): New insns for fixed-modes.
24751 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24752 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24753 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24754 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24755 implementations. Define to __builtin_avr_absFX,
24756 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24757 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24758 __builtin_avr_countlsFX, respectively.
24759 * config/avr/avr-c.c (target.h): Include it.
24760 (enum avr_builtin_id): New enum.
24761 (avr_resolve_overloaded_builtin): New static function.
24762 (avr_register_target_pragmas): Use it to set
24763 targetm.resolve_overloaded_builtin.
24764 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24765 tree nodes used by DEF_BUILTIN.
24766 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24767 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24768 <AVR_BUILTIN_xxBITS>: Same.
24770 2013-02-08 Richard Biener <rguenther@suse.de>
24772 * cfgloop.c (verify_loop_structure): Properly handle
24773 a loop exiting to another loop header.
24774 * ira-int.h (ira_loops): Remove.
24775 * ira.c (ira_loops): Remove.
24776 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24777 (do_reload): Use loop_optimizer_finalize.
24778 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24779 number_of_loops to access the loop tree.
24780 (more_one_region_p): Likewise.
24781 (finish_loop_tree_nodes): Likewise.
24782 (rebuild_regno_allocno_maps): Likewise.
24783 (mark_loops_for_removal): Likewise.
24784 (mark_all_loops_for_removal): Likewise.
24785 (remove_unnecessary_regions): Likewise.
24786 (ira_build): Likewise.
24787 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24789 2013-02-08 Richard Biener <rguenther@suse.de>
24791 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24792 * ipa-pure-const.c (analyze_function): Avoid calling
24793 mark_irreducible_loops twice.
24794 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24796 2013-02-07 David S. Miller <davem@davemloft.net>
24798 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24800 * var-tracking.c (vt_add_function_parameter): Test the presence of
24801 HAVE_window_save properly and do not remap argument registers when
24802 we have a leaf function.
24804 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24807 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24809 * config/i386/i386.c (ix86_print_operand): Ditto.
24811 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24813 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24815 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24817 PR rtl-optimization/56225
24818 * lra-constraints.c (process_alt_operands): Check that reload hard
24819 reg can hold value for strict_low_part.
24821 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24824 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24825 dwarf2out_end_function.
24826 (in_first_function_p, maybe_at_text_label_p,
24827 first_loclabel_num_not_at_text_label): New variables.
24828 (dwarf2out_var_location): In the first function find out
24829 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24830 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24833 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24835 PR rtl-optimization/56178
24836 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24837 SUBREG of a register. Tidy up related block of code.
24838 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24839 note if the source is a register or a SUBREG of a register.
24841 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24844 * config/rs6000/rs6000.md (ptrm): New mode attr.
24845 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24846 call_value_indirect_aix<pttrsize>,
24847 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24850 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24852 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24853 if -bnortl. Convert to strcmp and strncmp.
24855 2013-02-07 Alan Modra <amodra@gmail.com>
24858 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24859 addresses won't wrap when offsetting.
24860 (rs6000_secondary_reload): Provide secondary reloads needed for
24861 wrapping LO_SUM addresses.
24863 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24865 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24866 MACH, just __MACH__.
24868 2013-02-06 Richard Biener <rguenther@suse.de>
24870 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24871 instead of calling fix_loop_structure.
24873 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24875 PR middle-end/56217
24876 * omp-low.c (use_pointer_for_field): Return false if
24877 lower_send_shared_vars doesn't generate any copy-out code.
24879 2013-02-06 Tom de Vries <tom@codesourcery.com>
24881 PR rtl-optimization/56131
24882 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24883 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24884 of the label is NULL. Add comment.
24886 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24888 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24891 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24892 (STATIC_LIBTSAN_LIBS): Likewise.
24893 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24894 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24895 is defined, don't add anything else beyond that.
24896 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24897 (LINK_COMMAND_SPEC): Use them.
24899 PR tree-optimization/56205
24900 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24901 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24902 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24904 2013-02-05 Richard Biener <rguenther@suse.de>
24906 PR tree-optimization/53342
24907 PR tree-optimization/53185
24908 * tree-vectorizer.h (vect_check_strided_load): Remove.
24909 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24910 not disallow peeling for vectorized strided loads.
24911 (vect_check_strided_load): Make static and simplify.
24912 (vect_analyze_data_refs): Adjust.
24913 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24914 correctly when vectorizing strided loads.
24916 2013-02-05 Richard Biener <rguenther@suse.de>
24918 * doc/install.texi: Refer to ISL, not PPL.
24920 2013-02-05 Jan Hubicka <jh@suse.cz>
24922 PR tree-optimization/55789
24923 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24925 2013-02-05 Jan Hubicka <jh@suse.cz>
24927 PR tree-optimization/55789
24928 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24929 the dead call anyway.
24931 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24934 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24936 2013-02-04 Alexander Potapenko <glider@google.com>
24937 Jack Howarth <howarth@bromo.med.uc.edu>
24938 Jakub Jelinek <jakub@redhat.com>
24941 * config/darwin.c (sort_ctor_records): Stabilized qsort
24942 on constructor priority by using original position.
24943 (finalize_ctors): New routine to sort constructors by
24944 priority before use in assemble_integer.
24945 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24947 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24950 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24951 about visibility on artificial decls.
24952 * config/sol2.c (solaris_assemble_visibility): Likewise.
24954 2013-02-04 Kai Tietz <ktietz@redhat.com>
24957 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24958 argument and improve checking of return-argument types for 16-byte
24960 (ix86_function_value_1): Add additional valtype argument on call
24961 of function_value_64.
24962 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24963 handling infunction_value_64 function.
24965 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24967 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24969 2013-02-04 Richard Biener <rguenther@suse.de>
24971 PR tree-optimization/56188
24972 * tree-ssa-structalias.c (label_visit): Consider case with
24973 initially non-empty points-to set.
24974 (perform_var_substitution): Dump node mapping and clean up.
24976 2013-02-04 Richard Guenther <rguenther@suse.de>
24979 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24980 node prevail as last resort.
24981 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24982 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24984 2013-02-04 Richard Biener <rguenther@suse.de>
24986 PR tree-optimization/56113
24987 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24989 (equiv_class_lookup_or_add): ... this.
24990 (label_visit): Adjust and fix error in previous patch.
24991 (perform_var_substitution): Adjust.
24993 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24995 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24996 * config/sh/sh4-300.md: Likewise.
24997 * config/sh/sh4a.md: Likewise.
24998 * config/sh/constraints.md: Likewise.
24999 * config/sh/sh.md: Likewise.
25000 * config/sh/netbsd-elf.h: Likewise.
25001 * config/sh/predicates.md: Likewise.
25002 * config/sh/sh-protos.h: Likewise.
25003 * config/sh/ushmedia.h: Likewise.
25004 * config/sh/linux.h: Likewise.
25005 * config/sh/sh.c: Likewise.
25006 * config/sh/superh.h: Likewise.
25007 * config/sh/elf.h: Likewise.
25008 * config/sh/sh4.md: Likewise.
25009 * config/sh/sh.h: Likewise.
25011 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25013 * config/pa/constraints.md: Adjust unused letters. Change "T"
25014 constraint to match_test floating_point_store_memory_operand().
25015 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25016 (base14_operand): New.
25017 (floating_point_store_memory_operand): New.
25018 (integer_store_memory_operand): Revise to use base14_operand and
25019 reg_plus_base_memory_operand.
25020 (move_dest_operand): Allow symbolic_memory_operands.
25021 (symbolic_memory_operand): Check for LO_SOM.
25022 (symbolic_operand): Change default case to break.
25023 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25024 CONST_DOUBLE values to be reloaded by putting them into memory when
25025 the destination is a floating point register.
25026 (movdf): Remove code to handle CONST_DOUBLE.
25028 (reload_indf_r1): New.
25029 (reload_insf_r1): New.
25030 Consistently use "Q" and "T" constraints with integer and floating
25031 point move instructions, respectively.
25032 (movdi): Remove FAIL.
25033 Change predicate for source operand unamed DImode move from
25034 general_operand to move_src_operand.
25035 (umulsidi3): Change predicate for destination operand to
25037 Likewise for similar unamed patterns.
25038 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25039 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25040 (hppa_legitimize_address): Simplify mask calculation.
25041 (pa_emit_move_sequence): Revised handling of secondary reloads from
25042 REG+D addresses for floating point loads and stores. Directly handle
25043 loading CONST0_RTX (mode) to a floating point register.
25044 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25045 to floating point registers. Don't restrict secondary reloads to
25046 floating point registers to integer modes. Revise some comments and
25048 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25049 (pa_legitimate_address_p): New.
25050 (pa_legitimize_reload_address): New.
25051 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25052 (STRICT_REG_OK_FOR_BASE_P): New.
25053 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25054 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25056 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25057 Andrew Dixie <andrewd@gentrack.com>
25059 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25062 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25064 * expmed.c (extract_bit_field_1): Pass the full width of the
25065 structure to get_best_reg_extraction_insn.
25067 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25070 * configure.ac (use_cxa_atexit): Add AIX.
25071 * configure: Regenerate.
25073 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25075 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25078 * final.c (need_profile_function): New variable.
25079 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25080 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25081 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25082 notes, targetm.asm_out.function_prologue doesn't emit anything,
25083 HAVE_prologue and profiler should be emitted before prologue,
25084 set need_profile_function instead of emitting it.
25085 (final_scan_insn): If need_profile_function, emit
25086 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25087 NOTE_INSN_FUNCTION_BEG note.
25089 2013-02-01 Richard Henderson <rth@redhat.com>
25091 * config/rs6000/rs6000.md (smulditi3): New.
25094 * config/alpha/alpha.md (umulditi3): New.
25096 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25098 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25099 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25101 2013-02-01 Richard Biener <rguenther@suse.de>
25103 PR tree-optimization/56113
25104 * tree-ssa-structalias.c (label_visit): Reduce work for
25105 single-predecessor nodes.
25107 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25109 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25110 range isn't testing for zero.
25112 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25114 PR middle-end/56113
25115 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25117 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25118 Nick Clifton <nickc@redhat.com>
25120 * config/v850/constraints.md (Q): Define as a memory constraint.
25121 * config/v850/predicates.md (label_ref_operand): New predicate.
25122 (e3v5_shift_operand): New predicate.
25123 (ior_operator): New predicate.
25124 * config/v850/t-v850: Add e3v5 multilib.
25125 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25126 (v850_gen_movdi): Prototype.
25127 * config/v850/v850.c: Add support for e3v5 architecture.
25128 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25130 (construct_save_jarl): Add e3v5 long JARL support.
25131 (v850_adjust_insn_length): New function. Adjust length of call
25132 insns when using e3v5 instructions.
25133 (v850_gen_movdi): New function: Generate instructions to move a
25135 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25136 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25137 (TARGET_USE_FPU): Enable for e3v5.
25138 (CONST_OK_FOR_W): New macro.
25139 (ADJUST_INSN_LENGTH): Define.
25140 * config/v850/v850.md (UNSPEC_LOOP): Define.
25141 (attr cpu): Add v850e3v5.
25142 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25143 (movdi): New pattern.
25144 (movdi_internal): New pattern.
25145 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25146 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25147 (cstoresf4): Likewise.
25148 (cstoredf4): Likewise.
25149 (insv): New pattern.
25150 (rotlso3_a): New pattern.
25151 (rotlsi3_b): New pattern
25152 (rotlsi3_v850e3v5): New pattern.
25153 (doloop_begin): New pattern.
25154 (fix_loop_counter): New pattern.
25155 (doloop_end): New pattern.
25156 (branch_normal): Add e3v5 long branch support.
25157 (branch_invert): Likewise.
25158 (branch_z_normal): Likewise.
25159 (branch_z_invert): Likewise.
25160 (branch_nz_normal): Likewise.
25161 (branch_nz_invert): Likewise.
25162 (call_internal_short): Add e3v5 register-indirect JARL support.
25163 (call_internal_long): Likewise.
25164 (call_value_internal_short): Likewise.
25165 (call_value_internal_long): Likewise.
25166 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25167 (mloop): New option.
25168 * config.gcc: Add support for configuring v840e3v5 target.
25169 * doc/invoke.texi: Document new v850 specific command line options.
25171 2013-01-31 Paul Koning <ni1d@arrl.net>
25175 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25176 children if parent is a class.
25177 (prune_unused_types_prune): Don't add DW_AT_declaration.
25179 2013-01-31 Richard Biener <rguenther@suse.de>
25181 PR tree-optimization/56157
25182 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25183 match up operand with SLP child.
25185 2013-01-31 Jason Merrill <jason@redhat.com>
25188 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25189 parameters the first time.
25190 (gen_scheduled_generic_parms_dies): Check completeness here.
25192 2013-01-31 Richard Biener <rguenther@suse.de>
25194 PR middle-end/53073
25195 * common.opt (faggressive-loop-optimizations): New flag,
25196 enabled by default.
25197 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25198 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25199 infer_loop_bounds_from_undefined by it.
25201 2013-01-31 Richard Biener <rguenther@suse.de>
25203 PR tree-optimization/56150
25204 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25205 visit virtual operands.
25206 (find_uses_to_rename_bb): Likewise.
25208 2013-01-31 Richard Biener <rguenther@suse.de>
25210 PR tree-optimization/56150
25211 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25212 mixed store non-store stmts.
25214 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25217 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25218 LIBASAN_EARLY_SPEC is defined.
25219 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25220 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25222 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25225 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25226 invalid args instead of ICEing on it.
25227 (ix86_valid_target_attribute_tree): Return error_mark_node if
25228 ix86_valid_target_attribute_inner_p failed.
25229 (ix86_valid_target_attribute_p): Return false only if
25230 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25231 target("default") attribute.
25232 (sorted_attr_string): Change argument from const char * to tree,
25233 merge in all target attribute arguments rather than just one.
25234 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25235 instead of free. Avoid using strcat.
25236 (ix86_mangle_function_version_assembler_name): Mangle
25237 target("default") as if no target attribute is present. Adjust
25238 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25239 instead of xmalloc and XDELETEVEC instead of free.
25240 (ix86_function_versions): Don't return true if one of the decls
25241 doesn't have target attribute. If they don't and one of the decls
25242 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25243 sorted_attr_string caller. Use XDELETEVEC instead of free.
25244 (ix86_supports_function_versions): Remove.
25245 (make_name): Fix up formatting.
25246 (make_dispatcher_decl): Remove resolver_name and its initialization.
25247 Avoid leaking memory.
25248 (is_function_default_version): Return true if there is
25249 target("default") attribute rather than no target attribute at all.
25250 (make_resolver_func): Avoid leaking memory.
25251 (ix86_generate_version_dispatcher_body): Likewise.
25252 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25253 * target.def (supports_function_versions): Remove.
25254 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25255 * doc/tm.texi: Regenerated.
25257 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25259 PR rtl-optimization/56144
25260 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25261 for values with side effects.
25263 2013-01-30 Richard Biener <rguenther@suse.de>
25265 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25266 (sparseset_pop): Likewise.
25267 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25268 to be able to use quick_push in the worker loop.
25270 2013-01-30 Marek Polacek <polacek@redhat.com>
25272 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25274 2013-01-30 Richard Biener <rguenther@suse.de>
25277 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25279 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25281 PR tree-optimization/56064
25282 * fixed-value.c (fixed_from_double_int): New function.
25283 * fixed-value.h (fixed_from_double_int): New prototype.
25284 (const_fixed_from_double_int): New static inline function.
25285 * fold-const.c (native_interpret_fixed): New static function.
25286 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25287 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25288 (native_encode_fixed): New static function.
25289 (native_encode_expr) <FIXED_CST>: Use it.
25290 (native_interpret_int): Move double_int worker code to...
25291 * double-int.c (double_int::from_buffer): ...this new static method.
25292 * double-int.h (double_int::from_buffer): Prototype it.
25294 2013-01-30 Richard Biener <rguenther@suse.de>
25296 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25297 New pointer-map and obstack.
25298 (init_alias_vars): Allocate pointer-map and obstack.
25299 (delete_points_to_sets): Free them.
25300 (find_what_var_points_to): Cache result.
25301 (find_what_p_points_to): Adjust for changed interface of
25302 find_what_var_points_to.
25303 (compute_points_to_sets): Likewise.
25304 (ipa_pta_execute): Likewise.
25306 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25308 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25309 * configure: Regenerate.
25310 * config.in: Regenerate.
25311 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25312 #nobits/#progbits if supported.
25314 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25317 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25318 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25319 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25321 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25323 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25324 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25326 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25328 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25329 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25331 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25333 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25335 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25336 * config/arm/cortex-a7.md: New bypasses using
25337 arm_mac_accumulator_is_result.
25339 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25341 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25342 (cortex_a7_neon_mla): Likewise.
25343 (cortex_a7_fpfmad): New reservation.
25344 (cortex_a7_fpmacs): Use ffmas and update required units.
25345 (cortex_a7_fpmuld): Update required units and latency.
25346 (cortex_a7_fpmacd): Likewise.
25347 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25348 (cortex_a7_neon). Likewise.
25349 (bypass) Update participating units.
25351 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25353 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25354 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25356 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25357 (vfp_fmul): Use ffmad.
25358 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25359 (cortex_r4_fmacd): Use ffmad.
25360 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25361 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25362 (cortex_a9_fmacd): Use ffmad.
25363 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25364 (cortex_a8_vfp_macd): Use ffmad.
25365 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25366 (cortex_a5_fpmacd): Use ffmad.
25367 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25368 (cortex_a15_vfp_macd): Use ffmad.
25369 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25371 2013-01-29 Jason Merrill <jason@redhat.com>
25374 * varasm.c (default_assemble_visibility): Don't warn about
25375 visibility on artificial decls.
25377 2013-01-29 Richard Biener <rguenther@suse.de>
25379 PR tree-optimization/56113
25380 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25382 (label_visit): Free duplicate bitmaps and record the leader instead.
25383 (perform_var_substitution): Adjust.
25385 2013-01-29 Richard Biener <rguenther@suse.de>
25387 PR tree-optimization/55270
25388 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25389 the CFG, schedule loops for fixup.
25391 2013-01-29 Nick Clifton <nickc@redhat.com>
25393 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25396 2013-01-28 Leif Ekblad <leif@rdos.net>
25398 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25399 * config/i386/i386.h (TARGET_RDOS): New macro.
25400 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25401 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25402 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25403 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25404 DEFAULT_LARGE_SECTION_THRESHOLD.
25405 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25406 * config/i386/rdos.h: New file.
25407 * config/i386/rdos64.h: New file.
25409 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25412 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25415 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25417 PR rtl-optimization/56117
25418 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25419 call cselib_lookup_from_insn on the MEM before calling
25420 add_insn_mem_dependence.
25422 2013-01-28 Richard Biener <rguenther@suse.de>
25424 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25425 to a stmt that didn't have one.
25426 (copy_phis_for_bb): Likewise for PHI arguments.
25427 (copy_debug_stmt): Likewise for debug stmts.
25429 2013-01-28 Richard Biener <rguenther@suse.de>
25431 PR tree-optimization/56034
25432 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25433 (partition_builtin_p): Adjust.
25434 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25435 it is the last partition.
25436 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25437 up the vertex for the definition.
25438 (classify_partition): Classify whether a partition is a
25439 PKIND_REDUCTION, thus has uses outside of the loop.
25440 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25441 Merge all PKIND_REDUCTION partitions into the last partition.
25442 (tree_loop_distribution): Seed partitions from reductions as well.
25444 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25446 PR tree-optimization/56125
25447 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25448 pow(x,c) into sqrt(x) * powi(x, n/2) or
25449 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25450 optimizing for size.
25451 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25452 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25455 PR tree-optimization/56094
25456 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25457 to UNKNOWN_LOCATION while gimplifying expr.
25459 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25462 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25463 operand 0 in movabs insn template for -masm=intel asm alternative.
25464 (*movabs<mode>_2): Ditto for operand 1.
25466 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25469 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25472 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25474 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25477 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25479 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25480 Undef to avoid warning.
25482 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25484 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25485 * configure: Regenerate.
25487 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25489 PR tree-optimization/56098
25490 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25491 for stmts with volatile ops.
25492 (cond_store_replacement): Don't optimize if assign has volatile ops.
25493 (cond_if_else_store_replacement_1): Don't optimize if either
25494 then_assign or else_assign have volatile ops.
25495 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25498 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25500 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25502 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25504 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25505 missing ':' in asm example.
25507 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25509 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25510 entries into lane and laneq entries.
25511 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25512 Remove AdvSIMD scalar modes.
25513 (aarch64_sq<r>dmulh_laneq<mode>): New.
25514 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25516 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25517 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25518 * config/aarch64/iterators.md (VCOND): New.
25521 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25524 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25525 Add NULL LIBNAME argument to existing definitions.
25526 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25527 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25528 * config/avr/avr.c (DEF_BUILTIN): Same.
25529 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25530 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25531 implementation is available (DECL_ASSEMBLER_NAME is set).
25532 (avr_fold_absfx): New static function.
25533 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25534 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25535 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25536 AVR_BUILTIN_ABSLLK.
25537 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25538 (abshk, absk, abslk, absllk): Provide as static inline functions.
25540 2013-01-25 Marek Polacek <polacek@redhat.com>
25542 PR tree-optimization/56035
25543 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25545 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25547 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25548 (*movtf_internal_rex64): Add (!o,C) alternative
25549 (*movxf_internal_rex64): Ditto.
25550 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25552 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25554 * doc/invoke.texi: fix typo.
25555 * doc/objc.texi: fix typo.
25557 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25559 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25560 for the first two alternatives.
25562 2013-01-24 Diego Novillo <dnovillo@google.com>
25564 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25565 (ggc-zone.o): Remove.
25566 * configure.ac: Remove option --with-gc.
25567 * configure: Re-generate.
25568 * doc/install.texi: Remove documentation for --with-gc.
25569 * gengtype.c (write_enum_defn): Remove. Update all users.
25570 (write_Types_process_field): Remove generation of gt_e_* argument.
25571 (output_type_enum): Remove. Update all users.
25572 (write_enum_defn): Remove. Update all users.
25573 (enum alloc_zone): Remove. Update all users.
25574 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25575 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25576 Update all callers.
25577 (struct ptr_data): Remove field TYPE. Update all users.
25578 (gt_pch_note_object): Remove argument TYPE. Update all users.
25579 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25581 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25582 (struct alloc_zone): Remove.
25583 (ggc_internal_alloc_zone_stat): Remove.
25584 (ggc_internal_cleared_alloc_zone_stat): Remove.
25585 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25586 (ggc_pch_count_object): Remove last argument. Update all users.
25587 (ggc_pch_alloc_object): Remove last argument. Update all users.
25588 (struct alloc_zone): Remove.
25589 * ggc-zone.c: Remove.
25590 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25591 (struct alloc_zone): Remove.
25592 (ggc_alloc_typed_stat): Remove.
25593 (ggc_alloc_typed): Remove.
25594 (ggc_splay_alloc): Remove first argument.
25595 (rtl_zone): Remove. Update all users.
25596 (tree_zone): Remove. Update all users.
25597 (tree_id_zone): Remove. Update all users.
25598 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25599 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25600 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25601 * tree-ssanames.c: Remove references to zone allocator in comments.
25603 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25605 * config/avr/avr.c (avr_out_fract): Make register numbers that
25606 might be outside of source operand signed.
25608 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25610 * config/i386/constraints.md (Yf): New constraint.
25611 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25612 of f constraint to conditionaly disable x87 register preferences.
25613 (*movdf_internal): Ditto.
25614 (*movsf_internal): Ditto.
25616 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25618 PR inline-asm/55934
25619 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25620 that have operands with impossible constraints.
25621 Add a FIXME for a speed-up opportunity.
25622 * lra-constraints.c (process_alt_operands): Verify that a class
25623 selected from constraints on asms is valid for the operand mode.
25624 (curr_insn_transform): Remove incorrect comment.
25626 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25628 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25629 TOC operand is a valid symbol ref in the constant pool.
25631 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25633 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25635 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25638 * config/avr/stdfix.h: New file.
25639 * t-avr (stdfix-gcc.h): New rule to build it.
25640 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25642 2013-01-23 Kostya Serebryany <kcc@google.com>
25644 * config/darwin.h: remove dependency on
25645 CoreFoundation (asan on Mac OS).
25647 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25650 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25651 instead of cmpdi_operand for first comparison operand.
25652 Don't assert that comparison operands aren't both constants.
25654 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25656 * doc/install.texi (Downloading the Source): Update references to
25657 downloading separate components.
25659 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25661 * doc/extend.texi (__int128): Improve grammar.
25663 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25666 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25667 alternative to (o,r).
25668 (*movdi_internal_rex64): Remove (!o,n) alternative.
25669 (DImode immediate->memory splitter): Remove.
25670 (DImode immediate->memory peephole2): Remove.
25671 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25672 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25673 alternative to (!o,*r).
25674 (*movtf_internal_sse): New pattern.
25675 (*movxf_internal_rex64): New pattern.
25676 (*movxf_internal): Disable for TARGET_64BIT.
25677 (*movdf_internal_rex64): Remove (!o,F) alternative.
25679 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25681 PR middle-end/56074
25682 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25683 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25684 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25685 stmt locations where LOCATION_LOCUS of the stmt location is
25686 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25689 * config/i386/i386.md (UNSPEC_STOS): New.
25690 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25691 *strsetqi_1): Add UNSPEC_STOS.
25693 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25696 * doc/invoke.texi: Remove left over -Wsynth example.
25698 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25700 PR tree-optimization/56051
25701 * fold-const.c (fold_binary_loc): Don't fold
25702 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25703 a narrowing conversion, or widening conversion from signed
25706 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25708 PR rtl-optimization/56023
25709 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25710 dependent on debug instruction.
25712 2013-01-21 Martin Jambor <mjambor@suse.cz>
25714 PR middle-end/56022
25715 * function.c (allocate_struct_function): Call
25716 invoke_set_current_function_hook earlier.
25718 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25720 * reload1.c (init_reload): Only initialize reload_obstack
25721 during the first call.
25723 2013-01-21 Marek Polacek <polacek@redhat.com>
25725 * cfgloop.c (verify_loop_structure): Fix up grammar.
25727 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25729 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25730 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25732 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25735 * config/arm/marvell-pj4.md: Update copyright year.
25736 Fix up use of alu to alu_reg and simple_alu_imm.
25738 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25740 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25742 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25745 * lra-constraints.c (curr_insn_transform): Don't reuse original
25746 insn for secondary memory move when memory mode should be different.
25748 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25750 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25751 atomic_storedi_1): New patterns.
25753 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25755 btver2 pipeline descriptions.
25756 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25758 * config/i386/i386.md (btver2_decode): New type attributes.
25759 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25761 * config/i386/btver2.md: New file describing btver2 pipelines.
25763 2013-01-19 Andrew Pinski <apinski@cavium.com>
25765 PR tree-optimization/52631
25766 * tree-ssa-sccvn (visit_use): Before looking up the original
25767 statement, try looking up the simplified expression.
25769 2013-01-19 Anthony Green <green@moxielogic.com>
25771 * config/moxie/moxie.c (moxie_expand_prologue): Set
25772 current_function_static_stack_size.
25774 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25776 PR tree-optimization/56029
25777 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25778 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25780 2013-01-18 Sharad Singhai <singhai@google.com>
25782 PR tree-optimization/55995
25783 * dumpfile.c (dump_loc): Print location only if available.
25784 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25786 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25789 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25790 secondary memory move.
25791 (inherit_reload_reg): Use rclass instead of cl for
25792 check_secondary_memory_needed_p.
25794 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25796 PR middle-end/56015
25797 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25798 the case where writing real complex part of target modifies op1.
25800 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25802 * config/aarch64/aarch64-simd.md
25803 (aarch64_vcond_internal<mode>): Handle unordered cases.
25804 * config/aarch64/iterators.md (v_cmp_result): New.
25806 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25807 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25809 * config/arm/marvell-pj4.md: New file.
25810 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25811 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25812 (generic_vfp): Likewise.
25813 * config/arm/arm-cores.def: Add marvell-pj4.
25814 * config/arm/arm-tune.md: Regenerate.
25815 * config/arm/arm-tables.opt: Regenerate.
25816 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25817 * doc/invoke.texi: Document marvell-pj4.
25819 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25821 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25823 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25828 * var-tracking.c (negative_power_of_two_p): New.
25829 (global_get_addr_cache, local_get_addr_cache): New.
25830 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25831 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25833 (vt_stack_offset_p): Remove.
25834 (vt_canon_true_dep): Always canonicalize loc's address.
25835 (clobber_overlapping_mems): Make sure we have a MEM.
25836 (local_get_addr_clear_given_value): New.
25837 (val_reset): Clear local cached entries.
25838 (compute_bb_dataflow): Create and release the local cache.
25839 Disable duplicate MEMs clobbering.
25840 (emit_notes_in_bb): Clobber MEMs likewise.
25841 (vt_emit_notes): Create and release the local cache.
25842 (vt_initialize, vt_finalize): Create and release the global
25843 cache, respectively.
25844 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25846 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25848 PR libmudflap/53359
25849 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25850 not found in the symtab.
25852 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25855 PR rtl-optimization/55547
25856 PR rtl-optimization/53827
25859 * alias.c (offset_overlap_p): New, factored out of...
25860 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25861 the conservative special case for symbolic constants. Don't
25862 adjust zero sizes on alignment.
25864 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25866 PR rtl-optimization/52573
25867 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25868 REG_UNUSED for the same register.
25870 2013-01-17 Richard Biener <rguenther@suse.de>
25871 Marek Polacek <polacek@redhat.com>
25873 PR rtl-optimization/55833
25874 * loop-unswitch.c (unswitch_loops): Move loop verification...
25875 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25876 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25877 Set it to true when we're removing a loop from hierarchy tree in
25878 an irreducible region.
25879 (fix_bb_placements): Adjust caller.
25880 (fix_loop_placements): Likewise.
25882 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25884 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25885 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25886 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25887 Remove ID. Adjust comments.
25888 * config/avr/avr-c.c (avr_builtin_name): Remove.
25889 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25890 * config/avr/avr.c (avr_tolower): New static function.
25891 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25892 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25893 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25896 2013-01-17 Jan Hubicka <jh@suse.cz>
25898 PR tree-optimization/55273
25899 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25901 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25904 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25905 store through atomic_store<mode>_1.
25906 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25908 2013-01-17 Martin Jambor <mjambor@suse.cz>
25910 PR tree-optimizations/55264
25911 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25912 for virtual methods.
25913 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25914 virtual methods before inlining is over.
25915 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25917 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25920 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25922 PR rtl-optimization/56005
25923 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25924 pending reads for prefetch.
25926 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25928 * config/aarch64/aarch64.md
25929 (*cstoresi_neg_uxtw): New pattern.
25930 (*cmovsi_insn_uxtw): New pattern.
25931 (*<optab>si3_uxtw): New pattern.
25932 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25933 (*<optab>si3_insn_uxtw): New pattern.
25934 (*bswapsi2_uxtw): New pattern.
25936 2013-01-16 Richard Biener <rguenther@suse.de>
25938 * tree-inline.c (tree_function_versioning): Remove set but
25939 never used variable.
25941 2013-01-16 Richard Biener <rguenther@suse.de>
25943 PR tree-optimization/55964
25944 * tree-flow.h (rename_variables_in_loop): Remove.
25945 (rename_variables_in_bb): Likewise.
25946 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25947 (copy_loop_before): Adjust and delete update-ssa status.
25948 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25949 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25950 (rename_variables_in_loop): Remove.
25951 (slpeel_update_phis_for_duplicate_loop): Likewise.
25952 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25953 use available cfg machinery instead of duplicating it.
25954 Update PHI nodes and perform poor-mans SSA update here.
25955 (slpeel_tree_peel_loop_to_edge): Adjust.
25957 2013-01-16 Richard Biener <rguenther@suse.de>
25959 PR tree-optimization/54767
25960 PR tree-optimization/53465
25961 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25962 (vrp_visit_phi_node): For PHI arguments coming via backedges
25963 drop all symbolical range information.
25964 (execute_vrp): Compute backedges.
25966 2013-01-16 Richard Biener <rguenther@suse.de>
25968 * doc/install.texi: Update CLooG and ISL requirements to
25971 2013-01-16 Christian Bruel <christian.bruel@st.com>
25974 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25975 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25976 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25978 2013-01-16 DJ Delorie <dj@redhat.com>
25980 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25981 (UNSPECV_SP_SWITCH_E): New.
25982 (sp_switch_1): Change to an unspec.
25983 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25986 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25988 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25989 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25990 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25991 (expand_mem_thread_fence): Ditto.
25992 (expand_mem_signal_fence): Ditto.
25993 (expand_atomic_load): Ditto.
25994 (expand_atomic_store): Ditto.
25996 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25998 PR rtl-optimization/55547
25999 PR rtl-optimization/53827
26002 * alias.c (memrefs_conflict_p): Set sizes to negative after
26005 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26008 * function.c (thread_prologue_and_epilogue_insns): Always
26009 add crtl->drap_reg to set_up_by_prologue.set, even if
26010 stack_realign_drap is false.
26012 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26014 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26015 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26016 *call): Fix indention.
26018 2013-01-15 Tom de Vries <tom@codesourcery.com>
26021 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26024 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26026 PR rtl-optimization/55153
26027 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26029 2013-01-15 Martin Jambor <mjambor@suse.cz>
26031 PR tree-optimization/55920
26032 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26033 accesses as grp_to_be_debug_replaced.
26035 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26037 PR tree-optimization/55920
26038 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26039 there is non-useless type conversion needed from debug rhs to lhs,
26040 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26042 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26043 Mikael Pettersson <mikpe@it.uu.se>
26046 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26048 (ASM_OUTPUT_CASE_LABEL): Remove.
26049 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26050 * final.c (shorten_branches): Update alignment of labels before
26051 jump tables if CASE_VECTOR_SHORTEN_MODE.
26053 2013-01-15 Richard Biener <rguenther@suse.de>
26056 * system.h: Do not include gmp.h for building host tools.
26058 2013-01-15 Richard Biener <rguenther@suse.de>
26060 PR middle-end/55882
26061 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26062 account for bitpos when computing alignment.
26064 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26066 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26067 (ix86_target_macros_internal): Likewise.
26069 * config/i386/i386.c (m_CORE2I7): Removed.
26070 (m_CORE_HASWELL): New macro.
26071 (m_CORE_ALL): Likewise.
26072 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26073 (initial_ix86_arch_features): Likewise.
26074 (processor_target_table): Initializations for Core avx2.
26075 (cpu_names): New names "core-avx2".
26076 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26077 PROCESSOR_CORE_HASWELL.
26078 (ix86_issue_rate): New case.
26079 (ia32_multipass_dfa_lookahead): Likewise.
26080 (ix86_sched_init_global): Likewise.
26082 * config/i386/i386.h (TARGET_HASWELL): New macro.
26083 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26084 (processor_type): New PROCESSOR_HASWELL.
26086 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26088 PR tree-optimization/55955
26089 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26090 *SHIFT_EXPR and *ROTATE_EXPR codes.
26092 PR tree-optimization/48766
26093 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26094 -ftrapv disable -fwrapv.
26096 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26099 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26100 etc. to 1 and not to __flash.
26101 Use LL suffix for __INT24_MAX__ with -mint8.
26102 Use ULL suffix for __UINT24_MAX__ with -mint8.
26104 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26106 * config/avr/avr-arch.h
26107 (struct base_arch_s): Use typedef avr_arch_t instead.
26108 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26109 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26110 * config/avr/avr.c: Same.
26111 * config/avr/avr-devices.c: Same.
26112 * config/avr/driver-avr.c: Same.
26113 * config/avr/gen-avr-mmcu-texi.c: Same.
26114 * config/avr/avr-mcus.def: Adjust comment.
26116 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26118 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26119 * config/aarch64/iterators.md (VALLDI): New.
26121 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26122 Andi Kleen <ak@linux.intel.com>
26125 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26126 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26129 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26131 * config/avr/avr-stdint.h: Remove trailing blanks.
26132 * config/avr/avr-log.h: Same.
26133 * config/avr/avr-arch.h: Same.
26134 * config/avr/avr-devices.c: Same.
26135 * config/avr/avr-dimode.md: Same.
26136 * config/avr/predicates.md: Same.
26137 * config/avr/avr-c.c: Same. And fix typo.
26139 * config/avr/avr-protos.h: Same. And:
26140 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26141 (init_cumulative_args): Rename to avr_init_cumulative_args.
26142 (expand_prologue): Rename to avr_expand_prologue.
26143 (expand_epilogue): Rename to avr_expand_epilogue.
26144 (adjust_insn_length): Rename to avr_adjust_insn_length.
26145 (notice_update_cc): Rename to avr_notice_update_cc.
26146 (final_prescan_insn): Rename to avr_final_prescan_insn.
26147 * config/avr/avr.c: Same.
26148 * config/avr/avr.h: Same.
26149 * config/avr/avr.md: Remove trailing blanks.
26150 (prologue): Use avr_expand_prologue.
26151 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26153 2013-01-14 Richard Biener <rguenther@suse.de>
26155 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26156 verify_location, collect_subblocks): New functions.
26157 (verify_gimple_in_cfg): Verify that locations only reference
26158 BLOCKs in the functions BLOCK tree.
26160 2013-01-14 Richard Biener <rguenther@suse.de>
26162 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26164 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26166 (insert_out_of_ssa_copy_on_edge): Likewise.
26167 (rewrite_close_phi_out_of_ssa): Likewise.
26168 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26170 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26171 propagated constants.
26172 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26175 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26177 * config/avr/avr-modes.def: Add GPL copyright notice.
26179 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26181 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26182 MEMMODEL_MASK to determine memory model.
26183 (atomic_store<mode>): Ditto from operands[2].
26184 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26186 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26189 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26190 (fold_gimple_assign): Don't call unshare_expr here.
26191 (fold_ctor_reference): Call unshare_expr.
26193 2013-01-13 Terry Guo <terry.guo@arm.com>
26195 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26196 * doc/fragments.texi: Document MULTILIB_REUSE.
26197 * gcc.c (multilib_reuse): New internal spec.
26198 (set_multilib_dir): Also search multilib from multilib_reuse.
26199 * genmultilib (tmpmultilib3): Refactor code.
26200 (tmpmultilib4): Ditto.
26201 (multilib_reuse): New multilib argument.
26203 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26205 * Makefile.in: Update copyright.
26207 2013-01-12 Tom de Vries <tom@codesourcery.com>
26209 PR middle-end/55890
26210 * calls.c (expand_call): Check if arg_nr is valid.
26212 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26214 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26215 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26216 documentation. Add missing '__' in front of
26217 __builtin_ia32_packssdw256.
26219 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26222 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26223 NO_REGS for larl operands.
26224 (s390_reload_larl_operand): Use s390_load_address instead of
26227 2013-01-11 Richard Biener <rguenther@suse.de>
26229 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26230 (verify_node_sharing): ... here.
26231 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26233 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26235 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26236 Substitute TREECHECKING.
26237 * configure: Regenerate.
26238 * Makefile.in (TREECHECKING): New.
26240 2013-01-11 Richard Guenther <rguenther@suse.de>
26242 PR tree-optimization/44061
26243 * tree-vrp.c (extract_range_basic): Compute zero as
26244 value-range for __builtin_constant_p of function parameters.
26246 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26248 Update copyright years.
26250 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26252 PR rtl-optimization/55672
26253 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26254 const to be eliminable.
26256 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26258 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26259 * configure: Regenerate.
26261 2013-01-10 Richard Biener <rguenther@suse.de>
26263 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26265 2013-01-10 Richard Biener <rguenther@suse.de>
26268 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26269 locations for virtual PHI arguments.
26270 (rewrite_update_phi_arguments): Likewise.
26272 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26274 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26277 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26279 PR tree-optimization/55921
26280 * tree-complex.c (expand_complex_asm): New function.
26281 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26283 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26286 * config/s390/s390.c (s390_symref_operand_p)
26287 (s390_loadrelative_operand_p): Merge the two functions.
26288 (s390_check_qrst_address, print_operand_address): Add parameters
26289 to s390_loadrelative_operand_p invokation.
26290 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26291 (s390_reload_larl_operand, s390_secondary_reload): Use
26292 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26293 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26295 2013-01-09 Mike Stump <mikestump@comcast.net>
26297 * dse.c (record_store): Remove unnecessary assert.
26299 2013-01-09 Jan Hubicka <jh@suse.cz>
26301 PR tree-optimization/55569
26302 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26303 * cfgloop.h (scale_loop_profile): Likewise.
26305 2013-01-09 Jan Hubicka <jh@suse.cz>
26308 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26310 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26312 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26314 PR middle-end/55114
26315 * expr.h (maybe_emit_group_store): Declare.
26316 * expr.c (maybe_emit_group_store): New function.
26317 * builtins.c (expand_builtin_int_roundingfn): Call it.
26318 (expand_builtin_int_roundingfn_2): Likewise.
26320 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26322 PR rtl-optimization/55829
26323 * lra-constraints.c (match_reload): Add code for absent output.
26324 (curr_insn_transform): Add code for reloads of matched inputs
26327 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26329 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26330 attribute of movddup insn to DF.
26331 (*vec_interleave_lowv2df): Ditto.
26332 (vec_dupv2df): Ditto.
26334 2013-01-09 Jan Hubicka <jh@suse.cz>
26336 PR tree-optimiation/55875
26337 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26338 EVERY_ITERATION parameter.
26339 (number_of_iterations_exit): Check if exit is executed every iteration.
26340 (idx_infer_loop_bounds): Similarly here.
26341 (n_of_executions_at_most): Simplify
26342 to only test for cases where statement is dominated by the
26343 particular bound; handle correctly the "postdominance" test.
26344 (scev_probably_wraps_p): Use max loop iterations info
26345 as a global bound first.
26347 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26348 Nick Clifton <nickc@redhat.com>
26350 * config/v850/v850.md (cbranchsf4): New pattern.
26351 (cstoresf4): New pattern.
26352 (cbranchdf4): New pattern.
26353 (cstoredf4): New pattern.
26354 (movsicc): Disallow floating point comparisons.
26355 (cmpsf_le_insn): Fix order of operators.
26356 (cmpsf_lt_insn): Likewise.
26357 (cmpsf_eq_insn): Likewise.
26358 (cmpdf_le_insn): Likewise.
26359 (cmpdf_lt_insn): Likewise.
26360 (cmpdf_eq_insn): Likewise.
26361 (cmpsf_ge_insn): Use LE comparison.
26362 (cmpdf_ge_insn): Likewise.
26363 (cmpsf_gt_insn): Use LT comparison.
26364 (cmpdf_gt_insn): Likewise.
26365 (cmpsf_ne_insn): Delete pattern.
26366 (cmpdf_ne_insn): Delete pattern.
26367 * config/v850/v850.c (v850_gen_float_compare): Use
26368 gen_cmpdf_eq_insn for NE comparison.
26369 (v850_float_z_comparison_operator)
26370 (v850_float_nz_comparison_operator): Move from here ...
26371 * config/v850/predicates.md: ... to here. Move GT and GE
26372 comparisons into v850_float_z_comparison_operator.
26373 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26375 (v850_float_nz_comparison_operator): Likewise.
26377 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26379 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26380 with calls to gen_insvsi/gen_insvdi.
26382 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26384 * config/i386/i386.c (initial_ix86_tune_features): Set up
26385 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26387 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26388 Jakub Jelinek <jakub@redhat.com>
26390 PR tree-optimization/48189
26391 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26392 If nitercst is 0, don't predict the exit edge.
26394 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26396 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26397 in asm_fprintf with reg_names.
26398 (aarch64_print_operand_address): Likewise.
26399 (aarch64_return_addr): Likewise.
26400 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26402 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26404 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26405 (INT_U6_BITS): Likewise.
26406 * config/pa/predicates.md (uint6_operand): New predicate.
26407 (shift5_operand, shift6_operand): Likewise.
26408 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26410 (lshrdi3): Use shift6_operand.
26411 (shrpsi4, shrpdi4): New insn patterns.
26412 (extzv): Delete expander.
26413 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26414 predicates in unamed zero extract patterns. Tighten common constraint.
26415 (extv): Delete expander.
26416 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26417 predicates in unamed sign extract patterns. Tighten common constraint.
26418 (insv): Delete expander.
26419 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26420 predicates in unamed insert patterns. Tighten common constraint.
26421 Change uint32_operand predicate to uint6_operand predicate in unamed
26422 DImode pattern to insert constant values of type 1...1xxxx.
26424 2013-01-04 Jan Hubicka <jh@suse.cz>
26426 PR tree-optimization/55823
26427 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26430 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26431 Uros Bizjak <ubizjak@gmail.com>
26433 PR rtl-optimization/55845
26434 * df-problems.c (can_move_insns_across): Stop scanning at
26435 volatile_insn_p source instruction or give up if
26436 across_from .. across_to range contains any volatile_insn_p
26439 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26441 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26442 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26444 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26445 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26447 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26450 * asan.c (asan_clear_shadow): New function.
26451 (asan_emit_stack_protection): Use it.
26453 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26455 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26456 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26457 with tab instead of space.
26459 2013-01-08 Nick Clifton <nickc@redhat.com>
26461 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26462 register bank 0 at the start of an interrupt handler.
26463 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26466 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26468 * config/aarch64/aarch64-simd.md
26469 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26470 (aarch64_simd_bsl): Likewise.
26471 (aarch64_vcond_internal<mode>): Likewise.
26472 (vcond<mode><mode>): Likewise.
26473 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26474 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26476 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26478 * config/aarch64/aarch64-builtins.c
26479 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26481 2013-01-08 Martin Jambor <mjambor@suse.cz>
26484 * tree-sra.c (analyze_access_subtree): Return true also after
26485 potentially creating a debug-only replacement.
26487 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26489 PR middle-end/55890
26490 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26492 PR tree-optimization/54120
26493 * tree-vrp.c (range_fits_type_p): Don't allow
26494 src_precision < precision from signed vr to unsigned_p
26495 if vr->min or vr->max is negative.
26496 (simplify_float_conversion_using_ranges): Test can_float_p
26497 against CODE_FOR_nothing.
26499 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26500 Richard Biener <rguenther@suse.de>
26502 PR middle-end/55851
26503 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26504 types instead of just INTEGER_TYPE types.
26506 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26508 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26511 2013-01-07 Steve Ellcey <sellcey@mips.com>
26514 * config/mips/mips.opt: Change mad to mmad to match documentation.
26516 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26519 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26520 .progmemx.data now.
26522 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26525 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26526 (avr_addrspace_t): Add .section_name field.
26527 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26529 (avr_addrspace): Same. Initialize .section_name. Remove last
26530 NULL entry. Put __memx into .progmemx.data.
26531 (progmem_section_prefix): Remove.
26532 (avr_asm_init_sections): No need to initialize progmem_section.
26533 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26534 section name prefix.
26535 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26536 retrieve the progmem section.
26537 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26538 boundary to run over avr_addrspace[].
26539 (avr_register_target_pragmas): Ditto.
26541 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26543 * varasm.c (output_constant_def_contents): For asan_protect_global
26544 protected strings, adjust DECL_ALIGN if needed, before testing for
26546 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26547 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26549 (output_object_block): For asan protected decls, emit asan padding
26550 after their contents.
26551 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26552 (asan_finish_file): Test it here instead.
26554 2013-01-07 Nick Clifton <nickc@redhat.com>
26555 Matthias Klose <doko@debian.org>
26556 Doug Kwan <dougkwan@google.com>
26557 H.J. Lu <hongjiu.lu@intel.com>
26560 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26562 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26564 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26566 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26568 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26570 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26573 * doc/install.texi (Cross-Compiler-Specific Options): Document
26576 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26578 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26579 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26580 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26581 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26582 vqmovun_high_s64): Fix source operand number and update copyright.
26584 2013-01-07 Richard Biener <rguenther@suse.de>
26586 PR middle-end/55890
26587 * gimple.h (gimple_call_builtin_p): New overload.
26588 * gimple.c (validate_call): New function.
26589 (gimple_call_builtin_p): Likewise.
26590 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26591 Use gimple_call_builtin_p.
26592 (find_func_clobbers): Likewise.
26593 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26594 (strlen_optimize_stmt): Likewise.
26596 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26598 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26599 (vld1q_dup_*): Likewise.
26600 (vld1_*): Likewise.
26601 (vld1q_*): Likewise.
26602 (vld1_lane_*): Likewise.
26603 (vld1q_lane_*): Likewise.
26605 2013-01-07 Richard Biener <rguenther@suse.de>
26607 * lto-streamer.h (LTO_minor_version): Bump to 2.
26609 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26611 * config/aarch64/aarch64-protos.h
26612 (aarch64_const_double_zero_rtx_p): Rename to...
26613 (aarch64_float_const_zero_rtx_p): ...this.
26614 (aarch64_float_const_representable_p): New.
26615 (aarch64_output_simd_mov_immediate): Likewise.
26616 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26617 move immediate case.
26618 * config/aarch64/aarch64.c
26619 (aarch64_const_double_zero_rtx_p): Rename to...
26620 (aarch64_float_const_zero_rtx_p): ...this.
26621 (aarch64_print_operand): Allow printing of new constants.
26622 (aarch64_valid_floating_const): New.
26623 (aarch64_legitimate_constant_p): Check for valid floating-point
26625 (aarch64_simd_valid_immediate): Likewise.
26626 (aarch64_vect_float_const_representable_p): New.
26627 (aarch64_float_const_representable_p): Likewise.
26628 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26629 (aarch64_output_simd_mov_immediate): New.
26630 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26631 (*movdf_aarch64): Likewise.
26632 * config/aarch64/constraints.md (Ufc): New.
26633 (Y): call aarch64_float_const_zero_rtx.
26634 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26636 2013-01-07 Richard Biener <rguenther@suse.de>
26638 PR tree-optimization/55888
26639 PR tree-optimization/55862
26640 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26641 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26642 not if it is contained therein.
26644 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26646 * config/avr/t-avr: Typo.
26648 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26651 * config/avr/t-avr: Don't automatically rebuild
26652 $(srcdir)/config/avr/t-multilib
26653 $(srcdir)/config/avr/avr-tables.opt
26654 $(srcdir)/doc/avr-mmcu.texi
26655 (avr-mcus): New phony target to build them on request.
26656 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26657 * avr/avr-mcus.def: Adjust comments.
26659 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26661 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26663 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26665 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26667 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26669 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26671 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26673 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26674 to generate profiling.
26675 * config/rs6000/aix64.h (LIB_SPEC): Same.
26677 2013-01-04 Andrew Pinski <apinski@cavium.com>
26679 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26681 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26683 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26685 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26687 (ix86_expand_move): Ditto.
26688 (ix86_zero_extend_to_Pmode): Ditto.
26689 (ix86_expand_call): Ditto.
26690 (ix86_expand_special_args_builtin): Ditto.
26691 (ix86_expand_builtin): Ditto.
26693 2013-01-04 Richard Biener <rguenther@suse.de>
26695 PR tree-optimization/55862
26696 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26697 translating them through PHI nodes.
26699 2013-01-04 Martin Jambor <mjambor@suse.cz>
26701 PR tree-optimization/55755
26702 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26703 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26705 2013-01-04 Marek Polacek <polacek@redhat.com>
26707 PR middle-end/55859
26708 * opts.c (default_options_optimization): Clarify error message.
26710 2013-01-04 Richard Biener <rguenther@suse.de>
26712 PR middle-end/55863
26713 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26716 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26719 * config/pa/pa.md (movsi): Revert previous change.
26720 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26723 2013-01-03 Richard Henderson <rth@redhat.com>
26725 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26726 after eliminating TLS symbols.
26728 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26731 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26732 * graphite-poly.c (debug_gmp_value): Likewise.
26734 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26737 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26738 selected code model, define __code_mode_small__, __code_model_medium__,
26739 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26740 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26741 xchg temporary register with %k. Declare temporary register as
26743 [__x86_64__]: For medium and large code models, preserve %rbx register.
26745 2013-01-03 Richard Biener <rguenther@suse.de>
26747 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26748 (dump_subscript): Adjust.
26749 (finalize_ddr_dependent): Do not dump redundant info.
26750 (analyze_siv_subscript): Adjust.
26751 (subscript_dependence_tester): Likewise.
26752 (compute_affine_dependence): Likewise.
26754 2013-01-03 Richard Biener <rguenther@suse.de>
26757 2013-01-03 Richard Biener <rguenther@suse.de>
26759 PR tree-optimization/55857
26760 * tree-vect-stmts.c (vectorizable_load): Do not setup
26761 re-alignment for invariant loads.
26763 2013-01-02 Richard Biener <rguenther@suse.de>
26765 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26766 invariant load do not generate a vector load from the scalar location.
26768 2013-01-03 Richard Biener <rguenther@suse.de>
26770 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26771 for not vectorizing.
26772 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26773 not build INDIRECT_REFs, call get_name once only.
26774 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26775 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26777 2013-01-03 Richard Biener <rguenther@suse.de>
26779 PR tree-optimization/55857
26780 * tree-vect-stmts.c (vectorizable_load): Do not setup
26781 re-alignment for invariant loads.
26783 2013-01-03 Richard Biener <rguenther@suse.de>
26786 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26787 prefer a built-in decl.
26789 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26791 * gcc.c (process_command): Update copyright notice dates.
26792 * gcov.c (print_version): Likewise.
26793 * gcov-dump.c (print_version): Likewise.
26795 PR rtl-optimization/55838
26796 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26797 iv0.step, iv1.step and step.
26799 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26800 Marc Glisse <marc.glisse@inria.fr>
26802 PR tree-optimization/55832
26803 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26804 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26805 integer_{one,zero}_node.
26807 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26810 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26811 * var-tracking.c (reverse_op): Don't add reverse ops to
26812 VALUEs that have already
26813 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26815 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26817 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26819 2013-01-02 Teresa Johnson <tejohnson@google.com>
26821 * dumpfile.c (dump_loc): Print filename with location.
26822 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26823 new location_t parameter to emit complete unroll message with
26824 new dump framework.
26825 (canonicalize_loop_induction_variables): Compute loops location
26826 and pass to try_unroll_loop_completely.
26827 * loop-unroll.c (report_unroll_peel): New function.
26828 (peel_loops_completely): Use new dump format with location
26829 for main dumpfile message, and invoke report_unroll_peel on success.
26830 (decide_unrolling_and_peeling): Ditto.
26831 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26832 by report_unroll_peel.
26833 (decide_peel_completely): Ditto.
26834 (decide_unroll_constant_iterations): Ditto.
26835 (decide_unroll_runtime_iterations): Ditto.
26836 (decide_peel_simple): Ditto.
26837 (decide_unroll_stupid): Ditto.
26838 * cfgloop.c (get_loop_location): New function.
26839 * cfgloop.h (get_loop_location): Declare.
26841 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26843 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26846 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26848 PR middle-end/55198
26849 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26850 BLKmode objects when EXPAND_MEMORY is specified.
26852 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26854 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26856 (fold_builtin_cpu): Do not share cpu model decls across statements.
26858 2013-01-02 Jason Merrill <jason@redhat.com>
26861 * tree.c (build_array_type_1): Revert earlier change.
26863 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26865 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26867 * config/aarch64/aarch64-tune.md: Re-generate.
26869 2013-01-02 Richard Biener <rguenther@suse.de>
26871 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26872 invariant load do not generate a vector load from the scalar location.
26874 2013-01-02 Richard Biener <rguenther@suse.de>
26877 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26878 * configure: Regenerate.
26880 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26882 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26883 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26884 (expand_builtin_int_roundingfn_2): Keep the original target around
26885 for the fallback case.
26887 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26889 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26890 to be clear for sign changes.
26892 2013-01-01 Jan Hubicka <jh@suse.cz>
26894 * ipa-inline-analysis.c: Fix formatting.
26896 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26898 PR tree-optimization/55831
26899 * tree-vect-loop.c (get_initial_def_for_induction): Use
26900 gsi_after_labels instead of gsi_start_bb.
26902 Copyright (C) 2013 Free Software Foundation, Inc.
26904 Copying and distribution of this file, with or without modification,
26905 are permitted in any medium without royalty provided the copyright
26906 notice and this notice are preserved.