1 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
4 fields to the reg_addr array that describes the valid addressing
5 mode for any register, general purpose registers, floating point
6 registers, and Altivec registers.
7 (FIRST_RELOAD_REG_CLASS): Likewise.
8 (LAST_RELOAD_REG_CLASS): Likewise.
9 (struct reload_reg_map_type): Likewise.
10 (reload_reg_map_type): Likewise.
11 (RELOAD_REG_VALID): Likewise.
12 (RELOAD_REG_MULTIPLE): Likewise.
13 (RELOAD_REG_INDEXED): Likewise.
14 (RELOAD_REG_OFFSET): Likewise.
15 (RELOAD_REG_PRE_INCDEC): Likewise.
16 (RELOAD_REG_PRE_MODIFY): Likewise.
18 (mode_supports_pre_incdec_p): New helper functions to say whether
19 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
20 (mode_supports_pre_modify_p): Likewise.
21 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
22 print the valid address mode bits for each mode.
23 (rs6000_debug_print_mode): Likewise.
24 (rs6000_debug_reg_global): Likewise.
25 (rs6000_setup_reg_addr_masks): New function to set up the address
26 mask bits for each type.
27 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
28 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
29 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
30 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
31 PRE_MODIFY are supported.
32 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
33 registers, instead of {src,dest}_av_p.
34 (rs6000_print_options_internal): Tweak the debug output slightly.
36 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
38 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
41 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
43 * tree-flow.h (struct omp_region): Move to omp-low.c.
44 Remove omp_ prototypes and variables.
45 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
46 (copy_var_decl): Relocate prototype from tree-flow.h.
47 * gimple.c (copy_var_decl): Relocate from omp-low.c.
48 * tree.h: Move prototype to omp-low.h.
49 * omp-low.h: New File. Relocate prototypes here.
50 * omp-low.c (struct omp_region): Make local here.
51 (root_omp_region): Make static.
52 (copy_var_decl) Move to gimple.c.
53 (new_omp_region): Make static.
54 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
55 * tree-cfg.c: Include omp-low.h.
56 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
57 * gimplify.c: Include omp-low.h.
58 * tree-parloops.c: Likewise.
60 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
62 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
65 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
67 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
68 operands are in memory, prefer to force non-matched operand 1 to
71 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
74 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
75 restrict TImode addresses to single indirect registers if both
76 -mquad-memory and -mvsx-timode are used.
77 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
78 we should emit load/store quad. Remove using %y for quad memory
81 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
82 constraints to allow load/store quad on machines where TImode is
83 not allowed in VSX registers. Use 'n' instead of 'F' constraint
84 for TImode to load integer constants.
86 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
88 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
90 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
92 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
93 handling of STACK_REG.
95 2013-10-17 Richard Biener <rguenther@suse.de>
97 PR tree-optimization/58143
98 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
100 (rewrite_to_defined_overflow): Likewise.
101 (move_computations_dom_walker::before_dom): Rewrite stmts
102 with undefined signed overflow that are not always executed
103 into unsigned arithmetic.
105 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
108 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
109 explicit isa flag to be an options variable, instead of using
110 global_options_set. Remove define from rs6000.h.
111 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
113 * config/rs6000/rs6000.c (rs6000_option_override_internal):
114 Initialize rs6000_isa_flags_explicit.
115 (rs6000_function_specific_save): Add gcc_options* parameter, so
116 that the powerpc builds after the 2013-10-15 changes.
117 (rs6000_function_specific_restore): Likewise.
119 2013-10-16 DJ Delorie <dj@redhat.com>
121 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
122 op is a REG before checking REGNO.
123 (rl78_alloc_physical_registers): Verify pattern is a SET before
126 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
128 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
130 (vec_unpacks_lo_v4sf): Likewise.
131 (vec_unpacks_float_hi_v4si): Likewise.
132 (vec_unpacks_float_lo_v4si): Likewise.
133 (vec_unpacku_float_hi_v4si): Likewise.
134 (vec_unpacku_float_lo_v4si): Likewise.
136 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
138 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
139 (vsx_concat_v2sf): Likewise.
141 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
143 * config/aarch64/aarch64.md
144 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
146 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
148 PR tree-optimization/58697
149 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
150 estimated_loop_iterations_int.
151 (max_stmt_executions_int): Call get_max_loop_iterations_int.
152 (get_max_loop_iterations_int): New. HWINT version of
153 get_max_loop_iterations.
154 * cfgloop.h: Add prototypes.
155 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
156 * loop-unroll.c (decide_peel_once_rolling): Call
157 get_estimated_loop_iterations_int.
158 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
159 * tree-ssa-loop-niter.h: Tweak prototypes.
161 2013-10-16 David Malcolm <dmalcolm@redhat.com>
163 * gengtype-parse.c (struct_field_seq): Ignore access-control
164 keywords ("public:" etc).
166 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
168 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
169 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
171 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
173 * config/arm/arm.opt (mlra): New option.
174 * config/arm/arm.c (arm_lra_p): New function.
175 (TARGET_LRA_P): Define.
177 2013-10-16 Paulo Matos <pmatos@broadcom.com>
179 * tree-core.h (tree_code_name): Remove.
180 * tree.h (get_tree_code_name): New prototype.
181 * tree.c (tree_code_name): Make static.
182 (get_tree_code_name): New function.
183 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
184 tree_class_check_failed, tree_range_check_failed,
185 tree_not_class_check_failed, omp_clause_check_failed,
186 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
187 wrapper get_tree_code_name instead of calling tree_code_name directly.
188 * tree-vrp.c (dump_asserts_for): Likewise.
189 * tree-dump.c (dequeue_and_dump): Likewise.
190 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
191 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
192 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
193 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
194 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
195 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
196 dump_gimple_omp_for): Likewise.
197 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
198 * tree-ssa-pre.c (print_pre_expr): Likewise.
199 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
200 * print-tree.c (print_node_brief, print_node): Likewise.
201 * gimple.c (gimple_check_failed): Likewise.
202 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
203 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
204 * config/mep/mep.c (mep_validate_vliw): Likewise.
205 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
206 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
208 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
210 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
213 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
215 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
216 (MULTILIB_MATCHES): Add multilib for -march=v8.
218 2013-10-15 Sriraman Tallam <tmsriram@google.com>
221 * optc-save-gen.awk: Add extra parameter to the save and restore
223 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
224 * tree.c (build_optimization_node): New parameter. Add extra parameter
225 to call to cl_optimization_save.
226 (build_target_option_node): New parameter. Add extra parameter
227 to call to cl_target_option_save.
228 * tree.h (build_optimization_node): New parameter.
229 (build_target_option_node): New parameter.
230 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
231 build_optimization_node and build_target_option_node.
232 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
233 (handle_pragma_push_options): Ditto.
234 * toplev.c (process_options): Ditto.
235 * opts.c (init_options_struct): Check for opts_set non-null.
236 * target.def (target_option.save): New parameter.
237 (target_option.restore): New parameter.
239 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
240 (ix86_pragma_target_parse): Ditto.
241 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
243 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
244 to build_optimization_node and build_target_option_node.
245 (rs6000_valid_attribute_p): Ditto.
246 (rs6000_pragma_target_parse): Ditto.
247 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
249 * config/i386/i386.h:
250 TARGET_64BIT_P: New Macro
251 TARGET_MMX_P: New Macro.
252 TARGET_3DNOW_P: New Macro.
253 TARGET_3DNOW_A_P: New Macro.
254 TARGET_SSE_P: New Macro.
255 TARGET_SSE2_P: New Macro.
256 TARGET_SSE3_P: New Macro.
257 TARGET_SSSE3_P: New Macro.
258 TARGET_SSE4_1_P: New Macro.
259 TARGET_SSE4_2_P: New Macro.
260 TARGET_AVX_P: New Macro.
261 TARGET_AVX2_P: New Macro.
262 TARGET_AVX512F_P: New Macro.
263 TARGET_AVX512PF_P: New Macro.
264 TARGET_AVX512ER_P: New Macro.
265 TARGET_AVX512CD_P: New Macro.
266 TARGET_FMA_P: New Macro.
267 TARGET_SSE4A_P: New Macro.
268 TARGET_FMA4_P: New Macro.
269 TARGET_XOP_P: New Macro.
270 TARGET_LWP_P: New Macro.
271 TARGET_ABM_P: New Macro.
272 TARGET_BMI_P: New Macro.
273 TARGET_BMI2_P: New Macro.
274 TARGET_LZCNT_P: New Macro.
275 TARGET_TBM_P: New Macro.
276 TARGET_POPCNT_P: New Macro.
277 TARGET_SAHF_P: New Macro.
278 TARGET_MOVBE_P: New Macro.
279 TARGET_CRC32_P: New Macro.
280 TARGET_AES_P: New Macro.
281 TARGET_PCLMUL_P: New Macro.
282 TARGET_CMPXCHG16B_P: New Macro.
283 TARGET_FSGSBASE_P: New Macro.
284 TARGET_RDRND_P: New Macro.
285 TARGET_F16C_P: New Macro.
286 TARGET_RTM_P: New Macro.
287 TARGET_HLE_P: New Macro.
288 TARGET_RDSEED_P: New Macro.
289 TARGET_PRFCHW_P: New Macro.
290 TARGET_ADX_P: New Macro.
291 TARGET_FXSR_P: New Macro.
292 TARGET_XSAVE_P: New Macro.
293 TARGET_XSAVEOPT_P: New Macro.
294 TARGET_LP64_P: New Macro.
295 TARGET_X32_P: New Macro.
296 TARGET_FPMATH_DEFAULT_P: New Macro.
297 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
298 * config/i386/i386.c (ix86_option_override_internal): New parameters.
300 Change ix86_tune_string to access opts->x_ix86_tune_string.
301 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
302 Change ix86_arch_string to access opts->x_ix86_arch_string.
303 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
304 Change ix86_pmode to access opts->x_ix86_pmode.
305 Change ix86_abi to access opts->x_ix86_abi.
306 Change ix86_cmodel to access opts->x_ix86_cmodel.
307 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
308 Change ix86_isa_flags_explicit to access
309 opts->x_ix86_isa_flags_explicit.
310 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
311 Change ix86_regparm to access opts->x_ix86_regparm.
312 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
313 Change ix86_preferred_stack_boundary_arg to access
314 opts->x_ix86_preferred_stack_boundary_arg.
315 Change ix86_force_align_arg_pointer to access
316 opts->x_ix86_force_align_arg_pointer.
317 Change ix86_incoming_stack_boundar_arg to access
318 opts->x_ix86_incoming_stack_boundar_arg.
319 Change ix86_fpmath to access opts->x_ix86_fpmath.
320 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
321 Change ix86_recip_name to access opts->x_ix86_recip_name.
322 Change ix86_stack_protector_guard to access
323 opts->x_ix86_stack_protector_guard.
324 Change ix86_tune_memcpy_strategy to access
325 opts->x_ix86_tune_memcpy_strategy.
326 Change ix86_tune_memset_strategy to access
327 opts->x_ix86_tune_memset_strategy.
328 Change global_options to access opts.
329 Change global_options_set to access opts_set.
330 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
331 Change TARGET_MMX to TARGET_MMX_P (opts->...).
332 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
333 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
334 Change TARGET_SSE to TARGET_SSE_P (opts->...).
335 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
336 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
337 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
338 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
339 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
340 Change TARGET_AVX to TARGET_AVX_P (opts->...).
341 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
342 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
343 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
344 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
345 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
346 Change TARGET_FMA to TARGET_FMA_P (opts->...).
347 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
348 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
349 Change TARGET_XOP to TARGET_XOP_P (opts->...).
350 Change TARGET_LWP to TARGET_LWP_P (opts->...).
351 Change TARGET_ABM to TARGET_ABM_P (opts->...).
352 Change TARGET_BMI to TARGET_BMI_P (opts->...).
353 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
354 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
355 Change TARGET_TBM to TARGET_TBM_P (opts->...).
356 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
357 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
358 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
359 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
360 Change TARGET_AES to TARGET_AES_P (opts->...).
361 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
362 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
363 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
364 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
365 Change TARGET_F16C to TARGET_F16C_P (opts->...).
366 Change TARGET_RTM to TARGET_RTM_P (opts->...).
367 Change TARGET_HLE to TARGET_HLE_P (opts->...).
368 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
369 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
370 Change TARGET_ADX to TARGET_ADX_P (opts->...).
371 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
372 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
373 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
374 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
375 Change TARGET_X32 to TARGET_X32_P (opts->...).
376 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
377 Change TARGET_FLOAT_RETURNS_IN_80387 to
378 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
379 (ix86_function_specific_save): New parameter. Use opts-> fields
380 to replace global fields.
381 (ix86_function_specific_restore): Ditto.
382 (ix86_valid_target_attribute_inner_p): New parameters.
384 Fix call to ix86_handle_option and set_option.
385 (ix86_valid_target_attribute_tree): New parameters.
386 Change global_options to access opts.
387 Change global_options_set to access opts_set.
388 Fix call to ix86_valid_target_attribute_inner_p.
389 Change ix86_tune_string to access opts->x_ix86_tune_string.
390 Change ix86_arch_string to access opts->x_ix86_arch_string.
391 Change ix86_fpmath to access opts->x_ix86_fpmath
392 Fix call to ix86_option_override_internal.
393 Fix call to ix86_add_new_builtins.
394 Fix calls to build_optimization_node and build_target_option_node.
395 (ix86_valid_target_attribute_p): Remove access to global_options.
396 Use new gcc_options structure func_options.
397 Fix call to ix86_valid_target_attribute_tree.
398 Fix call to build_optimization_node.
399 (get_builtin_code_for_version): Fix call to
400 ix86_valid_target_attribute_tree.
402 2013-10-15 David Malcolm <dmalcolm@redhat.com>
404 * Makefile.in (PICFLAG): New.
405 (enable_host_shared): New.
406 (INTERNAL_CFLAGS): Use PICFLAG.
407 (LIBIBERTY): Use pic build of libiberty.a if configured with
408 --enable-host-shared.
409 * configure.ac: Add --enable-host-shared, setting up new
411 * configure: Regenerate.
412 * doc/install.texi (--enable-shared): Add note contrasting it with ...
413 (--enable-host-shared): New option.
415 2013-10-15 Richard Biener <rguenther@suse.de>
417 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
418 for built-in functions.
420 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
422 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
423 (optimize_range_tests_1): New function,
424 extracted from optimize_range_tests.
425 (optimize_range_tests_xor): Similarly.
426 (optimize_range_tests_diff): New function.
427 (optimize_range_tests): Use optimize_range_tests_1.
429 2013-10-15 Cong Hou <congh@google.com>
431 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
432 requirement of the reduction pattern so that one operand of the
433 reduction operation can come from outside of the loop.
435 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
437 * config/arm/neon-schedgen.ml: Remove.
438 * config/arm/cortex-a9-neon.md: Remove comment regarding
441 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
443 * config/arm/types: Remove old neon types.
445 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
447 * config/arm/cortex-a7.md
448 (cortex_a7_neon_type): New.
449 (cortex_a7_neon_mul): Update for new types.
450 (cortex_a7_neon_mla): Likewise.
451 (cortex_a7_neon): Likewise.
453 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
455 * config/arm/cortex-a15-neon.md
456 (cortex_a15_neon_type): New,
458 (cortex_a15_neon_int_1): Remove.
459 (cortex_a15_neon_int_2): Likewise.
460 (cortex_a15_neon_int_3): Likewise.
461 (cortex_a15_neon_int_4): Likewise.
462 (cortex_a15_neon_int_5): Likewise.
463 (cortex_a15_neon_vqneg_vqabs): Likewise.
464 (cortex_a15_neon_vmov): Likewise.
465 (cortex_a15_neon_vaba): Likewise.
466 (cortex_a15_neon_vaba_qqq): Likewise.
467 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
468 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
469 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
471 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
472 (cortex_a15_neon_mla_qqq_8_16): Likewise.
473 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
474 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
475 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
476 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
477 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
478 (cortex_a15_neon_shift_1): Likewise.
479 (cortex_a15_neon_shift_2): Likewise.
480 (cortex_a15_neon_shift_3): Likewise.
481 (cortex_a15_neon_vshl_ddd): Likewise.
482 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
483 (cortex_a15_neon_vsra_vrsra): Likewise.
484 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
485 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
486 (cortex_a15_neon_bp_3cycle): Likewise.
487 (cortex_a15_neon_ldm_2): Likewise.
488 (cortex_a15_neon_stm_2): Likewise.
489 (cortex_a15_neon_mcr): Likewise.
490 (cortex_a15_neon_mrc): Likewise.
491 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
492 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
493 (cortex_a15_neon_fp_vmul_ddd): Likewise.
494 (cortex_a15_neon_fp_vmul_qqd): Likewise.
495 (cortex_a15_neon_fp_vmla_ddd): Likewise.
496 (cortex_a15_neon_fp_vmla_qqq): Likewise.
497 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
498 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
499 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
500 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
501 (cortex_a15_neon_bp_simple): Likewise.
502 (cortex_a15_neon_bp_2cycle): Likewise.
503 (cortex_a15_neon_bp_3cycle): Likewise.
504 (cortex_a15_neon_vld1_1_2_regs): Likewise.
505 (cortex_a15_neon_vld1_3_4_regs): Likewise.
506 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
507 (cortex_a15_neon_vld2_4_regs): Likewise.
508 (cortex_a15_neon_vld3_vld4): Likewise.
509 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
510 (cortex_a15_neon_vst1_3_4_regs): Likewise.
511 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
512 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
513 (cortex_a15_neon_vst3_vst4): Rename to...
514 (cortex_a15_neon_vst4): This, update for new attributes.
515 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
516 (cortex_a15_neon_vld3_vld4_lane): Likewise.
517 (cortex_a15_neon_vst1_vst2_lane): Likewise.
518 (cortex_a15_neon_vst3_vst4_lane): Likewise.
519 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
520 (cortex_a15_neon_ldm_2): Likewise.
521 (cortex_a15_neon_stm_2): Likewise.
522 (cortex_a15_neon_mcr): Likewise.
523 (cortex_a15_neon_mcr_2_mcrr): Likewise.
524 (cortex_a15_neon_mrc): Likewise.
525 (cortex_a15_neon_mrrc): Likewise.
527 (cortex_a15_neon_abd): New.
528 (cortex_a15_neon_abd_q): Likewise.
529 (cortex_a15_neon_aba): Likewise.
530 (cortex_a15_neon_aba_q): Likewise.
531 (cortex_a15_neon_acc): Likewise.
532 (cortex_a15_neon_acc_q): Likewise.
533 (cortex_a15_neon_arith_basic): Likewise.
534 (cortex_a15_neon_arith_complex): Likewise.
535 (cortex_a15_neon_multiply): Likewise.
536 (cortex_a15_neon_multiply_q): Likewise.
537 (cortex_a15_neon_mla): Likewise.
538 (cortex_a15_neon_mla_q): Likewise.
539 (cortex_a15_neon_sat_mla_long): Likewise.
540 (cortex_a15_neon_shift_acc): Likewise.
541 (cortex_a15_neon_shift_imm_basic): Likewise.
542 (cortex_a15_neon_shift_imm_complex): Likewise.
543 (cortex_a15_neon_shift_reg_basic): Likewise.
544 (cortex_a15_neon_shift_reg_basic_q): Likewise.
545 (cortex_a15_neon_shift_reg_complex): Likewise.
546 (cortex_a15_neon_shift_reg_complex_q): Likewise.
547 (cortex_a15_neon_fp_negabs): Likewise
548 (cortex_a15_neon_fp_arith): Likewise
549 (cortex_a15_neon_fp_arith_q): Likewise
550 (cortex_a15_neon_fp_cvt_int): Likewise
551 (cortex_a15_neon_fp_cvt_int_q): Likewise
552 (cortex_a15_neon_fp_cvt_16): Likewise
553 (cortex_a15_neon_fp_mul): Likewise
554 (cortex_a15_neon_fp_mul_q): Likewise
555 (cortex_a15_neon_fp_mla): Likewise
556 (cortex_a15_neon_fp_mla_q): Likewise
557 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
558 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
559 (cortex_a15_neon_bitops): Likewise.
560 (cortex_a15_neon_bitops_q): Likewise.
561 (cortex_a15_neon_from_gp): Likewise.
562 (cortex_a15_neon_from_gp_q): Likewise.
563 (cortex_a15_neon_tbl3_tbl4): Likewise.
564 (cortex_a15_neon_zip_q): Likewise.
565 (cortex_a15_neon_to_gp): Likewise.
566 (cortex_a15_neon_load_a): Likewise.
567 (cortex_a15_neon_load_b): Likewise.
568 (cortex_a15_neon_load_c): Likewise.
569 (cortex_a15_neon_load_d): Likewise.
570 (cortex_a15_neon_load_e): Likewise.
571 (cortex_a15_neon_load_f): Likewise.
572 (cortex_a15_neon_store_a): Likewise.
573 (cortex_a15_neon_store_b): Likewise.
574 (cortex_a15_neon_store_c): Likewise.
575 (cortex_a15_neon_store_d): Likewise.
576 (cortex_a15_neon_store_e): Likewise.
577 (cortex_a15_neon_store_f): Likewise.
578 (cortex_a15_neon_store_g): Likewise.
579 (cortex_a15_neon_store_h): Likewise.
580 (cortex_a15_vfp_to_from_gp): Likewise.
582 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
584 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
586 (cortex_a9_neon_vshl_ddd): Remove.
587 (cortex_a9_neon_vst3_vst4): Likewise.
588 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
590 (cortex_a9_neon_bit_ops_q): New.
592 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
593 (cortex_a9_neon_int_2): Likewise.
594 (cortex_a9_neon_int_3): Likewise.
595 (cortex_a9_neon_int_4): Likewise.
596 (cortex_a9_neon_int_5): Likewise.
597 (cortex_a9_neon_vqneg_vqabs): Likewise.
598 (cortex_a9_neon_vmov): Likewise.
599 (cortex_a9_neon_vaba): Likewise.
600 (cortex_a9_neon_vaba_qqq): Likewise.
601 (cortex_a9_neon_shift_1): Likewise.
602 (cortex_a9_neon_shift_2): Likewise.
603 (cortex_a9_neon_shift_3): Likewise.
604 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
605 (cortex_a9_neon_vsra_vrsra): Likewise.
606 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
607 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
608 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
610 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
611 (cortex_a9_neon_mla_qqq_8_16): Likewise.
612 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
614 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
615 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
616 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
617 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
618 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
619 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
620 (cortex_a9_neon_fp_vsum): Likewise.
621 (cortex_a9_neon_fp_vmul_ddd): Likewise.
622 (cortex_a9_neon_fp_vmul_qqd): Likewise.
623 (cortex_a9_neon_fp_vmla_ddd): Likewise.
624 (cortex_a9_neon_fp_vmla_qqq): Likewise.
625 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
626 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
627 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
628 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
629 (cortex_a9_neon_bp_simple): Likewise.
630 (cortex_a9_neon_bp_2cycle): Likewise.
631 (cortex_a9_neon_bp_3cycle): Likewise.
632 (cortex_a9_neon_ldr): Likewise.
633 (cortex_a9_neon_str): Likewise.
634 (cortex_a9_neon_vld1_1_2_regs): Likewise.
635 (cortex_a9_neon_vld1_3_4_regs): Likewise.
636 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
637 (cortex_a9_neon_vld2_4_regs): Likewise.
638 (cortex_a9_neon_vld3_vld4): Likewise.
639 (cortex_a9_neon_vld1_vld2_lane): Likewise.
640 (cortex_a9_neon_vld3_vld4_lane): Likewise.
641 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
642 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
643 (cortex_a9_neon_vst1_3_4_regs): Likewise.
644 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
645 (cortex_a9_neon_vst1_vst2_lane): Likewise.
646 (cortex_a9_neon_vst3_vst4_lane): Likewise.
647 (cortex_a9_neon_mcr): Likewise.
648 (cortex_a9_neon_mcr_2_mcrr): Likewise.
649 (cortex_a9_neon_mrc): Likewise.
650 (cortex_a9_neon_mrrc): Likewise.
652 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
654 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
656 (cortex_a8_neon_vshl_ddd): Remove.
657 (cortex_a8_neon_vst3_vst4): Likewise.
658 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
660 (cortex_a8_neon_bit_ops_q): New.
662 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
663 (cortex_a8_neon_int_2): Likewise..
664 (cortex_a8_neon_int_3): Likewise.
665 (cortex_a8_neon_int_5): Likewise.
666 (cortex_a8_neon_vqneg_vqabs): Likewise.
667 (cortex_a8_neon_int_4): Likewise.
668 (cortex_a8_neon_vaba): Likewise.
669 (cortex_a8_neon_vaba_qqq): Likewise.
670 (cortex_a8_neon_shift_1): Likewise.
671 (cortex_a8_neon_shift_2): Likewise.
672 (cortex_a8_neon_shift_3): Likewise.
673 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
674 (cortex_a8_neon_vsra_vrsra): Likewise.
675 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
676 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
677 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
679 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
680 (cortex_a8_neon_mla_qqq_8_16): Likewise.
681 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
683 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
684 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
685 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
686 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
687 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
688 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
689 (cortex_a8_neon_fp_vsum): Likewise.
690 (cortex_a8_neon_fp_vmul_ddd): Likewise.
691 (cortex_a8_neon_fp_vmul_qqd): Likewise.
692 (cortex_a8_neon_fp_vmla_ddd): Likewise.
693 (cortex_a8_neon_fp_vmla_qqq): Likewise.
694 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
695 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
696 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
697 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
698 (cortex_a8_neon_bp_simple): Likewise.
699 (cortex_a8_neon_bp_2cycle): Likewise.
700 (cortex_a8_neon_bp_3cycle): Likewise.
701 (cortex_a8_neon_ldr): Likewise.
702 (cortex_a8_neon_str): Likewise.
703 (cortex_a8_neon_vld1_1_2_regs): Likewise.
704 (cortex_a8_neon_vld1_3_4_regs): Likewise.
705 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
706 (cortex_a8_neon_vld2_4_regs): Likewise.
707 (cortex_a8_neon_vld3_vld4): Likewise.
708 (cortex_a8_neon_vld1_vld2_lane): Likewise.
709 (cortex_a8_neon_vld3_vld4_lane): Likewise.
710 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
711 (cortex_a8_neon_vst1_3_4_regs): Likewise.
712 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
713 (cortex_a8_neon_vst1_vst2_lane): Likewise.
714 (cortex_a8_neon_vst3_vst4_lane): Likewise.
715 (cortex_a8_neon_mcr): Likewise.
716 (cortex_a8_neon_mcr_2_mcrr): Likewise.
717 (cortex_a8_neon_mrc): Likewise.
718 (cortex_a8_neon_mrrc): Likewise.
720 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
722 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
724 * config/aarch64/aarch64-simd.md (neon_type): Remove.
725 (aarch64_simd_dup<mode>): Add "type" attribute.
726 (aarch64_dup_lane<mode>): Likewise.
727 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
728 (*aarch64_simd_mov<mode>): Likewise.
729 (aarch64_simd_mov_from_<mode>low): Likewise.
730 (aarch64_simd_mov_from_<mode>high): Likewise.
731 (orn<mode>3): Likewise.
732 (bic<mode>3): Likewise.
733 (add<mode>3): Likewise.
734 (sub<mode>3): Likewise.
735 (mul<mode>3): Likewise.
736 (*aarch64_mul3_elt<mode>): Likewise.
737 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
738 (*aarch64_mul3_elt_to_128df): Likewise.
739 (*aarch64_mul3_elt_to_64v2df): Likewise.
740 (neg<mode>2): Likewise.
741 (abs<mode>2): Likewise.
742 (abd<mode>_3): Likewise.
743 (aba<mode>_3): Likewise.
744 (fabd<mode>_3): Likewise.
745 (*fabd_scalar<mode>3): Likewise.
746 (and<mode>3): Likewise.
747 (ior<mode>3): Likewise.
748 (xor<mode>3): Likewise.
749 (one_cmpl<mode>2): Likewise.
750 (aarch64_simd_vec_set<mode>): Likewise.
751 (aarch64_simd_lshr<mode>): Likewise.
752 (aarch64_simd_ashr<mode>): Likewise.
753 (aarch64_simd_imm_shl<mode>): Likewise.
754 (aarch64_simd_reg_sshl<mode): Likewise.
755 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
756 (aarch64_simd_reg_shl<mode>_signed): Likewise.
757 (aarch64_simd_vec_setv2di): Likewise.
758 (aarch64_simd_vec_set<mode>): Likewise.
759 (aarch64_mla<mode>): Likewise.
760 (*aarch64_mla_elt<mode>): Likewise.
761 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
762 (aarch64_mls<mode>): Likewise.
763 (*aarch64_mls_elt<mode>): Likewise.
764 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
765 (<su><maxmin><mode>3): Likewise.
766 (move_lo_quad_<mode>): Likewise.
767 (aarch64_simd_move_hi_quad_<mode>): Likewise.
768 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
769 (vec_pack_trunc_<mode>): Likewise.
770 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
771 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
772 (*aarch64_<su>mlal_lo<mode>): Likewise.
773 (*aarch64_<su>mlal_hi<mode>): Likewise.
774 (*aarch64_<su>mlsl_lo<mode>): Likewise.
775 (*aarch64_<su>mlsl_hi<mode>): Likewise.
776 (*aarch64_<su>mlal<mode>): Likewise.
777 (*aarch64_<su>mlsl<mode>): Likewise.
778 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
779 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
780 (add<mode>3): Likewise.
781 (sub<mode>3): Likewise.
782 (mul<mode>3): Likewise.
783 (div<mode>3): Likewise.
784 (neg<mode>2): Likewise.
785 (abs<mode>2): Likewise.
786 (fma<mode>4): Likewise.
787 (*aarch64_fma4_elt<mode>): Likewise.
788 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
789 (*aarch64_fma4_elt_to_128df): Likewise.
790 (*aarch64_fma4_elt_to_64v2df): Likewise.
791 (fnma<mode>4): Likewise.
792 (*aarch64_fnma4_elt<mode>): Likewise.
793 (*aarch64_fnma4_elt_<vswap_width_name><mode>
794 (*aarch64_fnma4_elt_to_128df): Likewise.
795 (*aarch64_fnma4_elt_to_64v2df): Likewise.
796 (<frint_pattern><mode>2): Likewise.
797 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
798 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
799 (vec_unpacks_lo_v4sf): Likewise.
800 (aarch64_float_extend_lo_v2df): Likewise.
801 (vec_unpacks_hi_v4sf): Likewise.
802 (aarch64_float_truncate_lo_v2sf): Likewise.
803 (aarch64_float_truncate_hi_v4sf): Likewise.
804 (aarch64_vmls<mode>): Likewise.
805 (<su><maxmin><mode>3): Likewise.
806 (<maxmin_uns><mode>3): Likewise.
807 (reduc_<sur>plus_<mode>): Likewise.
808 (reduc_<sur>plus_v2di): Likewise.
809 (reduc_<sur>plus_v2si): Likewise.
810 (reduc_<sur>plus_<mode>): Likewise.
811 (aarch64_addpv4sf): Likewise.
812 (clz<mode>2): Likewise.
813 (reduc_<maxmin_uns>_<mode>): Likewise.
814 (reduc_<maxmin_uns>_v2di): Likewise.
815 (reduc_<maxmin_uns>_v2si): Likewise.
816 (reduc_<maxmin_uns>_<mode>): Likewise.
817 (reduc_<maxmin_uns>_v4sf): Likewise.
818 (aarch64_simd_bsl<mode>_internal): Likewise.
819 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
820 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
821 (aarch64_get_lane<mode>): Likewise.
822 (*aarch64_combinez<mode>): Likewise.
823 (aarch64_combine<mode>): Likewise.
824 (aarch64_simd_combine<mode>): Likewise.
825 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
826 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
827 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
828 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
829 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
830 (aarch64_<sur>h<addsub><mode>): Likewise.
831 (aarch64_<sur><addsub>hn<mode>): Likewise.
832 (aarch64_<sur><addsub>hn2<mode>): Likewise.
833 (aarch64_pmul<mode>): Likewise.
834 (aarch64_<su_optab><optab><mode>): Likewise.
835 (aarch64_<sur>qadd<mode>): Likewise.
836 (aarch64_sqmovun<mode>): Likewise.
837 (aarch64_<sur>qmovn<mode>): Likewise.
838 (aarch64_s<optab><mode>): Likewise.
839 (aarch64_sq<r>dmulh<mode>): Likewise.
840 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
841 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
842 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
843 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
844 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
845 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
846 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
847 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
848 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
849 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
850 (aarch64_sqdmull<mode>): Likewise.
851 (aarch64_sqdmull_lane<mode>_internal): Likewise.
852 (aarch64_sqdmull_n<mode>): Likewise.
853 (aarch64_sqdmull2<mode>_internal): Likewise.
854 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
855 (aarch64_sqdmull2_n<mode>_internal): Likewise.
856 (aarch64_<sur>shl<mode>): Likewise.
857 (aarch64_<sur>q<r>shl<mode>
858 (aarch64_<sur>shll_n<mode>): Likewise.
859 (aarch64_<sur>shll2_n<mode>): Likewise.
860 (aarch64_<sur>shr_n<mode>): Likewise.
861 (aarch64_<sur>sra_n<mode>): Likewise.
862 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
863 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
864 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
865 (aarch64_cm<optab><mode>): Likewise.
866 (aarch64_cm<optab>di): Likewise.
867 (aarch64_cm<optab><mode>): Likewise.
868 (aarch64_cm<optab>di): Likewise.
869 (aarch64_cmtst<mode>): Likewise.
870 (aarch64_cmtstdi): Likewise.
871 (aarch64_cm<optab><mode>): Likewise.
872 (*aarch64_fac<optab><mode>): Likewise.
873 (aarch64_addp<mode>): Likewise.
874 (aarch64_addpdi): Likewise.
875 (sqrt<mode>2): Likewise.
876 (vec_load_lanesoi<mode>): Likewise.
877 (vec_store_lanesoi<mode>): Likewise.
878 (vec_load_lanesci<mode>): Likewise.
879 (vec_store_lanesci<mode>): Likewise.
880 (vec_load_lanesxi<mode>): Likewise.
881 (vec_store_lanesxi<mode>): Likewise.
882 (*aarch64_mov<mode>): Likewise.
883 (aarch64_ld2<mode>_dreg): Likewise.
884 (aarch64_ld2<mode>_dreg): Likewise.
885 (aarch64_ld3<mode>_dreg): Likewise.
886 (aarch64_ld3<mode>_dreg): Likewise.
887 (aarch64_ld4<mode>_dreg): Likewise.
888 (aarch64_ld4<mode>_dreg): Likewise.
889 (aarch64_tbl1<mode>): Likewise.
890 (aarch64_tbl2v16qi): Likewise.
891 (aarch64_combinev16qi): Likewise.
892 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
893 (aarch64_st2<mode>_dreg): Likewise.
894 (aarch64_st2<mode>_dreg): Likewise.
895 (aarch64_st3<mode>_dreg): Likewise.
896 (aarch64_st3<mode>_dreg): Likewise.
897 (aarch64_st4<mode>_dreg): Likewise.
898 (aarch64_st4<mode>_dreg): Likewise.
899 (*aarch64_simd_ld1r<mode>): Likewise.
900 (aarch64_frecpe<mode>): Likewise.
901 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
902 (aarch64_frecps<mode>): Likewise.
904 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
906 * config/arm/iterators.md (V_elem_ch): New.
908 (VQH_type): Likewise.
909 * config/arm/arm.md (is_neon_type): New.
910 (conds): Use is_neon_type.
911 (anddi3_insn): Update type attribute.
912 (xordi3_insn): Likewise.
913 (one_cmpldi2): Likewise.
914 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
915 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
916 (*movmisalign<mode>_neon_store): Likewise.
917 (*movmisalign<mode>_neon_load): Likewise.
918 (vec_set<mode>_internal): Likewise.
919 (vec_set<mode>_internal): Likewise.
920 (vec_setv2di_internal): Likewise.
921 (vec_extract<mode>): Likewise.
922 (vec_extract<mode>): Likewise.
923 (vec_extractv2di): Likewise.
924 (*add<mode>3_neon): Likewise.
925 (adddi3_neon): Likewise.
926 (*sub<mode>3_neon): Likewise.
927 (subdi3_neon): Likewise.
928 (fma<VCVTF:mode>4): Likewise.
929 (fma<VCVTF:mode>4_intrinsic): Likewise.
930 (*fmsub<VCVTF:mode>4): Likewise.
931 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
932 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
933 (ior<mode>3): Likewise.
934 (and<mode>3): Likewise.
935 (orn<mode>3_neon): Likewise.
936 (orndi3_neon): Likewise.
937 (bic<mode>3_neon): Likewise.
938 (bicdi3_neon): Likewise.
939 (xor<mode>3): Likewise.
940 (one_cmpl<mode>2): Likewise.
941 (abs<mode>2): Likewise.
942 (neg<mode>2): Likewise.
943 (negdi2_neon): Likewise.
944 (*umin<mode>3_neon): Likewise.
945 (*umax<mode>3_neon): Likewise.
946 (*smin<mode>3_neon): Likewise.
947 (*smax<mode>3_neon): Likewise.
948 (vashl<mode>3): Likewise.
949 (vashr<mode>3_imm): Likewise.
950 (vlshr<mode>3_imm): Likewise.
951 (ashl<mode>3_signed): Likewise.
952 (ashl<mode>3_unsigned): Likewise.
953 (neon_load_count): Likewise.
954 (ashldi3_neon_noclobber): Likewise.
955 (ashldi3_neon): Likewise.
956 (signed_shift_di3_neon): Likewise.
957 (unsigned_shift_di3_neon): Likewise.
958 (ashrdi3_neon_imm_noclobber): Likewise.
959 (lshrdi3_neon_imm_noclobber): Likewise.
960 (<shift>di3_neon): Likewise.
961 (widen_ssum<mode>3): Likewise.
962 (widen_usum<mode>3): Likewise.
963 (quad_halves_<code>v4si): Likewise.
964 (quad_halves_<code>v4sf): Likewise.
965 (quad_halves_<code>v8hi): Likewise.
966 (quad_halves_<code>v16qi): Likewise.
967 (reduc_splus_v2di): Likewise.
968 (neon_vpadd_internal<mode>): Likewise.
969 (neon_vpsmin<mode>): Likewise.
970 (neon_vpsmax<mode>): Likewise.
971 (neon_vpumin<mode>): Likewise.
972 (neon_vpumax<mode>): Likewise.
973 (*ss_add<mode>_neon): Likewise.
974 (*us_add<mode>_neon): Likewise.
975 (*ss_sub<mode>_neon): Likewise.
976 (*us_sub<mode>_neon): Likewise.
977 (neon_vadd<mode>_unspec): Likewise.
978 (neon_vaddl<mode>): Likewise.
979 (neon_vaddw<mode>): Likewise.
980 (neon_vhadd<mode>): Likewise.
981 (neon_vqadd<mode>): Likewise.
982 (neon_vaddhn<mode>): Likewise.
983 (neon_vmul<mode>): Likewise.
984 (neon_vfms<VCVTF:mode>): Likewise.
985 (neon_vmlal<mode>): Likewise.
986 (neon_vmls<mode>): Likewise.
987 (neon_vmlsl<mode>): Likewise.
988 (neon_vqdmulh<mode>): Likewise.
989 (neon_vqdmlal<mode>): Likewise.
990 (neon_vqdmlsl<mode>): Likewise.
991 (neon_vmull<mode>): Likewise.
992 (neon_vqdmull<mode>): Likewise.
993 (neon_vsub<mode>_unspec): Likewise.
994 (neon_vsubl<mode>): Likewise.
995 (neon_vsubw<mode>): Likewise.
996 (neon_vqsub<mode>): Likewise.
997 (neon_vhsub<mode>): Likewise.
998 (neon_vsubhn<mode>): Likewise.
999 (neon_vceq<mode>): Likewise.
1000 (neon_vcge<mode>): Likewise.
1001 (neon_vcgeu<mode>): Likewise.
1002 (neon_vcgt<mode>): Likewise.
1003 (neon_vcgtu<mode>): Likewise.
1004 (neon_vcle<mode>): Likewise.
1005 (neon_vclt<mode>): Likewise.
1006 (neon_vcage<mode>): Likewise.
1007 (neon_vcagt<mode>): Likewise.
1008 (neon_vtst<mode>): Likewise.
1009 (neon_vabd<mode>): Likewise.
1010 (neon_vabdl<mode>): Likewise.
1011 (neon_vaba<mode>): Likewise.
1012 (neon_vabal<mode>): Likewise.
1013 (neon_vmax<mode>): Likewise.
1014 (neon_vmin<mode>): Likewise.
1015 (neon_vpaddl<mode>): Likewise.
1016 (neon_vpadal<mode>): Likewise.
1017 (neon_vpmax<mode>): Likewise.
1018 (neon_vpmin<mode>): Likewise.
1019 (neon_vrecps<mode>): Likewise.
1020 (neon_vrsqrts<mode>): Likewise.
1021 (neon_vqabs<mode>): Likewise.
1022 (neon_vqneg<mode>): Likewise.
1023 (neon_vcls<mode>): Likewise.
1024 (clz<mode>2): Likewise.
1025 (popcount<mode>2): Likewise.
1026 (neon_vrecpe<mode>): Likewise.
1027 (neon_vrsqrte<mode>): Likewise.
1028 (neon_vget_lane<mode>_sext_internal): Likewise.
1029 (neon_vget_lane<mode>_zext_internal): Likewise.
1030 (neon_vdup_n<mode>): Likewise.
1031 (neon_vdup_n<mode>): Likewise.
1032 (neon_vdup_nv2di): Likewise.
1033 (neon_vdup_lane<mode>_interal): Likewise.
1034 (*neon_vswp<mode>): Likewise.
1035 (neon_vcombine<mode>): Likewise.
1036 (float<mode><V_cvtto>2): Likewise.
1037 (floatuns<mode><V_cvtto>2): Likewise.
1038 (fix_trunc<mode><V_cvtto>2): Likewise.
1039 (fixuns_trunc<mode><V_cvtto>2
1040 (neon_vcvt<mode>): Likewise.
1041 (neon_vcvt<mode>): Likewise.
1042 (neon_vcvtv4sfv4hf): Likewise.
1043 (neon_vcvtv4hfv4sf): Likewise.
1044 (neon_vcvt_n<mode>): Likewise.
1045 (neon_vcvt_n<mode>): Likewise.
1046 (neon_vmovn<mode>): Likewise.
1047 (neon_vqmovn<mode>): Likewise.
1048 (neon_vqmovun<mode>): Likewise.
1049 (neon_vmovl<mode>): Likewise.
1050 (neon_vmul_lane<mode>): Likewise.
1051 (neon_vmul_lane<mode>): Likewise.
1052 (neon_vmull_lane<mode>): Likewise.
1053 (neon_vqdmull_lane<mode>): Likewise.
1054 (neon_vqdmulh_lane<mode>): Likewise.
1055 (neon_vqdmulh_lane<mode>): Likewise.
1056 (neon_vmla_lane<mode>): Likewise.
1057 (neon_vmla_lane<mode>): Likewise.
1058 (neon_vmlal_lane<mode>): Likewise.
1059 (neon_vqdmlal_lane<mode>): Likewise.
1060 (neon_vmls_lane<mode>): Likewise.
1061 (neon_vmls_lane<mode>): Likewise.
1062 (neon_vmlsl_lane<mode>): Likewise.
1063 (neon_vqdmlsl_lane<mode>): Likewise.
1064 (neon_vext<mode>): Likewise.
1065 (neon_vrev64<mode>): Likewise.
1066 (neon_vrev32<mode>): Likewise.
1067 (neon_vrev16<mode>): Likewise.
1068 (neon_vbsl<mode>_internal): Likewise.
1069 (neon_vshl<mode>): Likewise.
1070 (neon_vqshl<mode>): Likewise.
1071 (neon_vshr_n<mode>): Likewise.
1072 (neon_vshrn_n<mode>): Likewise.
1073 (neon_vqshrn_n<mode>): Likewise.
1074 (neon_vqshrun_n<mode>): Likewise.
1075 (neon_vshl_n<mode>): Likewise.
1076 (neon_vqshl_n<mode>): Likewise.
1077 (neon_vqshlu_n<mode>): Likewise.
1078 (neon_vshll_n<mode>): Likewise.
1079 (neon_vsra_n<mode>): Likewise.
1080 (neon_vsri_n<mode>): Likewise.
1081 (neon_vsli_n<mode>): Likewise.
1082 (neon_vtbl1v8qi): Likewise.
1083 (neon_vtbl2v8qi): Likewise.
1084 (neon_vtbl3v8qi): Likewise.
1085 (neon_vtbl4v8qi): Likewise.
1086 (neon_vtbl1v16qi): Likewise.
1087 (neon_vtbl2v16qi): Likewise.
1088 (neon_vcombinev16qi): Likewise.
1089 (neon_vtbx1v8qi): Likewise.
1090 (neon_vtbx2v8qi): Likewise.
1091 (neon_vtbx3v8qi): Likewise.
1092 (neon_vtbx4v8qi): Likewise.
1093 (*neon_vtrn<mode>_insn): Likewise.
1094 (*neon_vzip<mode>_insn): Likewise.
1095 (*neon_vuzp<mode>_insn): Likewise.
1096 (neon_vld1<mode>): Likewise.
1097 (neon_vld1_lane<mode>): Likewise.
1098 (neon_vld1_lane<mode>): Likewise.
1099 (neon_vld1_dup<mode>): Likewise.
1100 (neon_vld1_dup<mode>): Likewise.
1101 (neon_vld1_dupv2di): Likewise.
1102 (neon_vst1<mode>): Likewise.
1103 (neon_vst1_lane<mode>): Likewise.
1104 (neon_vst1_lane<mode>): Likewise.
1105 (neon_vld2<mode>): Likewise.
1106 (neon_vld2<mode>): Likewise.
1107 (neon_vld2_lane<mode>): Likewise.
1108 (neon_vld2_lane<mode>): Likewise.
1109 (neon_vld2_dup<mode>): Likewise.
1110 (neon_vst2<mode>): Likewise.
1111 (neon_vst2<mode>): Likewise.
1112 (neon_vst2_lane<mode>): Likewise.
1113 (neon_vst2_lane<mode>): Likewise.
1114 (neon_vld3<mode>): Likewise.
1115 (neon_vld3qa<mode>): Likewise.
1116 (neon_vld3qb<mode>): Likewise.
1117 (neon_vld3_lane<mode>): Likewise.
1118 (neon_vld3_lane<mode>): Likewise.
1119 (neon_vld3_dup<mode>): Likewise.
1120 (neon_vst3<mode>): Likewise.
1121 (neon_vst3qa<mode>): Likewise.
1122 (neon_vst3qb<mode>): Likewise.
1123 (neon_vst3_lane<mode>): Likewise.
1124 (neon_vst3_lane<mode>): Likewise.
1125 (neon_vld4<mode>): Likewise.
1126 (neon_vld4qa<mode>): Likewise.
1127 (neon_vld4qb<mode>): Likewise.
1128 (neon_vld4_lane<mode>): Likewise.
1129 (neon_vld4_lane<mode>): Likewise.
1130 (neon_vld4_dup<mode>): Likewise.
1131 (neon_vst4<mode>): Likewise.
1132 (neon_vst4qa<mode>): Likewise.
1133 (neon_vst4qb<mode>): Likewise.
1134 (neon_vst4_lane<mode>): Likewise.
1135 (neon_vst4_lane<mode>): Likewise.
1136 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1137 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1138 (neon_vec_<US>mult_lo_<mode>): Likewise.
1139 (neon_vec_<US>mult_hi_<mode>): Likewise.
1140 (neon_vec_<US>shiftl_<mode>): Likewise.
1141 (neon_unpack<US>_<mode>): Likewise.
1142 (neon_vec_<US>mult_<mode>): Likewise.
1143 (vec_pack_trunc_<mode>): Likewise.
1144 (neon_vec_pack_trunc_<mode>): Likewise.
1145 (neon_vabd<mode>_2): Likewise.
1146 (neon_vabd<mode>_3): Likewise.
1148 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1150 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1151 (load_pair): Update type attribute.
1152 (store_pair): Update type attribute.
1153 * config/aarch64/iterators.md (q): New.
1155 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1157 * config/arm/types.md: Add new types for Neon insns.
1159 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1160 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1161 Sergey Lega <sergey.s.lega@intel.com>
1162 Anna Tikhonova <anna.tikhonova@intel.com>
1163 Ilya Tocar <ilya.tocar@intel.com>
1164 Andrey Turetskiy <andrey.turetskiy@intel.com>
1165 Ilya Verbin <ilya.verbin@intel.com>
1166 Kirill Yukhin <kirill.yukhin@intel.com>
1167 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1169 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1170 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1171 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1173 (srcp14<mode>): Ditto.
1174 (rsqrt14<mode>): Ditto.
1175 (rsqrt14<mode>): Ditto.
1176 (avx512f_vmscalef<mode>): Ditto.
1177 (avx512f_scalef<mode>): Ditto.
1178 (avx512f_getexp<mode>): Ditto.
1179 (avx512f_sgetexp<mode>): Ditto.
1180 (avx512f_fixupimm<mode>): Ditto.
1181 (avx512f_sfixupimm<mode>): Ditto.
1182 (avx512f_rndscale<mode>): Ditto.
1183 (*avx512er_exp2<mode>): Ditto.
1184 (*avx512er_rcp28<mode>): Ditto.
1185 (avx512er_rsqrt28<mode>): Ditto.
1186 (avx512f_getmant<mode>): Ditto.
1187 (avx512f_getmant<mode>): Ditto.
1188 (avx512f_rndscale<mode>): Fix formatting.
1190 2013-10-15 Martin Jambor <mjambor@suse.cz>
1192 * ipa-utils.h (ipa_edge_within_scc): Declare.
1193 * ipa-cp.c (edge_within_scc): Moved...
1194 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1196 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1197 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1198 Sergey Lega <sergey.s.lega@intel.com>
1199 Anna Tikhonova <anna.tikhonova@intel.com>
1200 Ilya Tocar <ilya.tocar@intel.com>
1201 Andrey Turetskiy <andrey.turetskiy@intel.com>
1202 Ilya Verbin <ilya.verbin@intel.com>
1203 Kirill Yukhin <kirill.yukhin@intel.com>
1204 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1206 * config/i386/predicates.md (const_8_to_15_operand): New.
1207 (const_16_to_31_operand): Ditto.
1208 * config/i386/sse.md (V8FI): New.
1210 (reduc_splus_v8df): Ditto.
1211 (reduc_splus_v16sf): Ditto.
1212 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1213 (vec_extract_hi_<mode>): Ditto.
1214 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1215 (vec_set_lo_<mode>): Ditto.
1216 (vec_set_hi_<mode>): Ditto.
1217 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1218 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1219 (avx512f_pshufd_1): Ditto.
1220 (avx512f_broadcast<mode>): Ditto.
1221 (avx512f_broadcast<mode>): Ditto.
1222 (define_split): Split vec_extract_lo into move.
1223 (ssequartermode): Ditto.
1224 (ssedoublemode): Extened with wider modes.
1225 (vec_extract_lo_<mode>): Ditto.
1227 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1228 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1229 Sergey Lega <sergey.s.lega@intel.com>
1230 Anna Tikhonova <anna.tikhonova@intel.com>
1231 Ilya Tocar <ilya.tocar@intel.com>
1232 Andrey Turetskiy <andrey.turetskiy@intel.com>
1233 Ilya Verbin <ilya.verbin@intel.com>
1234 Kirill Yukhin <kirill.yukhin@intel.com>
1235 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1237 * config/i386/predicates.md (register_or_constm1_operand): New.
1238 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1239 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1240 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1241 UNSPEC_SCATTER_PREFETCH
1243 (avx512f_ucmp<mode>3): Ditto.
1244 (avx512f_vternlog<mode>): Ditto.
1245 (avx512f_align<mode>): Ditto.
1246 (<shift_insn><mode>3): Ditto.
1247 (avx512f_<rotate>v<mode>): Ditto.
1248 (avx512f_<rotate><mode>): Ditto.
1249 (avx512f_eq<mode>3): Ditto.
1250 (avx512f_eq<mode>3_1): Ditto.
1251 (avx512f_gt<mode>3): Ditto.
1252 (avx512f_testm<mode>3): Ditto.
1253 (avx512f_testnm<mode>3): Ditto.
1254 (avx512pf_gatherpf<mode>): Ditto.
1255 (*avx512pf_gatherpf<mode>_mask): Ditto.
1256 (*avx512pf_gatherpf<mode>): Ditto.
1257 (avx512pf_scatterpf<mode>): Ditto.
1258 (*avx512pf_scatterpf<mode>_mask): Ditto.
1259 (*avx512pf_scatterpf<mode>): Ditto.
1260 (avx512f_vec_dup_gpr<mode>): Ditto.
1261 (clz<mode>2): Ditto.
1262 (conflict<mode>): Ditto.
1263 (REDUC_SMINMAX_MODE): Extened with wider modes.
1264 (reduc_<code>_<mode>): Ditto.
1265 (vlshr<mode>3): Ditto.
1266 (vashl<mode>3): Ditto.
1268 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1269 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1270 Sergey Lega <sergey.s.lega@intel.com>
1271 Anna Tikhonova <anna.tikhonova@intel.com>
1272 Ilya Tocar <ilya.tocar@intel.com>
1273 Andrey Turetskiy <andrey.turetskiy@intel.com>
1274 Ilya Verbin <ilya.verbin@intel.com>
1275 Kirill Yukhin <kirill.yukhin@intel.com>
1276 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1278 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1281 (avx512fmaskmode): Ditto.
1282 (bcstscalarsuff): Ditto.
1283 (avx512f_blendm<mode>): Ditto.
1284 (cmp_imm_predicate): Ditto.
1285 (avx512f_cmp<mode>3): Ditto.
1286 (avx512f_vec_dup<mode>): Ditto.
1287 (avx512f_vec_dup_mem<mode>): Ditto.
1288 (avx512f_vpermi2var<mode>3): Ditto.
1289 (avx512f_vpermt2var<mode>3): Ditto.
1290 (vec_init<mode>): Ditto.
1291 (avx512f_gathersi<mode>): Ditto.
1292 (*avx512f_gathersi<mode>): Ditto.
1293 (*avx512f_gathersi<mode>_2): Ditto.
1294 (avx512f_gatherdi<mode>): Ditto.
1295 (*avx512f_gatherdi<mode>): Ditto.
1296 (*avx512f_gatherdi<mode>_2): Ditto.
1297 (avx512f_scattersi<mode>): Ditto.
1298 (*avx512f_scattersi<mode>): Ditto.
1299 (avx512f_scatterdi<mode>): Ditto.
1300 (*avx512f_scatterdi<mode>): Ditto.
1301 (sseintprefix): Extened with wider modes.
1302 (VEC_GATHER_IDXSI): Ditto.
1303 (VEC_GATHER_IDXDI): Ditto.
1304 (VEC_GATHER_SRCDI): Ditto.
1306 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1307 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1309 * config/arm/t-aprofile: New file.
1310 * config.gcc: Handle --with-multilib-list option.
1312 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1314 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1317 2013-10-15 Richard Biener <rguenther@suse.de>
1319 * tree-loop-distribution.c (build_empty_rdg): Inline into
1321 (rdg_flag_vertex): Inline into single user.
1322 (rdg_flag_vertex_and_dependent): Likewise.
1323 (build_rdg_partition_for_vertex): Remove processed bitmap.
1324 (rdg_build_partitions): Simplify.
1326 2013-10-15 Richard Biener <rguenther@suse.de>
1328 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1329 Restructure forwarding through conversions and copies to
1330 avoid performing copy-propagation the wrong way. Adjust
1331 recursion invocations.
1332 (forward_propagate_addr_expr): Add argument stating if we
1333 are recursing from a single-use.
1334 (ssa_forward_propagate_and_combine): Adjust.
1336 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1338 * dumpfile.h (gcc::dump_manager): New class, to hold state
1339 relating to dumpfile management.
1340 (get_dump_file_name): Remove in favor of method of dump_manager.
1341 (dump_initialized_p): Likewise.
1342 (dump_start): Likewise.
1343 (dump_finish): Likewise.
1344 (dump_switch_p): Likewise.
1345 (dump_register): Likewise.
1346 (get_dump_file_info): Likewise.
1347 * context.c (gcc::context::context): Construct the dump_manager
1349 * context.h (gcc::context::get_dumps): New.
1350 (gcc::context::m_dumps): New.
1351 * coverage.c (coverage_init): Port to dump_manager API.
1352 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1353 (extra_dump_files_in_use): Likewise.
1354 (extra_dump_files_alloced): Likewise.
1355 (gcc::dump_manager::dump_manager): New.
1356 (dump_register): Convert to...
1357 (gcc::dump_manager::dump_register): ...method, replacing
1358 function-static next_dump with m_next_dump field.
1359 (get_dump_file_info): Convert to...
1360 (gcc::dump_manager::get_dump_file_info): ...method.
1361 (get_dump_file_name): Convert to...
1362 (gcc::dump_manager::get_dump_file_name): ...method.
1363 (dump_start): Convert to...
1364 (gcc::dump_manager::dump_start): ...method.
1365 (dump_finish): Convert to...
1366 (gcc::dump_manager::dump_finish): ...method.
1367 (dump_begin): Replace body with...
1368 (gcc::dump_manager::dump_begin): ...new method.
1369 (dump_phase_enabled_p): Convert to...
1370 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1371 (dump_phase_enabled_p): Convert to...
1372 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1373 (dump_initialized_p): Convert to...
1374 (gcc::dump_manager::dump_initialized_p): ...method.
1375 (dump_flag_name): Replace body with...
1376 (gcc::dump_manager::dump_flag_name): ...new method.
1377 (dump_enable_all): Convert to...
1378 (gcc::dump_manager::dump_enable_all): ...new method.
1379 (opt_info_enable_passes): Convert to...
1380 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1381 (dump_switch_p_1): Convert to...
1382 (gcc::dump_manager::dump_switch_p_1): ...new method.
1383 (dump_switch_p): Convert to...
1384 (gcc::dump_manager::dump_switch_p): ...new method.
1385 (opt_info_switch_p): Port to dump_manager API.
1386 (enable_rtl_dump_file): Likewise.
1387 * opts-global.c (handle_common_deferred_options): Port to new
1389 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1390 (pass_manager::register_one_dump_file): Likewise.
1391 (pass_manager::register_pass): Likewise.
1392 (pass_init_dump_file): Likewise.
1393 (pass_fini_dump_file): Likewise.
1394 * statistics.c (statistics_early_init): Likewise.
1396 2013-10-14 Richard Biener <rguenther@suse.de>
1398 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1399 iterative_hash_canonical_type, gimple_canonical_type_hash,
1400 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1401 gimple_register_canonical_type, print_gimple_types_stats,
1402 free_gimple_type_tables): Move to lto/lto.c
1403 (gt-gimple.h): Do not include.
1404 * gimple.h (gimple_register_canonical_type,
1405 print_gimple_types_stats, free_gimple_type_tables): Remove.
1406 * Makefile.in (GTFILES): Remove gimple.c.
1408 2013-10-14 Travis Snoozy <quandary@remstate.com>
1411 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1412 scanning for msp430x targets.
1414 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1417 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1418 (registers_ok_for_ldd_peep): Move around.
1419 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1420 * config/sparc/sparc.md (widening peepholes): Use it.
1422 2013-10-14 Richard Biener <rguenther@suse.de>
1426 * gimple.c (iterative_hash_canonical_type): Make sure to
1427 record the hash into the correct hashtable slot.
1429 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1431 PR rtl-optimization/58662
1432 * combine.c (try_combine): Take into account death nodes on I2 when
1433 splitting a PARALLEL of two independent SETs. Fix dump message.
1435 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1438 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1439 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1440 * config/sh/t-sh (sh_treg_combine.o): New entry.
1441 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1442 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1443 (register_sh_passes): New function. Register sh_treg_combine pass.
1444 (sh_option_override): Invoke it.
1445 (sh_canonicalize_comparison): Handle op0_preserve_value.
1446 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1447 opportunities. Canonicalize branch condition.
1448 (nott): Allow only if pseudos can be created for non-SH2A.
1450 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1453 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1454 (ix86_expand_movmem): Replace copy_addr_to_reg with
1455 ix86_copy_addr_to_reg.
1456 (ix86_expand_setmem): Likewise.
1458 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1460 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1461 provided by ix86_fp_compare_mode instead of CCFPUmode.
1463 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1465 * config/aarch64/arm_neon.h
1466 (vtbx<1,3>_<psu>8): Fix register constriants.
1468 2013-10-11 Jeff Law <law@redhat.com>
1470 PR tree-optimization/58640
1471 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1472 threading paths that cross over two loop entry points.
1474 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1476 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1477 handle vector float as well.
1478 (*vsx_le_perm_load_v4si): Likewise.
1479 (*vsx_le_perm_store_v2di): Likewise.
1480 (*vsx_le_perm_store_v4si): Likewise.
1482 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1484 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1485 directly to circumvent subtract from splat{31} workaround.
1486 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1488 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1489 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1490 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1491 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1492 separate big and little endian logic.
1493 (*altivec_vperm_<mode>_internal): New define_insn.
1494 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1495 separate big and little endian logic.
1496 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1497 (vec_permv16qi): Add little endian logic.
1499 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1501 * doc/extend.texi (returns_nonnull): Remove arguments.
1503 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1504 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1505 Sergey Lega <sergey.s.lega@intel.com>
1506 Anna Tikhonova <anna.tikhonova@intel.com>
1507 Ilya Tocar <ilya.tocar@intel.com>
1508 Andrey Turetskiy <andrey.turetskiy@intel.com>
1509 Ilya Verbin <ilya.verbin@intel.com>
1510 Kirill Yukhin <kirill.yukhin@intel.com>
1511 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1513 * config/i386/sse.md (VI48F_256_512): New.
1514 (avx2_permvar<mode>): Change to ...
1515 (<avx2_avx512f>_permvar<mode>): This.
1517 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1518 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1519 Sergey Lega <sergey.s.lega@intel.com>
1520 Anna Tikhonova <anna.tikhonova@intel.com>
1521 Ilya Tocar <ilya.tocar@intel.com>
1522 Andrey Turetskiy <andrey.turetskiy@intel.com>
1523 Ilya Verbin <ilya.verbin@intel.com>
1524 Kirill Yukhin <kirill.yukhin@intel.com>
1525 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1527 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1528 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1529 * config/i386/sse.md (VI4_AVX): New.
1530 (sf2simodelower): Ditto.
1531 (sse2_cvtps2dq): Change to ...
1532 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1534 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1535 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1536 Sergey Lega <sergey.s.lega@intel.com>
1537 Anna Tikhonova <anna.tikhonova@intel.com>
1538 Ilya Tocar <ilya.tocar@intel.com>
1539 Andrey Turetskiy <andrey.turetskiy@intel.com>
1540 Ilya Verbin <ilya.verbin@intel.com>
1541 Kirill Yukhin <kirill.yukhin@intel.com>
1542 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1544 * config/i386/sse.md (V_512): New.
1546 (vcond<V_512:mode><VF_512:mode>): Ditto.
1547 (vcond<V_512:mode><VI_512:mode>): Ditto.
1548 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1550 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1551 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1552 Sergey Lega <sergey.s.lega@intel.com>
1553 Anna Tikhonova <anna.tikhonova@intel.com>
1554 Ilya Tocar <ilya.tocar@intel.com>
1555 Andrey Turetskiy <andrey.turetskiy@intel.com>
1556 Ilya Verbin <ilya.verbin@intel.com>
1557 Kirill Yukhin <kirill.yukhin@intel.com>
1558 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1560 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1561 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1563 (fma<mode>4): Removed condition.
1564 (fms<mode>4): Ditto.
1565 (fnma<mode>4): Ditto.
1566 (fnms<mode>4): Ditto.
1567 (fma4i_fmadd_<mode>): Ditto.
1568 (*fma_fmadd_<mode>): Ditto.
1569 (*fma_fmsub_<mode>): Ditto.
1570 (*fma_fnmadd_<mode>): Ditto.
1571 (*fma_fnmsub_<mode>): Ditto.
1572 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1573 (*fma_fmaddsub_<mode>): Ditto.
1574 (*fma_fmsubadd_<mode>): Ditto.
1575 (*fmai_fmadd_<mode>): Ditto.
1576 (*fmai_fmsub_<mode>): Ditto.
1577 (*fmai_fnmadd_<mode>): Ditto.
1578 (*fmai_fnmsub_<mode>): Ditto.
1580 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1581 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1582 Sergey Lega <sergey.s.lega@intel.com>
1583 Anna Tikhonova <anna.tikhonova@intel.com>
1584 Ilya Tocar <ilya.tocar@intel.com>
1585 Andrey Turetskiy <andrey.turetskiy@intel.com>
1586 Ilya Verbin <ilya.verbin@intel.com>
1587 Kirill Yukhin <kirill.yukhin@intel.com>
1588 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1590 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1591 (VI124_256): Changed to ...
1592 (VI124_256_48_512): This.
1593 (ssepackmode): Extended with wider modes.
1594 (<code><mode>3): Changed iterator.
1595 (*avx2_<code><mode>3): Ditto.
1596 (vec_pack_trunc_<mode>): Ditto.
1598 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1599 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1600 Sergey Lega <sergey.s.lega@intel.com>
1601 Anna Tikhonova <anna.tikhonova@intel.com>
1602 Ilya Tocar <ilya.tocar@intel.com>
1603 Andrey Turetskiy <andrey.turetskiy@intel.com>
1604 Ilya Verbin <ilya.verbin@intel.com>
1605 Kirill Yukhin <kirill.yukhin@intel.com>
1606 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1608 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1609 (VI8F_256_512): Ditto.
1610 (abs<mode>2): Changed iterator.
1611 (avx2_perm<mode>): Changed to ...
1612 (<avx2_avx512f>_perm<mode>): This.
1613 (avx2_perm<mode>_1): Changed to ...
1614 (<avx2_avx512f>_perm<mode>_1): This.
1616 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1617 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1618 Sergey Lega <sergey.s.lega@intel.com>
1619 Anna Tikhonova <anna.tikhonova@intel.com>
1620 Ilya Tocar <ilya.tocar@intel.com>
1621 Andrey Turetskiy <andrey.turetskiy@intel.com>
1622 Ilya Verbin <ilya.verbin@intel.com>
1623 Kirill Yukhin <kirill.yukhin@intel.com>
1624 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1626 * config/i386/sse.md (VI48_AVX512F): New.
1627 (VI48_AVX2): Changed to ...
1628 (VI48_AVX2_48_AVX512F): This.
1629 (avx2_ashrv<mode>): Changed to ...
1630 (<avx2_avx512f>_ashrv<mode>): This.
1631 (avx2_<shift_insn>v<mode>): Changed to ...
1632 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1634 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1635 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1636 Sergey Lega <sergey.s.lega@intel.com>
1637 Anna Tikhonova <anna.tikhonova@intel.com>
1638 Ilya Tocar <ilya.tocar@intel.com>
1639 Andrey Turetskiy <andrey.turetskiy@intel.com>
1640 Ilya Verbin <ilya.verbin@intel.com>
1641 Kirill Yukhin <kirill.yukhin@intel.com>
1642 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1644 * config/i386/sse.md (VI4_AVX512F): New.
1645 (VI8_AVX2_AVX512F): Ditto.
1646 (mul<mode>3): Extended with wider modes.
1647 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1648 (mul<mode>3): Ditto.
1649 (vec_widen_<s>mult_odd_<mode>): Ditto.
1651 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1652 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1653 Sergey Lega <sergey.s.lega@intel.com>
1654 Anna Tikhonova <anna.tikhonova@intel.com>
1655 Ilya Tocar <ilya.tocar@intel.com>
1656 Andrey Turetskiy <andrey.turetskiy@intel.com>
1657 Ilya Verbin <ilya.verbin@intel.com>
1658 Kirill Yukhin <kirill.yukhin@intel.com>
1659 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1661 * config/i386/sse.md (VI2_AVX512F): New.
1662 (VI124_AVX512F): Ditto.
1663 (sseunpackmode): Extended with wider modes.
1664 (sseunpackfltmode): Ditto.
1665 (vec_unpacks_float_hi_<mode>): Ditto.
1666 (vec_unpacks_float_lo_<mode>): Ditto.
1667 (vec_unpacku_float_hi_<mode>): Ditto.
1668 (vec_unpacku_float_lo_<mode>): Ditto.
1669 (vec_unpacks_lo_<mode>): Ditto.
1670 (vec_unpacks_hi_<mode>): Ditto.
1671 (vec_unpacku_lo_<mode>): Ditto.
1672 (vec_unpacku_hi_<mode>): Ditto.
1674 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1675 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1676 Sergey Lega <sergey.s.lega@intel.com>
1677 Anna Tikhonova <anna.tikhonova@intel.com>
1678 Ilya Tocar <ilya.tocar@intel.com>
1679 Andrey Turetskiy <andrey.turetskiy@intel.com>
1680 Ilya Verbin <ilya.verbin@intel.com>
1681 Kirill Yukhin <kirill.yukhin@intel.com>
1682 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1684 * config/i386/i386.md (multdiv): New.
1685 (multdiv_mnemonic): Ditto.
1686 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1687 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1688 (<sse>_vmdiv<mode>3): Removed.
1690 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1691 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1692 Sergey Lega <sergey.s.lega@intel.com>
1693 Anna Tikhonova <anna.tikhonova@intel.com>
1694 Ilya Tocar <ilya.tocar@intel.com>
1695 Andrey Turetskiy <andrey.turetskiy@intel.com>
1696 Ilya Verbin <ilya.verbin@intel.com>
1697 Kirill Yukhin <kirill.yukhin@intel.com>
1698 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1700 * config/i386/sse.md (V): Extended with wider modes.
1702 (ssehalfvecmode): Ditto.
1704 (ssepackfltmode): Ditto.
1705 (avx_vec_concat<mode>): Ditto.
1706 (V_256_512): New iterator.
1707 (VF2_512_256): Ditto.
1708 (si2dfmode): New attribute.
1709 (si2dfmodelower): Ditto.
1711 (concat_tg_mode): Ditto.
1712 (floatv4siv4df2): Changed to ...
1713 (float<si2dfmodelower><mode>2): This.
1714 (avx_cvtps2pd256): Changed to ...
1715 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1716 (vec_pack_trunc_v4df): Changed to ...
1717 (vec_pack_trunc_<mode>): This.
1718 (avx_vpermil<mode>): Changed to ...
1719 (<sse2_avx_avx512f>_vpermil<mode>): This.
1720 (<fixsuffix>fix_truncv8dfv8si2): New.
1721 (vec_pack_sfix_trunc_v8df): Ditto.
1722 (avx512f_rndscale<mode>): Ditto.
1723 (avx512f_roundpd512): Ditto.
1724 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1726 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1727 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1728 Sergey Lega <sergey.s.lega@intel.com>
1729 Anna Tikhonova <anna.tikhonova@intel.com>
1730 Ilya Tocar <ilya.tocar@intel.com>
1731 Andrey Turetskiy <andrey.turetskiy@intel.com>
1732 Ilya Verbin <ilya.verbin@intel.com>
1733 Kirill Yukhin <kirill.yukhin@intel.com>
1734 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1736 * config/i386/i386.md (any_fix): New iterator.
1737 (fixsuffix): New attribute.
1738 * config/i386/sse.md (VF1): Extened with wider modes.
1742 (sseintvecmodelower): Ditto.
1743 (ssescalarmode): Ditto.
1744 (ssescalarnum): Ditto.
1746 (ssexmmmode): Ditto.
1747 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1748 (<sse>_rcp<mode>2): Change iterator.
1749 (rsqrt<mode>2): Ditto.
1750 (<sse>_rsqrt<mode>2): Ditto.
1751 (avx2_vec_dup<mode>): Ditto.
1752 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1753 (round<mode>2_sfix): Ditto.
1754 (avx2_pbroadcast<mode>): Ditto.
1755 (*andnot<mode>3): Handle XI mode.
1756 (*<code><mode>3): Ditto.
1757 (AVXTOSSEMODE): Removed.
1758 (avx_vpermil<mode>): Changed to ...
1759 (<sse2_avx_avx512f>_vpermil<mode>): This.
1761 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1762 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1763 Sergey Lega <sergey.s.lega@intel.com>
1764 Anna Tikhonova <anna.tikhonova@intel.com>
1765 Ilya Tocar <ilya.tocar@intel.com>
1766 Andrey Turetskiy <andrey.turetskiy@intel.com>
1767 Ilya Verbin <ilya.verbin@intel.com>
1768 Kirill Yukhin <kirill.yukhin@intel.com>
1769 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1771 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1772 (<sse>_comi): Ditto.
1773 (<sse>_ucomi): Ditto.
1774 (sse_cvtss2siq_2): Ditto.
1775 (sse2_cvtsd2si): Ditto.
1776 (sse2_cvtsd2siq): Ditto.
1777 (sse2_cvttsd2si): Ditto.
1778 (sse2_cvttsd2siq): Ditto.
1779 (<shift_insn><mode>3): Ditto.
1780 (sse2_cvtsi2sdq): Update constraint and prefix.
1781 (sse_cvtsi2ss): Update prefix.
1782 (sse_cvtsi2ssq): Ditto.
1784 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1786 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1787 ignore internal calls.
1789 2013-10-11 Richard Biener <rguenther@suse.de>
1791 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1792 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1793 TREE_OVERFLOW is set.
1795 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1797 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1799 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1800 * gimple.h: Likewise.
1802 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1804 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1807 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1809 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1810 * configure: Regenerate.
1812 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1814 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1815 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1816 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1817 similarly pp_right_paren.
1818 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1819 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1820 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1821 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1823 (omp_declare_simd_clauses_equal,
1824 omp_remove_redundant_declare_simd_attrs): New functions.
1825 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1826 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1827 * tree.h (OMP_LOOP_CHECK): Define.
1828 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1829 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1830 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1831 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1832 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1833 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1834 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1835 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1836 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1837 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1838 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1839 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1840 (omp_remove_redundant_declare_simd_attrs): New prototype.
1841 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1842 GIMPLE_OMP_TEAMS): New codes.
1843 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1844 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1846 (target_nesting_level): New variable.
1847 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1848 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1849 for collapse > 1 static schedule unless ordered.
1850 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1851 (determine_parallel_type): Adjust get_ws_args_for caller.
1852 (install_var_field): Handle mask & 4 for double indirection.
1853 (scan_sharing_clauses): Ignore shared clause on teams construct.
1854 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1855 (create_omp_child_function): If inside target or declare target
1856 constructs, set "omp declare target" attribute on the child function.
1857 (find_combined_for): New function.
1858 (scan_omp_parallel): Handle combined loops.
1859 (scan_omp_target, scan_omp_teams): New functions.
1860 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1861 restrictions and set ctx->cancellable for cancellable constructs.
1862 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1863 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1864 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1865 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1866 (omp_clause_aligned_alignment): New function.
1867 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
1868 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1869 on teams constructs. Handle user defined reductions and new
1871 (lower_reduction_clauses): Don't set placeholder to address of ref
1872 if it has already the right type.
1873 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1874 (expand_parallel_call): Use the new non-_start suffixed builtins,
1875 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1876 and GOMP_parallel_end after the call.
1877 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1878 (expand_omp_for_init_counts): Handle combined loops.
1879 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1881 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1882 end of cancellable loops.
1883 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1884 Likewise. Handle collapse > 1 loops.
1885 (expand_omp_simd): Handle combined loops.
1886 (expand_omp_for): Add inner_stmt argument, adjust callers of
1887 expand_omp_for* functions, use expand_omp_for_static*chunk even
1888 for collapse > 1 unless ordered.
1889 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1890 of cancellable sections.
1891 (expand_omp_single): Remove need_barrier variable, just rely on
1892 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1893 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1894 (expand_omp_atomic_load, expand_omp_atomic_store,
1895 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1896 (expand_omp_target): New function.
1897 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1898 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1899 (build_omp_regions_1): Immediately close region for
1900 GF_OMP_TARGET_KIND_UPDATE.
1901 (maybe_add_implicit_barrier_cancel): New function.
1902 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1904 (lower_omp_single): Likewise. Add clobber after the barrier.
1905 (lower_omp_taskgroup): New function.
1906 (lower_omp_for): Handle combined loops. Adjust
1907 lower_rec_input_clauses caller. Handle cancellation.
1908 (lower_depend_clauses): New function.
1909 (lower_omp_taskreg): Lower depend clauses. Adjust
1910 lower_rec_input_clauses caller. Add clobber after the call. Handle
1912 (lower_omp_target, lower_omp_teams): New functions.
1913 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1914 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1915 and GOMP_cancellation_point calls.
1916 (lower_omp): Fold stmts inside of target region.
1917 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1918 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1919 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1920 (BT_FN_VOID_OMPFN_PTR_UINT,
1921 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1922 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1923 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1924 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1925 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1926 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1927 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1928 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1929 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1930 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1931 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1932 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1933 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1934 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1935 BUILT_IN_GOMP_PARALLEL_END.
1936 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1937 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1938 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1939 GF_OMP_FOR_KIND_DISTRIBUTE.
1940 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1941 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1942 (dump_gimple_omp_return): Print lhs if it has any.
1943 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1944 gimple_omp_atomic_seq_cst_p.
1945 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1946 and GIMPLE_OMP_TEAMS.
1947 * langhooks.c (lhd_omp_mappable_type): New function.
1948 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1949 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1951 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1952 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1953 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
1954 (struct gimplify_omp_ctx): Add combined_loop field.
1955 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1956 on stmts inside of target region.
1957 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1958 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1959 ORT_TARGET and ORT_TARGET_DATA.
1960 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1962 (omp_notice_threadprivate_variable): Complain about threadprivate
1963 variables in target region.
1964 (omp_notice_variable): Complain about vars with non-mappable type
1965 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1966 (omp_check_private): Ignore ORT_TARGET* regions.
1967 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1968 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1969 (find_combined_omp_for): New function.
1970 (gimplify_omp_for): Handle gimplification of combined loops.
1971 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1973 (gimplify_omp_target_update): New function.
1974 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1975 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1976 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1977 (gimplify_body): If fndecl has "omp declare target" attribute, add
1978 implicit ORT_TARGET context around it.
1979 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1980 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1981 * tree-nested.c (convert_nonlocal_reference_stmt,
1982 convert_local_reference_stmt, convert_gimple_call): Handle
1983 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1984 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1985 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1986 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1987 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1988 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1989 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1990 BUILT_IN_GOMP_LOOP_END_CANCEL,
1991 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1992 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1993 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1994 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1995 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1996 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1997 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1998 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1999 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2000 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2001 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2002 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2003 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2004 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2005 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2006 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2007 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2008 gimple_build_omp_teams): New functions.
2009 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2010 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2011 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2012 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2013 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2014 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2015 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2016 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2017 GF_OMP_ATOMIC_SEQ_CST): New.
2018 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2019 gimple_build_omp_teams): New prototypes.
2020 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2021 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2022 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2023 GIMPLE_OMP_SINGLE as end of range.
2024 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2025 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2026 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2027 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2028 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2029 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2030 gimple_omp_target_kind, gimple_omp_target_set_kind,
2031 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2032 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2033 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2034 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2035 gimple_omp_teams_set_clauses): New inlines.
2036 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2037 and GIMPLE_OMP_TASKGROUP.
2038 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2039 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2040 enum omp_clause_proc_bind_kind): New.
2041 (union omp_clause_subcode): Add depend_kind, map_kind and
2042 proc_bind_kind fields.
2043 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2044 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2045 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2046 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2047 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2049 2013-10-10 Teresa Johnson <tejohnson@google.com>
2051 * predict.c (tree_estimate_probability): Add new parameter
2052 for estimate_bb_frequencies.
2053 (estimate_bb_frequencies): Add new parameter to force estimation.
2054 (rebuild_frequencies): When max frequency in function is small,
2055 recompute counts from frequencies.
2056 * predict.h (estimate_bb_frequencies): New parameter.
2058 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2060 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2061 optimizations are disabled.
2063 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2065 * coverage.c (coverage_finish): Fix leak of da_file_name.
2067 2013-10-10 Jan Hubicka <jh@suse.cz>
2069 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2070 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2073 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2076 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2077 if any labels are in FALLTHRU_BB, use a special label emitted
2078 immediately after the asm goto insn rather than label_rtx
2080 (expand_asm_stmt): Adjust caller.
2081 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2082 edge if the last insn in predecessor is a jump with single successor,
2083 but it isn't simplejump_p.
2085 2013-10-10 Richard Biener <rguenther@suse.de>
2087 PR tree-optimization/58656
2088 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2090 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2092 * gimplify.c: Include expr.h and tm_p.h for targets with special
2093 va-arg padding requirements.
2095 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2097 * tree-flow.h: Move some prototypes to gimple.h.
2098 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2099 * gimple.h: Relocate some prototypes from tree-flow.h
2100 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2102 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2103 (build_va_arg_indirect_ref): Relocate and make static.
2104 (std_gimplify_va_arg_expr): Relocate here.
2105 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2106 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2108 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2110 * doc/md.texi: Document the mnemonic attribute.
2112 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2115 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2116 (set_attr_alternative x ...) when searching for user defined
2119 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2121 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2123 2013-10-09 Easwaran Raman <eraman@google.com>
2125 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2126 * cfgexpand.c (defer_stack_allocation): ...use here
2127 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2129 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2131 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2132 (operand_equal_for_value_replacement): New function, extracted from
2133 value_replacement and enhanced to catch more cases.
2134 (value_replacement): Use operand_equal_for_value_replacement.
2136 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2138 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2139 get_max_loop_iterations.
2141 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2143 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2144 Place comment above function.
2146 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2148 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2149 are not related to tree-cfg.c.
2150 * tree-ssa-address.h: New file. Relocate prototypes.
2151 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2152 (addr_for_mem_ref): New. Combine call to get_address_description and
2153 return addr_for_mem_ref.
2154 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2155 * tree-ssa-live.h: Adjust prototypes.
2156 * passes.c: Include tree-ssa-live.h.
2157 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2158 * graphite.c (graphite_transform_loops): Make static.
2159 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2160 make_pass_graphite, pass_data_graphite_transforms,
2161 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2162 * ipa-pure-const.c (warn_function_noreturn): Make static.
2163 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2164 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2165 Relocate from tree-cfg.c
2166 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2168 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2169 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2170 Move to ipa-pure-const.c.
2171 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2172 Relocate from tree-optimize.c.
2173 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2174 make_pass_fixup_cfg): Move to tree-cfg.c.
2175 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2176 Relocate some prototypes.
2177 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2178 * tree-dump.c (dump_function_to_file): Remove prototype.
2179 Add tree-flow.h to the include file.
2180 * tree-dump.h: Remove prototype.
2181 * tree-parloops.h: New File. Add prototypes.
2182 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2183 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2184 from tree-ssa-loop.c.
2185 * tree-predcom.c (run_tree_predictive_commoning,
2186 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2187 Relocate here from tree-ssa-loop.c.
2188 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2189 ssa_name_values.release ().
2190 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2191 (ssa_name_values): Relocate from tree-flow.h.
2192 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2193 * tree-ssa-loop.c (run_tree_predictive_commoning,
2194 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2195 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2196 make_pass_graphite, pass_data_graphite_transforms,
2197 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2198 tree_parallelize_loops, pass_data_parallelize_loops,
2199 make_pass_parallelize_loops): Move to other files.
2200 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2202 * tree.h: Remove prototypes from tree-address.c.
2204 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2206 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2207 (struct int_tree_map): Move to tree-hasher.h
2208 (SCALE, LABEL, PERCENT): Move to gimple.h
2209 * tree-flow-inline.h: Delete. Move functions to other files.
2210 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2211 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2212 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2213 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2214 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2215 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2216 * tree-hasher.h: Don't include tree-flow.h.
2217 (struct int_tree_map): Relocate from tree-flow.h.
2218 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2219 tree-flow-inline.h and make static.
2220 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2222 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2223 tree-flow-inline.h and make static.
2224 * tree.h (is_global_var, may_be_aliased): Relocate from
2226 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2227 * value-prof.c: No longer include tree-flow-inline.h.
2228 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2230 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2232 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2233 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2234 (enum move_pos): Move to tree-ssa-loop-im.h
2235 * cfgloop.h: Move some prototypes.
2236 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2237 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2238 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2239 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2240 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2241 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2242 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2243 (enum move_pos): Relocate here.
2244 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2246 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2247 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2248 make_pass_lim): Relocate here from tree-ssa-loop.c.
2249 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2251 (loop_edge_to_cancel, unloop_loops): Make static.
2252 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2253 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2254 (tree_complete_unroll, gate_tree_complete_unroll,
2255 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2256 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2257 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2258 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2259 (stmt_invariant_in_loop_p): Remove unused function.
2260 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2261 * tree-ssa-loop-manip.h: New file. Add prototypes.
2262 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2263 (gcov_type_to_double_int): Move to cfgloop.h.
2264 (double_int_cmp, bound_index,
2265 estimate_numbers_of_iterations_loop): Make static.
2266 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2267 (max_loop_iterations): Factor out get_max_loop_iterations.
2268 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2270 * tree-ssa-loop-niter.h: New file. Add prototypes.
2271 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2272 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2273 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2274 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2275 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2276 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2277 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2278 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2279 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2280 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2281 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2282 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2283 pass_data_complete_unroll, make_pass_complete_unroll,
2284 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2285 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2286 tree-ssa-loop-ivcanon.c.
2287 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2288 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2289 tree-ssa-loop-prefetch.c.
2290 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2292 (get_lsm_tmp_name): Relocate and add suffix parameter.
2293 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2294 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2295 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2296 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2297 (get_estimated_loop_iterations): Factor out accessor from
2298 estimated_loop_iterations in tree-ssa-loop-niter.c.
2299 (get_max_loop_iterations): Factor out accessor from
2300 _max_loop_iterations in tree-ssa-niter.c.
2301 * loop-unroll.c (decide_unroll_constant_iterations,
2302 decide_unroll_runtime_iterations, decide_peel_simple,
2303 decide_unroll_stupid): Use new get_* accessors.
2305 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2307 PR tree-optimization/20318
2308 * doc/extend.texi (returns_nonnull): New function attribute.
2309 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2311 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2312 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2314 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2317 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2318 false if both components are bitfields.
2320 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2322 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2323 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2324 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2325 * config/aarch64/aarch64.h
2326 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2327 * config/aarch64/aarch64-simd-builtins.def
2328 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2330 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2332 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2334 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2336 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2338 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2340 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2341 (vneg_f64): New intrinsic.
2342 (vneg_s8): Asm replaced with C.
2343 (vneg_s16): Likewise.
2344 (vneg_s32): Likewise.
2345 (vneg_s64): New intrinsic.
2346 (vnegq_f32): Asm replaced with C.
2347 (vnegq_f64): Likewise.
2348 (vnegq_s8): Likewise.
2349 (vnegq_s16): Likewise.
2350 (vnegq_s32): Likewise.
2351 (vnegq_s64): Likewise.
2353 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2355 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2357 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2359 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2360 packed stack special handling.
2361 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2362 back to fixed stack slots for FPRs saved due to stdarg.
2364 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2366 * config/s390/s390.c (s390_frame_info): Restructure function.
2368 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2370 * config/s390/s390.c (struct s390_frame_layout): New field
2372 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2373 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2374 regs_ever_clobbered to char*.
2375 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2376 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2377 (s390_register_info): Enable FPR save slots. Move/Copy some
2378 functionality into ...
2379 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2380 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2382 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2383 not imply a stack frame.
2384 (s390_init_frame_layout): Remove variable clobbered_regs.
2385 (s390_update_register_info): Remove function.
2386 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2387 cannot be used for register renaming.
2388 (s390_hard_regno_scratch_ok): New function.
2389 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2390 (s390_initial_elimination_offset): Change offset calculation of
2391 the return address pointer.
2392 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2393 (restore_gprs): Set frame related flag.
2394 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2395 (s390_emit_prologue): Call s390_register_info instead of
2396 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2397 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2398 (s390_optimize_prologue): Call s390_optimize_register_info.
2399 Try to remove also FPR slot save/restore INSNs. Remove frame
2400 related flags from restore INSNs.
2402 2013-10-08 DJ Delorie <dj@redhat.com>
2404 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2407 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2408 avoid conflict with the MI use of %c.
2409 * config/rl78/rl78-real.md: change %c to %C throughout.
2410 * config/rl78/rl78-virt.md: Likewise.
2412 2013-10-08 Jan Hubicka <jh@suse.cz>
2414 * config/i386/i386.c (ix86_option_override_internal): Switch
2415 to SSE math for -ffast-math when target ISA supports SSE2.
2417 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2419 * tree-flow.h: Remove some prototypes.
2420 * tree.h: Remove some protypes, add a couple.
2421 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2422 using_eh_for_cleanups_p): Add interface routines for front ends.
2423 * tree-eh.h: New file. Add protoptyes.
2424 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2425 (add_stmt_to_eh_lp_fn): Make static.
2426 (lower_try_finally): Use new using_eh_for_cleanups_p.
2427 * emit-rtl.c: Include tree-eh.h.
2428 * gimple.h: Include tree-eh.h.
2430 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2432 PR tree-optimization/58480
2433 * tree-vrp.c (infer_nonnull_range): New function.
2434 (infer_value_range): Call infer_nonnull_range.
2436 2013-10-08 Dehao Chen <dehao@google.com>
2438 PR tree-optimization/58619
2439 * tree-inline.c (copy_phis_for_bb): Combine location data
2442 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2445 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2446 RTX_FRAME_RELATED_P on INSN.
2448 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2450 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2451 (altivec_expand_vec_perm_const): Call it.
2453 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2455 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2456 sequences for LE VSX loads and stores at expand time.
2457 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2459 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2460 (rs6000_gen_le_vsx_permute): New.
2461 (rs6000_gen_le_vsx_load): New.
2462 (rs6000_gen_le_vsx_store): New.
2463 (rs6000_gen_le_vsx_move): New.
2464 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2465 (*vsx_le_perm_load_v4si): New.
2466 (*vsx_le_perm_load_v8hi): New.
2467 (*vsx_le_perm_load_v16qi): New.
2468 (*vsx_le_perm_store_v2di): New.
2469 (*vsx_le_perm_store_v4si): New.
2470 (*vsx_le_perm_store_v8hi): New.
2471 (*vsx_le_perm_store_v16qi): New.
2472 (*vsx_xxpermdi2_le_<mode>): New.
2473 (*vsx_xxpermdi4_le_<mode>): New.
2474 (*vsx_xxpermdi8_le_V8HI): New.
2475 (*vsx_xxpermdi16_le_V16QI): New.
2476 (*vsx_lxvd2x2_le_<mode>): New.
2477 (*vsx_lxvd2x4_le_<mode>): New.
2478 (*vsx_lxvd2x8_le_V8HI): New.
2479 (*vsx_lxvd2x16_le_V16QI): New.
2480 (*vsx_stxvd2x2_le_<mode>): New.
2481 (*vsx_stxvd2x4_le_<mode>): New.
2482 (*vsx_stxvd2x8_le_V8HI): New.
2483 (*vsx_stxvd2x16_le_V16QI): New.
2485 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2487 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2489 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2491 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2492 loop to work also for 31bit ABI.
2493 Save the stack pointer for frame_size > 0.
2495 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2497 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2498 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2499 constraint letters from expanders.
2500 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2501 retry count to general_operand.
2502 ("tabort"): Give operand 0 a mode.
2503 ("tabort_1"): Add mode and constraint letter for operand 0.
2504 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2506 2013-10-04 Jeff Law <law@redhat.com>
2508 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2510 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2512 (thread_across_edge): Here. Call it.
2514 2013-10-04 Cary Coutant <ccoutant@google.com>
2516 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2517 discarding a location list expression (or a piece of one).
2519 2013-10-03 Jan Hubicka <jh@suse.cz>
2521 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2522 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2523 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2525 2013-10-03 Jan Hubicka <jh@suse.cz>
2527 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2528 accumulate-outgoing-args when producing unwind info.
2530 2013-10-03 Wei Mi <wmi@google.com>
2532 * lra-constraints.c (insert_move_for_subreg): New function
2533 extracted from simplify_operand_subreg.
2534 (simplify_operand_subreg): Add reload for paradoxical subreg.
2536 2013-10-03 Rong Xu <xur@google.com>
2538 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2539 the candidate of builtin_expect such that we should fix the
2540 size/time estimation.
2541 (estimate_function_body_sizes): Do the acutally size/time fix-up
2544 2013-10-03 Rong Xu <xur@google.com>
2546 * predict.c (tree_predict_by_opcode): Get the probability
2547 for builtin_expect from param builtin_expect_probability.
2548 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2549 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2550 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2552 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2555 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2556 * calls.c (alloca_call_p): Use get_callee_fndecl.
2557 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2558 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2560 (vrp_visit_stmt): Remove duplicated code.
2562 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2564 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2565 ceildf2, btruncdf2, instead of vsx_* name.
2567 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2568 iterators to only do V2DF and V4SF here. Move the DF code to
2569 rs6000.md where it is combined with SF mode. Replace <VSv> with
2570 just 'v' since only vector operations are handled with these insns
2571 after moving the DF support to rs6000.md.
2572 (vsx_sub<mode>3): Likewise.
2573 (vsx_mul<mode>3): Likewise.
2574 (vsx_div<mode>3): Likewise.
2575 (vsx_fre<mode>2): Likewise.
2576 (vsx_neg<mode>2): Likewise.
2577 (vsx_abs<mode>2): Likewise.
2578 (vsx_nabs<mode>2): Likewise.
2579 (vsx_smax<mode>3): Likewise.
2580 (vsx_smin<mode>3): Likewise.
2581 (vsx_sqrt<mode>2): Likewise.
2582 (vsx_rsqrte<mode>2): Likewise.
2583 (vsx_fms<mode>4): Likewise.
2584 (vsx_nfma<mode>4): Likewise.
2585 (vsx_copysign<mode>3): Likewise.
2586 (vsx_btrunc<mode>2): Likewise.
2587 (vsx_floor<mode>2): Likewise.
2588 (vsx_ceil<mode>2): Likewise.
2589 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2590 (vsx_sminsf3): Likewise.
2591 (vsx_fmadf4): Likewise.
2592 (vsx_fmsdf4): Likewise.
2593 (vsx_nfmadf4): Likewise.
2594 (vsx_nfmsdf4): Likewise.
2595 (vsx_cmpdf_internal1): Likewise.
2597 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2598 simpler to select whether a target has SPE or traditional floating
2599 point support in iterators.
2600 (TARGET_DF_SPE): Likewise.
2601 (TARGET_SF_FPR): Likewise.
2602 (TARGET_DF_FPR): Likewise.
2603 (TARGET_SF_INSN): Macros to say whether floating point support
2604 exists for a given operation for expanders.
2605 (TARGET_DF_INSN): Likewise.
2607 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2608 combining of SF/DF mode operations, using both traditional and VSX
2616 (abs<mode>2): Combine SF/DF modes using traditional floating point
2617 instructions. Add support for using the upper DF registers with
2618 VSX support, and SF registers with power8-vector support. Update
2619 expanders for operations supported by both the SPE and traditional
2620 floating point units.
2621 (abs<mode>2_fpr): Likewise.
2622 (nabs<mode>2): Likewise.
2623 (nabs<mode>2_fpr): Likewise.
2624 (neg<mode>2): Likewise.
2625 (neg<mode>2_fpr): Likewise.
2626 (add<mode>3): Likewise.
2627 (add<mode>3_fpr): Likewise.
2628 (sub<mode>3): Likewise.
2629 (sub<mode>3_fpr): Likewise.
2630 (mul<mode>3): Likewise.
2631 (mul<mode>3_fpr): Likewise.
2632 (div<mode>3): Likewise.
2633 (div<mode>3_fpr): Likewise.
2634 (sqrt<mode>3): Likewise.
2635 (sqrt<mode>3_fpr): Likewise.
2636 (fre<Fs>): Likewise.
2637 (rsqrt<mode>2): Likewise.
2638 (cmp<mode>_fpr): Likewise.
2639 (smax<mode>3): Likewise.
2640 (smin<mode>3): Likewise.
2641 (smax<mode>3_vsx): Likewise.
2642 (smin<mode>3_vsx): Likewise.
2643 (negsf2): Delete SF operations that are merged with DF.
2650 (fmasf4_fpr): Likewise.
2651 (fmssf4_fpr): Likewise.
2652 (nfmasf4_fpr): Likewise.
2653 (nfmssf4_fpr): Likewise.
2654 (sqrtsf2): Likewise.
2655 (rsqrtsf_internal1): Likewise.
2656 (smaxsf3): Likewise.
2657 (sminsf3): Likewise.
2658 (cmpsf_internal1): Likewise.
2659 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2660 (negdf2): Delete DF operations that are merged with SF.
2662 (nabsdf2): Likewise.
2668 (rsqrtdf_internal1): Likewise.
2669 (fmadf4_fpr): Likewise.
2670 (fmsdf4_fpr): Likewise.
2671 (nfmadf4_fpr): Likewise.
2672 (nfmsdf4_fpr): Likewise.
2673 (sqrtdf2): Likewise.
2674 (smaxdf3): Likewise.
2675 (smindf3): Likewise.
2676 (cmpdf_internal1): Likewise.
2677 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2678 (btrunc<mode>2): Delete separate expander, and combine with the
2679 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2680 (btrunc<mode>2_fpr): Likewise.
2681 (ceil<mode>2): Likewise.
2682 (ceil<mode>2_fpr): Likewise.
2683 (floor<mode>2): Likewise.
2684 (floor<mode>2_fpr): Likewise.
2685 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2686 Add support for using the upper registers with VSX and
2687 power8-vector. Move insns to be closer to the define_expands. On
2688 VSX systems, prefer the traditional form of FMA over the VSX
2689 version, since the traditional form allows the target not to
2690 overlap with the inputs.
2691 (fms<mode>4_fpr): Likewise.
2692 (nfma<mode>4_fpr): Likewise.
2693 (nfms<mode>4_fpr): Likewise.
2695 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2696 Richard Earnshaw <richard.earnshaw@arm.com>
2698 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2699 (struct mult_cost_table): Likewise.
2700 (struct mem_cost_table): Likewise.
2701 (struct fp_cost_table): Likewise.
2702 (struct vector_cost_table): Likewise.
2703 (cpu_cost_table): Likewise.
2704 * config/arm/arm.opt (mold-rts-costs): New option.
2705 (mnew-generic-costs): Likewise.
2706 * config/arm/arm.c (generic_extra_costs): New table.
2707 (cortexa15_extra_costs): Likewise.
2708 (arm_slowmul_tune): Use NULL as new costs.
2709 (arm_fastmul_tune): Likewise.
2710 (arm_strongarm_tune): Likewise.
2711 (arm_xscale_tune): Likewise.
2712 (arm_9e_tune): Likewise.
2713 (arm_v6t2_tune): Likewise.
2714 (arm_cortex_a5_tune): Likewise.
2715 (arm_cortex_a9_tune): Likewise.
2716 (arm_v6m_tune): Likewise.
2717 (arm_fa726te_tune): Likewise.
2718 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2719 (arm_cortex_tune): Use generict_extra_costs.
2720 (shifter_op_p): New function.
2721 (arm_unspec_cost): Likewise.
2722 (LIBCALL_COST): Define.
2723 (arm_new_rtx_costs): New function.
2724 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2725 table is available. Use old costs otherwise unless mnew-generic-costs
2727 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2728 (cpu_cost_table): Declare.
2730 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2733 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2734 (*subs_mul_imm_<mode>)
2735 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2736 (*sub_<shift>_<mode>)
2737 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2738 Remove k constraint.
2740 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2742 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2744 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2746 2013-10-02 Teresa Johnson <tejohnson@google.com>
2748 * predict.c (probably_never_executed): New function.
2749 (probably_never_executed_bb_p): Invoke probably_never_executed.
2750 (probably_never_executed_edge_p): Ditto.
2751 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2752 Treat profile insanities conservatively.
2754 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2756 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2758 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2760 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2761 use it. Use smaller increase for scratch. Don't increase reject
2762 for early clobber scratch.
2763 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2764 setting eliminated regs except setting fp from hfp.
2765 (lra_eliminate): Check lra_insn_recog_data on NULL.
2767 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2770 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2771 setting -mvsx-timode by default until the underlying problem is fixed.
2772 (RS6000_CPU, power7 defaults): Likewise.
2774 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2776 * config/x-linux (host-linux.o): Remove header dependencies.
2777 Use $(COMPILE) and $(POSTCOMPILE).
2778 * config/t-linux-android (linux-android.o): Ditto.
2780 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2782 * Makefile.in (expmed.o-warn): Remove.
2784 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2786 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2787 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2789 2013-10-02 Teresa Johnson <tejohnson@google.com>
2791 * dojump.c (do_jump_1): Divide probability between
2792 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2794 2013-10-02 Tom Tromey <tromey@redhat.com>
2796 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2798 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2800 * tree-flow.h: Remove some prototypes.
2801 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2802 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2803 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2804 mark_virtual_phi_result_for_renaming): Relocate here.
2805 * tree-into-ssa.h: Add prototypes.
2806 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2807 single_pred_before_succ_order.
2808 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2809 (nonfreeing_call_p) Move to gimple.c.
2810 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2812 * basic-block.h (single_pred_before_succ_order): Add prototype.
2813 * gimple.c (nonfreeing_call_p): Relocate here.
2814 * gimple.h: Add prototype.
2815 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2816 * tree-ssa-dom.h: New file. Relocate prototypes here.
2817 * tree-ssa.h: Include tree-ssa-dom.h.
2819 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2821 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2822 Use $(COMPILE) and $(POSTCOMPILE).
2824 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2826 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2828 * tree-flow.h: Remove some prototypes.
2829 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2830 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2831 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2832 propagate_tree_value*): Move from here to...
2833 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2834 propagate_tree_value*): Relocate here.
2835 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2836 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2837 * gimple-fold.c: Remove gimple-fold.h from include list.
2838 * tree-vrp.c: Remove gimple-fold.h from include list.
2839 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2840 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2841 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2842 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2843 * sese.c: Add tree-ssa-propagate.h to include list.
2845 2013-10-02 Richard Biener <rguenther@suse.de>
2847 * tree-loop-distribution.c: Include tree-vectorizer.h for
2849 (enum partition_kind): Remove PKIND_REDUCTION.
2850 (struct partition_s): Remove has_writes member, add reduction_p member.
2851 (partition_alloc): Adjust.
2852 (partition_builtin_p): Likewise.
2853 (partition_has_writes): Remove.
2854 (partition_reduction_p): New function.
2855 (partition_merge_into): Likewise.
2856 (generate_code_for_partition): Commonize builtin partition
2858 (rdg_cannot_recompute_vertex_p): Remove.
2859 (already_processed_vertex_p): Likewise.
2860 (rdg_flag_vertex): Do not set has_writes.
2861 (classify_partition): Adjust.
2862 (rdg_build_partitions): Do not set has_writes, treat all
2863 partitions as useful.
2864 (distribute_loop): Record number of library calls generated. Adjust.
2865 (tree_loop_distribution): Report number of loops and library
2866 calls generated as opt-info.
2868 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2870 * tree-flow.h: Include new .h files. Move prototypes.
2871 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2872 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2873 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2874 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2875 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2876 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2877 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2878 * tree.h (get_ref_base_and_extent): Move prototype out.
2879 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2881 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2882 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2883 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2884 * tree-scalar-evolution.c: Include tree.h.
2885 * sese.c: Include tree.h.
2886 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2887 * dwarf2out.c: Include tree-dfa.h.
2888 * tree-chrec.c: Include tree.h.
2889 * tree-data-ref.c: Include tree.h.
2891 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2893 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2896 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2898 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2899 Use $(COMPILE) and $(POSTCOMPILE).
2900 (sol2-cxx.o): Likewise.
2901 (sol2-stubs.o): Likewise.
2903 * config/x-solaris (host-solaris.o): Likewise.
2905 * config/sparc/t-sparc (sparc.o): Remove.
2906 (sparc-c.o): Remove header dependencies.
2907 Use $(COMPILE) and $(POSTCOMPILE).
2908 * config/sparc/x-sparc: Likewise.
2910 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2912 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2913 * config/arc/arc700.md: Likewise.
2914 * config/arc/arc-modes.def: Likewise.
2915 * config/arc/arc-simd.h: Likewise.
2916 * config/arc/t-arc-uClibc: Likewise.
2917 * config/arc/t-arc-newlib: Likewise.
2919 2013-10-02 Renlin Li <renlin.li@arm.com>
2921 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2923 (aarch64_expand_epilogue): Likewise.
2925 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2926 Yufeng Zhang <yufeng.zhang@arm.com>
2928 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2930 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2931 'base_in' represent a conversion and legal_cast_p_1 holds; set
2932 'base_in' with the returned value from get_unwidened.
2934 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2936 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2937 plus_constant is not used.
2939 2013-10-01 Wei Mi <wmi@google.com>
2941 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2942 * config/i386/i386.md: Add define_peephole2 to
2943 break partial reg stall for cvtss2sd/cvtsd2ss.
2945 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2947 * config/arc/arc.c (pass_arc_ifcvt::clone):
2948 Update for ctxt_ -> m_ctxt change.
2950 2013-10-01 Jeff Law <law@redhat.com>
2952 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2953 outgoing_edge and intermediate_edge fields. Instead store the path.
2954 (redirection_data::hash): Hash on the last edge's destination index.
2955 (redirection_data::equal): Check the entire thread path.
2956 (lookup_redirectio_data): Corresponding changes.
2957 (create_edge_and_update_destination_phis): Likewise.
2958 (thread_single_edge): Likewise.
2960 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2961 Diego Novillo <dnovillo@google.com>
2963 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2964 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2965 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2966 versions of these patterns.
2968 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2969 (__builtin_arc_aligned): Likewise.
2971 * config/arc/arc.md: Expand adc_0 comment stating the intended
2972 purpose and why it isn't ready.
2973 Replace commented out call_value_via_label_mixed with a
2974 plain comment about bl_s.
2976 * config/arc/arc.c (stdio.h): Don't include directly.
2977 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2978 Remove if (1) condition.
2979 (arc_encode_section_info): Fix comment.
2981 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2983 * config/arc/arc.c (arc_conditional_register_usage):
2984 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2985 Also set reg_alloc_order for DMA config regs.
2987 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2988 Jeremy Bennett <jeremy.bennett@embecosm.com>
2990 * doc/install.texi (--with-cpu): Mention ARC.
2991 (arc-*-elf32): New paragraph.
2992 (arc-linux-uclibc): Likewise.
2993 * doc/md.texi (Machine Constraints): Add ARC part.
2994 * doc/invoke.texi: (menu): Add ARC Options.
2995 (Machine Dependent Options) <ARC Options>: Add synopsis.
2996 (node ARC Options): Add.
2997 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2998 (ARC Built-in Functions): New section defining
2999 generic ARC built-in functions.
3000 (ARC SIMD Built-in Functions): New section defining SIMD specific
3002 (Declaring Attributes of Functions): Extended
3003 description of short_call and long_call attributes for ARC and
3004 added index entries.
3006 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3007 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3008 Joern Rennecke <joern.rennecke@embecosm.com>
3009 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3010 Brendan Kehoe <brendan@zen.org>
3011 Michael Eager <eager@eagercon.com>
3012 Simon Cook <simon.cook@embecosm.com>
3013 Jeremy Bennett <jeremy.bennett@embecosm.com>
3015 * config/arc, common/config/arc: New directories.
3017 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3018 Brendan Kehoe <brendan@zen.org>
3019 Simon Cook <simon.cook@embecosm.com>
3021 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3023 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3025 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3026 * tree-ssa-coalesce.h: New. Move prototype to here.
3027 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3028 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3029 (gimple_can_coalesce_p): Move to...
3030 * gimple.c (gimple_can_coalesce_p): Here.
3032 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3034 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3035 (dump_decl_set): Move to gimple.c.
3036 * gimple.h: Don't include tree-ssa-operands.h.
3037 (dump_decl_set): Add prototype.
3038 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3039 Move to gimple-ssa.h.
3040 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3041 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3042 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3043 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3044 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3046 (dump_decl_set): Relocate here.
3047 * gimple-ssa.h: New file.
3048 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3049 Relocate from gimple.h.
3050 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3051 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3052 swap_tree_operands and remove non-ssa path.
3053 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3054 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3056 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3057 vect_is_simple_reduction_1): Use swap_ssa_operands.
3058 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3059 (enum need_phi_state): Move to tree-into-ssa.c.
3060 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3061 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3062 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3063 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3064 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3065 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3066 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3067 num_imm_uses): Move to ssa-iterators.h.
3068 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3069 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3071 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3072 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3073 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3074 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3075 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3076 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3077 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3078 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3079 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3080 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3081 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3082 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3083 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3084 (set_phi_nodes): Move to tree-phinodes.h.
3085 * tree-ssa-operands.h (enum ssa_op_iter_type,
3086 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3087 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3088 (dump_decl_set): Remove prototype.
3089 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3090 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3091 (set_phi_nodes): Relocate from tree-flow-inline.h.
3092 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3094 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3095 include list. Temporarily add gimple.h to include list.
3096 * ssa-iterators.h: New file.
3097 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3098 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3099 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3100 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3101 Relocate from tree-ssa-operands.h.
3102 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3103 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3104 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3105 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3106 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3107 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3108 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3109 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3110 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3111 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3112 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3113 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3114 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3115 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3116 Relocate from tree-flow-inline.h.
3117 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3119 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3122 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3123 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3124 Insert '\t' to output template.
3125 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3126 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3127 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3128 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3130 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3132 * doc/install.texi (Host/target specific installation notes for GCC):
3133 Put @anchor before @heading.
3134 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3135 Use @email for email addresses.
3137 2013-10-01 Jeff Law <law@redhat.com>
3139 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3140 a vec. Only delete the path if we create one without successfully
3141 registering a jump thread.
3142 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3144 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3145 (paths): New vector of jump threading paths.
3146 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3147 (THREAD_PATH): New accessor macro for the entire thread path.
3148 (lookup_redirection_data): Get intermediate and final outgoing edge
3149 from the thread path.
3150 (create_edge_and_update_destination_phis): Copy the threading path.
3151 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3152 jump threading path.
3153 (ssa_redirect_edges): Get edges and block types from the jump threading
3154 path. Free the path vector.
3155 (thread_block): Get edges from the jump threading path. Look at the
3156 entire path to see if we thread to a loop exit. If we cancel a jump
3157 thread request, then free the path vector.
3158 (thread_single_edge): Get edges and block types from the jump threading
3159 path. Free the path vector.
3160 (thread_through_loop_header): Get edges and block types from the jump
3161 threading path. Free the path vector.
3162 (mark_threaded_blocks): Iterate over the vector of paths and store
3163 the path on the appropriate edge. Get edges and block types from the
3164 jump threading path.
3165 (mark_threaded_blocks): Get edges and block types from the jump
3166 threading path. Free the path vector.
3167 (thread_through_all_blocks): Use the vector of paths rather than
3168 a vector of 3-edge sets.
3169 (register_jump_thread): Accept pointer to a path vector rather
3170 than the path vector itself. Store the path vector for later use.
3173 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3174 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3177 * config/s390/s390.c (s390_split_branches): Modify check for table
3179 (s390_chunkify_start): Rearrange table jump insn check in order to
3180 deal with compare and branch insns correctly.
3182 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3186 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3187 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3188 define_insn_and_split.
3189 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3190 (shiftsi3_compare): New pattern.
3192 * config/arm/unspecs.md (UNSPEC_RRX): New.
3194 2013-10-01 Alan Modra <amodra@gmail.com>
3196 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3197 casing inout operands.
3199 2013-10-01 Richard Biener <rguenther@suse.de>
3201 PR tree-optimization/58553
3202 * tree-loop-distribution.c (struct partition_s): Add niter member.
3203 (classify_partition): Populate niter member for the partition
3204 and properly identify whether the relevant store happens before
3205 or after the loop exit.
3206 (generate_memset_builtin): Use niter member from the partition.
3207 (generate_memcpy_builtin): Likewise.
3209 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3211 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3212 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3214 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3216 * basic-block.h (edge_list): Prefix member names with "m_".
3217 * context.h (context): Likewise.
3218 * domwalk.h (dom_walker): Likewise.
3219 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3220 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3221 * hash-table.h (hash_table): Likewise.
3222 * machmode.h (bit_field_mode_iterator): Likewise.
3223 * pass_manager.h (pass_list): Likewise.
3224 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3225 * tree-pass.h (pass_data): Likewise.
3226 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3227 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3228 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3229 * asan.c (pass_data_asan): Update accordingly.
3230 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3231 (control_dependences::control_dependences): Likewise.
3232 (control_dependences::~control_dependences): Likewise.
3233 (control_dependences::~control_dependences): Likewise.
3234 (control_dependences::get_edges_dependent_on): Likewise.
3235 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3236 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3237 * context.c (gcc::context::context): Likewise.
3238 * cprop.c (pass_rtl_cprop::clone): Likewise.
3239 * domwalk.c (dom_walker::walk): Likewise.
3240 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3241 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3242 * mode-switching.c (pass_mode_switching::clone): Likewise.
3243 * passes.c (opt_pass::opt_pass): Likewise.
3244 (pass_manager::pass_manager): Likewise.
3245 * predict.c (pass_strip_predict_hints::clone): Likewise.
3246 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3247 (pass_split_all_insns::clone): Likewise.
3248 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3250 (bit_field_mode_iterator::next_mode): Likewise.
3251 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3252 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3253 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3254 * tree-complex.c (pass_lower_complex::clone): Likewise.
3255 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3256 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3257 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3258 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3259 (pass_fold_builtins::clone): Likewise.
3260 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3261 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3262 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3263 (pass_cd_dce::clone): Likewise.
3264 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3265 (pass_phi_only_cprop::clone): Likewise.
3266 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3267 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3268 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3269 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3270 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3271 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3272 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3273 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3274 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3275 * tree-vrp.c (pass_vrp::clone): Likewise.
3276 * tsan.c (pass_tsan::clone): Likewise.
3278 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3281 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3282 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3285 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3286 optimization, punt if sign_bit_p looked through any zero extension.
3288 2013-09-30 Teresa Johnson <tejohnson@google.com>
3290 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3291 Update redirected out edge count in joiner case.
3292 (ssa_redirect_edges): Common the joiner and non-joiner cases
3293 so that joiner case gets profile updates.
3295 2013-09-30 Richard Biener <rguenther@suse.de>
3297 PR tree-optimization/58554
3298 * tree-loop-distribution.c (classify_partition): Require
3299 unconditionally executed stores for memcpy and memset recognition.
3300 (tree_loop_distribution): Calculate dominance info.
3302 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3304 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3305 (NO_PROFILE_COUNTERS): Likewise.
3306 (PROFILE_HOOK): Likewise.
3307 (FUNCTION_PROFILER): Likewise.
3308 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3310 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3312 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3313 calls and defines in TARGET_MACHO conditional.
3314 (load_macho_picbase_di): Likewise.
3315 (reload_macho_picbase): Likewise.
3316 (reload_macho_picbase_si): Likewise.
3317 (reload_macho_picbase_di): Likewise.
3318 (nonlocal_goto_receiver): Likewise.
3320 2013-09-30 Nick Clifton <nickc@redhat.com>
3322 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3323 that use the MSP430X ISA.
3324 (msp430_option_override): Scan -mmcu command line option for any
3325 MCU name that supports the MSP430X ISA.
3326 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3327 -mmcu options which enable the MSP430X ISA.
3329 2013-09-30 Richard Biener <rguenther@suse.de>
3332 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3333 before looking for setjmp-like calls.
3335 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3338 * config/darwin-protos.h (machopic_get_function_picbase): New.
3339 * config/darwin.c (machopic_get_function_picbase): New.
3340 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3341 label for a new func. (load_macho_picbase_di): Likewise.
3342 (reload_macho_picbase): New expand.
3343 (reload_macho_picbase_si): New insn.
3344 (reload_macho_picbase_di): New insn.
3345 (nonlocal_goto_receiver): New define and split.
3346 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3347 (unspecv enum): Add UNSPECV_NLGR.
3349 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3351 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3352 that altivec registers are correctly sized on Darwin.
3354 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3356 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3357 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3358 * config/x-darwin (host-darwin.o): Likewise.
3359 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3360 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3361 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3363 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3365 * doc/invoke.texi: Fix usage of @tie{} command.
3367 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3369 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3371 2013-09-28 Jan Hubicka <jh@suse.cz>
3373 * config/i386/x86-tune.def: Add documentation for each of the options;
3376 2013-09-28 Jan Hubicka <jh@suse.cz>
3378 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3380 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3381 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3382 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3383 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3384 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3386 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3388 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3389 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3390 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3391 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3392 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3393 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3394 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3395 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3396 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3397 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3398 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3399 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3400 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3401 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3402 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3403 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3404 gimple.h, godump.c, graphite-clast-to-gimple.c,
3405 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3406 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3407 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3408 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3409 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3410 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3411 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3412 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3413 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3414 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3415 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3416 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3417 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3418 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3419 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3420 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3421 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3422 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3423 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3424 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3425 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3426 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3427 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3428 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3429 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3430 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3431 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3432 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3433 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3434 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3435 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3436 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3437 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3438 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3439 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3440 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3441 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3442 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3443 whitespace before "(".
3445 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3447 * expr.h (extract_bit_field): Remove packedp parameter.
3448 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3449 from forward declaration.
3450 (store_split_bit_field): Remove packedp arg from calls to
3451 extract_fixed_bit_field.
3452 (extract_bit_field_1): Remove packedp parameter and packedp
3453 argument from recursive calls and calls to extract_fixed_bit_field.
3454 (extract_bit_field): Remove packedp parameter and corresponding
3455 arg to extract_bit_field_1.
3456 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3458 (extract_split_bit_field): Remove packedp arg from call to
3459 extract_fixed_bit_field.
3460 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3461 (copy_blkmode_from_reg): Likewise.
3462 (copy_blkmode_to_reg): Likewise.
3463 (read_complex_part): Likewise.
3464 (store_field): Likewise.
3465 (expand_expr_real_1): Likewise.
3466 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3467 to extract_bit_field.
3468 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3469 call to extract_bit_field.
3470 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3471 call to extract_bit_field.
3472 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3473 and special packedp behavior from -fstrict-volatile-bitfields
3476 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3478 * lra-eliminations.c (init_elim_table): Guard value_p.
3480 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3482 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3483 DFmode, DImode, and SFmode in the upper VSX registers based on the
3484 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3485 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3486 of the VSX register handling.
3488 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3490 (zero_extendsidi2_lfiwzx): Add support for loading into the
3491 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3492 only do VSX memory options on Altivec registers.
3493 (extendsidi2_lfiwax): Likewise.
3494 (extendsfdf2_fpr): Likewise.
3495 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3496 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3497 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3498 (movdi_internal64): Likewise.
3500 2013-09-27 Xinliang David Li <davidxl@google.com>
3502 * opts.c (finish_options): Adjust parameters
3503 according to vect cost model.
3504 (common_handle_option): Set dynamic vect cost
3506 targhooks.c (default_add_stmt_cost): Compute stmt cost
3508 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3509 Use helper function.
3510 * tree-vectorizer.h (unlimited_cost_model): New function.
3511 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3512 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3514 (vect_enhance_data_refs_alignment): Ditto.
3515 * flag-types.h: New enum.
3516 * common/config/i386/i386-common.c (ix86_option_init_struct):
3517 No need to initialize vect_cost_model flag.
3518 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3521 2013-09-27 Diego Novillo <dnovillo@google.com>
3523 * gimple.h (enum ssa_mode): Remove.
3525 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3527 * cfgloop.h (number_of_loops): Fix typo in check for null.
3529 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3532 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3533 are children of outermost saved_cfun's loop, and set it up to
3534 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3535 if loop != loop0 and SESE region contains bbs that belong to loop0.
3537 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3539 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3540 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3541 (set_address_segment, set_address_base, set_address_index)
3542 (set_address_disp): Accept the argument unconditionally.
3543 (baseness): Remove must_be_base_p and must_be_index_p checks.
3544 (decompose_normal_address): Classify as much as possible in the
3547 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3549 * cse.c (count_reg_usage): Handle INT_LIST.
3550 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3551 * reginfo.c (reg_scan_mark_refs): Likewise.
3552 * reload1.c (eliminate_regs_1): Likewise.
3554 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3557 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3560 2013-09-27 Richard Biener <rguenther@suse.de>
3562 PR tree-optimization/58459
3563 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3564 restriction not propagating into loops.
3566 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3568 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3569 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3570 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3572 2013-09-26 Richard Biener <rguenther@suse.de>
3574 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3576 2013-09-26 Richard Biener <rguenther@suse.de>
3578 * alias.h (component_uses_parent_alias_set): Rename to ...
3579 (component_uses_parent_alias_set_from): ... this.
3580 * alias.c (component_uses_parent_alias_set): Rename to ...
3581 (component_uses_parent_alias_set_from): ... this and return
3583 (reference_alias_ptr_type_1): Use the result from
3584 component_uses_parent_alias_set_from instead of stripping
3585 components one at a time.
3586 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3588 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3590 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3591 Move prototypes to...
3592 * tree-ssa-ter.h: New File. Move prototypes here.
3593 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3594 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3595 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3596 from is_replaceable_p.
3597 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3598 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3599 (process_replaceable): Use ter_is_replaceable_p.
3600 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3601 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3602 newly refactored ssa_is_replaceable_p.
3603 * cfgexpand.c: Include tree-outof-ssa.h.
3604 * ssaexpand.h: Delete.
3606 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3608 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3609 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3611 (create_gimple_tmp): Delete.
3612 (get_expr_type, build_assign, build_type_cast): Move to...
3613 * gimple-builder.c: New File.
3614 (get_expr_type): Relocate from gimple.c.
3615 (build_assign, build_type_cast): Change to only create ssanames.
3616 * gimple.h: Move prototypes to...
3617 * gimple-builder.h: New File. Here.
3618 * tree-ssa.h: And here.
3619 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3620 count_uses_and_derefs): Relocate from gimple.c.
3621 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3622 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3623 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3624 gimple_replace_ssa_lhs.
3625 * asan.c: Include gimple-builder.h.
3626 * Makefile.in: Add gimple-builder.o.
3628 2013-09-26 Richard Biener <rguenther@suse.de>
3630 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3631 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3632 (loe_visit_block): Use gcc_checking_assert.
3633 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3634 gimple_assign_ssa_name_copy_p.
3635 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3638 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3640 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3641 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3643 2013-09-26 Richard Biener <rguenther@suse.de>
3645 PR tree-optimization/58539
3646 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3647 the fact that debug statements are not taking part in loop-closed
3650 2013-09-26 Nick Clifton <nickc@redhat.com>
3652 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3653 time warning message.
3654 (msp430_print_operand_raw): Delete unused letter parameter.
3655 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3656 (msp430_print_operand_address): New function.
3657 (msp430_print_operand): Move address printing code from here to
3659 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3661 (zero_extendpsisi2): Likewise.
3662 (extendpsisi2): New pattern.
3663 (andneghi3): New pattern.
3665 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3667 * config/aarch64/aarch64.opt (mlra): New option.
3668 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3669 (TARGET_LRA_P): Define.
3671 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3673 * expr.c (expand_assignment): Remove obsolete comment.
3675 2013-09-25 Jeff Law <law@redhat.com>
3677 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3678 tree-ssa-threadupdate.h.
3679 (register_jump_thread): Similarly.
3680 * tree-ssa-threadupdate.h: New header file.
3681 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3682 * tree-vrp.c: Likewise.
3683 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3684 (thread_around_empty_blocks): Change type of path vector argument to
3685 an edge,type pair from just an edge. Initialize both elements when
3686 appending to a jump threading path. Tweak references to elements
3688 (thread_across_edge): Similarly. Release memory for the elements
3690 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3691 (dump_jump_thread_path): New function broken out from
3692 register_jump_thread.
3693 (register_jump_thread): Use dump_jump_thread_path. Change type of
3694 path vector entries. Search the path for NULL edges and dump
3695 the path if one is found. Tweak the conversion of path to 3-edge
3696 form to use the block copy type information embedded in the path.
3698 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3700 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3702 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3703 Vladimir Makarov <vmakarov@redhat.com>
3705 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3706 from the least significant bit.
3707 (strip_address_mutations): Add bitfield operations handling.
3708 (must_be_index_p): Add shifting and rotate operations handling.
3709 (set_address_base): Use must_be_base_p predicate.
3710 (set_address_index): Use must_be_index_p predicate.
3712 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3713 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3714 Sergey Lega <sergey.s.lega@intel.com>
3715 Anna Tikhonova <anna.tikhonova@intel.com>
3716 Ilya Tocar <ilya.tocar@intel.com>
3717 Andrey Turetskiy <andrey.turetskiy@intel.com>
3718 Ilya Verbin <ilya.verbin@intel.com>
3719 Kirill Yukhin <kirill.yukhin@intel.com>
3720 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3722 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3724 (ix86_expand_vector_move_misalign): Support new unaligned load and
3725 stores and use new names.
3726 (CODE_FOR_sse2_storedqu): Rename to ...
3727 (CODE_FOR_sse2_storedquv16qi): ... this.
3728 (CODE_FOR_sse2_loaddqu): Rename to ...
3729 (CODE_FOR_sse2_loaddquv16qi): ... this.
3730 (CODE_FOR_avx_loaddqu256): Rename to ...
3731 (CODE_FOR_avx_loaddquv32qi): ... this.
3732 (CODE_FOR_avx_storedqu256): Rename to ...
3733 (CODE_FOR_avx_storedquv32qi): ... this.
3734 * config/i386/i386.md (fpint_logic): New.
3735 * config/i386/sse.md (VMOVE): Extend for AVX512.
3739 (VI_UNALIGNED_LOADSTORE): Ditto.
3740 (sse2_avx_avx512f): Ditto.
3741 (sse2_avx2): Extend for AVX512.
3742 (sse4_1_avx2): Ditto.
3743 (avx2_avx512f): New.
3744 (sse): Extend for AVX512.
3747 (avxsizesuffix): Ditto.
3748 (sseintvecmode): Ditto.
3750 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3751 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3752 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3753 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3754 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3755 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3756 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3757 (STORENT_MODE): Extend for AVX512.
3758 (*absneg<mode>2): Replace constraint "x" with "v".
3759 (*mul<mode>3): Ditto.
3760 (*ieee_smin<mode>3): Ditto.
3761 (*ieee_smax<mode>3): Ditto.
3762 (avx_cmp<mode>3): Replace VF with VF_128_256.
3763 (*<sse>_maskcmp<mode>3_comm): Ditto.
3764 (<sse>_maskcmp<mode>3): Ditto.
3765 (<sse>_andnot<mode>3): Extend for AVX512.
3766 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3767 (<code><mode>3, fpint_logic): New.
3768 (*<code><mode>3): Extend for AVX512.
3769 (avx512flogicsuff): New.
3770 (avx512f_<logic><mode>): Ditto.
3771 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3773 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3774 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3775 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3776 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3777 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3778 (xop_vpermil2<mode>3): Ditto.
3779 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3780 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3781 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3782 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3784 2013-09-25 Tom Tromey <tromey@redhat.com>
3786 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3787 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3788 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3789 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3790 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3791 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3792 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3793 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3794 (GRAPHITE_HTAB_H): Remove.
3796 2013-09-25 Tom Tromey <tromey@redhat.com>
3798 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3800 2013-09-25 Tom Tromey <tromey@redhat.com>
3802 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3804 2013-09-25 Tom Tromey <tromey@redhat.com>
3806 * config/i386/t-i386 (i386.o): Remove.
3807 (i386-c.o): Use COMPILE and POSTCOMPILE.
3809 2013-09-25 Tom Tromey <tromey@redhat.com>
3811 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3813 2013-09-25 Tom Tromey <tromey@redhat.com>
3815 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3816 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3817 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3818 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3819 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3820 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3821 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3822 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3823 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3824 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3825 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3826 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3827 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3828 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3829 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3830 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3831 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3832 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3833 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3834 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3835 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3836 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3837 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3838 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3839 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3840 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3841 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3842 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3843 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3844 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3845 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3846 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3847 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3848 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3849 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3850 (graphite-clast-to-gimple.o, graphite-dependences.o)
3851 (graphite-interchange.o, graphite-poly.o)
3852 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3853 (graphite-optimize-isl.o, tree-vect-loop.o)
3854 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3855 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3856 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3857 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3858 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3859 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3860 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3861 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3862 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3863 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3864 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3865 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3866 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3867 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3868 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3869 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3870 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3871 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3872 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3873 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3874 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3875 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3876 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3877 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3878 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3879 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3880 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3881 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3882 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3883 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3884 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3885 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3886 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3887 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3888 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3889 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3890 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3891 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3892 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3893 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3894 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3895 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3896 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3897 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3898 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3899 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3900 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3901 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3902 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3903 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3904 (gcov-dump.o): Remove.
3905 (default-c.o): Use COMPILE and POSTCOMPILE.
3906 (CFLAGS-gcc.o): New variable.
3907 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3909 2013-09-25 Tom Tromey <tromey@redhat.com>
3911 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3912 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3913 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3914 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3915 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3916 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3917 (c-family/array-notation-common.o, c-family/stub-objc.o)
3918 (c-family/c-ubsan.o): Remove.
3920 2013-09-25 Tom Tromey <tromey@redhat.com>
3922 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3924 2013-09-25 Tom Tromey <tromey@redhat.com>
3926 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3927 to add -DENABLE_SHARED_LIBGCC.
3928 (gcc.o): Don't use subshell.
3930 2013-09-25 Tom Tromey <tromey@redhat.com>
3932 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3933 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3934 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3935 * configure, config.in: Rebuild.
3937 2013-09-25 Tom Tromey <tromey@redhat.com>
3939 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3940 (COMPILE, POSTCOMPILE): New variables.
3941 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3942 (DEPFILES): New variable.
3943 Include ".Po" files.
3944 * configure.ac: Add checks for dependency checking.
3945 * configure, aclocal.m4: Regenerate.
3947 2013-09-25 Tom Tromey <tromey@redhat.com>
3949 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3950 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3952 2013-09-25 Tom Tromey <tromey@redhat.com>
3954 * Makefile.in (generated_files): Add options.h,
3955 target-hooks-def.h, insn-opinit.h,
3956 common/common-target-hooks-def.h, pass-instances.def,
3957 c-family/c-target-hooks-def.h.
3959 2013-09-25 Jeff Law <law@redhat.com>
3961 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3962 than foo[foo.length () - 1] to access last member in a vec.
3963 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3965 2013-09-25 Richard Biener <rguenther@suse.de>
3968 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3970 2013-09-25 Jan Hubicka <jh@suse.cz>
3972 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3975 2013-09-25 Marek Polacek <polacek@redhat.com>
3978 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3979 when determining the type name.
3981 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3983 * config/sh/sh.md: Fix formatting.
3985 2013-09-24 Xinliang David Li <davidxl@google.com>
3987 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3988 max peel iterations parameter.
3989 * param.def: New parameter.
3990 * doc/invoke.texi: Document New parameter.
3992 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3994 * gimple-pretty-print.c: Various whitespace tweaks.
3995 * tree-core.h: Likewise.
3996 * tree-pretty-print.c: Likewise.
3997 * tree-ssa-alias.c: Likewise.
3998 * tree-ssa-copy.c: Likewise.
3999 * tree-ssanames.c: Likewise.
4000 * tree-ssanames.h: Likewise.
4001 * tree-vrp.c: Likewise.
4003 2013-09-24 Alan Modra <amodra@gmail.com>
4007 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4008 for output operands that disallow regs. Don't use EXPAND_WRITE on
4011 2013-09-24 Richard Biener <rguenther@suse.de>
4014 * tree.c (reference_alias_ptr_type): Move ...
4015 * alias.c (reference_alias_ptr_type): ... here and implement
4016 in terms of the new reference_alias_ptr_type_1.
4017 (ref_all_alias_ptr_type_p): New helper.
4018 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4019 use ref_all_alias_ptr_type_p.
4020 (get_deref_alias_set): Add flag_strict_aliasing check here.
4021 (reference_alias_ptr_type_1): New function, split out from ...
4022 (get_alias_set): ... here.
4023 (alias_ptr_types_compatible_p): New function.
4024 * alias.h (reference_alias_ptr_type): Declare.
4025 (alias_ptr_types_compatible_p): Likewise.
4026 * tree.h (reference_alias_ptr_type): Remove.
4027 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4028 to compare MEM_REF alias types.
4030 2013-09-24 Richard Biener <rguenther@suse.de>
4032 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4034 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4036 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4037 reload helper function arrays into a single array reg_addr.
4038 (reload_fpr_gpr): Likewise.
4039 (reload_gpr_vsx): Likewise.
4040 (reload_vsx_gpr): Likewise.
4041 (struct rs6000_reg_addr): Likewise.
4042 (reg_addr): Likewise.
4043 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4044 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4045 (rs6000_init_hard_regno_mode_ok): Likewise.
4046 (rs6000_secondary_reload_direct_move): Likewise.
4047 (rs6000_secondary_reload): Likewise.
4049 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4050 constraints: wu, ww, and wy. Repurpose wv constraint added during
4051 power8 changes. Put wg constraint in alphabetical order.
4053 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4054 for future work to add ISA 2.07 VSX single precision support.
4055 (-mvsx-scalar-double): Change default from -1 to 1, update
4056 documentation comment.
4057 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4058 (-mupper-regs-df): New debug switch to control whether DF values
4059 can go in the traditional Altivec registers.
4060 (-mupper-regs-sf): New debug switch to control whether SF values
4061 can go in the traditional Altivec registers.
4063 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4065 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4066 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4067 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4068 (rs6000_debug_legitimate_address_p): Print if we are running
4069 before, during, or after reload.
4070 (rs6000_secondary_reload): Add a comment.
4071 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4073 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4074 constraints. Update documentation string.
4075 (wd constraint): Likewise.
4076 (wf constraint): Likewise.
4077 (wg constraint): Likewise.
4078 (wn constraint): Likewise.
4079 (ws constraint): Likewise.
4080 (wt constraint): Likewise.
4081 (wx constraint): Likewise.
4082 (wz constraint): Likewise.
4083 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4085 (ww constraint): Likewise.
4086 (wy constraint): Likewise.
4087 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4088 in the previous submissions.
4089 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4091 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4093 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4094 (REG_BR_PROB): Say that the probability is stored in an int_list.
4095 * reg-notes.def: Update commentary to mention INT_LIST.
4096 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4097 (INT_LIST): New rtx.
4098 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4099 * rtlanal.c (int_reg_note_p): New function.
4100 (alloc_reg_note): Assert that the note does not have an int argument.
4101 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4102 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4103 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4104 rather than an INSN_LIST. Handle INT_LIST.
4105 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4106 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4107 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4108 Manipulate them as ints rather than rtxes.
4109 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4110 * regmove.c (copy_src_to_dest): Likewise.
4111 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4113 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4114 into the cases that need it.
4115 * config/arm/arm.c (arm_unwind_emit): Likewise.
4117 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4118 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4119 * loop-doloop.c (add_test, doloop_modify): Likewise.
4120 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4121 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4122 * predict.c (combine_predictions_for_insn): Likewise.
4123 * print-rtl.c (print_rtx): Handle INT_LIST.
4124 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4125 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4126 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4127 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4128 (ix86_split_fp_branch, predict_jump): Likewise.
4129 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4130 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4131 * config/spu/spu.c (ea_load_store_inline): Likewise.
4133 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4134 value of a REG_BR_PROB note.
4135 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4136 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4137 * emit-rtl.c (try_split): Likewise.
4138 * predict.c (br_prob_note_reliable_p): Likewise.
4139 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4140 * reorg.c (mostly_true_jump): Likewise.
4141 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4142 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4143 * config/i386/i386.c (ix86_print_operand): Likewise.
4144 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4145 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4146 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4147 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4148 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4149 * config/sparc/sparc.c (output_cbranch): Likewise.
4150 * config/spu/spu.c (get_branch_target): Likewise.
4151 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4152 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4154 2013-09-23 Jan Hubicka <jh@suse.cz>
4156 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4158 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4159 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4161 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4162 (gimple_fold_call): Dump inconsistent devirtualizations; add
4163 sanity check for type based devirtualizations.
4164 * ipa-prop.c: Include ipa-utils.h
4165 (ipa_intraprocedural_devirtualization): Add sanity check.
4166 (try_make_edge_direct_virtual_call): Likewise.
4168 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4170 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4171 assignment statements.
4173 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4175 * gimple-pretty-print.c (dump_ssaname_info): New function.
4176 (dump_gimple_phi): Call it.
4177 (pp_gimple_stmt_1): Likewise.
4178 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4179 (range_info_def): Declare.
4180 * tree-pretty-print.c (pp_double_int): New function.
4181 (dump_generic_node): Call it.
4182 * tree-pretty-print.h (pp_double_int): Declare.
4183 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4184 * tree-ssanames.h (range_info_def): New structure.
4185 (value_range_type): Move definition here.
4186 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4188 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4190 (set_range_info): New function.
4191 (get_range_info): Likewise.
4192 (duplicate_ssa_name_range_info): Likewise.
4193 (duplicate_ssa_name_fn): Check pointer type and call
4194 duplicate_ssa_name_range_info.
4195 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4196 * tree-vrp.c (value_range_type): Remove definition, now in
4198 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4199 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4200 (SSA_NAME_RANGE_INFO): New macro.
4202 2013-09-23 Richard Biener <rguenther@suse.de>
4204 PR tree-optimization/58464
4205 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4206 (phi_trans_add): Change to add conditionally on being not present.
4207 (phi_translate_1): Remove recursion detection here.
4208 (phi_translate): Pre-seed the cache with NULL to catch
4209 recursion here in a more generic way.
4210 (bitmap_find_leader): Adjust comment.
4211 (get_representative_for): Dump value-numbers.
4212 (create_expression_by_pieces): Likewise.
4213 (insert_into_preds_of_block): Likewise.
4215 2013-09-23 Christian Bruel <christian.bruel@st.com>
4218 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4219 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4221 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4224 2013-09-20 Renlin Li <renlin.li@arm.com>
4226 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4228 (aarch64_expand_epilogue): Likewise.
4229 (aarch64_legitimize_reload_address): Likewise.
4231 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4233 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4234 NULL_TREE before pushing them onto the vector. Likewise for labels.
4236 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4238 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4239 comparison operators when -fno-trapping-math is in effect.
4240 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4241 comparison operators in TFmode and assert that unsupported operators
4243 (ia64_print_operand): Likewise.
4245 2013-09-21 Jan Hubicka <jh@suse.cz>
4247 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4248 (sse_typeless_stores): Enable for core
4249 (sse_load0_by_pxor): Likewise.
4250 (four_jump_limit): Disable for core.
4251 (pad_returns): Likewise.
4252 (avoid_vector_decode): Likewise.
4253 (fuse_cmp_and_branch): Enable for cores.
4254 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4256 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4259 * config/pa/pa.c (pa_option_override): Disable auto increment and
4260 decrement instructions until reload is completed.
4262 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4263 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4264 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4266 2013-09-20 DJ Delorie <dj@redhat.com>
4267 Nick Clifton <nickc@redhat.com>
4269 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4270 (need_to_save): Save bank 0 on interrupts.
4271 (characterize_address): Strip far address wrappers.
4272 (rl78_as_legitimate_address): Likewise.
4273 (transcode_memory_rtx): Likewise.
4274 (rl78_peep_movhi_p): Disable this peephole after devirt.
4275 (rl78_propogate_register_origins): Forget all origins when a
4277 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4278 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4280 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4281 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4282 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4284 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4286 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4288 * config/pa/pa.md: In "scc" insn patterns, change output template to
4289 handle const0_rtx in reg_or_0_operand operands.
4291 2013-09-20 Martin Husemann <martin@NetBSD.org>
4294 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4295 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4297 2013-09-20 Richard Biener <rguenther@suse.de>
4300 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4301 remembering SSA name version and block index.
4302 (new_scev_info_str): Adjust.
4303 (hash_scev_info): Likewise. Also hash the block index.
4304 (eq_scev_info): Adjust.
4305 (find_var_scev_info): Likewise.
4306 (struct instantiate_cache_entry): Remove.
4307 (struct instantiate_cache_type): Use a htab to map name, block
4309 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4310 (get_instantiated_value_entry): Likewise.
4311 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4312 (instantiate_scev_name): Adjust.
4314 2013-09-20 Jeff Law <law@redhat.com>
4316 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4318 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4320 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4321 Call aarch64_simd_expand_args to update op[argc].
4323 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4325 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4328 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4330 * gengtype.c (file_rules): Added rule for *.cc files.
4331 (get_output_file_with_visibility): Give fatal message when no
4334 2013-09-20 Renlin Li <renlin.li@arm.com>
4336 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4337 (aarch64_expand_epilogue): Likewise.
4338 (aarch64_legitimize_reload_address): Likewise.
4340 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4343 * expr.c (expand_assignment): Remove misalignp code path.
4345 2013-09-20 Marek Polacek <polacek@redhat.com>
4348 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4349 TYPE_PRECISION. Add asserts.
4351 2013-09-20 Richard Biener <rguenther@suse.de>
4353 PR tree-optimization/58453
4354 * tree-loop-distribution.c (distribute_loop): Apply the cost
4355 model for -ftree-loop-distribute-patterns, too.
4357 2013-09-20 Richard Biener <rguenther@suse.de>
4360 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4361 make type comparison less strict.
4363 2013-09-20 Alan Modra <amodra@gmail.com>
4365 * configure: Regenerate.
4366 * aclocal.m4: Regenerate.
4368 2013-09-20 Marek Polacek <polacek@redhat.com>
4371 * doc/extend.texi: Document that attribute used is meant to be used
4372 on variables with static storage duration.
4374 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4376 PR tree-optimization/58472
4377 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4378 simd_lane_access set inv_p = false.
4379 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4380 the simduid magic VAR_DECL.
4382 2013-09-19 Jan Hubicka <jh@suse.cz>
4384 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4386 2013-09-17 Jeff Law <law@redhat.com>
4388 * tree-ssa-dom.c (record_temporary_equivalences): New function
4389 split out of dom_opt_dom_walker::after_dom_children.
4390 (dom_opt_dom_walker::thread_across_edge): Move common code
4391 in here from dom_opt_dom_walker::after_dom_children.
4392 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4394 2013-09-19 Jan Hubicka <jh@suse.cz>
4396 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4397 (TARGET_GENERIC): Use PROCESOR_GENERIC
4398 (enum processor_type): Unify generic32 and 64.
4399 * i386.md (cpu): Likewise.
4400 * x86-tune.def (use_leave): Enable for generic32.
4401 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4402 * athlon.md: Change generic64 to generic in all occurences.
4403 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4404 (ix86_target_macros_internal): Likewise.
4405 * driver-i386.c (host_detect_local_cpu): Likewise.
4406 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4408 (generic_memcpy, generic_memset, generic_cost): This one.
4409 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4410 (m_GENERIC32, m_GENERIC64): Remove.
4411 (m_GENERIC): Turn into one flag.
4412 (processor_target): Unify generic tunnings.
4413 (ix86_option_override_internal): Replace generic32/64 by generic.
4414 (ix86_issue_rate): Likewise.
4415 (ix86_adjust_cost): Likewise.
4417 2013-09-19 Jan Hubicka <jh@suse.cz>
4419 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4420 of speculative flag.
4422 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4424 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4425 GOMP_sections_start, even if !exit_reachable.
4427 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4429 * lra-constraints.c (need_for_all_save_p): Use macro
4430 HARD_REGNO_CALL_PART_CLOBBERED.
4431 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4432 set up pseudo conflict hard regs.
4434 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4437 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4440 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4442 PR rtl-optimization/58438
4443 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4444 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4445 from previous subpasses.
4447 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4449 * arm.c (arm_get_frame_offsets): Validate architecture supports
4450 LDRD/STRD before accepting the tuning preference.
4451 (arm_expand_prologue): Likewise.
4452 (arm_expand_epilogue): Likewise.
4454 2013-09-18 Richard Biener <rguenther@suse.de>
4456 PR tree-optimization/58417
4457 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4458 have chrecs with symbols defined in the loop as operands.
4459 (chrec_fold_multiply): Likewise.
4460 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4461 parameters before folding binary operations.
4462 (struct instantiate_cache_entry_hasher): Remove.
4463 (struct instantiate_cache_type): Use a pointer-map.
4464 (instantiate_cache_type::instantiate_cache_type): New function.
4465 (instantiate_cache_type::get): Likewise.
4466 (instantiate_cache_type::set): Likewise.
4467 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4468 (get_instantiated_value_entry): Likewise.
4469 (global_cache): New global.
4470 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4471 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4472 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4473 (instantiate_scev_name): Adjust.
4474 (instantiate_scev): Construct global instead of local cache.
4475 (resolve_mixers): Likewise.
4477 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4478 Paolo Carlini <paolo.carlini@oracle.com>
4481 * doc/implement-cxx.texi: Fix references to the C++ standards.
4483 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4485 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4486 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4487 simd_lane_access drs, update also DR_ALIGNED_TO.
4489 2013-09-18 Marek Polacek <polacek@redhat.com>
4492 * doc/extend.texi: Document no_sanitize_undefined attribute.
4493 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4494 no_sanitize_undefined attribute.
4496 2013-09-18 Nick Clifton <nickc@redhat.com>
4498 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4499 (ASM_DECLARE_FUNCTION_NAME): Define.
4501 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4503 * compare-elim.c (find_comparison_dom_walker): New class
4504 (find_comparisons_in_bb): Rename to
4505 find_comparison_dom_walker::before_dom_children
4506 (find_comparisons): Adjust
4507 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4509 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4510 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4511 (init_walk_dominator_tree): Remove declaration.
4512 (fini_walk_dominator_tree): Remove declaration.
4513 * fwprop.c (single_def_use_dom_walker): New class
4514 (single_def_use_enter_block): Convert to
4515 single_def_use_dom_walker::before_dom_children.
4516 (single_def_use_leave_block): Convert to
4517 single_def_use_dom_walker::after_dom_children.
4518 (build_single_def_use_links): Adjust.
4519 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4521 (find_candidates_in_block): Convert to
4522 find_candidates_dom_walker::before_dom_children.
4523 (execute_strength_reduction): Adjust.
4524 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4525 (sese_dom_walker): New class.
4526 (sese_dom_walker::sese_dom_walker): New constructor.
4527 (sese_dom_walker::~sese_dom_walker): New destructor.
4528 (build_sese_conditions_before): Convert to
4529 sese_dom_walker::before_dom_children.
4530 (build_sese_conditions_after): Convert to
4531 sese_dom_walker::after_dom_children.
4532 (build_poly_scop): Adjust
4533 * tree-into-ssa.c (rewrite_dom_walker): New class
4534 (rewrite_enter_block): Convert to
4535 rewrite_dom_walker::before_dom_children.
4536 (rewrite_leave_block): Convert to
4537 rewrite_dom_walker::after_dom_children.
4538 (rewrite_update_dom_walker): New class.
4539 (rewrite_update_enter_block): Convert to
4540 rewrite_update_dom_walker::before_dom_children.
4541 (rewrite_update_leave_block): Convert to
4542 rewrite_update_dom_walker::after_dom_children.
4543 (rewrite_blocks, rewrite_into_ssa): Adjust.
4544 (mark_def_dom_walker): New class.
4545 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4546 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4547 (mark_def_sites_blocks): Convert to
4548 mark_def_dom_walker::before_dom_children.
4549 (mark_def_site_blocks): Remove.
4550 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4551 (tree_ssa_dominator_optimize): Adjust.
4552 (dom_thread_across_edge): Convert to method
4553 dom_opt_dom_walker::thread_across_edge.
4554 (dom_opt_enter_block): Convert to member function
4555 dom_opt_dom_walker::before_dom_children.
4556 (dom_opt_leave_block): Convert to member function
4557 dom_opt_dom_walker::after_dom_children.
4558 * tree-ssa-dse.c (dse_dom_walker): New class.
4559 (dse_enter_block): Convert to member function
4560 dse_dom_walker::before_dom_children.
4561 (tree_ssa_dse): Adjust.
4562 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4563 (determine_invariantness_stmt): Convert to method
4564 invariantness_dom_walker::before_dom_children.
4565 (determine_invariantness): Remove
4566 (move_computations_dom_walker): New class.
4567 (move_computations_stmt): Convert to method
4568 move_computations_dom_walker::before_dom_children.
4569 (move_computations, tree_ssa_lim): Adjust.
4570 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4571 (nt_init_block): Make method
4572 notrappping_dom_walker::before_dom_children.
4573 (nt_fini_block): Make
4574 method nontrapping_dom_walker::after_dom_children.
4575 (get_non_trapping): Adjust.
4576 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4577 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4578 (eliminate_leave_block): Make method.
4579 eliminate_dom_walker::after_dom_children.
4581 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4582 (strlen_enter_block): Make method
4583 strlen_dom_walker::before_dom_children.
4584 (strlen_leave_block): Make
4585 method strlen_dom_walker::after_dom_children.
4586 (tree_ssa_strlen): Adjust.
4587 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4588 (tree_ssa_uncprop): Adjust.
4589 (uncprop_leave_block): Make method
4590 uncprop_dom_walker::after_dom_children.
4591 (uncprop_leave_block): Make method
4592 uncprop_dom_walker::before_dom_children.
4594 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4596 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4597 branch in same basic block. Check both src and dest of the move insn.
4599 2013-09-17 Nick Clifton <nickc@redhat.com>
4601 * config/rl78/rl78-real.md (bf): New pattern.
4603 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4604 (rl78_print_operand): Do not put a # before a %B.
4605 * config/rl78/rl78.opt: Tweak doc strings.
4607 2013-09-17 DJ Delorie <dj@redhat.com>
4609 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4610 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4611 (rl78_option_override): Likewise, if -mallregs.
4612 (is_virtual_register): Likewise.
4613 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4614 (REGNO_OK_FOR_BASE_P): Likewise.
4615 * config/rl78/rl78.opt (-mallregs): New.
4617 2013-09-17 Nick Clifton <nickc@redhat.com>
4619 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4620 For interrupt functions: save all call clobbered registers if the
4621 interrupt handler is not a leaf function.
4622 (rl78_expand_prologue): Always recompute the frame information.
4623 For interrupt functions: only select bank 0 if one of the bank 0
4624 registers is going to be psuhed.
4626 2013-09-17 DJ Delorie <dj@redhat.com>
4628 * config/rl78/constraints.md: For each W* constraint, rename to C*
4629 and create a W* constraint that checks for an optional ES: prefix
4631 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4632 (es_addr): New. Used to wrap far addresses.
4633 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4634 (rl78_es_base): New.
4635 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4636 wrapped far addresses.
4637 (rl78_print_operand_1): Unwrap far addresses before processing.
4638 (rl78_lo16): Wrap far addresses in unspecs.
4639 (rl78_es_addr): New.
4640 (rl78_es_base): New.
4641 (insn_ok_now): Check for not-yet-wrapped far addresses.
4642 (transcode_memory_rtx): Properly re-wrap far addresses.
4644 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4646 * config/sparc/t-rtems: Add leon3 multilibs.
4648 2013-09-17 Cong Hou <congh@google.com>
4650 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4651 when checking the dot production pattern. The type of rhs operand
4652 of multiply is now checked correctly.
4654 2013-09-17 Jeff Law <law@redhat.com>
4656 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4657 edge implied equivalences into successor phis.
4658 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4659 here from tree-ssa-threadedge.c.
4660 (mark_threaded_blocks): When threading through a joiner, if both
4661 successors of the joiner's clone reach the same block, verify the
4662 PHI arguments are equal. If not, cancel the jump threading request.
4663 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4664 tree-ssa-threadupdate.c
4665 (thread_across_edge): Don't check PHI argument equality when
4666 threading through joiner block here.
4668 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4670 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4671 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4672 (warn_uninit, warn_uninitialized_vars,
4673 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4674 Move to tree-ssa-uninit.c.
4675 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4676 (has_undefined_value_p): New. Pass dependant parts of
4677 ssa_undefined_value_p.
4678 (uninit_undefined_value_p): Use has_undefined_value_p.
4679 (warn_uninit, warn_uninitialized_vars,
4680 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4681 Move from tree-ssa.c.
4682 * tree-ssa.h: Adjust prototypes.
4684 2013-09-17 Jan Hubicka <jh@suse.cz>
4687 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4688 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4689 FUNCTION_NOT_OPTIMIZED.
4690 * ipa-inline-analysis.c (compute_inline_parameters): Function
4691 not optimized is not inlinable unless it is alwaysinline.
4692 (inline_analyze_function): Force calls in not optimized
4693 function not inlinable.
4695 2013-09-17 Jan Hubicka <jh@suse.cz>
4698 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4700 * ipa.c (function_and_variable_visibility): Likewise.
4701 * ipa-profile.c (ipa_profile): Likewise.
4703 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4706 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4707 attribute, and don't inline the resolver in this case.
4709 2013-09-17 Teresa Johnson <tejohnson@google.com>
4711 * coverage.c (get_coverage_counts): Add missing newline.
4713 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4715 PR tree-optimization/58088
4716 * fold-const.c (mask_with_trailing_zeros): New function.
4717 (fold_binary_loc): Make sure we don't recurse infinitely
4718 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4719 Use mask_with_trailing_zeros where appropriate.
4721 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4723 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4724 of 'prev' var to get better distance estimation.
4726 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4728 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4729 * tree-inline.c (is_parameter_of): New predicate.
4730 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4731 a parameter has been remapped.
4732 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4733 (optimize_inline_calls): Initialize transform_parameter.
4734 (copy_gimple_seq_and_replace_locals): Likewise.
4735 (tree_function_versioning): Likewise.
4736 (maybe_inline_call_in_expr): Likewise.
4738 2013-09-17 Nick Clifton <nickc@redhat.com>
4740 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4741 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4743 (is_attr_func): New function.
4744 (msp430_is_interrupt_func): New function.
4745 (is_naked_func): New function.
4746 (is_reentrant_func): New function.
4747 (is_critical_func): New function.
4748 (msp430_start_function): Add annotations for function attributes.
4749 (msp430_attr): New function.
4750 (msp430_attribute_table): New.
4751 (msp430_function_section): New function.
4752 (TARGET_ASM_FUNCTION_SECTION): Define.
4753 (msp430_builtin): New enum.
4754 (msp430_init_builtins): New function.
4755 (msp430_builtin_devl): New function.
4756 (msp430_expand_builtin): New function.
4757 (TARGET_INIT_BUILTINS): Define.
4758 (TARGET_EXPAND_BUILTINS): Define.
4759 (TARGET_BUILTIN_DECL): Define.
4760 (msp430_expand_prologue): Add support for naked, interrupt,
4761 critical and reentrant functions.
4762 (msp430_expand_epilogue): Likewise.
4763 (msp430_print_operand): Handle 'O' character.
4764 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4766 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4767 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4768 (pushm): Use a 'n' rather than an 'i' constraint.
4769 (msp_return): Add generation of the interrupt return instruction.
4770 (disable_interrupts): New pattern.
4771 (enable_interrupts): New pattern.
4772 (push_intr_state): New pattern.
4773 (pop_intr_state): New pattern.
4774 (bic_SR): New pattern.
4775 (bis_SR): New pattern.
4776 * doc/extend.texi: Document MSP430 function attributes and builtin
4779 2013-09-17 Richard Biener <rguenther@suse.de>
4781 PR tree-optimization/58432
4782 * tree-loop-distribution.c (tree_loop_distribution): Also
4783 scan PHIs for outside loop uses and seed a partition from them.
4785 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4787 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4788 (restructure_reference): Call backtrace_base_for_ref.
4790 2013-09-17 Alan Modra <amodra@gmail.com>
4793 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4796 2013-09-16 DJ Delorie <dj@redhat.com>
4798 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4799 vregs location for RL78/G10.
4800 (rl78_expand_prologue): Avoid SEL on G10.
4801 (rl78_expand_epilogue): Likewise.
4802 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4803 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4804 __RL78_G10__ when appropriate.
4805 (ASM_SPEC): Pass -mg10 along to the assembler.
4806 * config/rl78/rl78.md (sel_rb): Disable for G10.
4807 * config/rl78/rl78.opt: Add -mg10 option.
4808 * config/rl78/t-rl78: Add -mg10 multilib.
4810 2013-09-16 Xinliang David Li <davidxl@google.com>
4812 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4813 * omp-low.c (omp_max_vf): Ditto.
4814 (expand_omp_simd): Ditto.
4815 * tree-vectorizer.c (vectorize_loops): Ditto.
4816 (gate_vect_slp): Ditto.
4817 (gate_increase_alignment): Ditto.
4818 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4819 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4820 (gate_tree_vectorize): Name change.
4821 (tree_vectorize): Ditto.
4822 (pass_vectorize::gate): Call new function.
4823 (pass_vectorize::execute): Ditto.
4824 * opts.c: O3 default setting change.
4825 (finish_options): Check new flag.
4826 * doc/invoke.texi: Document new flags.
4827 * common.opt: New flags.
4829 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4831 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4832 (Condition Code): Don't reference it.
4833 * doc/tm.texi: Regenerate.
4835 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4838 * lra-constraints.c (undo_optional_reloads): Consider all optional
4839 reload even if it did not get a hard reg.
4841 2013-09-16 Teresa Johnson <tejohnson@google.com>
4843 * dumpfile.c (dump_loc): Remove newline emission.
4844 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4845 emission to dump_printf_loc calls where missing.
4846 (vect_mark_for_runtime_alias_test): Ditto.
4847 (vect_analyze_data_ref_dependence): Ditto.
4848 (vect_analyze_data_ref_dependences): Ditto.
4849 (vect_slp_analyze_data_ref_dependence): Ditto.
4850 (vect_slp_analyze_data_ref_dependences): Ditto.
4851 (vect_compute_data_ref_alignment): Ditto.
4852 (vect_update_misalignment_for_peel): Ditto.
4853 (vect_verify_datarefs_alignment): Ditto.
4854 (vector_alignment_reachable_p): Ditto.
4855 (vect_get_data_access_cost): Ditto.
4856 (vect_enhance_data_refs_alignment): Ditto.
4857 (vect_find_same_alignment_drs): Ditto.
4858 (vect_analyze_data_refs_alignment): Ditto.
4859 (vect_analyze_group_access): Ditto.
4860 (vect_analyze_data_ref_access): Ditto.
4861 (vect_analyze_data_ref_accesses): Ditto.
4862 (vect_prune_runtime_alias_test_list): Ditto.
4863 (vect_analyze_data_refs): Ditto.
4864 (vect_create_addr_base_for_vector_ref): Ditto.
4865 (vect_create_data_ref_ptr): Ditto.
4866 (vect_grouped_store_supported): Ditto.
4867 (vect_grouped_load_supported): Ditto.
4868 * value-prof.c (check_counter): Ditto.
4869 (check_ic_target): Ditto.
4870 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4871 (vect_recog_widen_mult_pattern): Ditto.
4872 (vect_recog_widen_sum_pattern): Ditto.
4873 (vect_recog_over_widening_pattern): Ditto.
4874 (vect_recog_widen_shift_pattern): Ditto.
4875 (vect_recog_rotate_pattern): Ditto.
4876 (vect_recog_vector_vector_shift_pattern): Ditto.
4877 (vect_recog_divmod_pattern): Ditto.
4878 (vect_recog_mixed_size_cond_pattern): Ditto.
4879 (vect_recog_bool_pattern): Ditto.
4880 (vect_pattern_recog_1): Ditto.
4881 (vect_pattern_recog): Ditto.
4882 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4883 (vect_is_simple_iv_evolution): Ditto.
4884 (vect_analyze_scalar_cycles_1): Ditto.
4885 (vect_get_loop_niters): Ditto.
4886 (vect_analyze_loop_1): Ditto.
4887 (vect_analyze_loop_form): Ditto.
4888 (vect_analyze_loop_operations): Ditto.
4889 (vect_analyze_loop_2): Ditto.
4890 (vect_analyze_loop): Ditto.
4891 (report_vect_op): Ditto.
4892 (vect_is_slp_reduction): Ditto.
4893 (vect_is_simple_reduction_1): Ditto.
4894 (vect_get_known_peeling_cost): Ditto.
4895 (vect_estimate_min_profitable_iters): Ditto.
4896 (vect_model_reduction_cost): Ditto.
4897 (vect_model_induction_cost): Ditto.
4898 (get_initial_def_for_induction): Ditto.
4899 (vect_create_epilog_for_reduction): Ditto.
4900 (vectorizable_reduction): Ditto.
4901 (vectorizable_induction): Ditto.
4902 (vectorizable_live_operation): Ditto.
4903 (vect_loop_kill_debug_uses): Ditto.
4904 (vect_transform_loop): Ditto.
4905 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4906 (vect_stmt_relevant_p): Ditto.
4907 (process_use): Ditto.
4908 (vect_mark_stmts_to_be_vectorized): Ditto.
4909 (vect_model_simple_cost): Ditto.
4910 (vect_model_promotion_demotion_cost): Ditto.
4911 (vect_model_store_cost): Ditto.
4912 (vect_get_store_cost): Ditto.
4913 (vect_model_load_cost): Ditto.
4914 (vect_get_load_cost): Ditto.
4915 (vect_init_vector_1): Ditto.
4916 (vect_get_vec_def_for_operand): Ditto.
4917 (vect_finish_stmt_generation): Ditto.
4918 (vectorizable_call): Ditto.
4919 (vectorizable_conversion): Ditto.
4920 (vectorizable_assignment): Ditto.
4921 (vectorizable_shift): Ditto.
4922 (vectorizable_operation): Ditto.
4923 (vectorizable_store): Ditto.
4924 (vectorizable_load): Ditto.
4925 (vectorizable_condition): Ditto.
4926 (vect_analyze_stmt): Ditto.
4927 (vect_transform_stmt): Ditto.
4928 (vect_is_simple_use): Ditto.
4929 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4930 (vect_can_advance_ivs_p): Ditto.
4931 (vect_update_ivs_after_vectorizer): Ditto.
4932 (vect_do_peeling_for_loop_bound): Ditto.
4933 (vect_gen_niters_for_prolog_loop): Ditto.
4934 (vect_update_inits_of_drs): Ditto.
4935 (vect_create_cond_for_alias_checks): Ditto.
4936 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4937 (vect_build_slp_tree_1): Ditto.
4938 (vect_supported_load_permutation_p): Ditto.
4939 (vect_analyze_slp_instance): Ditto.
4940 (vect_analyze_slp): Ditto.
4941 (vect_make_slp_decision): Ditto.
4942 (vect_detect_hybrid_slp): Ditto.
4943 (vect_bb_vectorization_profitable_p): Ditto.
4944 (vect_slp_analyze_bb_1): Ditto.
4945 (vect_update_slp_costs_according_to_vf): Ditto.
4946 (vect_get_mask_element): Ditto.
4947 (vect_transform_slp_perm_load): Ditto.
4948 (vect_schedule_slp_instance): Ditto.
4949 (vect_schedule_slp): Ditto.
4950 (vect_slp_transform_bb): Ditto.
4951 * profile.c (read_profile_edge_counts): Ditto.
4952 (compute_branch_probabilities): Ditto.
4953 * coverage.c (get_coverage_counts): Ditto.
4955 2013-09-16 Diego Novillo <dnovillo@google.com>
4957 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4959 2013-09-16 Jan Hubicka <jh@suse.cz>
4961 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4962 abstract functions; for static functions check the presence of body.
4964 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4966 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4967 * config/aarch64/aarch64-simd.md
4968 (aarch64_mla_elt<mode>): New.
4969 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4970 (aarch64_mls_elt<mode>): Likewise.
4971 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4972 (aarch64_fma4_elt<mode>): Likewise.
4973 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4974 (aarch64_fma4_elt_to_128v2df): Likewise.
4975 (aarch64_fma4_elt_to_64df): Likewise.
4976 (fnma<mode>4): Likewise.
4977 (aarch64_fnma4_elt<mode>): Likewise.
4978 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4979 (aarch64_fnma4_elt_to_128v2df): Likewise.
4980 (aarch64_fnma4_elt_to_64df): Likewise.
4981 * config/aarch64/iterators.md (VDQSF): New.
4982 * config/aarch64/arm_neon.h
4983 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4984 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4986 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4988 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4989 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4990 (aarch64_mul3_elt_to_128df): Likewise.
4991 (aarch64_mul3_elt_to_64v2df): Likewise.
4992 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4994 (VMUL_CHANGE_NLANES) Likewise.
4997 * config/aarch64/arm_neon.h
4998 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5000 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5002 * config/aarch64/arm_neon.h
5003 (vcvtx_high_f32_f64): Fix parameters.
5005 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5006 Uros Bizjak <ubizjak@gmail.com>
5008 * config/alpha.c: Include tree-ssanames.h.
5010 2013-09-16 Richard Biener <rguenther@suse.de>
5012 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5013 (dot_rdg_1): Handle control_dd.
5014 (create_edge_for_control_dependence): New function.
5015 (create_rdg_edges): Add control dependences if asked for.
5016 (build_rdg): Likewise.
5017 (generate_loops_for_partition): If there are not necessary
5018 control stmts remove all their dependencies.
5019 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5020 (distribute_loop): Pass on control dependences.
5021 (tree_loop_distribution): Compute control dependences and remove
5022 restriction on number of loop nodes.
5024 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5026 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5029 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5031 * cse.c (try_const_anchors): Punt on CC modes.
5033 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5035 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5037 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5040 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5043 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5046 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5048 (darwin_objc1_section): Likewise.
5049 (darwin_file_end): Emit Image Info section when required.
5051 2013-09-14 Jan Hubicka <jh@suse.cz>
5053 * tree-into-ssa.c (gate_into_ssa): New.
5054 (pass_data_build_ssa): Use it.
5055 * cgraph.h (expand_thunk): Update prototype.
5056 * cgraphunit.c (analyze_function): Expand thunks early.
5057 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5058 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5059 set lowered flag; do not add new function.
5060 (assemble_thunks_and_aliases): Update.
5061 * tree-ssa.c (gate_init_datastructures): New gate.
5062 (pass_data_init_datastructures): Use it.
5064 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5067 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5068 xmm register set described in the psABI.
5070 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5072 * dwarf2out.c (should_emit_struct_debug): Add check
5073 for type_decl variable is not NULL.
5075 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5077 * config.gcc: Use new winnt-c.c target hooks
5078 * config/t-winnt: New file
5079 * config/winnt-c.c: New file
5080 * doc/tm.texi.in: Document new hook
5081 * doc/tm.texi: Regenerated
5083 2013-09-13 Jan Hubicka <jh@suse.cz>
5086 * ipa-inline.c (check_callers): New function.
5087 (check_caller_edge): Remove.
5088 (want_inline_function_to_all_callers_p): Also permit alises that are
5090 (inline_to_all_callers): Terminate the walk when devirtualization
5091 introduce new calls.
5093 2013-09-13 Jan Hubicka <jh@suse.cz>
5095 * ipa-inline-analysis.c (struct growth_data): Add node.
5096 (do_estimate_growth_1): Fix detection of recursion.
5098 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5100 PR tree-optimization/58392
5101 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5102 to avoid shadowing of outer loop variable. If
5103 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5104 replace_by_duplicate_decl simduid of loops that have it set and
5105 set dest_cfun->has_simduid_loops and/or
5106 dest_cfun->has_force_vect_loops.
5107 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5108 instead of maybe_lookup_decl.
5109 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5110 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5111 call. Copy over force_vect and copy and remap simduid. Set
5112 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5113 (copy_cfg_body): Remove blocks_to_copy argument. Use
5114 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5115 caller. Don't set cfun->has_simduid_loops and/or
5116 cfun->has_force_vect_loops here.
5117 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5119 (expand_call_inline, tree_function_versioning): Adjust copy_body
5122 2013-09-13 Martin Jambor <mjambor@suse.cz>
5125 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5126 the assert if the edge was a speculative one.
5128 2013-09-13 Richard Biener <rguenther@suse.de>
5130 * tree-data-ref.h (known_dependences_p): Move here ...
5131 * tree-loop-distribution.c (known_dependences_p): ... from here.
5132 (dump_rdg_component, debug_rdg_component): Remove.
5134 (generate_loops_for_partition): Use gimple_uid instead of
5135 relying on matching stmt visit order.
5136 (rdg_build_partitions): Take starting stmt vector.
5137 (ldist_gen): Merge into ...
5138 (distribute_loop): ... this function. Do not compute starting
5140 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5142 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5144 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5145 Set type attribute correctly. Set predicable_short_it attribute.
5146 (cmpsi_shiftsi): Remove %? from output template.
5148 2013-09-13 Richard Biener <rguenther@suse.de>
5150 * tree-loop-distribution.c (struct rdg_component,
5151 rdg_defs_used_in_other_loops_p, free_rdg_components,
5152 rdg_build_components): Remove.
5153 (stmts_from_loop): Do not record virtual PHIs.
5154 (generate_loops_for_partition): Skip virtual PHIs.
5155 (build_rdg_partition_for_component): Rename to ...
5156 (build_rdg_partition_for_vertex): ... this and adjust.
5157 (rdg_build_partitions): Take a vector of starting vertices
5158 instead of components. Remove unnecessary leftover handling.
5159 (ldist_gen): Do not build components or record other stores.
5160 (distribute_loop): Do not distribute loops containing stmts
5163 2013-09-13 Christian Bruel <christian.bruel@st.com>
5166 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5168 2013-09-13 Kai Tietz <ktietz@redhat.com>
5170 * config.gcc: Separate cases for mingw and cygwin targets,
5171 and add 64-bit cygwin target case.
5173 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5174 dll-export inline-functions.
5175 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5178 2013-09-13 Jeff Law <law@redhat.com>
5182 2013-09-06 Jeff Law <law@redhat.com>
5184 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5185 edge implied equivalences into successor phis.
5187 2013-09-12 DJ Delorie <dj@redhat.com>
5189 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5192 2013-09-12 Brooks Moses <bmoses@google.com>
5195 * Makefile.in: Do not install driver binaries in $(target)/bin.
5197 2013-09-12 DJ Delorie <dj@redhat.com>
5199 * config/rl78/rl78.opt (mrelax): New.
5200 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5201 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5203 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5205 (rl78_expand_epilogue): Likewise.
5207 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5210 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5211 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5212 pointer elimination with using remove_reg_equal_offset_note.
5214 2013-09-12 DJ Delorie <dj@redhat.com>
5216 * config/msp430/: New port.
5217 * config.gcc (msp430): Added.
5218 * doc/invoke.texi: Document MSP430 options.
5219 * doc/install.texi: Document msp430-elf
5220 * doc/md.texi: Document msp430-elf
5221 * doc/contrib.texi: Document msp430-elf
5223 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5226 2013-09-12 Martin Jambor <mjambor@suse.cz>
5229 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5230 reference descriptor is NULL.
5231 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5234 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5236 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5237 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5238 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5239 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5240 make_temp_ssa_name): move to tree-ssanames.h
5241 * tree-ssa-alias.h: Move prototype.
5242 * tree-ssa.h: Include tree-ssanames.h.
5243 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5244 * tree-ssanames.h: New. Move items from tree-flow*.h
5245 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5247 2013-09-12 Richard Biener <rguenther@suse.de>
5249 PR tree-optimization/58404
5250 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5251 propagate non-invariant addresses into dereferences wrapped
5252 in component references.
5254 2013-09-12 Richard Biener <rguenther@suse.de>
5256 PR tree-optimization/58402
5257 * passes.def: Move pass_late_warn_uninitialized later.
5259 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5261 * tree-ssa.h: New. Move content from tree-flow.h and
5263 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5264 Move prototypes belonging to tree-ssa.c.
5265 * tree-flow-inline.h (redirect_edge_var_map_def,
5266 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5268 * gimple.h: Adjust prototypes.
5269 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5271 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5272 * tree.h: Move prototype to tree-ssa.h.
5273 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5274 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5275 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5276 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5277 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5278 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5279 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5280 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5281 graphite-blocking.c, graphite-clast-to-gimple.c,
5282 graphite-dependences.c, graphite-interchange.c,
5283 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5284 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5285 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5286 ipa-reference.c, ipa-split.c, ipa-utils.c,
5287 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5288 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5289 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5290 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5291 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5292 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5293 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5294 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5295 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5296 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5297 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5298 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5299 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5300 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5301 value-prof.c, var-tracking.c,
5302 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5304 2013-09-12 Richard Biener <rguenther@suse.de>
5306 PR tree-optimization/58396
5307 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5308 (build_rdg): Take a loop-nest parameter, fix memleaks.
5309 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5311 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5313 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5316 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5318 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5319 The FMA instruction names should have a 'v' prefix.
5321 2013-09-12 Richard Biener <rguenther@suse.de>
5323 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5324 (dot_rdg): Use popen instead of system in optional code.
5325 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5326 (already_processed_vertex_p): Adjust.
5327 (has_anti_or_output_dependence, predecessor_has_mem_write,
5328 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5329 rdg_flag_uses): Remove.
5330 (rdg_flag_vertex): Simplify.
5331 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5333 (build_rdg_partition_for_component): Process the first vertex
5334 of a component only.
5335 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5337 2013-09-12 Alan Modra <amodra@gmail.com>
5339 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5341 2013-09-11 DJ Delorie <dj@redhat.com>
5342 Nick Clifton <nickc@redhat.com>
5344 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5345 (rl78_stack_based_mem): New.
5346 * config/rl78/constraints.md (Iv08): New.
5355 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5357 (movsi): Change from expand to insn-and-split.
5358 (ashrsi3): Clobber AX.
5362 * config/rl78/rl78.md (CC_REG): Fix.
5363 (addsi3): Allow memory and immediate operands.
5364 (addsi3_internal): Split into...
5365 (addsi3_internal_virt): ...new, and ...
5366 (addsi3_internal_real): ...new.
5368 (subsi3_internal_virt): New.
5369 (subsi3_internal_real): New.
5370 (mulsi3): Add memory operand.
5371 (mulsi3_rl78): Likewise.
5372 (mulsi3_g13): Likewise.
5373 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5374 (cbranchqi4_real): Add more constraint options.
5375 (cbranchhi4_real): Expand pattern.
5376 (cbranchhi4_real_signed): New.
5377 (cbranchhi4_real_inverted): New.
5378 (cbranchsi4_real_lt): New.
5379 (cbranchsi4_real_ge): New.
5380 (cbranchsi4_real_signed): New.
5381 (cbranchsi4_real): New.
5383 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5385 (lshrsi3_virt): Likewise.
5386 (ashlsi3_virt): Likewise.
5387 (cbranchqi4_virt_signed): New.
5388 (cbranchhi4_virt_signed): New.
5389 (cbranchsi4_virt): New.
5390 * config/rl78/rl78.c: Whitespace fixes throughout.
5391 (move_elim_pass): New.
5392 (pass_data_rl78_move_elim): New.
5393 (pass_rl78_move_elim): New.
5394 (make_pass_rl78_move_elim): New.
5395 (rl78_devirt_info): Run devirt earlier.
5396 (rl78_move_elim_info): New.
5397 (rl78_asm_file_start): Register it.
5398 (rl78_split_movsi): New.
5399 (rl78_as_legitimate_address): Allow virtual base registers when
5401 (rl78_addr_space_convert): Remove spurious debug stuff.
5402 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5403 (rl78_print_operand): More cases for not printing '#'.
5404 (rl78_expand_compare): Remove most of the logic.
5405 (content_memory): New.
5406 (clear_content_memory): New.
5407 (get_content_index): New.
5408 (get_content_name): New.
5409 (display_content_memory): New.
5410 (update_content): New.
5411 (record_content): New.
5412 (already_contains): New.
5413 (insn_ok_now): Re-recog insns with virtual registers.
5414 (add_postponed_content_update): New.
5415 (process_postponed_content_update): New.
5416 (gen_and_emit_move): New.
5417 (transcode_memory_rtx): Record new location content.
5418 Use gen_and_emit_move.
5419 (force_into_acc): New.
5420 (move_to_acc): Use gen_and_emit_move.
5421 (move_from_acc): Likewise.
5422 (move_acc_to_reg): Likewise.
5423 (move_to_x): Likewise.
5424 (move_to_hl): Likewise.
5425 (move_to_de): Likewise.
5426 (rl78_alloc_physical_registers_op1): Record location content.
5427 (has_constraint): New.
5428 (rl78_alloc_physical_registers_op2): Record location content.
5430 (rl78_alloc_physical_registers_ro1): Likewise.
5431 (rl78_alloc_physical_registers_cmp): Likewise.
5432 (rl78_alloc_physical_registers_umul): Likewise.
5433 (rl78_alloc_address_registers_macax): New.
5434 (rl78_alloc_physical_registers): Initialize and set location
5435 content memory as needed.
5436 (rl78_reorg): Make sure split2 is called.
5437 (rl78_rtx_costs): New.
5439 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5441 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5442 for (not (neg ...)) and (neg (not ...)) cases.
5444 2013-09-11 Richard Biener <rguenther@suse.de>
5447 * passes.def: Split critical edges before late uninit warning passes.
5448 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5450 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5452 PR tree-optimization/58385
5453 * fold-const.c (build_range_check): If both low and high are NULL,
5454 use omit_one_operand_loc to preserve exp side-effects.
5456 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5458 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5460 2013-09-11 Richard Biener <rguenther@suse.de>
5462 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5463 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5464 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5465 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5466 stmts_from_loop, known_dependences_p, build_empty_rdg,
5467 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5468 * tree-loop-distribution.c: ... here.
5469 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5470 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5471 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5472 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5473 * tree-loop-distribution.c: ... here.
5474 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5475 (struct partition_s): Add loops member.
5476 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5477 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5478 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5480 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5481 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5482 Sergey Lega <sergey.s.lega@intel.com>
5483 Anna Tikhonova <anna.tikhonova@intel.com>
5484 Ilya Tocar <ilya.tocar@intel.com>
5485 Andrey Turetskiy <andrey.turetskiy@intel.com>
5486 Ilya Verbin <ilya.verbin@intel.com>
5487 Kirill Yukhin <kirill.yukhin@intel.com>
5488 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5490 * config/i386/constraints.md (k): New.
5492 * config/i386/i386.c (const regclass_map): Add new mask registers.
5493 (dbx_register_map): Ditto.
5494 (dbx64_register_map): Ditto.
5495 (svr4_dbx_register_map): Ditto.
5496 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5498 (ix86_preferred_reload_class): Disable constants for mask registers.
5499 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5500 (ix86_hard_regno_mode_ok): Support new mask registers.
5501 (x86_order_regs_for_local_alloc): Ditto.
5502 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5503 (FIXED_REGISTERS): Add new mask registers.
5504 (CALL_USED_REGISTERS): Ditto.
5505 (REG_ALLOC_ORDER): Ditto.
5506 (VALID_MASK_REG_MODE): New.
5507 (FIRST_MASK_REG): Ditto.
5508 (LAST_MASK_REG): Ditto.
5509 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5510 (MAYBE_MASK_CLASS_P): New.
5511 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5512 (REG_CLASS_CONTENTS): Ditto.
5513 (MASK_REGNO_P): New.
5514 (ANY_MASK_REG_P): Ditto.
5515 (HI_REGISTER_NAMES): Add new mask registers.
5516 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5517 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5519 (attribute "type"): Add mskmov, msklog.
5520 (attribute "length_immediate"): Support them.
5521 (attribute "memory"): Ditto.
5522 (attribute "prefix_0f"): Ditto.
5523 (*movhi_internal): Support new mask registers.
5524 (*movqi_internal): Ditto.
5525 (define_split): Split out clobber pattern is a logic
5526 insn on mask registers.
5527 (*k<logic><mode>): New.
5528 (*andhi_1): Extend to support mask regs.
5529 (*andqi_1): Extend to support mask regs.
5531 (define_split): Split and-not to and and not if operands
5533 (*<code><mode>_1): Separate HI mode to new pattern...
5534 (*<code>hi_1): This.
5535 (*<code>qi_1): Extend to support mask regs.
5537 (kortestzhi): Ditto.
5538 (kortestchi): Ditto.
5540 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5541 (*one_cmplhi2_1): ...Here, now with mask registers support.
5542 (*one_cmplqi2_1): Support new mask registers.
5543 (HI/QImode arithmetics splitter): Don't split if mask registers
5545 (HI/QImode not splitter): Ditto.
5546 * config/i386/predicated.md (mask_reg_operand): New.
5547 (general_reg_operand): Ditto.
5549 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5551 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5552 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5554 2013-09-10 Jeff Law <law@redhat.com>
5556 PR tree-optimization/58380
5557 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5558 of threading through a buried loop header.
5560 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5561 return value for single successor case.
5563 2013-09-10 Jan Hubicka <jh@suse.cz>
5565 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5567 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5570 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5571 support conditional execution.
5572 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5574 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5576 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5578 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5579 Check destination too to check move insn.
5580 (undo_optional_reloads): Add check that the original peudo did not
5581 changed its allocation and the optional reload was inherited on last
5582 inheritance pass. Break loop after deciding to keep optional reload.
5583 (lra_undo_inheritance): Add check that inherited pseudo still in
5586 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5588 * config/aarch64/aarch64.md (generic_sched): New.
5589 * config/aarch64/aarch64-generic.md (load): Make conditional
5590 on generic_sched attribute.
5591 (nonload): Likewise.
5593 2013-09-10 Jan Hubicka <jh@suse.cz>
5595 * lto-cgraph.c: Include ipa-utils.h.
5596 (compute_ltrans_boundary): Also add possible targets into the boundary.
5598 2013-09-10 Jan Hubicka <jh@suse.cz>
5600 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5601 VAR_DECL of vtable rather than full expression.
5603 2013-09-10 Jan Hubicka <jh@suse.cz>
5604 Paolo Carlini <paolo.carlini@oracle.com>
5606 * cgraphunit.c (analyze_functions): Save input_location, set it
5607 to UNKNOWN_LOCATION and restore it at the end.
5609 2013-09-10 Martin Jambor <mjambor@suse.cz>
5611 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5612 represented by a thunk.
5614 2013-09-10 Jeff Law <law@redhat.com>
5616 PR tree-optimization/58343
5617 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5618 jump threading requests through loop headers buried in the middle
5619 of a jump threading path.
5621 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5622 in return value/type.
5624 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5626 PR rtl-optimization/58365
5627 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5628 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5631 2013-09-10 Richard Biener <rguenther@suse.de>
5633 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5634 * tree-data-ref.c (create_rdg_vertices): Collect all data
5635 references, signal failure to the caller, use data-ref API.
5636 (build_rdg): Compute data references only once. Maintain lifetime
5637 of data references and data dependences from within RDG.
5638 (free_rdg): Free dependence relations.
5639 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5640 inventing extra dependences.
5641 (distribute_loop): Update for RDG API changes.
5643 2013-09-10 Kai Tietz <ktietz@redhat.com>
5645 * doc/invoke.texi (fms-extensions): Document changed
5646 behavior for ms-abi targets.
5647 * config/i386/i386.c (ix86_option_override_internal):
5648 Set default value of option -fms-extension for ms-abi targets.
5650 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5652 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5654 2013-09-10 Alan Modra <amodra@gmail.com>
5657 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5659 2013-09-10 Alan Modra <amodra@gmail.com>
5661 * config/rs6000/predicates.md (add_cint_operand): New.
5662 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5663 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5664 using add_cint_operand.
5665 (largetoc_high_plus_aix): Likewise.
5667 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5669 PR tree-optimization/58364
5670 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5671 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5672 the current range can't be an unconditional true or false.
5674 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5676 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5678 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5680 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5682 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5685 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5687 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5689 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5690 NO_REGS for immediate that can't be moved directly into FP_REGS.
5692 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5694 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5695 comparison with negated operand.
5696 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5699 2013-09-09 Richard Biener <rguenther@suse.de>
5702 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5703 of a subloop record all its block as affecting loop-closed SSA form.
5705 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5707 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5708 of an rtx/bitpos pair.
5709 (store_fixed_bit_field): Update accordingly.
5711 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5713 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5715 * builtins.c (expand_errno_check): Likewise.
5716 * dwarf2cfi.c (init_return_column_size): Likewise.
5717 * except.c (sjlj_mark_call_sites): Likewise.
5718 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5719 * lra-constraints.c (emit_inc): Likewise.
5720 * ree.c (combine_set_extension): Likewise.
5721 * regmove.c (fixup_match_2): Likewise.
5722 * reload1.c (inc_for_reload): Likewise.
5724 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5726 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5727 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5728 Use gen_int_mode with the mode of the associated simplify_* call.
5729 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5730 * expmed.c (expand_shift_1): Likewise.
5731 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5732 * loop-iv.c (iv_number_of_iterations): Likewise.
5733 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5734 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5736 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5738 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5739 of the associated expand_* call.
5740 (asan_emit_stack_protection): Likewise.
5741 * builtins.c (round_trampoline_addr): Likewise.
5742 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5743 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5744 (emit_store_flag): Likewise.
5745 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5746 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5748 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5749 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5750 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5751 * modulo-sched.c (generate_prolog_epilog): Likewise.
5752 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5753 (expand_ctz, expand_ffs, expand_unop): Likewise.
5755 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5757 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5758 of the associated gen_rtx_* call.
5759 * caller-save.c (init_caller_save): Likewise.
5760 * combine.c (find_split_point, make_extraction): Likewise.
5761 (make_compound_operation): Likewise.
5762 * dwarf2out.c (mem_loc_descriptor): Likewise.
5763 * explow.c (plus_constant, probe_stack_range): Likewise.
5764 * expmed.c (expand_mult_const): Likewise.
5765 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5766 * reload1.c (init_reload): Likewise.
5767 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5768 * var-tracking.c (adjust_mems): Likewise.
5769 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5770 rather than gen_rtx_fmt_ee.
5772 2013-09-09 Jan Hubicka <jh@suse.cz>
5775 * value-prof.c (gimple_ic): Copy also abnormal edges.
5777 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5779 * asan.c (asan_shadow_cst): Use gen_int_mode.
5781 2013-09-08 Jan Hubicka <jh@suse.cz>
5783 * ipa-profile.c: Add toplevel comment.
5784 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5785 (contains_hot_call_p): New function.
5786 (ipa_propagate_frequency): Set frequencies based on counts when
5788 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5789 profile; do not tamper with profile after inlining if it is read.
5791 2013-09-08 Jan Hubicka <jh@suse.cz>
5793 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5796 2013-09-08 Jan Hubicka <jh@suse.cz>
5798 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5801 2013-09-08 Jeff Law <law@redhat.com>
5804 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5807 2013-09-08 Andi Kleen <ak@linux.intel.com>
5809 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5811 2013-09-08 Jan Hubicka <jh@suse.cz>
5813 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5815 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5817 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5818 for non-debug insns.
5819 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5820 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5821 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5824 2013-09-08 Jan Hubicka <jh@suse.cz>
5826 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5827 targets and devirtualize to BUILT_IN_UNREACHABLE.
5828 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5829 * ipa.c (walk_polymorphic_call_targets): New function.
5830 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5831 functions; use the new timevar.
5832 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5833 was removed from the program.
5834 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5835 not consider it in the walk when its vtable is dead.
5836 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5838 (devirt_variable_node_removal_hook): New function.
5839 (possible_polymorphic_call_targets): Also register
5840 devirt_variable_node_removal_hook.
5841 (ipa_devirt): Do not do non-speculative devirtualization.
5842 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5844 2013-09-08 Jan Hubicka <jh@suse.cz>
5846 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5847 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5848 varpool_remove_variable_insertion_hook): Declare.
5849 * varpool.c (varpool_node_hook_list): New structure.
5850 (first_varpool_node_removal_hook,
5851 first_varpool_variable_insertion_hook): New variables.
5852 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5853 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5854 varpool_remove_variable_insertion_hook,
5855 varpool_call_variable_insertion_hooks): New functions.
5856 (varpool_remove_node): Use it.
5858 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5861 * diagnostic.c (diagnostic_build_prefix): When s.file is
5862 "<built-in>" don't output line and column numbers.
5864 2013-09-06 Jan Hubicka <jh@suse.cz>
5866 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5867 * lto-streamer-out.c: Stream thunks, too.
5868 * lto-streamer-in.c (input_function): Pop cfun here
5869 (lto_read_body): Instead of here.
5871 2013-09-06 Caroline Tice <cmtice@google.com>
5873 * doc/install.texi: Add documentation for the --enable-vtable-verify
5874 and the --disable-libvtv configure options.
5876 2013-09-06 Jeff Law <law@redhat.com>
5878 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5879 edge implied equivalences into successor phis.
5881 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5883 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5885 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5887 * resource.c (mark_target_live_regs): Compute resources taking
5888 into account if a call is predicated or not.
5890 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5892 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5893 the compiler in the function names.
5895 2013-09-06 Jan Hubicka <jh@suse.cz>
5898 * ipa-inline.c (has_caller_p): New function.
5899 (want_inline_function_to_all_callers_p): Use it.
5900 (sum_callers, inline_to_all_callers): Break out from ...
5901 (ipa_inline): ... here.
5903 2013-09-06 Jan Hubicka <jh@suse.cz>
5905 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5906 only when AVX is enabled.
5908 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5910 * config/aarch64/aarch64.md
5911 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5912 is fpsimd_<load/store>2.
5913 (load_pair<mode>): Likewise.
5914 (store_pair<mode>): Likewise.
5916 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5918 * config/arm/types.md (type): Add "mrs" type.
5919 * config/aarch64/aarch64.md
5920 (aarch64_load_tp_hard): Make type "mrs".
5922 (load_tp_hard): Make type "mrs".
5923 * config/arm/cortex-a15.md: Update with new attributes.
5924 * config/arm/cortex-a5.md: Update with new attributes.
5925 * config/arm/cortex-a53.md: Update with new attributes.
5926 * config/arm/cortex-a7.md: Update with new attributes.
5927 * config/arm/cortex-a8.md: Update with new attributes.
5928 * config/arm/cortex-a9.md: Update with new attributes.
5929 * config/arm/cortex-m4.md: Update with new attributes.
5930 * config/arm/cortex-r4.md: Update with new attributes.
5931 * config/arm/fa526.md: Update with new attributes.
5932 * config/arm/fa606te.md: Update with new attributes.
5933 * config/arm/fa626te.md: Update with new attributes.
5934 * config/arm/fa726te.md: Update with new attributes.
5936 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5938 * config/aarch64/aarch64.md
5939 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5940 (*movtf_aarch64): Likewise.
5942 (thumb1_movdi_insn): Use "multiple" for type where more than one
5943 instruction is used for a move.
5944 (*arm32_movhf): Likewise.
5945 (*thumb_movdf_insn): Likewise.
5947 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5949 * config/arm/types.md (type): Rename fcpys to fmov.
5951 (*arm_movsi_vfp): Rename type fcpys as fmov.
5952 (*thumb2_movsi_vfp): Likewise
5953 (*movhf_vfp_neon): Likewise
5954 (*movhf_vfp): Likewise
5955 (*movsf_vfp): Likewise
5956 (*thumb2_movsf_vfp): Likewise
5957 (*movsfcc_vfp): Likewise
5958 (*thumb2_movsfcc_vfp): Likewise
5959 * config/aarch64/aarch64-simd.md
5960 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5961 * config/aarch64/aarch64.md
5962 (*movsi_aarch64): Replace type mov_reg with fmovs.
5963 (*movdi_aarch64): Likewise
5964 (*movsf_aarch64): Likewise
5965 (*movdf_aarch64): Likewise
5967 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5968 * config/arm/iwmmxt.md
5969 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5970 * config/arm/arm1020e.md: Update with new attributes.
5971 * config/arm/cortex-a15-neon.md: Update with new attributes.
5972 * config/arm/cortex-a5.md: Update with new attributes.
5973 * config/arm/cortex-a53.md: Update with new attributes.
5974 * config/arm/cortex-a7.md: Update with new attributes.
5975 * config/arm/cortex-a8-neon.md: Update with new attributes.
5976 * config/arm/cortex-a9.md: Update with new attributes.
5977 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5978 * config/arm/cortex-r4f.md: Update with new attributes.
5979 * config/arm/marvell-pj4.md: Update with new attributes.
5980 * config/arm/vfp11.md: Update with new attributes.
5982 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5984 * config/aarch64/aarch64.md
5985 (*madd<mode>): Fix type attribute.
5986 (*maddsi_uxtw): Likewise.
5987 (*msub<mode>): Likewise.
5988 (*msubsi_uxtw): Likewise.
5989 (<su_optab>maddsidi4): Likewise.
5990 (<su_optab>msubsidi4): Likewise.
5992 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5994 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5995 * config/arm/arm.md (core_cycles): Remove fdiv.
5996 * config/arm/vfp.md:
5997 (*sqrtsf2_vfp): Update for attribute changes.
5998 (*sqrtdf2_vfp): Likewise.
5999 * config/aarch64/aarch64.md:
6000 (sqrt<mode>2): Update for attribute changes.
6001 * config/arm/arm1020e.md: Update with new attributes.
6002 * config/arm/cortex-a15-neon.md: Update with new attributes.
6003 * config/arm/cortex-a5.md: Update with new attributes.
6004 * config/arm/cortex-a53.md: Update with new attributes.
6005 * config/arm/cortex-a7.md: Update with new attributes.
6006 * config/arm/cortex-a8-neon.md: Update with new attributes.
6007 * config/arm/cortex-a9.md: Update with new attributes.
6008 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6009 * config/arm/cortex-r4f.md: Update with new attributes.
6010 * config/arm/marvell-pj4.md: Update with new attributes.
6011 * config/arm/vfp11.md: Update with new attributes.
6013 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6015 * config/arm/types.md
6016 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6017 * config/aarch64/aarch64.md
6018 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6020 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6021 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6022 (float<GPI:mode><GPF:mode>2): Likewise.
6024 (*truncsisf2_vfp): Update with new attributes.
6025 (*truncsidf2_vfp): Likewise.
6026 (fixuns_truncsfsi2): Likewise.
6027 (fixuns_truncdfsi2): Likewise.
6028 (*floatsisf2_vfp): Likewise.
6029 (*floatsidf2_vfp): Likewise.
6030 (floatunssisf2): Likewise.
6031 (floatunssidf2): Likewise.
6032 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6033 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6034 * config/arm/arm1020e.md: Update with new attributes.
6035 * config/arm/cortex-a15-neon.md: Update with new attributes.
6036 * config/arm/cortex-a5.md: Update with new attributes.
6037 * config/arm/cortex-a53.md: Update with new attributes.
6038 * config/arm/cortex-a7.md: Update with new attributes.
6039 * config/arm/cortex-a8-neon.md: Update with new attributes.
6040 * config/arm/cortex-a9.md: Update with new attributes.
6041 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6042 * config/arm/cortex-r4f.md: Update with new attributes.
6043 * config/arm/marvell-pj4.md: Update with new attributes.
6044 * config/arm/vfp11.md: Update with new attributes.
6046 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6048 * config/aarch64/arm_neon.h
6049 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6050 (vqtbx<1,2,3,4><q>_s8): Likewise.
6052 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6054 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6055 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6056 (add<mode>3): Add type attribute.
6057 (add<mode>3): Likewise.
6058 (usadd<mode>3): Likewise.
6059 (ssadd<mode>3): Likewise.
6060 (sub<mode>3): Likewise.
6061 (sub<mode>3): Likewise.
6062 (ussub<mode>3): Likewise.
6063 (sssub<mode>3): Likewise.
6064 (ssmulsa3): Likewise.
6065 (usmulusa3): Likewise.
6066 (arm_usatsihi): Likewise.
6068 (*movdi_vfp): Add types for all instructions.
6069 (*movdi_vfp_cortexa8): Likewise.
6070 (*movhf_vfp_neon): Likewise.
6071 (*movhf_vfp): Likewise.
6072 (*movdf_vfp): Likewise.
6073 (*thumb2_movdf_vfp): Likewise.
6074 (*thumb2_movdfcc_vfp): Likewise.
6075 * config/arm/arm.md: Add type attribute to all insn patterns.
6076 (*thumb1_adddi3): Add type attribute.
6077 (*arm_adddi3): Likewise.
6078 (*adddi_sesidi_di): Likewise.
6079 (*adddi_zesidi_di): Likewise.
6080 (*thumb1_addsi3): Likewise.
6081 (addsi3_compare0): Likewise.
6082 (*addsi3_compare0_scratch): Likewise.
6083 (*compare_negsi_si): Likewise.
6084 (cmpsi2_addneg): Likewise.
6085 (*addsi3_carryin_<optab>): Likewise.
6086 (*addsi3_carryin_alt2_<optab>): Likewise.
6087 (*addsi3_carryin_clobercc_<optab>): Likewise.
6088 (*subsi3_carryin): Likewise.
6089 (*subsi3_carryin_const): Likewise.
6090 (*subsi3_carryin_compare): Likewise.
6091 (*subsi3_carryin_compare_const): Likewise.
6092 (*arm_subdi3): Likewise.
6093 (*thumb_subdi3): Likewise.
6094 (*subdi_di_zesidi): Likewise.
6095 (*subdi_di_sesidi): Likewise.
6096 (*subdi_zesidi_di): Likewise.
6097 (*subdi_sesidi_di): Likewise.
6098 (*subdi_zesidi_ze): Likewise.
6099 (thumb1_subsi3_insn): Likewise.
6100 (*arm_subsi3_insn): Likewise.
6101 (*anddi3_insn): Likewise.
6102 (*anddi_zesidi_di): Likewise.
6103 (*anddi_sesdi_di): Likewise.
6104 (*ne_zeroextracts): Likewise.
6105 (*ne_zeroextracts): Likewise.
6106 (*ite_ne_zeroextr): Likewise.
6107 (*ite_ne_zeroextr): Likewise.
6108 (*anddi_notdi_di): Likewise.
6109 (*anddi_notzesidi): Likewise.
6110 (*anddi_notsesidi): Likewise.
6111 (andsi_notsi_si): Likewise.
6112 (thumb1_bicsi3): Likewise.
6113 (*iordi3_insn): Likewise.
6114 (*iordi_zesidi_di): Likewise.
6115 (*iordi_sesidi_di): Likewise.
6116 (*thumb1_iorsi3_insn): Likewise.
6117 (*xordi3_insn): Likewise.
6118 (*xordi_zesidi_di): Likewise.
6119 (*xordi_sesidi_di): Likewise.
6120 (*arm_xorsi3): Likewise.
6121 (*andsi_iorsi3_no): Likewise.
6122 (*smax_0): Likewise.
6123 (*smax_m1): Likewise.
6124 (*arm_smax_insn): Likewise.
6125 (*smin_0): Likewise.
6126 (*arm_smin_insn): Likewise.
6127 (*arm_umaxsi3): Likewise.
6128 (*arm_uminsi3): Likewise.
6129 (*minmax_arithsi): Likewise.
6130 (*minmax_arithsi_): Likewise.
6131 (*satsi_<SAT:code>): Likewise.
6132 (arm_ashldi3_1bit): Likewise.
6133 (arm_ashrdi3_1bit): Likewise.
6134 (arm_lshrdi3_1bit): Likewise.
6135 (*arm_negdi2): Likewise.
6136 (*thumb1_negdi2): Likewise.
6137 (*arm_negsi2): Likewise.
6138 (*thumb1_negsi2): Likewise.
6139 (*negdi_extendsid): Likewise.
6140 (*negdi_zero_extend): Likewise.
6141 (*arm_abssi2): Likewise.
6142 (*thumb1_abssi2): Likewise.
6143 (*arm_neg_abssi2): Likewise.
6144 (*thumb1_neg_abss): Likewise.
6145 (one_cmpldi2): Likewise.
6146 (extend<mode>di2): Likewise.
6147 (*compareqi_eq0): Likewise.
6148 (*arm_extendhisi2addsi): Likewise.
6149 (*arm_movdi): Likewise.
6150 (*thumb1_movdi_insn): Likewise.
6151 (*arm_movt): Likewise.
6152 (*thumb1_movsi_insn): Likewise.
6153 (pic_add_dot_plus_four): Likewise.
6154 (pic_add_dot_plus_eight): Likewise.
6155 (tls_load_dot_plus_eight): Likewise.
6156 (*thumb1_movhi_insn): Likewise.
6157 (*thumb1_movsf_insn): Likewise.
6158 (*movdf_soft_insn): Likewise.
6159 (*thumb_movdf_insn): Likewise.
6160 (cbranchsi4_insn): Likewise.
6161 (cbranchsi4_scratch): Likewise.
6162 (*negated_cbranchsi4): Likewise.
6163 (*tbit_cbranch): Likewise.
6164 (*tlobits_cbranch): Likewise.
6165 (*tstsi3_cbranch): Likewise.
6166 (*cbranchne_decr1): Likewise.
6167 (*addsi3_cbranch): Likewise.
6168 (*addsi3_cbranch_scratch): Likewise.
6169 (*arm_cmpdi_insn): Likewise.
6170 (*arm_cmpdi_unsig): Likewise.
6171 (*arm_cmpdi_zero): Likewise.
6172 (*thumb_cmpdi_zero): Likewise.
6173 (*deleted_compare): Likewise.
6174 (*mov_scc): Likewise.
6175 (*mov_negscc): Likewise.
6176 (*mov_notscc): Likewise.
6177 (*cstoresi_eq0_thumb1_insn): Likewise.
6178 (cstoresi_nltu_thumb1): Likewise.
6179 (cstoresi_ltu_thu): Likewise.
6180 (thumb1_addsi3_addgeu): Likewise.
6181 (*arm_jump): Likewise.
6182 (*thumb_jump): Likewise.
6183 (*check_arch2): Likewise.
6184 (arm_casesi_internal): Likewise.
6185 (thumb1_casesi_dispatch): Likewise.
6186 (*arm_indirect_jump): Likewise.
6187 (*thumb1_indirect_jump): Likewise.
6189 (*and_scc): Likewise.
6190 (*ior_scc): Likewise.
6191 (*compare_scc): Likewise.
6192 (*cond_move): Likewise.
6193 (*cond_arith): Likewise.
6194 (*cond_sub): Likewise.
6195 (*cmp_ite0): Likewise.
6196 (*cmp_ite1): Likewise.
6197 (*cmp_and): Likewise.
6198 (*cmp_ior): Likewise.
6199 (*ior_scc_scc): Likewise.
6200 (*ior_scc_scc_cmp): Likewise.
6201 (*and_scc_scc): Likewise.
6202 (*and_scc_scc_cmp): Likewise.
6203 (*and_scc_scc_nod): Likewise.
6204 (*negscc): Likewise.
6205 (movcond_addsi): Likewise.
6206 (movcond): Likewise.
6207 (*ifcompare_plus_move): Likewise.
6208 (*if_plus_move): Likewise.
6209 (*ifcompare_move_plus): Likewise.
6210 (*if_move_plus): Likewise.
6211 (*ifcompare_arith_arith): Likewise.
6212 (*if_arith_arith): Likewise.
6213 (*ifcompare_arith_move): Likewise.
6214 (*if_arith_move): Likewise.
6215 (*ifcompare_move_arith): Likewise.
6216 (*if_move_arith): Likewise.
6217 (*ifcompare_move_not): Likewise.
6218 (*if_move_not): Likewise.
6219 (*ifcompare_not_move): Likewise.
6220 (*if_not_move): Likewise.
6221 (*ifcompare_shift_move): Likewise.
6222 (*if_shift_move): Likewise.
6223 (*ifcompare_move_shift): Likewise.
6224 (*if_move_shift): Likewise.
6225 (*ifcompare_shift_shift): Likewise.
6226 (*ifcompare_not_arith): Likewise.
6227 (*ifcompare_arith_not): Likewise.
6228 (*if_arith_not): Likewise.
6229 (*ifcompare_neg_move): Likewise.
6230 (*if_neg_move): Likewise.
6231 (*ifcompare_move_neg): Likewise.
6232 (*if_move_neg): Likewise.
6233 (prologue_thumb1_interwork): Likewise.
6234 (*cond_move_not): Likewise.
6235 (*sign_extract_onebit): Likewise.
6236 (*not_signextract_onebit): Likewise.
6237 (stack_tie): Likewise.
6238 (align_4): Likewise.
6239 (align_8): Likewise.
6240 (consttable_end): Likewise.
6241 (consttable_1): Likewise.
6242 (consttable_2): Likewise.
6243 (consttable_4): Likewise.
6244 (consttable_8): Likewise.
6245 (consttable_16): Likewise.
6246 (*thumb1_tablejump): Likewise.
6247 (prefetch): Likewise.
6248 (force_register_use): Likewise.
6249 (thumb_eh_return): Likewise.
6250 (load_tp_hard): Likewise.
6251 (load_tp_soft): Likewise.
6252 (tlscall): Likewise.
6253 (*arm_movtas_ze): Likewise.
6254 (*arm_rev): Likewise.
6255 (*arm_revsh): Likewise.
6256 (*arm_rev16): Likewise.
6257 * config/arm/thumb2.md
6258 (*thumb2_smaxsi3): Likewise.
6259 (*thumb2_sminsi3): Likewise.
6260 (*thumb32_umaxsi3): Likewise.
6261 (*thumb2_uminsi3): Likewise.
6262 (*thumb2_negdi2): Likewise.
6263 (*thumb2_abssi2): Likewise.
6264 (*thumb2_neg_abss): Likewise.
6265 (*thumb2_movsi_insn): Likewise.
6266 (tls_load_dot_plus_four): Likewise.
6267 (*thumb2_movhi_insn): Likewise.
6268 (*thumb2_mov_scc): Likewise.
6269 (*thumb2_mov_negs): Likewise.
6270 (*thumb2_mov_negs): Likewise.
6271 (*thumb2_mov_nots): Likewise.
6272 (*thumb2_mov_nots): Likewise.
6273 (*thumb2_movsicc_): Likewise.
6274 (*thumb2_movsfcc_soft_insn): Likewise.
6275 (*thumb2_indirect_jump): Likewise.
6276 (*thumb2_and_scc): Likewise.
6277 (*thumb2_ior_scc): Likewise.
6278 (*thumb2_ior_scc_strict_it): Likewise.
6279 (*thumb2_cond_move): Likewise.
6280 (*thumb2_cond_arith): Likewise.
6281 (*thumb2_cond_ari): Likewise.
6282 (*thumb2_cond_sub): Likewise.
6283 (*thumb2_negscc): Likewise.
6284 (*thumb2_movcond): Likewise.
6285 (thumb2_casesi_internal): Likewise.
6286 (thumb2_casesi_internal_pic): Likewise.
6287 (*thumb2_alusi3_short): Likewise.
6288 (*thumb2_mov<mode>_shortim): Likewise.
6289 (*thumb2_addsi_short): Likewise.
6290 (*thumb2_subsi_short): Likewise.
6291 (thumb2_addsi3_compare0): Likewise.
6292 (*thumb2_cbz): Likewise.
6293 (*thumb2_cbnz): Likewise.
6294 (*thumb2_one_cmplsi2_short): Likewise.
6295 (*thumb2_negsi2_short): Likewise.
6296 (*orsi_notsi_si): Likewise.
6297 * config/arm/arm1020e.md: Update with new attributes.
6298 * config/arm/arm1026ejs.md: Update with new attributes.
6299 * config/arm/arm1136jfs.md: Update with new attributes.
6300 * config/arm/arm926ejs.md: Update with new attributes.
6301 * config/arm/cortex-a15.md: Update with new attributes.
6302 * config/arm/cortex-a5.md: Update with new attributes.
6303 * config/arm/cortex-a53.md: Update with new attributes.
6304 * config/arm/cortex-a7.md: Update with new attributes.
6305 * config/arm/cortex-a8.md: Update with new attributes.
6306 * config/arm/cortex-a9.md: Update with new attributes.
6307 * config/arm/cortex-m4.md: Update with new attributes.
6308 * config/arm/cortex-r4.md: Update with new attributes.
6309 * config/arm/fa526.md: Update with new attributes.
6310 * config/arm/fa606te.md: Update with new attributes.
6311 * config/arm/fa626te.md: Update with new attributes.
6312 * config/arm/fa726te.md: Update with new attributes.
6314 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6316 * config/aarch64/aarch64-simd.md
6317 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6318 <vwx> iterator to ensure correct register choice.
6319 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6320 (aarch64_sqdmull_n<mode>): Likewise.
6321 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6322 * config/aarch64/arm_neon.h
6323 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6324 (vml<as><q>_n_<su>16): Likewise.
6325 (vml<as>l_high_lane<q>_<su>16): Likewise.
6326 (vml<as>l_high_n_<su>16): Likewise.
6327 (vml<as>l_lane<q>_<su>16): Likewise.
6328 (vml<as>l_n_<su>16): Likewise.
6329 (vmul<q>_lane<q>_<su>16): Likewise.
6330 (vmul<q>_n_<su>16): Likewise.
6331 (vmull_lane<q>_<su>16): Likewise.
6332 (vmull_n_<su>16): Likewise.
6333 (vmull_high_lane<q>_<su>16): Likewise.
6334 (vmull_high_n_<su>16): Likewise.
6335 (vqrdmulh<q>_n_s16): Likewise.
6337 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6339 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6340 have the correct lane parameter.
6342 2013-09-06 Richard Biener <rguenther@suse.de>
6344 * cfganal.c (control_dependences::~control_dependences):
6345 Properly free all of the vector.
6347 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6350 * config/i386/i386.c (ix86_conditional_register_usage):
6351 Proper initialize extended SSE registers.
6353 2013-09-06 Jan Hubicka <jh@suse.cz>
6355 PR tree-optimization/58311
6356 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6358 2013-09-06 Jan Hubicka <jh@suse.cz>
6360 * Makefile.in (tree-sra.o): Update dependencies.
6361 * tree-sra.c: Include ipa-utils.h
6362 (scan_function): Use recursive_call_p.
6363 (has_caller_p): New function.
6364 (cgraph_for_node_and_aliases): Count also callers of aliases.
6366 2013-09-06 Jan Hubicka <jh@suse.cz>
6369 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6370 * tree-tailcall.c: Include ipa-utils.h.
6371 (find_tail_calls): Use it.
6372 * ipa-pure-const.c (check_call): Likewise.
6373 * ipa-utils.c (recursive_call_p): New function.
6374 * ipa-utils.h (recursive_call_p): Dclare.
6375 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6376 (symtab_semantically_equivalent_p): New function.
6377 * Makefile.in (tree-tailcall.o): Update dependencies.
6379 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6381 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6384 2013-09-06 Richard Biener <rguenther@suse.de>
6386 * lto-streamer.h (lto_global_var_decls): Remove.
6387 * Makefile.in (OBJS): Remove lto-symtab.o.
6388 (lto-symtab.o): Remove.
6389 (GTFILES): Remove lto-symtab.c
6390 * lto-symtab.c: Move to lto/
6392 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6394 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6395 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6396 (UNSPEC_FPINT_RINT): New constant definitions.
6397 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6398 definition with 2 attributes.
6399 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6400 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6403 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6405 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6406 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6407 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6408 attribute to "z196_alone".
6409 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6412 2013-09-06 Richard Biener <rguenther@suse.de>
6414 * basic-block.h (class control_dependences): New.
6415 * tree-ssa-dce.c (control_dependence_map): Remove.
6417 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6418 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6419 find_pdom, find_control_dependence, find_all_control_dependences):
6421 (mark_control_dependent_edges_necessary,
6422 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6423 tree_dce_done, perform_tree_ssa_dce): Adjust.
6424 * cfganal.c (set_control_dependence_map_bit,
6425 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6426 find_all_control_dependences): Move from tree-ssa-dce.c and
6427 implement as methods of control_dependences class.
6428 (control_dependences::control_dependences): New.
6429 (control_dependences::~control_dependences): Likewise.
6430 (control_dependences::get_edges_dependent_on): Likewise.
6431 (control_dependences::get_edge): Likewise.
6433 2013-09-04 Jan Hubicka <jh@suse.cz>
6435 * tree.c (types_same_for_odr): Drop overactive check.
6436 * ipa-devirt.c (hash_type_name): Likewise.
6438 2013-09-04 Jan Hubicka <jh@suse.cz>
6440 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6441 (analyze_functions): ... here.
6443 2013-09-04 Jan Hubicka <jh@suse.cz>
6446 * cgraphunit.c (analyze_functions): Clear AUX fields
6447 after processing; initialize assembler name has.
6449 2013-09-05 Jeff Law <law@redhat.com>
6451 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6452 from thread_around_empty_block. Record threading path into PATH.
6453 Recurse if threading through the initial block is successful.
6454 (thread_across_edge): Corresponding changes to slightly simplify.
6456 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6458 * config/aarch64/aarch64.md
6459 (type): Remove frecpe, frecps, frecpx.
6460 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6461 fix to be a TARGET_SIMD instruction.
6462 (aarch64_frecps): Remove.
6463 * config/aarch64/aarch64-simd.md
6464 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6465 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6467 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6468 Sofiane Naci <sofiane.naci@arm.com>
6470 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6471 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6472 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6473 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6474 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6475 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6476 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6477 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6478 into "clz, "rbit". Rename "shift" to "shift_imm".
6479 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6480 changes. Update for attribute changes all occurrences of arlo_* and
6482 * config/arm/arm-fixed.md: Update for attribute changes
6483 all occurrences of arlo_* types.
6484 * config/arm/thumb2.md: Update for attribute changes all occurrences
6486 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6487 (cortexa7_older_only): Likewise.
6488 (cortexa7_younger): Likewise.
6489 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6490 (1020alu_shift_op): Likewise.
6491 (1020alu_shift_reg_op): Likewise.
6492 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6493 (alu_shift_op): Likewise.
6494 (alu_shift_reg_op): Likewise.
6495 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6496 (11_alu_shift_op): Likewise.
6497 (11_alu_shift_reg_op): Likewise.
6498 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6499 (9_alu_shift_reg_op): Likewise.
6500 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6502 (cortex_a15_alu_shift): Likewise.
6503 (cortex_a15_alu_shift_reg): Likewise.
6504 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6506 (cortex_a5_alu_shift): Likewise.
6507 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6509 (cortex_a53_alu_shift): Likewise.
6510 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6512 (cortex_a7_alu_reg): Likewise.
6513 (cortex_a7_alu_shift): Likewise.
6514 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6516 (cortex_a8_alu_shift): Likewise.
6517 (cortex_a8_alu_shift_reg): Likewise.
6518 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6519 (cortex_a9_dp_shift): Likewise.
6520 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6522 * config/arm/cortex-r4.md
6523 (cortex_r4_alu): Update for attribute changes.
6524 (cortex_r4_mov): Likewise.
6525 (cortex_r4_alu_shift_reg): Likewise.
6526 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6527 (526_alu_shift_op): Likewise.
6528 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6529 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6530 (626te_alu_shift_op): Likewise.
6531 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6532 (726te_alu_shift_op): Likewise.
6533 (726te_alu_shift_reg_op): Likewise.
6534 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6535 (mp626_alu_shift_op): Likewise.
6536 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6537 (pj4_alu_conds): Likewise.
6538 (pj4_shift): Likewise.
6539 (pj4_shift_conds): Likewise.
6540 (pj4_alu_shift): Likewise.
6541 (pj4_alu_shift_conds): Likewise.
6542 * config/aarch64/aarch64.md: Update for attribute change
6543 all occurrences of arlo_* and shift* types.
6545 2013-09-05 Mike Stump <mikestump@comcast.net>
6547 * tree.h: Move documentation for tree_function_decl to tree-core.h
6548 with the declaration.
6550 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6553 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6554 looking for widest mode.
6556 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6558 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6560 2013-09-05 Richard Biener <rguenther@suse.de>
6562 PR tree-optimization/58137
6563 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6564 Do not create vectors of pointers.
6565 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6566 types for the components of the vector initializer.
6567 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6568 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6570 2013-09-05 Martin Jambor <mjambor@suse.cz>
6572 * ipa-prop.c (remove_described_reference): Accept missing references,
6573 return false if that hppens, otherwise return true.
6574 (cgraph_node_for_jfunc): New function.
6575 (try_decrement_rdesc_refcount): Likewise.
6576 (try_make_edge_direct_simple_call): Use them.
6577 (ipa_edge_removal_hook): Remove references from rdescs.
6578 (ipa_edge_duplication_hook): Clone rdescs and their references
6579 when the new edge has the same caller as the old one.
6580 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6581 reference before removing any edges.
6583 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6585 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6587 * thumb2.md (thumb2_storewb_parisi): New pattern.
6589 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6591 * config/aarch64/aarch64-option-extensions.def: Add
6592 AARCH64_OPT_EXTENSION of 'crc'.
6593 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6594 (AARCH64_ISA_CRC): Ditto.
6595 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6596 description of the CRC extension.
6598 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6600 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6601 * config/rs6000/linux.h: Ditto.
6602 * alpha/linux.h: Ditto.
6603 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6604 no_c99_libc_has_function.
6605 * config/c6x/uclinux-elf.h: Ditto.
6606 * config/lm32/uclinux-elf.h: Ditto.
6607 * config/m68k/uclinux.h: Ditto.
6608 * config/moxie/uclinux.h: Ditto.
6609 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6610 (crisv32-*-linux*, cris-*-linux*): Ditto.
6611 * config/bfin/bfin.c: Include "tm_p.h".
6613 2013-09-05 Richard Biener <rguenther@suse.de>
6615 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6616 check for a definition without a basic-block.
6618 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6619 Sofiane Naci <sofiane.naci@arm.com>
6621 * config/aarch64/aarch64.md
6622 (*movti_aarch64): Rename r_2_f and f_2_r.
6623 (*movsf_aarch64): Likewise.
6624 (*movdf_aarch64): Likewise.
6625 (*movtf_aarch64): Likewise.
6626 (aarch64_movdi_<mode>low): Likewise.
6627 (aarch64_movdi_<mode>high): Likewise.
6628 (aarch64_mov<mode>high_di): Likewise.
6629 (aarch64_mov<mode>low_di): Likewise.
6630 (aarch64_movtilow_tilow): Likewise.
6631 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6632 values to config/arm/types.md
6633 (attribute "conds"): Update for attribute change.
6634 (anddi3_insn): Likewise.
6635 (iordi3_insn): Likewise.
6636 (xordi3_insn): Likewise.
6637 (one_cmpldi2): Likewise.
6638 * config/arm/types.md (type): Add Neon types.
6639 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6640 use "type" attribute.
6641 (movmisalign<mode>_neon_store): Likewise.
6642 (movmisalign<mode>_neon_load): Likewise.
6643 (vec_set<mode>_internal): Likewise.
6644 (vec_setv2di_internal): Likewise.
6645 (vec_extract<mode>): Likewise.
6646 (vec_extractv2di): Likewise.
6647 (add<mode>3_neon): Likewise.
6648 (adddi3_neon): Likewise.
6649 (sub<mode>3_neon): Likewise.
6650 (subdi3_neon): Likewise.
6651 (mul<mode>3_neon): Likewise.
6652 (mul<mode>3add<mode>_neon): Likewise.
6653 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6654 (fma<VCVTF:mode>4)): Likewise.
6655 (fma<VCVTF:mode>4_intrinsic): Likewise.
6656 (fmsub<VCVTF:mode>4)): Likewise.
6657 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6658 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6659 (ior<mode>3): Likewise.
6660 (and<mode>3): Likewise.
6661 (anddi3_neon): Likewise.
6662 (orn<mode>3_neon): Likewise.
6663 (orndi3_neon): Likewise.
6664 (bic<mode>3_neon): Likewise.
6665 (bicdi3_neon): Likewise.
6666 (xor<mode>3): Likewise.
6667 (one_cmpl<mode>2): Likewise.
6668 (abs<mode>2): Likewise.
6669 (neg<mode>2): Likewise.
6670 (umin<mode>3_neon): Likewise.
6671 (umax<mode>3_neon): Likewise.
6672 (smin<mode>3_neon): Likewise.
6673 (smax<mode>3_neon): Likewise.
6674 (vashl<mode>3): Likewise.
6675 (vashr<mode>3_imm): Likewise.
6676 (vlshr<mode>3_imm): Likewise.
6677 (ashl<mode>3_signed): Likewise.
6678 (ashl<mode>3_unsigned): Likewise.
6679 (neon_load_count): Likewise.
6680 (ashldi3_neon_noclobber): Likewise.
6681 (signed_shift_di3_neon): Likewise.
6682 (unsigned_shift_di3_neon): Likewise.
6683 (ashrdi3_neon_imm_noclobber): Likewise.
6684 (lshrdi3_neon_imm_noclobber): Likewise.
6685 (widen_ssum<mode>3): Likewise.
6686 (widen_usum<mode>3): Likewise.
6687 (quad_halves_<code>v4si): Likewise.
6688 (quad_halves_<code>v4sf): Likewise.
6689 (quad_halves_<code>v8hi): Likewise.
6690 (quad_halves_<code>v16qi): Likewise.
6691 (reduc_splus_v2di): Likewise.
6692 (neon_vpadd_internal<mode>): Likewise.
6693 (neon_vpsmin<mode>): Likewise.
6694 (neon_vpsmax<mode>): Likewise.
6695 (neon_vpumin<mode>): Likewise.
6696 (neon_vpumax<mode>): Likewise.
6697 (ss_add<mode>_neon): Likewise.
6698 (us_add<mode>_neon): Likewise.
6699 (ss_sub<mode>_neon): Likewise.
6700 (us_sub<mode>_neon): Likewise.
6701 (neon_vadd<mode>_unspec): Likewise.
6702 (neon_vaddl<mode>): Likewise.
6703 (neon_vaddw<mode>): Likewise.
6704 (neon_vhadd<mode>): Likewise.
6705 (neon_vqadd<mode>): Likewise.
6706 (neon_vaddhn<mode>): Likewise.
6707 (neon_vmul<mode>): Likewise.
6708 (neon_vmla<mode>): Likewise.
6709 (neon_vmlal<mode>): Likewise.
6710 (neon_vmls<mode>): Likewise.
6711 (neon_vmlsl<mode>): Likewise.
6712 (neon_vqdmulh<mode>): Likewise.
6713 (neon_vqdmlal<mode>): Likewise.
6714 (neon_vqdmlsl<mode>): Likewise.
6715 (neon_vmull<mode>): Likewise.
6716 (neon_vqdmull<mode>): Likewise.
6717 (neon_vsub<mode>_unspec): Likewise.
6718 (neon_vsubl<mode>): Likewise.
6719 (neon_vsubw<mode>): Likewise.
6720 (neon_vqsub<mode>): Likewise.
6721 (neon_vhsub<mode>): Likewise.
6722 (neon_vsubhn<mode>): Likewise.
6723 (neon_vceq<mode>): Likewise.
6724 (neon_vcge<mode>): Likewise.
6725 (neon_vcgeu<mode>): Likewise.
6726 (neon_vcgt<mode>): Likewise.
6727 (neon_vcgtu<mode>): Likewise.
6728 (neon_vcle<mode>): Likewise.
6729 (neon_vclt<mode>): Likewise.
6730 (neon_vcage<mode>): Likewise.
6731 (neon_vcagt<mode>): Likewise.
6732 (neon_vtst<mode>): Likewise.
6733 (neon_vabd<mode>): Likewise.
6734 (neon_vabdl<mode>): Likewise.
6735 (neon_vaba<mode>): Likewise.
6736 (neon_vabal<mode>): Likewise.
6737 (neon_vmax<mode>): Likewise.
6738 (neon_vmin<mode>): Likewise.
6739 (neon_vpaddl<mode>): Likewise.
6740 (neon_vpadal<mode>): Likewise.
6741 (neon_vpmax<mode>): Likewise.
6742 (neon_vpmin<mode>): Likewise.
6743 (neon_vrecps<mode>): Likewise.
6744 (neon_vrsqrts<mode>): Likewise.
6745 (neon_vqabs<mode>): Likewise.
6746 (neon_vqneg<mode>): Likewise.
6747 (neon_vcls<mode>): Likewise.
6748 (clz<mode>2): Likewise.
6749 (popcount<mode>2): Likewise.
6750 (neon_vrecpe): Likewise.
6751 (neon_vrsqrte): Likewise.
6752 (neon_vget_lane<mode>_sext_internal): Likewise.
6753 (neon_vget_lane<mode>_zext_internal): Likewise.
6754 (neon_vdup_n<mode>): Likewise.
6755 (neon_vdup_nv2di): Likewise.
6756 (neon_vdpu_lane<mode>_internal): Likewise.
6757 (neon_vswp<mode>): Likewise.
6758 (float<mode><V_cvtto>2): Likewise.
6759 (floatuns<mode><V_cvtto>2): Likewise.
6760 (fix_trunc<mode><V_cvtto>)2): Likewise
6761 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6762 (neon_vcvt<mode>): Likewise.
6763 (neon_vcvtv4sfv4hf): Likewise.
6764 (neon_vcvtv4hfv4sf): Likewise.
6765 (neon_vcvt_n<mode>): Likewise.
6766 (neon_vmovn<mode>): Likewise.
6767 (neon_vqmovn<mode>): Likewise.
6768 (neon_vqmovun<mode>): Likewise.
6769 (neon_vmovl<mode>): Likewise.
6770 (neon_vmul_lane<mode>): Likewise.
6771 (neon_vmull_lane<mode>): Likewise.
6772 (neon_vqdmull_lane<mode>): Likewise.
6773 (neon_vqdmulh_lane<mode>): Likewise.
6774 (neon_vmla_lane<mode>): Likewise.
6775 (neon_vmlal_lane<mode>): Likewise.
6776 (neon_vqdmlal_lane<mode>): Likewise.
6777 (neon_vmls_lane<mode>): Likewise.
6778 (neon_vmlsl_lane<mode>): Likewise.
6779 (neon_vqdmlsl_lane<mode>): Likewise.
6780 (neon_vext<mode>): Likewise.
6781 (neon_vrev64<mode>): Likewise.
6782 (neon_vrev32<mode>): Likewise.
6783 (neon_vrev16<mode>): Likewise.
6784 (neon_vbsl<mode>_internal): Likewise.
6785 (neon_vshl<mode>): Likewise.
6786 (neon_vqshl<mode>): Likewise.
6787 (neon_vshr_n<mode>): Likewise.
6788 (neon_vshrn_n<mode>): Likewise.
6789 (neon_vqshrn_n<mode>): Likewise.
6790 (neon_vqshrun_n<mode>): Likewise.
6791 (neon_vshl_n<mode>): Likewise.
6792 (neon_vqshl_n<mode>): Likewise.
6793 (neon_vqshlu_n<mode>): Likewise.
6794 (neon_vshll_n<mode>): Likewise.
6795 (neon_vsra_n<mode>): Likewise.
6796 (neon_vsri_n<mode>): Likewise.
6797 (neon_vsli_n<mode>): Likewise.
6798 (neon_vtbl1v8qi): Likewise.
6799 (neon_vtbl2v8qi): Likewise.
6800 (neon_vtbl3v8qi): Likewise.
6801 (neon_vtbl4v8qi): Likewise.
6802 (neon_vtbx1v8qi): Likewise.
6803 (neon_vtbx2v8qi): Likewise.
6804 (neon_vtbx3v8qi): Likewise.
6805 (neon_vtbx4v8qi): Likewise.
6806 (neon_vtrn<mode>_internal): Likewise.
6807 (neon_vzip<mode>_internal): Likewise.
6808 (neon_vuzp<mode>_internal): Likewise.
6809 (neon_vld1<mode>): Likewise.
6810 (neon_vld1_lane<mode>): Likewise.
6811 (neon_vld1_dup<mode>): Likewise.
6812 (neon_vld1_dupv2di): Likewise.
6813 (neon_vst1<mode>): Likewise.
6814 (neon_vst1_lane<mode>): Likewise.
6815 (neon_vld2<mode>): Likewise.
6816 (neon_vld2_lane<mode>): Likewise.
6817 (neon_vld2_dup<mode>): Likewise.
6818 (neon_vst2<mode>): Likewise.
6819 (neon_vst2_lane<mode>): Likewise.
6820 (neon_vld3<mode>): Likewise.
6821 (neon_vld3qa<mode>): Likewise.
6822 (neon_vld3qb<mode>): Likewise.
6823 (neon_vld3_lane<mode>): Likewise.
6824 (neon_vld3_dup<mode>): Likewise.
6825 (neon_vst3<mode>): Likewise.
6826 (neon_vst3qa<mode>): Likewise.
6827 (neon_vst3qb<mode>): Likewise.
6828 (neon_vst3_lane<mode>): Likewise.
6829 (neon_vld4<mode>): Likewise.
6830 (neon_vld4qa<mode>): Likewise.
6831 (neon_vld4qb<mode>): Likewise.
6832 (neon_vld4_lane<mode>): Likewise.
6833 (neon_vld4_dup<mode>): Likewise.
6834 (neon_vst4<mode>): Likewise.
6835 (neon_vst4qa<mode>): Likewise.
6836 (neon_vst4qb<mode>): Likewise.
6837 (neon_vst4_lane<mode>): Likewise.
6838 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6839 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6840 (neon_vec_<US>mult_lo_<mode>): Likewise.
6841 (neon_vec_<US>mult_hi_<mode>): Likewise.
6842 (neon_vec_<US>shiftl_<mode>): Likewise.
6843 (neon_unpack<US>_<mode>): Likewise.
6844 (neon_vec_<US>mult_<mode>): Likewise.
6845 (vec_pack_trunc_<mode>): Likewise.
6846 (neon_vec_pack_trunk_<mode>): Likewise.
6847 (neon_vabd<mode>_2): Likewise.
6848 (neon_vabd<mode>_3): Likewise.
6849 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6850 (thumb2_movsi_vfp): Likewise.
6851 (movdi_vfp): Likewise.
6852 (movdi_vfp_cortexa8): Likewise.
6853 (movhf_vfp_neon): Likewise.
6854 (movhf_vfp): Likewiwse.
6855 (movsf_vfp): Likewiwse.
6856 (thumb2_movsf_vfp): Likewiwse.
6857 (movdf_vfp): Likewise.
6858 (thumb2_movdf_vfp): Likewise.
6859 (movsfcc_vfp): Likewise.
6860 (thumb2_movsfcc_vfp): Likewise.
6861 (movdfcc_vfp): Likewise.
6862 (thumb2_movdfcc_vfp): Likewise.
6863 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6864 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6865 (v10_v2c): Likewise.
6866 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6868 (cortex_a15_neon_int_2): Likewise.
6869 (cortex_a15_neon_int_3): Likewise.
6870 (cortex_a15_neon_int_4): Likewise.
6871 (cortex_a15_neon_int_5): Likewise.
6872 (cortex_a15_neon_vqneg_vqabs): Likewise.
6873 (cortex_a15_neon_vmov): Likewise.
6874 (cortex_a15_neon_vaba): Likewise.
6875 (cortex_a15_neon_vaba_qqq): Likewise.
6876 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6877 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6878 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6879 scalar_64_32_long_scalar): Likewise.
6880 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6881 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6882 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6883 lotype_qdd_64_32_long): Likewise.
6884 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6885 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6886 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6887 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6888 (cortex_a15_neon_shift_1): Likewise.
6889 (cortex_a15_neon_shift_2): Likewise.
6890 (cortex_a15_neon_shift_3): Likewise.
6891 (cortex_a15_neon_vshl_ddd): Likewise.
6892 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6893 (cortex_a15_neon_vsra_vrsra): Likewise.
6894 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6895 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6896 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6897 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6898 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6899 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6900 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6901 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6902 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6903 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6904 (cortex_a15_neon_bp_simple): Likewise.
6905 (cortex_a15_neon_bp_2cycle): Likewise.
6906 (cortex_a15_neon_bp_3cycle): Likewise.
6907 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6908 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6909 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6910 (cortex_a15_neon_vld2_4_regs): Likewise.
6911 (cortex_a15_neon_vld3_vld4): Likewise.
6912 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6913 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6914 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6915 (cortex_a15_neon_vst3_vst4): Likewise.
6916 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6917 (cortex_a15_neon_vld3_vld4_lane" 10
6918 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6919 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6920 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6921 (cortex_a15_neon_ldm_2): Likewise.0
6922 (cortex_a15_neon_stm_2): Likewise.
6923 (cortex_a15_neon_mcr): Likewise.
6924 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6925 (cortex_a15_neon_mrc): Likewise.
6926 (cortex_a15_neon_mrrc): Likewise.
6927 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6929 (cortex_a15_alu_shift): Likewise.
6930 (cortex_a15_alu_shift_reg): Likewise.
6931 (cortex_a15_mult32): Likewise.
6932 (cortex_a15_mult64): Likewise.
6933 (cortex_a15_block): Likewise.
6934 (cortex_a15_branch): Likewise.
6935 (cortex_a15_load1): Likewise.
6936 (cortex_a15_load3): Likewise.
6937 (cortex_a15_store1): Likewise.
6938 (cortex_a15_store3): Likewise.
6939 (cortex_a15_call): Likewise.
6940 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6941 (cortex_a5_f2r): Likewise.
6942 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6944 (cortex_a53_f2r): Likewise.
6945 * config/arm/cortex-a7.md
6946 (cortex_a7_branch): Update for attribute change.
6947 (cortex_a7_call): Likewise.
6948 (cortex_a7_alu_imm): Likewise.
6949 (cortex_a7_alu_reg): Likewise.
6950 (cortex_a7_alu_shift): Likewise.
6951 (cortex_a7_mul): Likewise.
6952 (cortex_a7_load1): Likewise.
6953 (cortex_a7_store1): Likewise.
6954 (cortex_a7_load2): Likewise.
6955 (cortex_a7_store2): Likewise.
6956 (cortex_a7_load3): Likewise.
6957 (cortex_a7_store3): Likewise.
6958 (cortex_a7_load4): Likewise.
6959 (cortex_a7_store4): Likewise.
6960 (cortex_a7_fpalu): Likewise.
6961 (cortex_a7_fconst): Likewise.
6962 (cortex_a7_fpmuls): Likewise.
6963 (cortex_a7_neon_mul): Likewise.
6964 (cortex_a7_fpmacs): Likewise.
6965 (cortex_a7_neon_mla: Likewise.
6966 (cortex_a7_fpmuld: Likewise.
6967 (cortex_a7_fpmacd: Likewise.
6968 (cortex_a7_fpfmad: Likewise.
6969 (cortex_a7_fdivs: Likewise.
6970 (cortex_a7_fdivd: Likewise.
6971 (cortex_a7_r2f: Likewise.
6972 (cortex_a7_f2r: Likewise.
6973 (cortex_a7_f_flags: Likewise.
6974 (cortex_a7_f_loads: Likewise.
6975 (cortex_a7_f_loadd: Likewise.
6976 (cortex_a7_f_stores: Likewise.
6977 (cortex_a7_f_stored: Likewise.
6978 (cortex_a7_neon): Likewise.
6979 * config/arm/cortex-a8-neon.md
6980 (cortex_a8_neon_mrc): Update for attribute change.
6981 (cortex_a8_neon_mrrc): Likewise.
6982 (cortex_a8_neon_int_1): Likewise.
6983 (cortex_a8_neon_int_2): Likewise.
6984 (cortex_a8_neon_int_3): Likewise.
6985 (cortex_a8_neon_int_4): Likewise.
6986 (cortex_a8_neon_int_5): Likewise.
6987 (cortex_a8_neon_vqneg_vqabs): Likewise.
6988 (cortex_a8_neon_vmov): Likewise.
6989 (cortex_a8_neon_vaba): Likewise.
6990 (cortex_a8_neon_vaba_qqq): Likewise.
6991 (cortex_a8_neon_vsma): Likewise.
6992 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6993 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6994 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6995 long_scalar): Likewise.
6996 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6997 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6998 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6999 long_scalar_qdd_64_32_long): Likewise.
7000 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7001 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7002 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7003 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7004 (cortex_a8_neon_shift_1): Likewise.
7005 (cortex_a8_neon_shift_2): Likewise.
7006 (cortex_a8_neon_shift_3): Likewise.
7007 (cortex_a8_neon_vshl_ddd): Likewise.
7008 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7009 (cortex_a8_neon_vsra_vrsra): Likewise.
7010 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7011 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7012 (cortex_a8_neon_fp_vsum): Likewise.
7013 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7014 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7015 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7016 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7017 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7018 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7019 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7020 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7021 (cortex_a8_neon_bp_simple): Likewise.
7022 (cortex_a8_neon_bp_2cycle): Likewise.
7023 (cortex_a8_neon_bp_3cycle): Likewise.
7024 (cortex_a8_neon_ldr): Likewise.
7025 (cortex_a8_neon_str): Likewise.
7026 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7027 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7028 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7029 (cortex_a8_neon_vld2_4_regs): Likewise.
7030 (cortex_a8_neon_vld3_vld4): Likewise.
7031 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7032 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7033 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7034 (cortex_a8_neon_vst3_vst4): Likewise.
7035 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7036 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7037 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7038 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7039 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7040 (cortex_a8_neon_mcr): Likewise.
7041 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7042 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7043 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7045 (ca9_neon_mrrc): Likewise.
7046 (cortex_a9_neon_int_1): Likewise.
7047 (cortex_a9_neon_int_2): Likewise.
7048 (cortex_a9_neon_int_3): Likewise.
7049 (cortex_a9_neon_int_4): Likewise.
7050 (cortex_a9_neon_int_5): Likewise.
7051 (cortex_a9_neon_vqneg_vqabs): Likewise.
7052 (cortex_a9_neon_vmov): Likewise.
7053 (cortex_a9_neon_vaba): Likewise.
7054 (cortex_a9_neon_vaba_qqq): Likewise.
7055 (cortex_a9_neon_vsma): Likewise.
7056 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7057 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7058 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7059 long_scalar): Likewise.
7060 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7061 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7062 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7063 long_scalar_qdd_64_32_long): Likewise.
7064 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7065 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7066 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7067 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7068 (cortex_a9_neon_shift_1): Likewise.
7069 (cortex_a9_neon_shift_2): Likewise.
7070 (cortex_a9_neon_shift_3): Likewise.
7071 (cortex_a9_neon_vshl_ddd): Likewise.
7072 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7073 (cortex_a9_neon_vsra_vrsra): Likewise.
7074 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7075 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7076 (cortex_a9_neon_fp_vsum): Likewise.
7077 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7078 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7079 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7080 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7081 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7082 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7083 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7084 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7085 (cortex_a9_neon_bp_simple): Likewise.
7086 (cortex_a9_neon_bp_2cycle): Likewise.
7087 (cortex_a9_neon_bp_3cycle): Likewise.
7088 (cortex_a9_neon_ldr): Likewise.
7089 (cortex_a9_neon_str): Likewise.
7090 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7091 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7092 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7093 (cortex_a9_neon_vld2_4_regs): Likewise.
7094 (cortex_a9_neon_vld3_vld4): Likewise.
7095 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7096 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7097 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7098 (cortex_a9_neon_vst3_vst4): Likewise.
7099 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7100 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7101 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7102 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7103 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7104 (cortex_a9_neon_mcr): Likewise.
7105 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7106 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7107 (cortex_a9_fps): Likewise.
7108 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7110 (cortex_m4_fmuls): Likewise.
7111 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7113 (cortex_r4_mrc): Likewise.
7114 * config/arm/iterators.md: Update comment referring to neon_type.
7115 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7116 (iwmmxt_movsi_insn): Likewise.
7117 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7119 (pj4_core_to_vfp): Likewise.
7120 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7122 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7123 (vfp_fstore): Likewise.
7124 * doc/md.texi: Change references to neon_type to refer to type.
7126 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7128 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7130 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7132 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7135 2013-09-04 Easwaran Raman <eraman@google.com>
7138 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7139 (build_and_add_sum): Use it.
7140 (appears_later_in_bb): Simplify code.
7142 2013-09-04 Teresa Johnson <tejohnson@google.com>
7144 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7146 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7148 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7150 2013-09-04 Jan Hubicka <jh@suse.cz>
7152 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7153 * ipa-devirt.c: Include diganostic.h
7154 (odr_type_d): Add types and types_set.
7155 (hash_type_name): Work for types with vtables during LTO.
7156 (odr_hasher::remove): Fix comment; destroy types_set.
7157 (add_type_duplicate): New function,
7158 (get_odr_type): Use it.
7159 (dump_type_inheritance_graph): Dump type duplicates.
7160 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7162 * tree.c (types_same_for_odr): Give exact answers on types with
7165 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7167 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7168 explaining their differences.
7170 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7172 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7174 2013-09-03 Jeff Law <law@redhat.com>
7176 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7177 when not threading through a joiner block. Pass joiner/no joiner
7178 state to register_jump_thread.
7179 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7180 state from argument rather than implying on path length.
7181 Dump the entire jump thread path into debugging dump.
7182 * tree-flow.h (register_jump_thread): Update prototype.
7184 2013-08-29 Xinliang David Li <davidxl@google.com>
7186 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7187 Remove a trivial gcc_assert.
7189 2013-08-29 Xinliang David Li <davidxl@google.com>
7191 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7192 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7193 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7194 Delay base decl alignment adjustment.
7195 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7196 * tree-vectorizer.h: New data structure.
7197 (set_dr_misalignment): New function.
7198 (dr_misalignment): Ditto.
7199 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7200 (vectorizable_load): Ditto.
7201 (ensure_base_align): New function.
7202 (vectorize_loops): Add dbg_cnt support.
7203 (execute_vect_slp): Ditto.
7204 * dbgcnt.def: New debug counter.
7205 * Makefile: New dependency.
7207 2013-09-03 Meador Inge <meadori@codesourcery.com>
7211 2013-08-30 Meador Inge <meadori@codesourcery.com>
7213 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7215 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7217 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7218 function descriptor.
7220 2013-09-03 Richard Biener <rguenther@suse.de>
7222 * tree-affine.c (add_elt_to_tree): Fix association issue,
7223 avoid useless converts and make sure to always return a
7224 properly typed result.
7226 2013-09-03 Richard Biener <rguenther@suse.de>
7229 * fold-const.c (negate_expr_p): Fix division case.
7230 (negate_expr): Likewise.
7232 2013-09-03 Richard Biener <rguenther@suse.de>
7235 * tree-streamer-out.c: Include tm.h.
7236 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7238 2013-09-03 Jan Hubicka <jh@suse.cz>
7240 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7242 2013-09-03 Alan Modra <amodra@gmail.com>
7244 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7245 multilibs to big-endian target and vice versa.
7246 * config/rs6000/t-linux64: Use := assignment on all vars.
7247 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7248 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7249 * config/rs6000/t-linux64le: New file.
7250 * config/rs6000/t-linux64bele: New file.
7251 * config/rs6000/t-linux64lebe: New file.
7253 2013-09-02 Jan Hubicka <jh@suse.cz>
7255 * ipa-inline-transform.c (inline_transform): Do not
7256 optimize_inline_calls when not optimizing.
7258 2013-09-02 Jan Hubicka <jh@suse.cz>
7260 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7261 duplicated nodes for assembler names.
7262 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7263 hard registers from assembler name hash.
7265 2013-09-02 Jan Hubicka <jh@suse.cz>
7267 * ipa-split.c (execute_split_functions): Split externally visible
7268 functions called once.
7270 2013-09-02 Martin Jambor <mjambor@suse.cz>
7273 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7274 linked list. When finding the correct duplicate, also consider also
7275 the caller in additon to its inlined_to node.
7277 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7279 * config/aarch64/aarch64-simd-builtins.def
7280 (dup_lane_scalar): Remove.
7281 * config/aarch64/aarch64-simd.md
7282 (aarch64_simd_dup): Add 'w->w' alternative.
7283 (aarch64_dup_lane<mode>): Allow for VALL.
7284 (aarch64_dup_lane_scalar<mode>): Remove.
7285 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7286 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7287 (aarch64_get_lane_unsigned<mode>): Likewise.
7288 (aarch64_get_lane<mode>): Likewise.
7289 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7290 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7291 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7292 (VCON): Change container of V2SF.
7293 (vswap_width_name): Likewise.
7294 * config/aarch64/arm_neon.h
7295 (__aarch64_vdup_lane_any): New.
7296 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7297 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7298 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7300 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7303 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7304 the source or the destination is a single hard register.
7306 2013-09-02 Richard Biener <rguenther@suse.de>
7309 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7312 2013-09-02 Richard Biener <rguenther@suse.de>
7314 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7315 arithmetic to sizetype.
7317 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7319 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7320 Find auto-increment use both before and after candidate.
7322 2013-09-02 Marek Polacek <polacek@redhat.com>
7324 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7326 2013-09-01 Jan Hubicka <jh@suse.cz>
7328 * Makefile.in: Add ipa-profile.o
7329 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7330 * cgraph.c (struct cgraph_propagate_frequency_data,
7331 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7332 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7333 * cgraph.h (cgraph_propagate_frequency): Remove.
7334 * ipa-inline-analysis.c: Include ipa-utils.h;
7335 drop duplicated cfgloop.h.
7336 (inline_update_callee_summaries): Update.
7337 * ipa-profile.c: New file.
7338 * ipa-utils.h (ipa_propagate_frequency): Declare.
7339 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7340 data-streamer.h, value-prof.h.
7341 (symtab_remove_unreachable_nodes): Update profile.
7342 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7343 account_time_size, cmp_counts, dump_histogram,
7344 ipa_profile_generate_summary, ipa_profile_write_summary,
7345 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7346 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7347 Move to ipa-profile.c.
7349 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7351 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7353 2013-09-01 Jan Hubicka <jh@suse.cz>
7355 * common.opt (fdevirtualize-speculatively): New function.
7356 * invoke.texi (fdevirtualize-speculatively): Document.
7357 * ipa-devirt.c: Include ipa-inline.h
7358 (likely_target_p): New function.
7359 (ipa_devirt): New function.
7360 (gate_ipa_devirt): New function.
7361 (pass_data_ipa_devirt): New static var.
7362 (pass_ipa_devirt): Likewise.
7363 (make_pass_ipa_devirt): New function.
7364 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7365 (common_handle_option): Disable devirtualization when
7366 value range profiling is available.
7367 * passes.def (pass_ipa_devirt): Add.
7368 * timever.def (TV_IPA_DEVIRT): New timevar.
7369 * tree-pass.h (make_pass_ipa_devirt):
7371 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7373 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7374 include sanitize(undefined).
7376 2013-08-31 Diego Novillo <dnovillo@google.com>
7378 * Makefile.in (TREE_CORE_H): Define.
7380 (GTFILES): Add tree-core.h.
7381 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7383 * tree-core.h: New file.
7384 Move all data structures, enum, typedefs, global
7385 declarations and constants from ...
7388 2013-08-31 Jan Hubicka <jh@suse.cz>
7390 * bulitins.c (expand_builtin): Do not early exit for gcov
7391 instrumented functions.
7393 2013-08-31 Marek Polacek <polacek@redhat.com>
7395 * ubsan.c: Include tm_p.h.
7397 2013-08-31 Jan Hubicka <jh@suse.cz>
7399 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7402 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7403 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7404 * tree-cfg.c (verify_gimple_label): ... here.
7405 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7406 (ipa_merge_profiles): New function.
7407 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7408 (lto_input_function_body): Likewise.
7409 * ipa-utils.h (ipa_merge_profiles): Declare.
7410 * lto-streamer.h (lto_input_function_body): Update prototype.
7411 (emit_label_in_global_context_p): Remove.
7412 * lto-symtab.c: Include ipa-utils.h
7413 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7415 2013-08-31 Jan Hubicka <jh@suse.cz>
7417 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7419 2013-08-31 Jan Hubicka <jh@suse.cz>
7421 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7423 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7425 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7426 "cmp" RTX before signed_comparison_operator check to account
7429 2013-08-30 Jan Hubicka <jh@suse.cz>
7431 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7432 (detect_type_change_1): Rename to ...
7433 (detect_type_change): ... this one; early return on non-polymorphic
7435 (detect_type_change_ssa): Add comp_type parameter; update
7436 use of detect_type_change.
7437 (compute_complex_assign_jump_func): Add param_type parameter;
7438 update use of detect_type_change_ssa.
7439 (compute_complex_ancestor_jump_func): Likewise.
7440 (ipa_get_callee_param_type): New function.
7441 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7442 update calls to the jump function computation functions.
7444 2013-08-30 Teresa Johnson <tejohnson@google.com>
7445 Steven Bosscher <steven@gcc.gnu.org>
7447 * cfgrtl.c (fixup_new_cold_bb): New routine.
7448 (commit_edge_insertions): Invoke fixup_partitions.
7449 (find_partition_fixes): New routine.
7450 (fixup_partitions): Ditto.
7451 (verify_hot_cold_block_grouping): Update comments.
7452 (rtl_verify_edges): Invoke find_partition_fixes.
7453 (rtl_verify_bb_pointers): Update comments.
7454 (rtl_verify_bb_layout): Ditto.
7455 * basic-block.h (probably_never_executed_edge_p): Declare.
7456 (fixup_partitions): Ditto.
7457 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7458 * bb-reorder.c (sanitize_hot_paths): New function.
7459 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7461 * predict.c (probably_never_executed_edge_p): New routine.
7462 * cfg.c (check_bb_profile): Add partition insanity warnings.
7464 2013-08-30 Meador Inge <meadori@codesourcery.com>
7466 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7468 2013-08-30 Marek Polacek <polacek@redhat.com>
7470 * Makefile.in (ubsan.o): Add.
7471 (c-family/c-ubsan.o): Add.
7472 (builtins.o): Add ubsan.h dependency.
7473 * ubsan.h: New file.
7474 * ubsan.c: New file.
7475 * common.opt: Add -fsanitize=undefined option.
7476 (flag_sanitize): Add variable.
7477 (fsanitize=): Add option. Add Driver.
7478 (fsanitize=thread): Remove option.
7479 (fsanitize=address): Likewise.
7480 (static-libubsan): New option.
7481 * doc/invoke.texi: Document the new flag and -static-libubsan.
7482 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7483 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7484 * builtin-attrs.def (ATTR_COLD): Define.
7485 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7486 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7487 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7488 * flag-types.h (sanitize_code): New enum.
7489 * opts.c (common_handle_option): Parse command line arguments
7490 of -fsanitize=. Add -fsanitize=unreachable option.
7491 * varasm.c (get_variable_section): Adjust.
7492 (assemble_noswitch_variable): Likewise.
7493 (assemble_variable): Likewise.
7494 (output_constant_def_contents): Likewise.
7495 (categorize_decl_for_section): Likewise.
7496 (place_block_symbol): Likewise.
7497 (output_object_block): Likewise.
7498 * builtins.def: Likewise.
7499 * toplev.c (compile_file): Likewise.
7500 (process_options): Likewise.
7501 * cppbuiltin.c: Likewise.
7502 * tsan.c (tsan_pass): Likewise.
7503 (tsan_gate): Likewise.
7504 (tsan_gate_O0): Likewise.
7505 * cfgexpand.c (partition_stack_vars): Likewise.
7506 (expand_stack_vars): Likewise.
7507 (defer_stack_allocation): Likewise.
7508 (expand_used_vars): Likewise.
7509 * cfgcleanup.c (old_insns_match_p): Likewise.
7510 * asan.c (asan_finish_file): Likewise.
7511 (asan_instrument): Likewise.
7512 (gate_asan): Likewise.
7513 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7514 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7515 (asan_global_struct): Use pointer_sized_int_node instead
7516 calling build_nonstandard_integer_type.
7517 (initialize_sanitizer_builtins): Likewise.
7518 (asan_finish_file): Likewise.
7519 * gcc.c: Document %{%:function(args):X}.
7520 (static_spec_functions): Add sanitize.
7521 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7522 store funcval != NULL there.
7523 (do_spec_1): Adjust handle_spec_function caller.
7524 (handle_braces): Allow %:function(args) as condition.
7525 (sanitize_spec_function): New function.
7526 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7527 (LIBUBSAN_SPEC): Likewise.
7528 (LIBUBSAN_EARLY_SPEC): Likewise.
7529 (SANITIZER_SPEC): Handle libubsan.
7530 (SANITIZER_EARLY_SPEC): Likewise.
7531 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7532 instead of fsanitize=address.
7533 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7534 instead of fsanitize=address*.
7535 * builtins.c: Include ubsan.h.
7536 (fold_builtin_0): Instrument __builtin_unreachable.
7537 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7538 instead of flag_asan.
7539 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7540 (pointer_sized_int_node): Define.
7541 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7543 2013-08-30 Mike Stump <mikestump@comcast.net>
7545 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7548 2013-08-29 Jan Hubicka <jh@suse.cz>
7550 * cgraph.c (cgraph_function_body_availability): Handle weakref
7552 * passes.def: Remove pass_fixup_cfg.
7553 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7554 track when we need to remove functions.
7555 (gate_ipa_inline): Execute inlining always; add comment why.
7556 (pass_data_ipa_inline): Remove TODO_remove_functions.
7557 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7558 do not produce summaries.
7559 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7560 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7561 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7562 constant pool and vtable.
7564 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7566 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7567 New arm_neon.h's internal macros to specify 64-bit constants.
7568 Avoid using stdint.h's macros.
7570 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7572 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7573 before applying REG_P.
7575 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7577 PR tree-optimization/58277
7578 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7579 after seeing too many stmts with vdef in between dombb and current
7580 bb, invalidate everything.
7582 2013-08-30 Richard Biener <rguenther@suse.de>
7584 * fold-const.c (fold_single_bit_test): Fix overflow test.
7586 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7588 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7589 and which can live in a register, always retrieve the value on entry.
7590 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7591 passed by invisible reference specially.
7592 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7593 (vt_add_function_parameter): Correctly deal with a parameter passed by
7594 invisible reference.
7596 2013-08-30 Jan Hubicka <jh@suse.cz>
7598 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7600 2013-08-30 Richard Biener <rguenther@suse.de>
7602 PR tree-optimization/58228
7603 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7604 allow invariant loads in nested loop vectorization.
7606 2013-08-30 Richard Biener <rguenther@suse.de>
7608 PR tree-optimization/58223
7609 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7610 (has_anti_or_output_dependence): ... this and adjust to also
7611 look for output dependences.
7612 (mark_nodes_having_upstream_mem_writes): Adjust.
7613 (rdg_flag_uses): Likewise.
7615 2013-08-30 Richard Biener <rguenther@suse.de>
7617 PR tree-optimization/58010
7618 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7619 assert that we have a loop-closed PHI.
7621 2013-08-29 Jan Hubicka <jh@suse.cz>
7623 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7624 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7625 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7626 (lto_free_function_in_decl_state_for_node): New function.
7628 2013-08-29 Xinliang David Li <davidxl@google.com>
7630 * loop-unroll.c (report_unroll_peel): Minor message change.
7631 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7632 Emit alignment peeling message with default -fopt-info.
7633 (vect_loop_versioning): Emit loop version info message.
7634 * tree-vectorizer.c (vectorize_loops): Minor message change.
7635 (execute_vect_slp): Ditto.
7637 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7639 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7640 of the clone from the DECL_NAME of the original function.
7642 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7644 * passes.c (register_pass): Add overload.
7645 * tree-pass.h (register_pass): Forward declare it. Add comment.
7647 2013-08-29 Jan Hubicka <jh@suse.cz>
7649 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7650 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7651 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7652 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7653 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7654 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7655 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7656 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7658 2013-08-29 Teresa Johnson <tejohnson@google.com>
7660 * dumpfile.c (dump_loc): Output column number.
7661 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7662 * doc/invoke.texi: Document optall -fopt-info flag.
7663 * profile.c (read_profile_edge_counts): Use new dump framework.
7664 (compute_branch_probabilities): Ditto.
7665 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7666 when pass not in any opt group.
7667 * pass_manager.h (pass_manager::get_pass_profile): New method.
7668 * value-prof.c (check_counter): Use new dump framework.
7669 (check_ic_target): Ditto.
7670 * coverage.c (get_coverage_counts): Ditto.
7671 (coverage_init): Setup new dump framework.
7673 2013-08-29 Richard Biener <rguenther@suse.de>
7675 PR tree-optimization/58246
7676 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7677 handle the dominance check inside a basic-block.
7679 2013-08-29 Richard Biener <rguenther@suse.de>
7682 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7683 of default defs that appear in abnormal PHI nodes.
7685 2013-08-29 Richard Biener <rguenther@suse.de>
7687 PR tree-optimization/57685
7688 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7689 single-use operands to avoid exponential complexity.
7691 2013-08-28 Dehao Chen <dehao@google.com>
7693 * ipa-inline.c (edge_badness): Fix integer underflow.
7695 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7697 * gtm-builtins.def (_ITM_free): Declare leaf.
7699 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7702 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7703 (*tls_local_dynamic_base_64_largepic): Likewise.
7704 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7705 Remove predicate from call operand.
7706 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7707 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7709 2013-08-28 Jeff Law <law@redhat.com>
7711 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7712 checks for the number of predecessors and successors allowed.
7713 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7714 which require copying a joiner block if there is a request which
7715 is a subpath that requires no joiner block copying.
7717 2013-08-28 Jan Hubicka <jh@suse.cz>
7719 * lto-streamer-out.c (DFS_write_tree_body): Drop
7720 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7721 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7722 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7723 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7724 Do not read DECL_ERROR_ISSUED.
7725 (unpack_ts_decl_with_vis_value_fields): Do not read
7727 (lto_input_ts_binfo_tree_pointers): Do not read
7728 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7729 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7730 write DECL_ERROR_ISSUED..
7731 (pack_ts_decl_with_vis_value_fields): Do not write
7733 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7734 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7735 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7736 * tree.h (tree_decl_common): Update comment.
7737 (DECL_ERROR_ISSUED): Remove.
7739 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7742 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7744 2013-08-28 Jan Hubicka <jh@suse.cz>
7746 * builtins.def (free): Declare leaf.
7748 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7751 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7752 * configure: Regenerate.
7754 2013-08-27 Martin Jambor <mjambor@suse.cz>
7756 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7757 (ipa_ancestor_jf_data): Likewise.
7758 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7759 (ipa_get_jf_pass_through_type_preserved): New function.
7760 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7761 (ipa_get_jf_ancestor_type_preserved): New function.
7762 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7763 (ipa_get_jf_ancestor_result): Likewise.
7764 (propagate_vals_accross_pass_through): Use
7765 ipa_get_jf_pass_through_result to do all the value mappings.
7766 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7767 type_preserved flag.
7768 (ipa_set_jf_cst_copy): New function.
7769 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7770 (ipa_set_jf_arith_pass_through): Likewise.
7771 (ipa_set_ancestor_jf): Likewise.
7772 (compute_complex_assign_jump_func): Set type_preserved instead of
7774 (ipa_compute_jump_functions_for_edge): Likewise.
7775 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7776 (update_jump_functions_after_inlining): Update type_preserved.
7777 Explicitely create jump functions when combining one with pass_through.
7778 (ipa_write_jump_function): Stream the type_preserved flags.
7779 (ipa_read_jump_function): Likewise.
7781 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7782 Aldy Hernandez <aldyh@redhat.com>
7784 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7785 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7786 * function.h (struct function): Add has_force_vect_loops and
7788 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7789 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7791 * gimple.def: Update CLAUSES comments.
7792 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7793 (gimple_build_omp_for): Add argument to prototype.
7794 (gimple_omp_for_kind): New.
7795 (gimple_omp_for_set_kind): New.
7796 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7797 GOVD_DATA_SHARE_CLASS.
7798 (enum omp_region_type): Add ORT_SIMD.
7799 (gimple_add_tmp_var): Handle ORT_SIMD.
7800 (gimplify_var_or_parm_decl): Same.
7801 (is_gimple_stmt): Same.
7802 (omp_firstprivatize_variable): Same.
7803 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7804 (omp_notice_variable): Handle ORT_SIMD.
7805 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7806 (omp_check_private): Handle ORT_SIMD.
7807 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7809 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7810 Handle OMP_CLAUSE_LASTPRIVATE.
7811 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7813 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7814 (gimplify_expr): Handle OMP_SIMD.
7815 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7816 (expand_GOMP_SIMD_VF): New.
7817 (expand_GOMP_SIMD_LAST_LANE): New.
7818 * internal-fn.def (GOMP_SIMD_LANE): New.
7819 (GOMP_SIMD_VF): New.
7820 (GOMP_SIMD_LAST_LANE): New.
7821 * omp-low.c: Include target.h.
7822 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7824 (check_omp_nesting_restrictions): Same.
7826 (lower_rec_simd_input_clauses): New.
7827 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7829 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7830 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7831 (expand_omp_build_assign): New.
7832 (expand_omp_for_init_counts): New.
7833 (expand_omp_for_init_vars): New.
7834 (extract_omp_for_update_vars): New.
7835 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7836 and rewrite accordingly.
7837 (expand_omp_simd): New.
7838 (expand_omp_for): Use expand_omp_simd.
7839 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7840 (lower_omp_for): Do not lower the body.
7841 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7843 * tree-flow.h (find_omp_clause): Remove prototype.
7844 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7845 forcing vectorization of the loop, or if flag_tree_vectorize.
7846 (gate_tree_if_conversion): Similarly.
7847 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7848 gimple_build_omp_for.
7849 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7850 * tree-parloops (create_parallel_loop): Pass kind argument to
7851 gimple_build_omp_for.
7852 * tree-pretty-print.c (dump_omp_clause): Add cases for
7853 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7854 OMP_CLAUSE__SIMDUID_.
7855 (dump_generic_node): Handle OMP_SIMD.
7856 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7857 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7858 unroll OMP_SIMD loops here.
7859 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7860 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7862 (vect_analyze_data_refs): Handle simd loops.
7863 * tree-vect-loop.c (vectorizable_live_operation): Handle
7865 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7866 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7867 (vectorizable_load): Same.
7868 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7869 (struct simduid_to_vf): New.
7870 (simduid_to_vf::hash): New.
7871 (simduid_to-vf::equal): New.
7872 (struct simd_array_to_simduid): New.
7873 (simd_array_to_simduid::hash): New.
7874 (simd_array_to_simduid::equal): New.
7875 (adjust_simduid_builtins): New.
7876 (struct note_simd_array_uses_struct): New.
7877 (note_simd_array_uses_cb): New.
7878 (note_simd_array_uses): New.
7879 (vectorize_loops): Handle simd hints and adjust simd builtins
7881 * tree-vectorizer.h (struct _stmt_vec_info): Add
7882 simd_lane_access_p field.
7883 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7884 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7885 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7886 (omp_clause_code_name): Same.
7887 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7888 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7889 * tree.def (OMP_SIMD): New entry.
7890 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7891 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7892 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7893 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7894 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7895 (OMP_CLAUSE_LINEAR_STEP): New.
7896 (OMP_CLAUSE_SAFELEN_EXPR): New.
7897 (OMP_CLAUSE__SIMDUID__DECL): New.
7898 (find_omp_clause): New prototype.
7900 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7902 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7903 Haswell processor detection.
7905 2013-08-27 Christian Widmer <shadow@umbrox.de>
7908 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7909 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7910 AVX2 capable processors.
7912 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7914 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7915 of vget_low_* with implementations in terms of other intrinsics.
7917 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7920 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7921 values to -1, 0 and 1.
7923 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7925 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7926 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7927 (<optab><mode>3_insn): Remove.
7928 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7929 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7930 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7931 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7932 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7933 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7934 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7935 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7936 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7937 (ror<mode>3_insn): Likewise.
7938 * config/aarch64/predicates.md (aarch64_simd_register): New.
7940 2013-08-27 Richard Biener <rguenther@suse.de>
7942 PR tree-optimization/57521
7943 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7944 one edge is non-critical.
7945 (find_phi_replacement_condition): Make sure to use a non-critical
7946 edge. Cleanup and remove old bug workarounds.
7947 (bb_postdominates_preds): Remove.
7948 (if_convertible_loop_p_1): Do not compute post-dominators.
7949 (combine_blocks): Do not free post-dominators.
7950 (main_tree_if_conversion): Likewise.
7951 (pass_data_if_conversion): Add TODO_verify_ssa.
7953 2013-08-27 DJ Delorie <dj@redhat.com>
7955 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7957 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7959 * function.c (assign_parm_find_data_types): Set passed_mode and
7960 nominal_mode to the TYPE_MODE of nominal_type for the built
7961 pointer type in case of the struct-pass-by-reference.
7963 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7965 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7966 (UINT16_TYPE): Change default to "unsigned int".
7968 * config/avr/avr.opt (mfract-convert-truncate): New option.
7969 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7970 is set, round negative fractional integers according to n1169
7971 when converting to integer types.
7973 2013-08-26 Jan Hubicka <jh@suse.cz>
7975 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7976 methods can not be called indirectly when their address is not taken.
7978 2013-08-26 Jan Hubicka <jh@suse.cz>
7980 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7983 2013-08-26 Jan Hubicka <jh@suse.cz>
7985 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7988 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7990 * reload.c (find_valid_class): Allow classes that do not include
7991 FIRST_PSEUDO_REGISTER - 1.
7993 2013-08-26 Jan Hubicka <jh@suse.cz>
7995 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7996 fix edge count/frequency when speculation failed; fix type check
7997 for the direct call.
7999 2013-08-26 Jan Hubicka <jh@suse.cz>
8001 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8003 2013-08-26 Jan Hubicka <jh@suse.cz>
8005 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8006 to be changed by edge redirection.
8008 2013-08-26 Jan Hubicka <jh@suse.cz>
8010 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8011 formating; add sanity check.
8012 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8013 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8014 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8015 (resolve_noninline_speculation): Update callee keys, too.
8017 2013-08-26 Jan Hubicka <jh@suse.cz>
8019 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8020 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8022 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8024 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8027 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8029 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8030 (x86_64_elf_select_section): Ditto.
8031 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8033 (ix86_pass_by_reference): Ditto.
8034 (output_set_got): Ditto.
8035 (ix86_unary_operator_ok): Ditto.
8036 (ix86_expand_builtin): Ditto.
8038 2013-08-23 Jan Hubicka <jh@suse.cz>
8040 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8042 2013-08-23 Jan Hubicka <jh@suse.cz>
8044 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8045 (tree_decl_with_vis): Add FINAL field.
8047 2013-08-23 Jeff Law <law@redhat.com>
8049 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8050 the debugging dump when the expression is fully redundant.
8052 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8054 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8055 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8056 * pretty-print.c (pp_formatted_text_data): Likewise.
8057 (pp_write_text_to_stream): Likewise.
8058 (pp_write_text_as_dot_label_to_stream): Likewise.
8059 (pp_append_r): Likewise.
8060 (pp_format): Likewise.
8061 (pp_flush): Likewise.
8062 (pp_clear_output_area): Likewise.
8063 (pp_append_text): Likewise.
8064 (pp_formatted_text): Likewise.
8065 (pp_remaining_character_count_for_line): Likewise.
8066 (pp_newline): Likewise.
8067 (pp_character): Likewise.
8068 (output_buffer::~output_buffer): Define.
8069 (pretty_printer::~pretty_printer): Destruct output buffer.
8070 * pretty-print.h (output_buffer::~output_buffer): Declare.
8071 (pretty_printer::~pretty_printer): Declare virtual.
8073 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8076 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8077 HOST_WIDE_INT_M1U): New macros.
8078 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8079 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8080 unsigned -1 for lshift.
8081 * cse.c (cse_insn): Likewise.
8082 * double-int.c (rshift_double, lshift_double): Likewise.
8083 * builtins.c (fold_builtin_bitop): Likewise.
8084 * combine.c (force_to_mode): Likewise.
8085 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8086 * simplify-rtx.c (simplify_const_unary_operation,
8087 simplify_const_binary_operation): Likewise.
8088 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8089 check_va_list_escapes): Likewise.
8090 * rtlanal.c (nonzero_bits1): Likewise.
8091 * expmed.c (expand_smod_pow2): Likewise.
8092 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8094 2013-08-23 Jan Hubicka <jh@suse.cz>
8096 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8097 as having address taken.
8099 2013-08-23 Jan Hubicka <jh@suse.cz>
8101 * ipa-utils.h (method_class_type): Declare.
8102 * ipa-devirt.c (method_class_type): Export.
8104 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8105 do not walk base classes of anonymous types.
8107 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8109 PR rtl-optimization/58220
8111 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8112 handle SEQUENCE insns properly.
8114 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8116 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8118 (pp_newline_and_indent): Likewise.
8119 (pp_separate_with): Likewise.
8120 * pretty-print.c (pp_newline_and_flush): Define.
8121 (pp_newline_and_indent): Likewise.
8122 (pp_separate_with): Likewise.
8124 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8127 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8128 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8130 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8132 * config/i386/predicates.md (ext_sse_reg_operand): New.
8133 * config/i386/i386.md (*movti_internal): Use
8134 predicate to determine if EVEX is needed.
8135 (*movsi_internal): Ditto.
8136 (*movdf_internal): Ditto.
8137 (*movsf_internal): Ditto.
8138 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8140 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8142 PR tree-optimization/58209
8143 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8144 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8145 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8146 emit POINTER_PLUS_EXPR.
8147 (create_tailcall_accumulator): For pointer result type accumulate in
8150 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8152 * configure.ac: Add backslashes missing from the last change.
8153 * configure: Regenerate.
8155 2013-08-22 Jan Hubicka <jh@susue.cz>
8157 * ipa.c (function_and_variable_visibility): First remember function
8158 was global and then make it local.
8160 2013-08-22 Julian Brown <julian@codesourcery.com>
8162 * configure.ac: Add aarch64 to list of arches which use "nop" in
8164 * configure: Regenerate.
8166 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8168 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8169 gnu_libc_has_function.
8170 * config/s390/tpf.h: Likewise.
8172 2013-08-22 Jan Hubicka <jh@susue.cz>
8174 * timevar.c (validate_phases): Add cast.
8176 2013-08-22 Jan Hubicka <jh@susue.cz>
8178 * timevar.c (validate_phases): Use size_t for memory.
8179 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8181 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8183 * pretty-print.h (output_buffer::output_buffer): Declare.
8184 (pretty_printer::pretty_printer): Likewise.
8185 (pp_construct): Remove.
8186 * pretty-print.c (output_buffer::output_buffer): Define.
8187 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8188 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8189 (print_gimple_expr): Likewise.
8190 (print_gimple_seq): Likewise.
8191 (gimple_dump_bb): Likewise.
8192 * sched-vis.c (dump_value_slim): Likewise.
8193 (dump_insn_slim): Likewise.
8194 (dump_rtl_slim): Likewise.
8195 (str_pattern_slim): Likewise.
8196 * tree-mudflap.c (mf_varname_tree): Likewise.
8197 * graph.c (print_graph_cfg): Likewise.
8198 (start_graph_dump): Likewise.
8199 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8201 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8202 pretty printer initialization.
8203 * coretypes.h (diagnostic_context): Remove superflous type alias
8205 (pretty_printer): Likewise. Declare directly as a class.
8206 (pretty_print_info): Remove declaration as class.
8207 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8208 and pp_clear_output_area.
8209 (asan_add_global): Likewise.
8211 2013-08-22 Jan Hubicka <jh@suse.cz>
8213 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8214 * ipa-utils.h (update_type_inheritance_graph): Declare.
8215 (possible_polymorphic_call_target_p): Declare.
8216 (possible_polymorphic_call_target_p): New.
8217 * ipa-devirt.c: Update toplevel comments.
8218 (cached_polymorphic_call_targets): Move up.
8219 (odr_type_d): Move ID down.
8220 (polymorphic_type_binfo_p): Update comment.
8221 (odr_hasher::remove): Likewise;
8222 (get_odr_type): Set anonymous_namespace.
8223 (dump_odr_type): Dump it.
8224 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8225 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8226 (record_binfo): Add comment.
8227 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8229 (devirt_node_removal_hook): Do not iCE when cache is freed.
8230 (possible_polymorphic_call_target_p): New predicate.
8231 (update_type_inheritance_graph): New function.
8233 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8234 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8235 Sergey Lega <sergey.s.lega@intel.com>
8236 Anna Tikhonova <anna.tikhonova@intel.com>
8237 Ilya Tocar <ilya.tocar@intel.com>
8238 Andrey Turetskiy <andrey.turetskiy@intel.com>
8239 Ilya Verbin <ilya.verbin@intel.com>
8240 Kirill Yukhin <kirill.yukhin@intel.com>
8241 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8243 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8244 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8245 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8246 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8247 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8248 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8249 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8250 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8251 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8252 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8253 OPT_mavx512pf, OPT_mavx512er cases.
8254 * config/i386/constraints.md (v): New constraint.
8255 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8256 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8257 (bit_AVX512CD): New.
8258 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8259 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8260 * config/i386/i386-c.c (ix86_target_macros_internal):
8261 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8263 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8264 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8265 * config/i386/i386.c (regclass_map, dbx_register_map)
8266 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8267 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8268 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8270 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8271 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8272 -mavx512pf options. Fix formatting.
8273 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8274 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8275 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8276 -mavx512cd, -mavx512pf options.
8277 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8278 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8279 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8281 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8282 (ix86_set_reg_reg_cost): Ditto.
8283 (x86_order_regs_for_local_alloc): Ditto.
8284 (MAX_VECT_LEN): Extend to 64-byte.
8285 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8286 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8287 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8288 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8289 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8290 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8291 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8292 (SSE_REG_MODE_P): Support new modes.
8293 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8294 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8295 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8296 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8297 (REG_CLASS_CONTENTS): Add new registers.
8298 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8299 (EXT_REX_SSE_REGNO_P): New.
8300 (HI_REGISTER_NAMES): Add new registers.
8301 * config/i386/i386.md: Define constants for new registers.
8302 (mode): Add new 512-bit modes.
8303 (prefix): Support evex prefix.
8304 (isa): Support avx512f, noavx512f, fma_avx512f.
8305 (ssemodesuffix): Add new 512-bit modes.
8307 (*movxi_internal_avx512f): Ditto.
8308 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8310 (*movsi_internal): Likewise.
8311 (*movdf_internal): Likewise.
8312 (*movsf_internal): Likewise.
8313 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8314 (<code><mode>3): Likewise.
8315 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8317 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8318 with the new constraint "v".
8319 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8321 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8322 with the new constraint "v".
8323 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8324 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8325 (*<plusminus_insn><mode>3): Likewise.
8326 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8327 (*mul<mode>3): Likewise.
8328 (<sse>_vmmul<mode>3): Likewise.
8329 (<sse>_div<mode>3): Likewise.
8330 (<sse>_vmdiv<mode>3): Likewise.
8331 (<sse>_sqrt<mode>2): Likewise.
8332 (<sse>_vmsqrt<mode>2): Likewise.
8333 (*<code><mode>3_finite): Likewise.
8334 (*<code><mode>3) <smaxmin>: Likewise.
8335 (<sse>_vm<code><mode>3): Likewise.
8336 (*<code><mode>3) <any_logic>: Likewise.
8337 (*fma_fmadd_<mode>): Likewise.
8338 (*fma_fmsub_<mode>): Likewise.
8339 (*fma_fnmadd_<mode>): Likewise.
8340 (*fma_fnmsub_<mode>): Likewise.
8341 (*fma_fmaddsub_<mode>): Likewise.
8342 (*fma_fmsubadd_<mode>): Likewise.
8343 (*fmai_fmadd_<mode>): Likewise.
8344 (*fmai_fmsub_<mode>): Likewise.
8345 (*fmai_fnmadd_<mode>): Likewise.
8346 (*fmai_fnmsub_<mode>): Likewise.
8347 (sse_cvtsi2ss): Likewise.
8348 (sse_cvtsi2ssq): Likewise.
8349 (sse_cvtss2si): Likewise.
8350 (sse_cvtss2si_2): Likewise.
8351 (sse_cvtss2siq): Likewise.
8352 (sse_cvtss2siq_2): Likewise.
8353 (sse_cvttss2si): Likewise.
8354 (sse_cvtss2siq_2): Likewise.
8355 (float<sseintvecmodelower><mode>2): Likewise.
8356 (sse2_cvtsd2si_2): Likewise.
8357 (sse2_cvtsd2siq_2): Likewise.
8358 (*<plusminus_insn><mode>3): Likewise.
8359 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8360 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8361 (ashr<mode>3): Likewise.
8362 (<shift_insn><mode>3): Likewise.
8363 (avx2_<code><mode>3): Likewise.
8364 (*avx2_<code><mode>3): Likewise.
8365 (*andnot<mode>3): Likewise.
8366 (*<code><mode>3) <any_logic>: Likewise.
8367 (abs<mode>2): Likewise.
8368 (avx2_permvar<mode>): Likewise.
8369 (avx2_perm<mode>_1): Likewise.
8370 (*avx_vpermilp<mode>): Likewise.
8371 (avx_vpermilvar<mode>3): Likewise.
8372 (avx2_ashrv<mode>): Likewise.
8373 (avx2_<shift_insn>v<mode>): Likewise.
8374 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8376 * doc/rtl.texi: Document XImode.
8378 2013-08-21 Jeff Law <law@redhat.com>
8380 * tree-flow.h (register_jump_thread): Pass vector of edges
8381 instead of each important edge.
8382 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8383 thread path into a vector and pass that to register_jump_thread.
8384 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8385 passed in edge vector to the current 3-edge form.
8388 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8390 * dce.c (fini_dce): Call df_analyze again just in case
8391 delete_unmarked_insns removed anything.
8393 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8395 * reload.h (struct reg_equivs): Rename to ..
8396 (struct reg_equivs_s): .. this.
8398 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8400 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8402 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8404 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8406 2013-08-21 Jeff Law <law@redhat.com>
8408 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8409 simplify assignments too. If the RHS collapses to a singleton
8410 range, then return the value for the range.
8412 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8414 * config/i386/sse.md (V16): Rename to...
8416 (mov<mode>): Update iterator name.
8417 (*mov<mode>_internal): Ditto.
8418 (push<mode>1): Ditto.
8419 (movmisalign<mode>): Ditto.
8421 2013-08-20 Jan Hubicka <jh@suse.cz>
8424 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8425 entry for direct edges.
8426 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8428 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8430 Revert my last two changes, r201865 and r201864:
8433 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8435 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8436 instances can own GC refs.
8438 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8439 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8440 (gcc::context::gt_pch_nx): Likewise.
8441 (gcc::context::gt_pch_nx): Likewise.
8442 * ggc.h (gt_ggc_mx <T>): New.
8443 (gt_pch_nx_with_op <T>): New.
8444 (gt_pch_nx <T>): New.
8445 * passes.c (opt_pass::gt_ggc_mx): New.
8446 (opt_pass::gt_pch_nx): New.
8447 (opt_pass::gt_pch_nx_with_op): New.
8448 (pass_manager::gt_ggc_mx): New.
8449 (pass_manager::gt_pch_nx): New.
8450 (pass_manager::gt_pch_nx_with_op): New.
8451 (pass_manager::operator new): Use
8452 ggc_internal_cleared_alloc_stat rather than xcalloc.
8453 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8454 (pass_manager::gt_ggc_mx): New.
8455 (pass_manager::gt_pch_nx): New.
8456 (pass_manager::gt_pch_nx_with_op): New.
8457 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8458 (opt_pass::operator new): New.
8459 (opt_pass::gt_ggc_mx): New.
8460 (opt_pass::gt_pch_nx): New.
8461 (opt_pass::gt_pch_nx_with_op): New.
8464 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8466 * Makefile.in (GTFILES): Add context.h.
8467 * context.c (gcc::context::operator new): New.
8468 (gcc::context::gt_ggc_mx): New.
8469 (gcc::context::gt_pch_nx): New.
8470 (gcc::context::gt_pch_nx): New.
8471 * context.h (gcc::context): Add GTY((user)) marking.
8472 (gcc::context::operator new): New.
8473 (gcc::context::gt_ggc_mx): New.
8474 (gcc::context::gt_pch_nx): New.
8475 (gcc::context::gt_pch_nx): New.
8476 (g): Add GTY marking.
8477 (gt_ggc_mx (gcc::context *)): New.
8478 (gt_pch_nx (gcc::context *)): New.
8479 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8480 void *cookie)): New.
8481 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8483 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8485 * dce.c (fini_dce): Call df_analyze again just in case
8486 delete_unmarked_insns removed anything.
8488 2013-08-20 Teresa Johnson <tejohnson@google.com>
8490 PR rtl-optimizations/57451
8491 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8492 from crossing split section boundaries.
8494 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8496 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8497 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8498 with MULTLIB_DEFAULTS.
8500 2013-08-20 Nick Clifton <nickc@redhat.com>
8502 * target.def (narrow_volatile_bitfield): Note that the default
8503 value is false, not !TARGET_STRICT_ALIGN.
8504 * doc/tm.texi: Regenerate.
8506 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8508 Fix LIB_SPEC for systems without libpthread.
8510 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8511 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8513 * config/i386/linux-common.h: Likewise.
8514 * config/mips/linux-common.h: Likewise.
8516 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8518 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8521 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8523 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8524 instances can own GC refs.
8526 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8527 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8528 (gcc::context::gt_pch_nx): Likewise.
8529 (gcc::context::gt_pch_nx): Likewise.
8530 * ggc.h (gt_ggc_mx <T>): New.
8531 (gt_pch_nx_with_op <T>): New.
8532 (gt_pch_nx <T>): New.
8533 * passes.c (opt_pass::gt_ggc_mx): New.
8534 (opt_pass::gt_pch_nx): New.
8535 (opt_pass::gt_pch_nx_with_op): New.
8536 (pass_manager::gt_ggc_mx): New.
8537 (pass_manager::gt_pch_nx): New.
8538 (pass_manager::gt_pch_nx_with_op): New.
8539 (pass_manager::operator new): Use
8540 ggc_internal_cleared_alloc_stat rather than xcalloc.
8541 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8542 (pass_manager::gt_ggc_mx): New.
8543 (pass_manager::gt_pch_nx): New.
8544 (pass_manager::gt_pch_nx_with_op): New.
8545 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8546 (opt_pass::operator new): New.
8547 (opt_pass::gt_ggc_mx): New.
8548 (opt_pass::gt_pch_nx): New.
8549 (opt_pass::gt_pch_nx_with_op): New.
8551 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8553 * Makefile.in (GTFILES): Add context.h.
8554 * context.c (gcc::context::operator new): New.
8555 (gcc::context::gt_ggc_mx): New.
8556 (gcc::context::gt_pch_nx): New.
8557 (gcc::context::gt_pch_nx): New.
8558 * context.h (gcc::context): Add GTY((user)) marking.
8559 (gcc::context::operator new): New.
8560 (gcc::context::gt_ggc_mx): New.
8561 (gcc::context::gt_pch_nx): New.
8562 (gcc::context::gt_pch_nx): New.
8563 (g): Add GTY marking.
8564 (gt_ggc_mx (gcc::context *)): New.
8565 (gt_pch_nx (gcc::context *)): New.
8566 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8567 void *cookie)): New.
8568 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8570 2013-08-20 Alan Modra <amodra@gmail.com>
8573 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8574 (rs6000_emit_epilogue): Likewise.
8576 2013-08-19 Dehao Chen <dehao@google.com>
8578 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8580 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8581 Jakub Jelinek <jakub@redhat.com>
8583 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8584 (BUILT_IN_FABSD64): Likewise.
8585 (BUILT_IN_FABSD128): Likewise.
8586 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8587 (fold_builtin_1): Likewise.
8588 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8589 destination and source operands.
8590 (*abstd2_fpr): Likewise.
8591 (*nabstd2_fpr): Likewise.
8593 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8595 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8598 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8600 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8603 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8605 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8606 * builtins.c (default_libc_has_function): New.
8607 (gnu_libc_has_function): Ditto.
8608 (no_c99_libc_has_function): Ditto.
8609 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8610 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8611 (fold_builtin_sincos): Likewise.
8612 (fold_builtin_cexp): Likewise.
8613 * builtins.def (DEF_C94_BUILTIN): Likewise.
8614 (DEF_C99_BUILTIN): Likewise.
8615 (DEF_C99_C90RES_BUILTIN): Likewise.
8616 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8617 definitions to using this define.
8618 * config/darwin-protos.h (darwin_libc_has_function): New.
8619 * config/darwin.c (darwin_libc_has_function): Ditto.
8620 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8621 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8622 * config/darwin.h: Ditto.
8623 * config/elfos.h: Ditto.
8624 * config/freebsd.h: Ditto.
8625 * config/i386/cygming.h: Ditto.
8626 * config/i386/djgpp.h: Ditto.
8627 * config/i386/i386-interix.h: Ditto.
8628 * config/microblaze/microblaze.h: Ditto.
8629 * config/mmix/mmix.h: Ditto.
8630 * config/gnu-user.h: Ditto.
8631 * config/ia64/hpux.h: Ditto.
8632 * config/pa/pa-hpux.h: Ditto.
8633 * config/pdp11/pdp11.h: Ditto.
8634 * config/picochip/picochip.h: Ditto.
8635 * config/linux.h: Ditto.
8636 * config/netbsd.h: Ditto.
8637 * config/openbsd.h: Ditto.
8638 * config/rs6000/aix43.h: Ditto.
8639 * config/rs6000/aix51.h: Ditto.
8640 * config/rs6000/aix52.h: Ditto.
8641 * config/rs6000/aix53.h: Ditto.
8642 * config/rs6000/aix61.h: Ditto.
8643 * config/rs6000/darwin.h: Ditto.
8644 * config/rs6000/linux.h: Ditto.
8645 * config/rs6000/linux64.h: Ditto.
8646 * config/s390/tpf.h: Ditto.
8647 * config/sol2-10.h: Ditto.
8648 * config/sol2.h: Ditto.
8649 * config/vms/vms.h: Ditto.
8650 * config/vxworks.h: Ditto.
8651 * config/linux-android.c (linux_android_libc_has_function):
8652 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8653 * config/linux-protos.h (linux_android_libc_has_function):
8655 * config/i386/i386.c (ix86_libc_has_function): New.
8656 * config/i386/i386-protos.h
8657 (ix86_libc_has_function): New declaration.
8658 * config/i386/i386.md
8659 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8660 ("isinf<mode>2): Likewise.
8661 * convert.c (convert_to_integer): Using new target hook
8662 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8663 TARGET_C99_FUNCTIONS.
8664 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8665 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8666 * coretypes.h (function_class): New enum for different
8667 classes of functions.
8668 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8669 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8670 (TARGET_HAS_SINCOS): Likewise.
8671 (TARGET_LIBC_HAS_FUNCTION): New.
8672 * doc/tm.texi: Regenerated.
8673 * targhooks.h (default_libc_has_function): New declaration.
8674 (no_c99_libc_has_function): Ditto.
8675 (gnu_libc_has_function): Ditto.
8676 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8677 and TARGET_HAS_SINCOS.
8679 2013-08-18 Jan Hubicka <jh@suse.cz>
8681 * Makeifle-in (ipa-devirt.o): New.
8682 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8683 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8684 (analyze_functions): Look into possible targets of polymorphic call.
8685 * dumpfile.c (dump_files): Add type-inheritance dump.
8686 * dumpfile.h (TDI_inheritance): New.
8687 * ipa-devirt.c: New file.
8688 * ipa-utils.h (odr_type_d): Forward declare.
8689 (odr_type): New type.
8690 (build_type_inheritance_graph): Declare.
8691 (possible_polymorphic_call_targets): Declare and introduce inline
8692 variant when only edge is pased.
8693 (dump_possible_polymorphic_call_targets): Likewise.
8694 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8695 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8696 (types_same_for_odr): ... here.
8697 * tree.h (type_in_anonymous_namespace_p): Declare.
8699 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8701 PR tree-optimization/58006
8702 * tree-parloops.c (take_address_of): Don't ICE if get_name
8704 (eliminate_local_variables_stmt): Remove clobber stmts.
8706 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8708 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8709 error message is issued for an alias to undefined symbol.
8711 2013-08-18 Jan Hubicka <jh@suse.cz>
8713 * cgraph.c (cgraph_create_indirect_edge): Discover
8714 polymorphic calls and record basic info into indirect_info.
8715 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8716 devirtualization, ignore objc function calls.
8717 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8718 call with no parm index info.
8719 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8720 * tree.c (virtual_method_call_p): New function.
8721 * tree.h (virtual_method_call_p): Declare.
8723 2013-08-16 Jan Hubicka <jh@suse.cz>
8726 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8728 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8730 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8733 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8735 * gengtype.c (type_for_name): Add special-case support for
8736 locating types within the "gcc::" namespace.
8737 (open_base_files): Emit a "using namespace gcc" directive.
8739 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8742 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8743 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8745 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8746 array instead of each individual operand as a separate argument.
8747 (emit_fusion_gpr_load): Likewise.
8748 (expand_fusion_gpr_load): Add new function declaration.
8750 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8751 signature to have the operands passed as an array, instead of as
8752 separate arguments. Allow ZERO_EXTEND to be in the memory
8753 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8754 depend on the register live/dead flags when peepholes are run.
8755 (expand_fusion_gpr_load): New function to be called from the
8756 peephole2 pass, to change the register that addis sets to be the
8758 (emit_fusion_gpr_load): Change the calling signature to have the
8759 operands passed as an array, instead of as separate arguments.
8760 Allow ZERO_EXTEND to be in the memory address, and also
8761 SIGN_EXTEND if -mpower8-fusion-sign.
8763 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8765 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8766 adjust the register addis loads up in the peephole2 pass. Do not
8767 depend on the register live/dead state when the peephole pass is done.
8769 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8771 * gengtype.c (create_user_defined_type): Ensure that the kind
8772 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8773 declaration is seen before the GTY((user)) marking.
8775 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8778 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8780 2013-08-16 Jan Hubicka <jh@suse.cz>
8782 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8783 arugment expected_type.
8784 (gimple_fold_call): Use it.
8785 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8786 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8787 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8788 (try_make_edge_direct_virtual_call): Likewise.
8789 * tree.c (obj_type_ref_class): New.
8790 * tree.h (obj_type_ref_class): Use it.
8792 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8794 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8795 (rtl_slim_pp): Likewise.
8796 (init_rtl_slim_pretty_print): Likewise.
8797 (dump_value_slim): Don't call it. Use local pretty printer.
8798 (dump_insn_slim): Likewise.
8799 (dump_rtl_slim): Likewise.
8800 (str_pattern_slim): Likewise.
8801 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8804 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8806 PR tree-optimization/58164
8807 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8808 walk gimple_goto_dest of GIMPLE_GOTO.
8810 PR tree-optimization/58165
8811 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8812 bi_call must be the last stmt in a bb, don't split_block, instead
8813 use fallthru edge from it and give up if there is none.
8814 Release conds vector when returning early.
8816 2013-08-14 Xinliang David Li <davidxl@google.com>
8818 * config/i386/i386.c (ix86_option_override_internal):
8819 Remove unused variable and field.
8821 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8824 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8825 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8826 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8827 and Linux, correct BLKmode alignment when 128-bit alignment is
8828 required and compatibility flag is not set.
8829 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8830 for zero-size arguments when compatibility flag is not set.
8832 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8834 PR tree-optimization/58145
8835 * tree-sra.c (build_ref_for_offset): If prev_base has
8836 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8838 2013-08-14 Xinliang David Li <davidxl@google.com>
8840 * config/i386/i386.c (ix86_option_override_internal):
8841 Fix uninitialized variable error.
8843 2013-08-14 Xinliang David Li <davidxl@google.com>
8845 * config/i386/i386.opt: Define two new options.
8846 * config/i386/x86-tune.def: Add arch selector field in macros.
8847 * config/i386/i386.h: Adjust macro definition.
8848 * config/i386/i386.c (ix86_option_override_internal):
8850 (parse_mtune_ctrl_str): New function.
8851 (set_ix86_tune_features): New function.
8852 (ix86_function_specific_restore): Call the new helper function.
8854 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8856 PR rtl-optimization/57662
8857 * sel-sched.c (code_motion_process_successors): When the current insn
8858 is removed after the recursive traversal, break from the loop.
8859 Add comments and debug printouts.
8861 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8862 Alexandre Oliva <aoliva@redhat.com>
8865 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8866 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8867 there also UNSPEC_PLTOFF.
8869 2013-08-14 Marek Polacek <polacek@redhat.com>
8871 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8872 NUM_CONDITIONS bit positions.
8874 2013-08-13 Cary Coutant <ccoutant@google.com>
8876 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8877 (attr_checksum): Hash vector contents instead of pointer.
8878 (attr_checksum_ordered): Likewise.
8880 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8882 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8883 when Pmode != word_mode. Add length_address attribute.
8884 (sse3_monitor_<mode>): Merge from sse3_monitor and
8885 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8886 Pmode != word_mode. Update insn length attribute.
8887 * config/i386/i386.c (ix86_option_override_internal): Update
8888 ix86_gen_monitor selection for merged sse3_monitor insn.
8890 2013-08-13 Julian Brown <julian@codesourcery.com>
8892 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8893 perform invalid legitimization on greater-than-word-size modes for
8896 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8898 * ira.c (setup_class_translate_array): Use aclass instead of cl
8899 for classes not fully covered by allocno classes.
8901 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8903 PR tree-optimization/57661
8904 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8905 * tree-inline.c (tree_function_versioning): Initialize it.
8906 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8907 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8908 that is not being copied.
8911 * asan.c (instrument_strlen_call): Fix typo in comment.
8912 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8914 2013-08-13 Steve Ellcey <sellcey@mips.com>
8916 * config/mips/mips.md (prefetch): Use lw instead of ld on
8917 loongson in 32bit mode.
8919 2013-08-13 Nick Clifton <nickc@redhat.com>
8921 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8923 2013-08-13 Jan Hubicka <jh@suse.cz>
8925 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8926 introduced edge; fix typo in sanity check.
8927 (cgraph_resolve_speculation): Export; improve diagnostic.
8928 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8929 speculation at type mismatch.
8930 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8931 (cgraph_resolve_speculation): Declare.
8932 (symtab_can_be_discarded): New function.
8933 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8934 * ipa-inline-transform.c (speculation_removed): New global var.
8935 (clone_inlined_nodes): See if speculation can be removed.
8936 (inline_call): If speculations was removed, we growths may not match.
8937 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8938 (speculation_useful_p): New function.
8939 (resolve_noninline_speculation): New function.
8940 (inline_small_functions): Resolve useless speculations.
8941 * ipa-inline.h (speculation_useful_p): Declare
8942 * ipa.c (can_replace_by_local_alias): Simplify.
8943 (ipa_profile): Produce speculative calls in non-lto, too;
8944 add simple cost model; produce local aliases.
8946 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8948 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8951 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8953 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8955 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8957 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8958 iterator instead of X87MODEF.
8960 2013-08-12 Perez Read <netfirewall@gmail.com>
8963 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8964 operand 0 for intel asm alternative.
8965 (*movabs<mode>_2): Ditto for operand 1.
8967 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8969 * config/aarch64/arm_none.h
8970 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8972 2013-08-12 Nick Clifton <nickc@redhat.com>
8974 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8976 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8978 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8979 expand for QI/HImode operand to produce more effictive code for
8980 unsigned char(short) --> float(double) conversion.
8982 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8984 * doc/invoke.texi: Mention that -ftls-model does not force the final
8987 2013-08-12 Marek Polacek <polacek@redhat.com>
8988 Marc Glisse <marc.glisse@inria.fr>
8990 PR tree-optimization/57980
8991 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8992 when creating -1 constant.
8994 2013-08-10 Jan Hubicka <jh@suse.cz>
8996 Workaround binutils PR14342.
8997 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8998 (gimple_init_edge_profiler): Likewise.
8999 (gimple_gen_ic_func_profiler): Likewise.
9001 2013-08-09 Jan Hubicka <jh@suse.cz>
9003 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9005 2013-08-09 Xinliang David Li <davidxl@google.com>
9007 * config/i386/stringop.def: New file.
9008 * config/i386/stringop.opt: New file.
9009 * config/i386/i386-opts.h: Include stringopt.def.
9010 * config/i386/i386.opt: Include stringopt.opt.
9011 * config/i386/i386.c (ix86_option_override_internal):
9012 Override default size based stringop inline strategies with options.
9013 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9016 2013-08-09 Jan Hubicka <jh@suse.cz>
9018 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9020 2013-08-09 Jan Hubicka <jh@suse.cz>
9022 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9024 (dump_cgraph_node): Dump profile-id.
9025 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9026 and common_target_probability.
9027 * lto-cgraph.c (lto_output_edge): Stream common targets.
9028 (lto_output_node): Stream profile ids.
9029 (input_node): Stream profile ids.
9030 (input_edge): Stream common targets.
9031 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9032 * ipa.c: Include value-prof.h
9033 (ipa_profile_generate_summary): Turn indirect call statement histograms
9034 into common targets.
9035 (ipa_profile): Turn common targets into speculative edges.
9037 2013-08-09 Jan Hubicka <jh@suse.cz>
9039 * cgraph.h (cgraph_node): Add profile_id.
9040 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9041 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9042 (del_node_map): Update.
9043 (find_func_by_funcdef_no): Replace by ...
9044 (find_func_by_profile_id): ... this one.
9045 (gimple_ic_transform): Do not remove useful histograms when
9046 speculation is not done; dump info when indirect call removal
9048 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9049 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9050 (__gcov_indirect_call_profiler_v2): .. this one.
9051 * profile.h (init_node_map): Update.
9052 * coverage.c (coverage_compute_profile_id): New function.
9053 * coverage.h (coverage_compute_profile_id): Declare.
9054 * tree-profile.c (init_ic_make_global_vars): Make
9055 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9056 (gimple_init_edge_profiler): Update prototype of
9057 __gcov_indirect_call_profiler.
9058 (gimple_gen_ic_func_profiler): Simplify.
9059 (tree_profiling): Use init_node_map
9061 2013-08-09 Jan Hubicka <jh@suse.cz>
9063 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9064 non-speculative refs.
9065 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9066 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9067 (cgraph_set_call_stmt): Likewise.
9068 (cgraph_create_edge_1): Fix release checking compilatoin;
9070 (cgraph_free_edge): Free indirect info.
9071 (cgraph_turn_edge_to_speculative): New function.
9072 (cgraph_speculative_call_info): New function.
9073 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9074 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9075 (dump_cgraph_node): Dump speculation.
9076 (verify_edge_count_and_frequency): Accept speculative edges.
9077 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9078 (verify_cgraph_node): Handle speculation.
9079 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9080 (cgraph_set_call_stmt): Update prototype.
9081 (cgraph_make_edge_direct): Update prototype.
9082 (cgraph_speculative_call_info): Declare.
9083 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9084 to change; update call of ipa_find_references.
9085 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9086 lto_stmt_uid and speculative flags.
9087 (ipa_dump_references): Dump speculation.
9088 (ipa_clone_references): Clone speculative flag.
9089 (ipa_clone_referring): Likewise.
9090 (ipa_clone_ref): New function.
9091 (ipa_find_reference): Look into lto_stmt_uids
9092 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9093 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9094 (ipa_find_reference): Update declaration.
9095 (ipa_clone_ref): Declare.
9096 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9097 stream speculative flag.
9098 (lto_output_ref): Stream statements uids and speculation.
9099 (input_ref): Likewise.
9100 (input_edge): Stream speuclation.
9101 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9102 (cgraph_set_call_stmt_including_clones): Handle speculation.
9103 * ipa-inline.c (heap_edge_removal_hook): New function.
9104 (inline_small_functions): Register it.
9105 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9106 also initialize refs.
9107 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9109 (try_make_edge_direct_simple_call): Likewise.
9110 (try_make_edge_direct_simple_call): Likewise.
9111 (update_indirect_edges_after_inlining): Likewise.
9112 (remove_described_reference): Look proper lto_stmt_uid.
9113 (propagate_controlled_uses): Likewise.
9114 (propagate_controlled_uses): Liekwise.
9115 * tree-inline.c (copy_bb): Copy speculative edges.
9116 (redirect_all_calls): New function.
9117 (copy_cfg_body): Do redirection after loop info is updated.
9118 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9120 2013-08-09 Jan Hubicka <jh@suse.cz>
9122 * lto-streamer-out.c (output_function): Renumber PHIs.
9123 * lto-streamer-in.c (input_function): Likewise.
9125 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9127 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9128 (get_lane_unsigned): Likewise.
9129 (dup_lane_scalar): Likewise.
9130 (get_lane): enable for VALL.
9131 * config/aarch64/aarch64-simd.md
9132 (aarch64_dup_lane_scalar<mode>): Remove.
9133 (aarch64_get_lane_signed<mode>): Likewise.
9134 (aarch64_get_lane_unsigned<mode>): Likewise.
9135 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9136 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9137 (aarch64_get_lane<mode>): Enable for all vector modes.
9138 (aarch64_get_lanedi): Remove misleading constraints.
9139 * config/aarch64/arm_neon.h
9140 (__aarch64_vget_lane_any): Define.
9141 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9142 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9143 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9144 * config/aarch64/iterators.md (VDQQH): New.
9148 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9150 * configure.ac: Add GAS check for LEON instructions on SPARC.
9151 * configure: Regenerate.
9152 * config.in: Likewise.
9153 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9155 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9156 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9158 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9159 (AS_LEON_FLAG): New macro.
9160 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9161 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9162 Deal with LEON and LEON3 for the memory model.
9163 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9164 (atomic_compare_and_swap<mode>_1): Likewise.
9165 (*atomic_compare_and_swap<mode>_1): Likewise.
9167 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9169 * config/arm/neon.md (vcond): Fix floating-point vector
9170 comparisons against 0.
9172 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9174 * lra-constraints.c (emit_spill_move): Remove assert.
9175 (process_alt_operands): Add more debugging
9176 output. Increase reject for spilling into memory. Decrease
9177 reject for reloading scratch.
9178 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9180 2013-08-08 Steve Ellcey <sellcey@mips.com>
9182 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9183 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9184 micromips incompatible. Add nan2008.
9185 (MULTILIB_DIRNAMES): Add nan2008.
9186 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9187 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9188 and micromips incompatible. Add nan2008.
9189 (MULTILIB_DIRNAMES): Add nan2008.
9190 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9192 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9194 PR rtl-optimization/58079
9195 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9196 simplify_comparison has widened a comparison with an integer.
9198 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9200 * config/arm/neon.md (movmisalign<mode>): Disable when we
9201 don't allow unaligned accesses.
9202 (*movmisalign<mode>_neon_store): Likewise.
9203 (*movmisalign<mode>_neon_load): Likewise.
9204 (*movmisalign<mode>_neon_store): Likewise.
9205 (*movmisalign<mode>_neon_load): Likewise.
9207 2013-08-08 Jan Hubicka <jh@suse.cz>
9209 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9210 (make_pass_rebuild_cgraph_edges): Also clear references.
9211 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9212 * ipa-inline-transform.c (inline_transform): Remove all references
9214 * cgraphunit.c (expand_function): Remove all references after
9216 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9217 (ipa_find_reference): Rewrite to iterator.
9218 (remove_stmt_references): Likewise.
9219 (ipa_clear_stmts_in_references): New function.
9220 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9221 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9223 * ipa-split.c (split_function): Remove references in split function.
9225 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9228 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9229 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9231 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9234 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9235 suggested mode for the assignment isn't compatible with the
9238 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9241 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9243 2013-08-07 Xinliang David Li <davidxl@google.com>
9245 * config/i386/i386.opt: New option -mtune-ctrl=.
9246 * config/i386/x86-tune.def: New file.
9247 * config/i386/i386.h: include x86-tune.def.
9248 * config/i386/i386.c (ix86_option_override_internal):
9249 Parsing -mtune-ctrl= option and set tune features.
9251 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9254 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9255 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9258 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9260 * rtl.h (update_alignments): Declare.
9261 * final.c (grow_label_align): New function extracted from...
9262 (shorten_branches): ...here. Call it.
9263 (update_alignments): New function.
9264 * reorg.c (sibling_labels): New variable.
9265 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9266 the new label along with it onto the sibling_labels vector.
9267 (fill_simple_delay_slots): Adjust call to get_label_before.
9268 (fill_slots_from_thread): Likewise.
9269 (relax_delay_slots): Likewise.
9270 (make_return_insns): Likewise.
9271 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9273 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9275 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9276 document its semantics.
9277 (diagnostic_report_diagnostic): Adjust accordingly.
9279 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9281 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9282 (sparc_option_override): ...and port to new C++ pass API.
9283 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9285 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9287 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9289 2013-08-06 Caroline Tice <cmtice@google.com>
9291 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9292 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9293 * tree-pass.h: Add pass_vtable_verify.
9294 * varasm.c (assemble_variable): Add code to properly set the comdat
9295 section and name for the .vtable_map_vars section.
9296 (assemble_vtyv_preinit_initializer): New function.
9297 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9299 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9300 * vtable-verify.c: New file.
9301 * vtable-verify.h: New file.
9302 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9303 initialiation levels.
9304 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9305 * passes.def: Insert pass_vtable_verify.
9306 * aclocal.m4: Reorder includes.
9307 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9308 -fvtv-counts options.
9309 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9310 as appropriate, if -fvtable-verify=... is used.
9311 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9312 -fvtable-verify=... is used.
9313 * Makefile.in (OBJS): Add vtable-verify.o to list.
9314 (vtable-verify.o): Add new build rule.
9315 (GTFILES): Add vtable-verify.c to list.
9316 * common.opt (fvtable-verify=): New flag.
9317 (vtv_priority): Values for fvtable-verify= flag.
9318 (fvtv-counts): New flag.
9319 (fvtv-debug): New flag.
9320 * tree.h (save_vtable_map_decl): New extern function decl.
9322 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9324 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9325 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9326 (pass_data_rl78_devirt): ...new pass_data instance and...
9327 (make_pass_rl78_devirt): ...new function.
9328 (rl78_asm_file_start): Port pass registration to new C++ API.
9330 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9332 * coretypes.h (rtl_opt_pass): Add.
9333 (gcc::context): Add.
9334 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9335 (epiphany_init): Port to new C++ pass API.
9336 (epiphany_optimize_mode_switching): Likewise.
9337 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9338 (pass_manager::get_pass_mode_switching): New.
9339 (pass_manager::get_pass_peephole2): New.
9340 * mode-switching.c (pass_mode_switching): Add clone method.
9341 * recog.c (pass_peephole2): Add clone method.
9342 (pass_split_all_insns): Add clone method.
9344 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9346 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9347 (mips_option_override): ...here, porting to new C++ API for passes.
9349 2013-08-06 Jan Hubicka <jh@suse.cz>
9351 * cgraph.c (cgraph_get_body): New function based on lto.c
9353 * cgraph.h (cgraph_get_body): Declare.
9354 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9356 * cgraphunit.c (expand_function): Get body prior expanding.
9357 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9358 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9360 * passes.c (do_per_function_toporder): Get body.
9361 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9362 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9364 2013-08-06 Martin Jambor <mjambor@suse.cz>
9367 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9368 is not re-finalized. Rename second parameter to no_collect.
9370 2013-08-06 Martin Jambor <mjambor@suse.cz>
9373 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9374 MEM_REF has proper alignment information.
9376 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9379 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9381 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9382 rtx (*) (rtx, ...) with insn_gen_fn.
9383 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9384 insn_gen_fn::stored_funcptr. Add initializer braces.
9386 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9388 Rewrite how instances of passes are cloned to remove assumptions
9389 about their sizes (thus allowing pass subclasses to have
9390 additional data fields, albeit non-GC-managed ones at this point).
9392 * passes.c (make_pass_instance): Now that passes have clone
9393 methods, rewrite this function to eliminate XNEW and memcpy
9394 calls that used hardcoded sizes. Since this function no longer
9395 creates pass instances, rename it to...
9396 (add_pass_instance): ...this. Document the old way that passes were
9397 numbered and flagged, and rework this function to continue using it.
9398 (next_pass_1): Add an initial_pass argument for use by
9400 (position_pass): When adding multiple instances of a pass, use
9401 the pass's clone method, rather than relying on the XNEW/memcpy
9402 within the former make_pass_instance (now add_pass_instance).
9403 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9404 the initial instance of the current pass within the pass manager.
9406 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9408 This is the automated part of the conversion of passes from C
9409 structs to C++ classes.
9411 Patch autogenerated by refactor_passes.py from
9412 https://github.com/davidmalcolm/gcc-refactoring-scripts
9413 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9415 * asan.c (pass_asan): Convert from a global struct to a subclass of
9416 gimple_opt_pass along with...
9417 (pass_data_asan): ...new pass_data instance and...
9418 (make_pass_asan): ...new function.
9419 (pass_asan_O0): Convert from a global struct to a subclass of
9420 gimple_opt_pass along with...
9421 (pass_data_asan_O0): ...new pass_data instance and...
9422 (make_pass_asan_O0): ...new function.
9423 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9424 subclass of rtl_opt_pass along with...
9425 (pass_data_inc_dec): ...new pass_data instance and...
9426 (make_pass_inc_dec): ...new function.
9427 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9428 a subclass of rtl_opt_pass along with...
9429 (pass_data_reorder_blocks): ...new pass_data instance and...
9430 (make_pass_reorder_blocks): ...new function.
9431 (pass_duplicate_computed_gotos): Convert from a global struct to a
9432 subclass of rtl_opt_pass along with...
9433 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9434 (make_pass_duplicate_computed_gotos): ...new function.
9435 (pass_partition_blocks): Convert from a global struct to a subclass of
9436 rtl_opt_pass along with...
9437 (pass_data_partition_blocks): ...new pass_data instance and...
9438 (make_pass_partition_blocks): ...new function.
9439 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9440 struct to a subclass of rtl_opt_pass along with...
9441 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9443 (make_pass_branch_target_load_optimize1): ...new function.
9444 (pass_branch_target_load_optimize2): Convert from a global struct to a
9445 subclass of rtl_opt_pass along with...
9446 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9448 (make_pass_branch_target_load_optimize2): ...new function.
9449 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9450 of rtl_opt_pass along with...
9451 (pass_data_jump): ...new pass_data instance and...
9452 (make_pass_jump): ...new function.
9453 (pass_jump2): Convert from a global struct to a subclass of
9454 rtl_opt_pass along with...
9455 (pass_data_jump2): ...new pass_data instance and...
9456 (make_pass_jump2): ...new function.
9457 * cfgexpand.c (pass_expand): Convert from a global struct to a
9458 subclass of rtl_opt_pass along with...
9459 (pass_data_expand): ...new pass_data instance and...
9460 (make_pass_expand): ...new function.
9461 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9462 of rtl_opt_pass along with...
9463 (pass_data_free_cfg): ...new pass_data instance and...
9464 (make_pass_free_cfg): ...new function.
9465 (pass_into_cfg_layout_mode): Convert from a global struct to a
9466 subclass of rtl_opt_pass along with...
9467 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9468 (make_pass_into_cfg_layout_mode): ...new function.
9469 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9470 subclass of rtl_opt_pass along with...
9471 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9472 (make_pass_outof_cfg_layout_mode): ...new function.
9473 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9474 struct to a subclass of gimple_opt_pass along with...
9475 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9476 (make_pass_build_cgraph_edges): ...new function.
9477 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9478 subclass of gimple_opt_pass along with...
9479 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9480 (make_pass_rebuild_cgraph_edges): ...new function.
9481 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9482 subclass of gimple_opt_pass along with...
9483 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9485 (make_pass_remove_cgraph_callee_edges): ...new function.
9486 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9487 struct to a subclass of rtl_opt_pass along with...
9488 (pass_data_stack_adjustments): ...new pass_data instance and...
9489 (make_pass_stack_adjustments): ...new function.
9490 * combine.c (pass_combine): Convert from a global struct to a subclass
9491 of rtl_opt_pass along with...
9492 (pass_data_combine): ...new pass_data instance and...
9493 (make_pass_combine): ...new function.
9494 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9495 global struct to a subclass of rtl_opt_pass along with...
9496 (pass_data_compare_elim_after_reload): ...new pass_data instance
9498 (make_pass_compare_elim_after_reload): ...new function.
9499 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9500 of rtl_opt_pass along with...
9501 (pass_data_rtl_cprop): ...new pass_data instance and...
9502 (make_pass_rtl_cprop): ...new function.
9503 * cse.c (pass_cse): Convert from a global struct to a subclass of
9504 rtl_opt_pass along with...
9505 (pass_data_cse): ...new pass_data instance and...
9506 (make_pass_cse): ...new function.
9507 (pass_cse2): Convert from a global struct to a subclass of
9508 rtl_opt_pass along with...
9509 (pass_data_cse2): ...new pass_data instance and...
9510 (make_pass_cse2): ...new function.
9511 (pass_cse_after_global_opts): Convert from a global struct to a
9512 subclass of rtl_opt_pass along with...
9513 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9514 (make_pass_cse_after_global_opts): ...new function.
9515 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9516 of rtl_opt_pass along with...
9517 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9518 (make_pass_ud_rtl_dce): ...new function.
9519 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9520 rtl_opt_pass along with...
9521 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9522 (make_pass_fast_rtl_dce): ...new function.
9523 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9524 a subclass of rtl_opt_pass along with...
9525 (pass_data_df_initialize_opt): ...new pass_data instance and...
9526 (make_pass_df_initialize_opt): ...new function.
9527 (pass_df_initialize_no_opt): Convert from a global struct to a
9528 subclass of rtl_opt_pass along with...
9529 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9530 (make_pass_df_initialize_no_opt): ...new function.
9531 (pass_df_finish): Convert from a global struct to a subclass of
9532 rtl_opt_pass along with...
9533 (pass_data_df_finish): ...new pass_data instance and...
9534 (make_pass_df_finish): ...new function.
9535 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9536 rtl_opt_pass along with...
9537 (pass_data_rtl_dse1): ...new pass_data instance and...
9538 (make_pass_rtl_dse1): ...new function.
9539 (pass_rtl_dse2): Convert from a global struct to a subclass of
9540 rtl_opt_pass along with...
9541 (pass_data_rtl_dse2): ...new pass_data instance and...
9542 (make_pass_rtl_dse2): ...new function.
9543 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9544 subclass of rtl_opt_pass along with...
9545 (pass_data_dwarf2_frame): ...new pass_data instance and...
9546 (make_pass_dwarf2_frame): ...new function.
9547 * except.c (pass_set_nothrow_function_flags): Convert from a global
9548 struct to a subclass of rtl_opt_pass along with...
9549 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9551 (make_pass_set_nothrow_function_flags): ...new function.
9552 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9553 subclass of rtl_opt_pass along with...
9554 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9556 (make_pass_convert_to_eh_region_ranges): ...new function.
9557 * final.c (pass_compute_alignments): Convert from a global struct to a
9558 subclass of rtl_opt_pass along with...
9559 (pass_data_compute_alignments): ...new pass_data instance and...
9560 (make_pass_compute_alignments): ...new function.
9561 (pass_final): Convert from a global struct to a subclass of
9562 rtl_opt_pass along with...
9563 (pass_data_final): ...new pass_data instance and...
9564 (make_pass_final): ...new function.
9565 (pass_shorten_branches): Convert from a global struct to a subclass of
9566 rtl_opt_pass along with...
9567 (pass_data_shorten_branches): ...new pass_data instance and...
9568 (make_pass_shorten_branches): ...new function.
9569 (pass_clean_state): Convert from a global struct to a subclass of
9570 rtl_opt_pass along with...
9571 (pass_data_clean_state): ...new pass_data instance and...
9572 (make_pass_clean_state): ...new function.
9573 * function.c (pass_instantiate_virtual_regs): Convert from a global
9574 struct to a subclass of rtl_opt_pass along with...
9575 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9576 (make_pass_instantiate_virtual_regs): ...new function.
9577 (pass_leaf_regs): Convert from a global struct to a subclass of
9578 rtl_opt_pass along with...
9579 (pass_data_leaf_regs): ...new pass_data instance and...
9580 (make_pass_leaf_regs): ...new function.
9581 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9582 subclass of rtl_opt_pass along with...
9583 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9585 (make_pass_thread_prologue_and_epilogue): ...new function.
9586 (pass_match_asm_constraints): Convert from a global struct to a
9587 subclass of rtl_opt_pass along with...
9588 (pass_data_match_asm_constraints): ...new pass_data instance and...
9589 (make_pass_match_asm_constraints): ...new function.
9590 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9591 subclass of rtl_opt_pass along with...
9592 (pass_data_rtl_fwprop): ...new pass_data instance and...
9593 (make_pass_rtl_fwprop): ...new function.
9594 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9595 rtl_opt_pass along with...
9596 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9597 (make_pass_rtl_fwprop_addr): ...new function.
9598 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9599 rtl_opt_pass along with...
9600 (pass_data_rtl_pre): ...new pass_data instance and...
9601 (make_pass_rtl_pre): ...new function.
9602 (pass_rtl_hoist): Convert from a global struct to a subclass of
9603 rtl_opt_pass along with...
9604 (pass_data_rtl_hoist): ...new pass_data instance and...
9605 (make_pass_rtl_hoist): ...new function.
9606 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9607 subclass of gimple_opt_pass along with...
9608 (pass_data_lower_cf): ...new pass_data instance and...
9609 (make_pass_lower_cf): ...new function.
9610 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9611 from a global struct to a subclass of gimple_opt_pass along with...
9612 (pass_data_strength_reduction): ...new pass_data instance and...
9613 (make_pass_strength_reduction): ...new function.
9614 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9615 of rtl_opt_pass along with...
9616 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9617 (make_pass_rtl_ifcvt): ...new function.
9618 (pass_if_after_combine): Convert from a global struct to a subclass of
9619 rtl_opt_pass along with...
9620 (pass_data_if_after_combine): ...new pass_data instance and...
9621 (make_pass_if_after_combine): ...new function.
9622 (pass_if_after_reload): Convert from a global struct to a subclass of
9623 rtl_opt_pass along with...
9624 (pass_data_if_after_reload): ...new pass_data instance and...
9625 (make_pass_if_after_reload): ...new function.
9626 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9627 a subclass of rtl_opt_pass along with...
9628 (pass_data_initialize_regs): ...new pass_data instance and...
9629 (make_pass_initialize_regs): ...new function.
9630 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9631 of ipa_opt_pass_d along with...
9632 (pass_data_ipa_cp): ...new pass_data instance and...
9633 (make_pass_ipa_cp): ...new function.
9634 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9635 global struct to a subclass of gimple_opt_pass along with...
9636 (pass_data_inline_parameters): ...new pass_data instance and...
9637 (make_pass_inline_parameters): ...new function.
9638 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9639 subclass of gimple_opt_pass along with...
9640 (pass_data_early_inline): ...new pass_data instance and...
9641 (make_pass_early_inline): ...new function.
9642 (pass_ipa_inline): Convert from a global struct to a subclass of
9643 ipa_opt_pass_d along with...
9644 (pass_data_ipa_inline): ...new pass_data instance and...
9645 (make_pass_ipa_inline): ...new function.
9646 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9647 struct to a subclass of gimple_opt_pass along with...
9648 (pass_data_local_pure_const): ...new pass_data instance and...
9649 (make_pass_local_pure_const): ...new function.
9650 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9651 ipa_opt_pass_d along with...
9652 (pass_data_ipa_pure_const): ...new pass_data instance and...
9653 (make_pass_ipa_pure_const): ...new function.
9654 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9655 to a subclass of ipa_opt_pass_d along with...
9656 (pass_data_ipa_reference): ...new pass_data instance and...
9657 (make_pass_ipa_reference): ...new function.
9658 * ipa-split.c (pass_split_functions): Convert from a global struct to
9659 a subclass of gimple_opt_pass along with...
9660 (pass_data_split_functions): ...new pass_data instance and...
9661 (make_pass_split_functions): ...new function.
9662 (pass_feedback_split_functions): Convert from a global struct to a
9663 subclass of gimple_opt_pass along with...
9664 (pass_data_feedback_split_functions): ...new pass_data instance and...
9665 (make_pass_feedback_split_functions): ...new function.
9666 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9667 global struct to a subclass of simple_ipa_opt_pass along with...
9668 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9670 (make_pass_ipa_function_and_variable_visibility): ...new function.
9671 (pass_ipa_free_inline_summary): Convert from a global struct to a
9672 subclass of simple_ipa_opt_pass along with...
9673 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9674 (make_pass_ipa_free_inline_summary): ...new function.
9675 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9676 subclass of ipa_opt_pass_d along with...
9677 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9679 (make_pass_ipa_whole_program_visibility): ...new function.
9680 (pass_ipa_profile): Convert from a global struct to a subclass of
9681 ipa_opt_pass_d along with...
9682 (pass_data_ipa_profile): ...new pass_data instance and...
9683 (make_pass_ipa_profile): ...new function.
9684 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9685 ipa_opt_pass_d along with...
9686 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9687 (make_pass_ipa_cdtor_merge): ...new function.
9688 * ira.c (pass_ira): Convert from a global struct to a subclass of
9689 rtl_opt_pass along with...
9690 (pass_data_ira): ...new pass_data instance and...
9691 (make_pass_ira): ...new function.
9692 (pass_reload): Convert from a global struct to a subclass of
9693 rtl_opt_pass along with...
9694 (pass_data_reload): ...new pass_data instance and...
9695 (make_pass_reload): ...new function.
9696 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9697 subclass of rtl_opt_pass along with...
9698 (pass_data_cleanup_barriers): ...new pass_data instance and...
9699 (make_pass_cleanup_barriers): ...new function.
9700 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9701 of rtl_opt_pass along with...
9702 (pass_data_loop2): ...new pass_data instance and...
9703 (make_pass_loop2): ...new function.
9704 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9705 rtl_opt_pass along with...
9706 (pass_data_rtl_loop_init): ...new pass_data instance and...
9707 (make_pass_rtl_loop_init): ...new function.
9708 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9709 rtl_opt_pass along with...
9710 (pass_data_rtl_loop_done): ...new pass_data instance and...
9711 (make_pass_rtl_loop_done): ...new function.
9712 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9713 subclass of rtl_opt_pass along with...
9714 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9715 (make_pass_rtl_move_loop_invariants): ...new function.
9716 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9717 rtl_opt_pass along with...
9718 (pass_data_rtl_unswitch): ...new pass_data instance and...
9719 (make_pass_rtl_unswitch): ...new function.
9720 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9721 subclass of rtl_opt_pass along with...
9722 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9724 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9725 (pass_rtl_doloop): Convert from a global struct to a subclass of
9726 rtl_opt_pass along with...
9727 (pass_data_rtl_doloop): ...new pass_data instance and...
9728 (make_pass_rtl_doloop): ...new function.
9729 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9730 a subclass of rtl_opt_pass along with...
9731 (pass_data_lower_subreg): ...new pass_data instance and...
9732 (make_pass_lower_subreg): ...new function.
9733 (pass_lower_subreg2): Convert from a global struct to a subclass of
9734 rtl_opt_pass along with...
9735 (pass_data_lower_subreg2): ...new pass_data instance and...
9736 (make_pass_lower_subreg2): ...new function.
9737 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9738 struct to a subclass of ipa_opt_pass_d along with...
9739 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9740 (make_pass_ipa_lto_gimple_out): ...new function.
9741 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9742 of ipa_opt_pass_d along with...
9743 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9744 (make_pass_ipa_lto_finish_out): ...new function.
9745 * mode-switching.c (pass_mode_switching): Convert from a global struct
9746 to a subclass of rtl_opt_pass along with...
9747 (pass_data_mode_switching): ...new pass_data instance and...
9748 (make_pass_mode_switching): ...new function.
9749 * modulo-sched.c (pass_sms): Convert from a global struct to a
9750 subclass of rtl_opt_pass along with...
9751 (pass_data_sms): ...new pass_data instance and...
9752 (make_pass_sms): ...new function.
9753 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9754 subclass of gimple_opt_pass along with...
9755 (pass_data_expand_omp): ...new pass_data instance and...
9756 (make_pass_expand_omp): ...new function.
9757 (pass_lower_omp): Convert from a global struct to a subclass of
9758 gimple_opt_pass along with...
9759 (pass_data_lower_omp): ...new pass_data instance and...
9760 (make_pass_lower_omp): ...new function.
9761 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9762 of gimple_opt_pass along with...
9763 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9764 (make_pass_diagnose_omp_blocks): ...new function.
9765 * passes.c (pass_early_local_passes): Convert from a global struct to
9766 a subclass of simple_ipa_opt_pass along with...
9767 (pass_data_early_local_passes): ...new pass_data instance and...
9768 (make_pass_early_local_passes): ...new function.
9769 (pass_all_early_optimizations): Convert from a global struct to a
9770 subclass of gimple_opt_pass along with...
9771 (pass_data_all_early_optimizations): ...new pass_data instance and...
9772 (make_pass_all_early_optimizations): ...new function.
9773 (pass_all_optimizations): Convert from a global struct to a subclass
9774 of gimple_opt_pass along with...
9775 (pass_data_all_optimizations): ...new pass_data instance and...
9776 (make_pass_all_optimizations): ...new function.
9777 (pass_all_optimizations_g): Convert from a global struct to a subclass
9778 of gimple_opt_pass along with...
9779 (pass_data_all_optimizations_g): ...new pass_data instance and...
9780 (make_pass_all_optimizations_g): ...new function.
9781 (pass_rest_of_compilation): Convert from a global struct to a subclass
9782 of rtl_opt_pass along with...
9783 (pass_data_rest_of_compilation): ...new pass_data instance and...
9784 (make_pass_rest_of_compilation): ...new function.
9785 (pass_postreload): Convert from a global struct to a subclass of
9786 rtl_opt_pass along with...
9787 (pass_data_postreload): ...new pass_data instance and...
9788 (make_pass_postreload): ...new function.
9789 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9790 subclass of rtl_opt_pass along with...
9791 (pass_data_gcse2): ...new pass_data instance and...
9792 (make_pass_gcse2): ...new function.
9793 * postreload.c (pass_postreload_cse): Convert from a global struct to
9794 a subclass of rtl_opt_pass along with...
9795 (pass_data_postreload_cse): ...new pass_data instance and...
9796 (make_pass_postreload_cse): ...new function.
9797 * predict.c (pass_profile): Convert from a global struct to a subclass
9798 of gimple_opt_pass along with...
9799 (pass_data_profile): ...new pass_data instance and...
9800 (make_pass_profile): ...new function.
9801 (pass_strip_predict_hints): Convert from a global struct to a subclass
9802 of gimple_opt_pass along with...
9803 (pass_data_strip_predict_hints): ...new pass_data instance and...
9804 (make_pass_strip_predict_hints): ...new function.
9805 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9806 of rtl_opt_pass along with...
9807 (pass_data_peephole2): ...new pass_data instance and...
9808 (make_pass_peephole2): ...new function.
9809 (pass_split_all_insns): Convert from a global struct to a subclass of
9810 rtl_opt_pass along with...
9811 (pass_data_split_all_insns): ...new pass_data instance and...
9812 (make_pass_split_all_insns): ...new function.
9813 (pass_split_after_reload): Convert from a global struct to a subclass
9814 of rtl_opt_pass along with...
9815 (pass_data_split_after_reload): ...new pass_data instance and...
9816 (make_pass_split_after_reload): ...new function.
9817 (pass_split_before_regstack): Convert from a global struct to a
9818 subclass of rtl_opt_pass along with...
9819 (pass_data_split_before_regstack): ...new pass_data instance and...
9820 (make_pass_split_before_regstack): ...new function.
9821 (pass_split_before_sched2): Convert from a global struct to a subclass
9822 of rtl_opt_pass along with...
9823 (pass_data_split_before_sched2): ...new pass_data instance and...
9824 (make_pass_split_before_sched2): ...new function.
9825 (pass_split_for_shorten_branches): Convert from a global struct to a
9826 subclass of rtl_opt_pass along with...
9827 (pass_data_split_for_shorten_branches): ...new pass_data instance
9829 (make_pass_split_for_shorten_branches): ...new function.
9830 * ree.c (pass_ree): Convert from a global struct to a subclass of
9831 rtl_opt_pass along with...
9832 (pass_data_ree): ...new pass_data instance and...
9833 (make_pass_ree): ...new function.
9834 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9835 subclass of rtl_opt_pass along with...
9836 (pass_data_stack_regs): ...new pass_data instance and...
9837 (make_pass_stack_regs): ...new function.
9838 (pass_stack_regs_run): Convert from a global struct to a subclass of
9839 rtl_opt_pass along with...
9840 (pass_data_stack_regs_run): ...new pass_data instance and...
9841 (make_pass_stack_regs_run): ...new function.
9842 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9843 subclass of rtl_opt_pass along with...
9844 (pass_data_cprop_hardreg): ...new pass_data instance and...
9845 (make_pass_cprop_hardreg): ...new function.
9846 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9847 subclass of rtl_opt_pass along with...
9848 (pass_data_reginfo_init): ...new pass_data instance and...
9849 (make_pass_reginfo_init): ...new function.
9850 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9851 of rtl_opt_pass along with...
9852 (pass_data_regmove): ...new pass_data instance and...
9853 (make_pass_regmove): ...new function.
9854 * regrename.c (pass_regrename): Convert from a global struct to a
9855 subclass of rtl_opt_pass along with...
9856 (pass_data_regrename): ...new pass_data instance and...
9857 (make_pass_regrename): ...new function.
9858 * reorg.c (pass_delay_slots): Convert from a global struct to a
9859 subclass of rtl_opt_pass along with...
9860 (pass_data_delay_slots): ...new pass_data instance and...
9861 (make_pass_delay_slots): ...new function.
9862 (pass_machine_reorg): Convert from a global struct to a subclass of
9863 rtl_opt_pass along with...
9864 (pass_data_machine_reorg): ...new pass_data instance and...
9865 (make_pass_machine_reorg): ...new function.
9866 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9867 of rtl_opt_pass along with...
9868 (pass_data_sched): ...new pass_data instance and...
9869 (make_pass_sched): ...new function.
9870 (pass_sched2): Convert from a global struct to a subclass of
9871 rtl_opt_pass along with...
9872 (pass_data_sched2): ...new pass_data instance and...
9873 (make_pass_sched2): ...new function.
9874 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9875 to a subclass of rtl_opt_pass along with...
9876 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9877 (make_pass_stack_ptr_mod): ...new function.
9878 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9879 to a subclass of rtl_opt_pass along with...
9880 (pass_data_rtl_store_motion): ...new pass_data instance and...
9881 (make_pass_rtl_store_motion): ...new function.
9882 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9883 of gimple_opt_pass along with...
9884 (pass_data_tracer): ...new pass_data instance and...
9885 (make_pass_tracer): ...new function.
9886 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9887 to a subclass of gimple_opt_pass along with...
9888 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9889 (make_pass_diagnose_tm_blocks): ...new function.
9890 (pass_lower_tm): Convert from a global struct to a subclass of
9891 gimple_opt_pass along with...
9892 (pass_data_lower_tm): ...new pass_data instance and...
9893 (make_pass_lower_tm): ...new function.
9894 (pass_tm_init): Convert from a global struct to a subclass of
9895 gimple_opt_pass along with...
9896 (pass_data_tm_init): ...new pass_data instance and...
9897 (make_pass_tm_init): ...new function.
9898 (pass_tm_mark): Convert from a global struct to a subclass of
9899 gimple_opt_pass along with...
9900 (pass_data_tm_mark): ...new pass_data instance and...
9901 (make_pass_tm_mark): ...new function.
9902 (pass_tm_edges): Convert from a global struct to a subclass of
9903 gimple_opt_pass along with...
9904 (pass_data_tm_edges): ...new pass_data instance and...
9905 (make_pass_tm_edges): ...new function.
9906 (pass_tm_memopt): Convert from a global struct to a subclass of
9907 gimple_opt_pass along with...
9908 (pass_data_tm_memopt): ...new pass_data instance and...
9909 (make_pass_tm_memopt): ...new function.
9910 (pass_ipa_tm): Convert from a global struct to a subclass of
9911 simple_ipa_opt_pass along with...
9912 (pass_data_ipa_tm): ...new pass_data instance and...
9913 (make_pass_ipa_tm): ...new function.
9914 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9915 subclass of gimple_opt_pass along with...
9916 (pass_data_call_cdce): ...new pass_data instance and...
9917 (make_pass_call_cdce): ...new function.
9918 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9919 subclass of gimple_opt_pass along with...
9920 (pass_data_build_cfg): ...new pass_data instance and...
9921 (make_pass_build_cfg): ...new function.
9922 (pass_split_crit_edges): Convert from a global struct to a subclass of
9923 gimple_opt_pass along with...
9924 (pass_data_split_crit_edges): ...new pass_data instance and...
9925 (make_pass_split_crit_edges): ...new function.
9926 (pass_warn_function_return): Convert from a global struct to a
9927 subclass of gimple_opt_pass along with...
9928 (pass_data_warn_function_return): ...new pass_data instance and...
9929 (make_pass_warn_function_return): ...new function.
9930 (pass_warn_function_noreturn): Convert from a global struct to a
9931 subclass of gimple_opt_pass along with...
9932 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9933 (make_pass_warn_function_noreturn): ...new function.
9934 (pass_warn_unused_result): Convert from a global struct to a subclass
9935 of gimple_opt_pass along with...
9936 (pass_data_warn_unused_result): ...new pass_data instance and...
9937 (make_pass_warn_unused_result): ...new function.
9938 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9939 a subclass of gimple_opt_pass along with...
9940 (pass_data_merge_phi): ...new pass_data instance and...
9941 (make_pass_merge_phi): ...new function.
9942 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9943 a subclass of gimple_opt_pass along with...
9944 (pass_data_lower_complex): ...new pass_data instance and...
9945 (make_pass_lower_complex): ...new function.
9946 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9947 gimple_opt_pass along with...
9948 (pass_data_lower_complex_O0): ...new pass_data instance and...
9949 (make_pass_lower_complex_O0): ...new function.
9950 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9951 subclass of gimple_opt_pass along with...
9952 (pass_data_lower_eh): ...new pass_data instance and...
9953 (make_pass_lower_eh): ...new function.
9954 (pass_refactor_eh): Convert from a global struct to a subclass of
9955 gimple_opt_pass along with...
9956 (pass_data_refactor_eh): ...new pass_data instance and...
9957 (make_pass_refactor_eh): ...new function.
9958 (pass_lower_resx): Convert from a global struct to a subclass of
9959 gimple_opt_pass along with...
9960 (pass_data_lower_resx): ...new pass_data instance and...
9961 (make_pass_lower_resx): ...new function.
9962 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9963 of gimple_opt_pass along with...
9964 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9965 (make_pass_lower_eh_dispatch): ...new function.
9966 (pass_cleanup_eh): Convert from a global struct to a subclass of
9967 gimple_opt_pass along with...
9968 (pass_data_cleanup_eh): ...new pass_data instance and...
9969 (make_pass_cleanup_eh): ...new function.
9970 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9971 to a subclass of simple_ipa_opt_pass along with...
9972 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9973 (make_pass_ipa_lower_emutls): ...new function.
9974 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9975 a subclass of gimple_opt_pass along with...
9976 (pass_data_if_conversion): ...new pass_data instance and...
9977 (make_pass_if_conversion): ...new function.
9978 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9979 subclass of gimple_opt_pass along with...
9980 (pass_data_build_ssa): ...new pass_data instance and...
9981 (make_pass_build_ssa): ...new function.
9982 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9983 global struct to a subclass of gimple_opt_pass along with...
9984 (pass_data_loop_distribution): ...new pass_data instance and...
9985 (make_pass_loop_distribution): ...new function.
9986 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9987 subclass of gimple_opt_pass along with...
9988 (pass_data_mudflap_1): ...new pass_data instance and...
9989 (make_pass_mudflap_1): ...new function.
9990 (pass_mudflap_2): Convert from a global struct to a subclass of
9991 gimple_opt_pass along with...
9992 (pass_data_mudflap_2): ...new pass_data instance and...
9993 (make_pass_mudflap_2): ...new function.
9994 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9995 subclass of gimple_opt_pass along with...
9996 (pass_data_mudflap_1): ...new pass_data instance and...
9997 (make_pass_mudflap_1): ...new function.
9998 (pass_mudflap_2): Convert from a global struct to a subclass of
9999 gimple_opt_pass along with...
10000 (pass_data_mudflap_2): ...new pass_data instance and...
10001 (make_pass_mudflap_2): ...new function.
10002 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10003 gimple_opt_pass along with...
10004 (pass_data_nrv): ...new pass_data instance and...
10005 (make_pass_nrv): ...new function.
10006 (pass_return_slot): Convert from a global struct to a subclass of
10007 gimple_opt_pass along with...
10008 (pass_data_return_slot): ...new pass_data instance and...
10009 (make_pass_return_slot): ...new function.
10010 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10011 to a subclass of gimple_opt_pass along with...
10012 (pass_data_object_sizes): ...new pass_data instance and...
10013 (make_pass_object_sizes): ...new function.
10014 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10015 global struct to a subclass of gimple_opt_pass along with...
10016 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10018 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10019 (pass_fixup_cfg): Convert from a global struct to a subclass of
10020 gimple_opt_pass along with...
10021 (pass_data_fixup_cfg): ...new pass_data instance and...
10022 (make_pass_fixup_cfg): ...new function.
10023 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10024 (make_pass_mudflap_1): ...new function.
10025 (pass_mudflap_2): Replace declaration with that of...
10026 (make_pass_mudflap_2): ...new function.
10027 (pass_asan): Replace declaration with that of...
10028 (make_pass_asan): ...new function.
10029 (pass_asan_O0): Replace declaration with that of...
10030 (make_pass_asan_O0): ...new function.
10031 (pass_tsan): Replace declaration with that of...
10032 (make_pass_tsan): ...new function.
10033 (pass_tsan_O0): Replace declaration with that of...
10034 (make_pass_tsan_O0): ...new function.
10035 (pass_lower_cf): Replace declaration with that of...
10036 (make_pass_lower_cf): ...new function.
10037 (pass_refactor_eh): Replace declaration with that of...
10038 (make_pass_refactor_eh): ...new function.
10039 (pass_lower_eh): Replace declaration with that of...
10040 (make_pass_lower_eh): ...new function.
10041 (pass_lower_eh_dispatch): Replace declaration with that of...
10042 (make_pass_lower_eh_dispatch): ...new function.
10043 (pass_lower_resx): Replace declaration with that of...
10044 (make_pass_lower_resx): ...new function.
10045 (pass_build_cfg): Replace declaration with that of...
10046 (make_pass_build_cfg): ...new function.
10047 (pass_early_tree_profile): Replace declaration with that of...
10048 (make_pass_early_tree_profile): ...new function.
10049 (pass_cleanup_eh): Replace declaration with that of...
10050 (make_pass_cleanup_eh): ...new function.
10051 (pass_sra): Replace declaration with that of...
10052 (make_pass_sra): ...new function.
10053 (pass_sra_early): Replace declaration with that of...
10054 (make_pass_sra_early): ...new function.
10055 (pass_early_ipa_sra): Replace declaration with that of...
10056 (make_pass_early_ipa_sra): ...new function.
10057 (pass_tail_recursion): Replace declaration with that of...
10058 (make_pass_tail_recursion): ...new function.
10059 (pass_tail_calls): Replace declaration with that of...
10060 (make_pass_tail_calls): ...new function.
10061 (pass_tree_loop): Replace declaration with that of...
10062 (make_pass_tree_loop): ...new function.
10063 (pass_tree_loop_init): Replace declaration with that of...
10064 (make_pass_tree_loop_init): ...new function.
10065 (pass_lim): Replace declaration with that of...
10066 (make_pass_lim): ...new function.
10067 (pass_tree_unswitch): Replace declaration with that of...
10068 (make_pass_tree_unswitch): ...new function.
10069 (pass_predcom): Replace declaration with that of...
10070 (make_pass_predcom): ...new function.
10071 (pass_iv_canon): Replace declaration with that of...
10072 (make_pass_iv_canon): ...new function.
10073 (pass_scev_cprop): Replace declaration with that of...
10074 (make_pass_scev_cprop): ...new function.
10075 (pass_empty_loop): Replace declaration with that of...
10076 (make_pass_empty_loop): ...new function.
10077 (pass_record_bounds): Replace declaration with that of...
10078 (make_pass_record_bounds): ...new function.
10079 (pass_graphite): Replace declaration with that of...
10080 (make_pass_graphite): ...new function.
10081 (pass_graphite_transforms): Replace declaration with that of...
10082 (make_pass_graphite_transforms): ...new function.
10083 (pass_if_conversion): Replace declaration with that of...
10084 (make_pass_if_conversion): ...new function.
10085 (pass_loop_distribution): Replace declaration with that of...
10086 (make_pass_loop_distribution): ...new function.
10087 (pass_vectorize): Replace declaration with that of...
10088 (make_pass_vectorize): ...new function.
10089 (pass_slp_vectorize): Replace declaration with that of...
10090 (make_pass_slp_vectorize): ...new function.
10091 (pass_complete_unroll): Replace declaration with that of...
10092 (make_pass_complete_unroll): ...new function.
10093 (pass_complete_unrolli): Replace declaration with that of...
10094 (make_pass_complete_unrolli): ...new function.
10095 (pass_parallelize_loops): Replace declaration with that of...
10096 (make_pass_parallelize_loops): ...new function.
10097 (pass_loop_prefetch): Replace declaration with that of...
10098 (make_pass_loop_prefetch): ...new function.
10099 (pass_iv_optimize): Replace declaration with that of...
10100 (make_pass_iv_optimize): ...new function.
10101 (pass_tree_loop_done): Replace declaration with that of...
10102 (make_pass_tree_loop_done): ...new function.
10103 (pass_ch): Replace declaration with that of...
10104 (make_pass_ch): ...new function.
10105 (pass_ccp): Replace declaration with that of...
10106 (make_pass_ccp): ...new function.
10107 (pass_phi_only_cprop): Replace declaration with that of...
10108 (make_pass_phi_only_cprop): ...new function.
10109 (pass_build_ssa): Replace declaration with that of...
10110 (make_pass_build_ssa): ...new function.
10111 (pass_build_alias): Replace declaration with that of...
10112 (make_pass_build_alias): ...new function.
10113 (pass_build_ealias): Replace declaration with that of...
10114 (make_pass_build_ealias): ...new function.
10115 (pass_dominator): Replace declaration with that of...
10116 (make_pass_dominator): ...new function.
10117 (pass_dce): Replace declaration with that of...
10118 (make_pass_dce): ...new function.
10119 (pass_dce_loop): Replace declaration with that of...
10120 (make_pass_dce_loop): ...new function.
10121 (pass_cd_dce): Replace declaration with that of...
10122 (make_pass_cd_dce): ...new function.
10123 (pass_call_cdce): Replace declaration with that of...
10124 (make_pass_call_cdce): ...new function.
10125 (pass_merge_phi): Replace declaration with that of...
10126 (make_pass_merge_phi): ...new function.
10127 (pass_split_crit_edges): Replace declaration with that of...
10128 (make_pass_split_crit_edges): ...new function.
10129 (pass_pre): Replace declaration with that of...
10130 (make_pass_pre): ...new function.
10131 (pass_profile): Replace declaration with that of...
10132 (make_pass_profile): ...new function.
10133 (pass_strip_predict_hints): Replace declaration with that of...
10134 (make_pass_strip_predict_hints): ...new function.
10135 (pass_lower_complex_O0): Replace declaration with that of...
10136 (make_pass_lower_complex_O0): ...new function.
10137 (pass_lower_complex): Replace declaration with that of...
10138 (make_pass_lower_complex): ...new function.
10139 (pass_lower_vector): Replace declaration with that of...
10140 (make_pass_lower_vector): ...new function.
10141 (pass_lower_vector_ssa): Replace declaration with that of...
10142 (make_pass_lower_vector_ssa): ...new function.
10143 (pass_lower_omp): Replace declaration with that of...
10144 (make_pass_lower_omp): ...new function.
10145 (pass_diagnose_omp_blocks): Replace declaration with that of...
10146 (make_pass_diagnose_omp_blocks): ...new function.
10147 (pass_expand_omp): Replace declaration with that of...
10148 (make_pass_expand_omp): ...new function.
10149 (pass_expand_omp_ssa): Replace declaration with that of...
10150 (make_pass_expand_omp_ssa): ...new function.
10151 (pass_object_sizes): Replace declaration with that of...
10152 (make_pass_object_sizes): ...new function.
10153 (pass_strlen): Replace declaration with that of...
10154 (make_pass_strlen): ...new function.
10155 (pass_fold_builtins): Replace declaration with that of...
10156 (make_pass_fold_builtins): ...new function.
10157 (pass_stdarg): Replace declaration with that of...
10158 (make_pass_stdarg): ...new function.
10159 (pass_early_warn_uninitialized): Replace declaration with that of...
10160 (make_pass_early_warn_uninitialized): ...new function.
10161 (pass_late_warn_uninitialized): Replace declaration with that of...
10162 (make_pass_late_warn_uninitialized): ...new function.
10163 (pass_cse_reciprocals): Replace declaration with that of...
10164 (make_pass_cse_reciprocals): ...new function.
10165 (pass_cse_sincos): Replace declaration with that of...
10166 (make_pass_cse_sincos): ...new function.
10167 (pass_optimize_bswap): Replace declaration with that of...
10168 (make_pass_optimize_bswap): ...new function.
10169 (pass_optimize_widening_mul): Replace declaration with that of...
10170 (make_pass_optimize_widening_mul): ...new function.
10171 (pass_warn_function_return): Replace declaration with that of...
10172 (make_pass_warn_function_return): ...new function.
10173 (pass_warn_function_noreturn): Replace declaration with that of...
10174 (make_pass_warn_function_noreturn): ...new function.
10175 (pass_cselim): Replace declaration with that of...
10176 (make_pass_cselim): ...new function.
10177 (pass_phiopt): Replace declaration with that of...
10178 (make_pass_phiopt): ...new function.
10179 (pass_forwprop): Replace declaration with that of...
10180 (make_pass_forwprop): ...new function.
10181 (pass_phiprop): Replace declaration with that of...
10182 (make_pass_phiprop): ...new function.
10183 (pass_tree_ifcombine): Replace declaration with that of...
10184 (make_pass_tree_ifcombine): ...new function.
10185 (pass_dse): Replace declaration with that of...
10186 (make_pass_dse): ...new function.
10187 (pass_nrv): Replace declaration with that of...
10188 (make_pass_nrv): ...new function.
10189 (pass_rename_ssa_copies): Replace declaration with that of...
10190 (make_pass_rename_ssa_copies): ...new function.
10191 (pass_sink_code): Replace declaration with that of...
10192 (make_pass_sink_code): ...new function.
10193 (pass_fre): Replace declaration with that of...
10194 (make_pass_fre): ...new function.
10195 (pass_check_data_deps): Replace declaration with that of...
10196 (make_pass_check_data_deps): ...new function.
10197 (pass_copy_prop): Replace declaration with that of...
10198 (make_pass_copy_prop): ...new function.
10199 (pass_vrp): Replace declaration with that of...
10200 (make_pass_vrp): ...new function.
10201 (pass_uncprop): Replace declaration with that of...
10202 (make_pass_uncprop): ...new function.
10203 (pass_return_slot): Replace declaration with that of...
10204 (make_pass_return_slot): ...new function.
10205 (pass_reassoc): Replace declaration with that of...
10206 (make_pass_reassoc): ...new function.
10207 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10208 (make_pass_rebuild_cgraph_edges): ...new function.
10209 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10210 (make_pass_remove_cgraph_callee_edges): ...new function.
10211 (pass_build_cgraph_edges): Replace declaration with that of...
10212 (make_pass_build_cgraph_edges): ...new function.
10213 (pass_local_pure_const): Replace declaration with that of...
10214 (make_pass_local_pure_const): ...new function.
10215 (pass_tracer): Replace declaration with that of...
10216 (make_pass_tracer): ...new function.
10217 (pass_warn_unused_result): Replace declaration with that of...
10218 (make_pass_warn_unused_result): ...new function.
10219 (pass_diagnose_tm_blocks): Replace declaration with that of...
10220 (make_pass_diagnose_tm_blocks): ...new function.
10221 (pass_lower_tm): Replace declaration with that of...
10222 (make_pass_lower_tm): ...new function.
10223 (pass_tm_init): Replace declaration with that of...
10224 (make_pass_tm_init): ...new function.
10225 (pass_tm_mark): Replace declaration with that of...
10226 (make_pass_tm_mark): ...new function.
10227 (pass_tm_memopt): Replace declaration with that of...
10228 (make_pass_tm_memopt): ...new function.
10229 (pass_tm_edges): Replace declaration with that of...
10230 (make_pass_tm_edges): ...new function.
10231 (pass_split_functions): Replace declaration with that of...
10232 (make_pass_split_functions): ...new function.
10233 (pass_feedback_split_functions): Replace declaration with that of...
10234 (make_pass_feedback_split_functions): ...new function.
10235 (pass_strength_reduction): Replace declaration with that of...
10236 (make_pass_strength_reduction): ...new function.
10237 (pass_ipa_lower_emutls): Replace declaration with that of...
10238 (make_pass_ipa_lower_emutls): ...new function.
10239 (pass_ipa_function_and_variable_visibility): Replace declaration with
10241 (make_pass_ipa_function_and_variable_visibility): ...new function.
10242 (pass_ipa_tree_profile): Replace declaration with that of...
10243 (make_pass_ipa_tree_profile): ...new function.
10244 (pass_early_local_passes): Replace declaration with that of...
10245 (make_pass_early_local_passes): ...new function.
10246 (pass_ipa_whole_program_visibility): Replace declaration with that
10248 (make_pass_ipa_whole_program_visibility): ...new function.
10249 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10250 (make_pass_ipa_lto_gimple_out): ...new function.
10251 (pass_ipa_increase_alignment): Replace declaration with that of...
10252 (make_pass_ipa_increase_alignment): ...new function.
10253 (pass_ipa_inline): Replace declaration with that of...
10254 (make_pass_ipa_inline): ...new function.
10255 (pass_ipa_free_lang_data): Replace declaration with that of...
10256 (make_pass_ipa_free_lang_data): ...new function.
10257 (pass_ipa_free_inline_summary): Replace declaration with that of...
10258 (make_pass_ipa_free_inline_summary): ...new function.
10259 (pass_ipa_cp): Replace declaration with that of...
10260 (make_pass_ipa_cp): ...new function.
10261 (pass_ipa_reference): Replace declaration with that of...
10262 (make_pass_ipa_reference): ...new function.
10263 (pass_ipa_pure_const): Replace declaration with that of...
10264 (make_pass_ipa_pure_const): ...new function.
10265 (pass_ipa_pta): Replace declaration with that of...
10266 (make_pass_ipa_pta): ...new function.
10267 (pass_ipa_lto_finish_out): Replace declaration with that of...
10268 (make_pass_ipa_lto_finish_out): ...new function.
10269 (pass_ipa_tm): Replace declaration with that of...
10270 (make_pass_ipa_tm): ...new function.
10271 (pass_ipa_profile): Replace declaration with that of...
10272 (make_pass_ipa_profile): ...new function.
10273 (pass_ipa_cdtor_merge): Replace declaration with that of...
10274 (make_pass_ipa_cdtor_merge): ...new function.
10275 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10277 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10278 (pass_init_datastructures): Replace declaration with that of...
10279 (make_pass_init_datastructures): ...new function.
10280 (pass_fixup_cfg): Replace declaration with that of...
10281 (make_pass_fixup_cfg): ...new function.
10282 (pass_expand): Replace declaration with that of...
10283 (make_pass_expand): ...new function.
10284 (pass_instantiate_virtual_regs): Replace declaration with that of...
10285 (make_pass_instantiate_virtual_regs): ...new function.
10286 (pass_rtl_fwprop): Replace declaration with that of...
10287 (make_pass_rtl_fwprop): ...new function.
10288 (pass_rtl_fwprop_addr): Replace declaration with that of...
10289 (make_pass_rtl_fwprop_addr): ...new function.
10290 (pass_jump): Replace declaration with that of...
10291 (make_pass_jump): ...new function.
10292 (pass_jump2): Replace declaration with that of...
10293 (make_pass_jump2): ...new function.
10294 (pass_lower_subreg): Replace declaration with that of...
10295 (make_pass_lower_subreg): ...new function.
10296 (pass_cse): Replace declaration with that of...
10297 (make_pass_cse): ...new function.
10298 (pass_fast_rtl_dce): Replace declaration with that of...
10299 (make_pass_fast_rtl_dce): ...new function.
10300 (pass_ud_rtl_dce): Replace declaration with that of...
10301 (make_pass_ud_rtl_dce): ...new function.
10302 (pass_rtl_dce): Replace declaration with that of...
10303 (make_pass_rtl_dce): ...new function.
10304 (pass_rtl_dse1): Replace declaration with that of...
10305 (make_pass_rtl_dse1): ...new function.
10306 (pass_rtl_dse2): Replace declaration with that of...
10307 (make_pass_rtl_dse2): ...new function.
10308 (pass_rtl_dse3): Replace declaration with that of...
10309 (make_pass_rtl_dse3): ...new function.
10310 (pass_rtl_cprop): Replace declaration with that of...
10311 (make_pass_rtl_cprop): ...new function.
10312 (pass_rtl_pre): Replace declaration with that of...
10313 (make_pass_rtl_pre): ...new function.
10314 (pass_rtl_hoist): Replace declaration with that of...
10315 (make_pass_rtl_hoist): ...new function.
10316 (pass_rtl_store_motion): Replace declaration with that of...
10317 (make_pass_rtl_store_motion): ...new function.
10318 (pass_cse_after_global_opts): Replace declaration with that of...
10319 (make_pass_cse_after_global_opts): ...new function.
10320 (pass_rtl_ifcvt): Replace declaration with that of...
10321 (make_pass_rtl_ifcvt): ...new function.
10322 (pass_into_cfg_layout_mode): Replace declaration with that of...
10323 (make_pass_into_cfg_layout_mode): ...new function.
10324 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10325 (make_pass_outof_cfg_layout_mode): ...new function.
10326 (pass_loop2): Replace declaration with that of...
10327 (make_pass_loop2): ...new function.
10328 (pass_rtl_loop_init): Replace declaration with that of...
10329 (make_pass_rtl_loop_init): ...new function.
10330 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10331 (make_pass_rtl_move_loop_invariants): ...new function.
10332 (pass_rtl_unswitch): Replace declaration with that of...
10333 (make_pass_rtl_unswitch): ...new function.
10334 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10335 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10336 (pass_rtl_doloop): Replace declaration with that of...
10337 (make_pass_rtl_doloop): ...new function.
10338 (pass_rtl_loop_done): Replace declaration with that of...
10339 (make_pass_rtl_loop_done): ...new function.
10340 (pass_web): Replace declaration with that of...
10341 (make_pass_web): ...new function.
10342 (pass_cse2): Replace declaration with that of...
10343 (make_pass_cse2): ...new function.
10344 (pass_df_initialize_opt): Replace declaration with that of...
10345 (make_pass_df_initialize_opt): ...new function.
10346 (pass_df_initialize_no_opt): Replace declaration with that of...
10347 (make_pass_df_initialize_no_opt): ...new function.
10348 (pass_reginfo_init): Replace declaration with that of...
10349 (make_pass_reginfo_init): ...new function.
10350 (pass_inc_dec): Replace declaration with that of...
10351 (make_pass_inc_dec): ...new function.
10352 (pass_stack_ptr_mod): Replace declaration with that of...
10353 (make_pass_stack_ptr_mod): ...new function.
10354 (pass_initialize_regs): Replace declaration with that of...
10355 (make_pass_initialize_regs): ...new function.
10356 (pass_combine): Replace declaration with that of...
10357 (make_pass_combine): ...new function.
10358 (pass_if_after_combine): Replace declaration with that of...
10359 (make_pass_if_after_combine): ...new function.
10360 (pass_ree): Replace declaration with that of...
10361 (make_pass_ree): ...new function.
10362 (pass_partition_blocks): Replace declaration with that of...
10363 (make_pass_partition_blocks): ...new function.
10364 (pass_match_asm_constraints): Replace declaration with that of...
10365 (make_pass_match_asm_constraints): ...new function.
10366 (pass_regmove): Replace declaration with that of...
10367 (make_pass_regmove): ...new function.
10368 (pass_split_all_insns): Replace declaration with that of...
10369 (make_pass_split_all_insns): ...new function.
10370 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10371 (make_pass_fast_rtl_byte_dce): ...new function.
10372 (pass_lower_subreg2): Replace declaration with that of...
10373 (make_pass_lower_subreg2): ...new function.
10374 (pass_mode_switching): Replace declaration with that of...
10375 (make_pass_mode_switching): ...new function.
10376 (pass_sms): Replace declaration with that of...
10377 (make_pass_sms): ...new function.
10378 (pass_sched): Replace declaration with that of...
10379 (make_pass_sched): ...new function.
10380 (pass_ira): Replace declaration with that of...
10381 (make_pass_ira): ...new function.
10382 (pass_reload): Replace declaration with that of...
10383 (make_pass_reload): ...new function.
10384 (pass_clean_state): Replace declaration with that of...
10385 (make_pass_clean_state): ...new function.
10386 (pass_branch_prob): Replace declaration with that of...
10387 (make_pass_branch_prob): ...new function.
10388 (pass_value_profile_transformations): Replace declaration with that
10390 (make_pass_value_profile_transformations): ...new function.
10391 (pass_postreload_cse): Replace declaration with that of...
10392 (make_pass_postreload_cse): ...new function.
10393 (pass_gcse2): Replace declaration with that of...
10394 (make_pass_gcse2): ...new function.
10395 (pass_split_after_reload): Replace declaration with that of...
10396 (make_pass_split_after_reload): ...new function.
10397 (pass_branch_target_load_optimize1): Replace declaration with that
10399 (make_pass_branch_target_load_optimize1): ...new function.
10400 (pass_thread_prologue_and_epilogue): Replace declaration with that
10402 (make_pass_thread_prologue_and_epilogue): ...new function.
10403 (pass_stack_adjustments): Replace declaration with that of...
10404 (make_pass_stack_adjustments): ...new function.
10405 (pass_peephole2): Replace declaration with that of...
10406 (make_pass_peephole2): ...new function.
10407 (pass_if_after_reload): Replace declaration with that of...
10408 (make_pass_if_after_reload): ...new function.
10409 (pass_regrename): Replace declaration with that of...
10410 (make_pass_regrename): ...new function.
10411 (pass_cprop_hardreg): Replace declaration with that of...
10412 (make_pass_cprop_hardreg): ...new function.
10413 (pass_reorder_blocks): Replace declaration with that of...
10414 (make_pass_reorder_blocks): ...new function.
10415 (pass_branch_target_load_optimize2): Replace declaration with that
10417 (make_pass_branch_target_load_optimize2): ...new function.
10418 (pass_leaf_regs): Replace declaration with that of...
10419 (make_pass_leaf_regs): ...new function.
10420 (pass_split_before_sched2): Replace declaration with that of...
10421 (make_pass_split_before_sched2): ...new function.
10422 (pass_compare_elim_after_reload): Replace declaration with that of...
10423 (make_pass_compare_elim_after_reload): ...new function.
10424 (pass_sched2): Replace declaration with that of...
10425 (make_pass_sched2): ...new function.
10426 (pass_stack_regs): Replace declaration with that of...
10427 (make_pass_stack_regs): ...new function.
10428 (pass_stack_regs_run): Replace declaration with that of...
10429 (make_pass_stack_regs_run): ...new function.
10430 (pass_df_finish): Replace declaration with that of...
10431 (make_pass_df_finish): ...new function.
10432 (pass_compute_alignments): Replace declaration with that of...
10433 (make_pass_compute_alignments): ...new function.
10434 (pass_duplicate_computed_gotos): Replace declaration with that of...
10435 (make_pass_duplicate_computed_gotos): ...new function.
10436 (pass_variable_tracking): Replace declaration with that of...
10437 (make_pass_variable_tracking): ...new function.
10438 (pass_free_cfg): Replace declaration with that of...
10439 (make_pass_free_cfg): ...new function.
10440 (pass_machine_reorg): Replace declaration with that of...
10441 (make_pass_machine_reorg): ...new function.
10442 (pass_cleanup_barriers): Replace declaration with that of...
10443 (make_pass_cleanup_barriers): ...new function.
10444 (pass_delay_slots): Replace declaration with that of...
10445 (make_pass_delay_slots): ...new function.
10446 (pass_split_for_shorten_branches): Replace declaration with that of...
10447 (make_pass_split_for_shorten_branches): ...new function.
10448 (pass_split_before_regstack): Replace declaration with that of...
10449 (make_pass_split_before_regstack): ...new function.
10450 (pass_convert_to_eh_region_ranges): Replace declaration with that
10452 (make_pass_convert_to_eh_region_ranges): ...new function.
10453 (pass_shorten_branches): Replace declaration with that of...
10454 (make_pass_shorten_branches): ...new function.
10455 (pass_set_nothrow_function_flags): Replace declaration with that of...
10456 (make_pass_set_nothrow_function_flags): ...new function.
10457 (pass_dwarf2_frame): Replace declaration with that of...
10458 (make_pass_dwarf2_frame): ...new function.
10459 (pass_final): Replace declaration with that of...
10460 (make_pass_final): ...new function.
10461 (pass_rtl_seqabstr): Replace declaration with that of...
10462 (make_pass_rtl_seqabstr): ...new function.
10463 (pass_release_ssa_names): Replace declaration with that of...
10464 (make_pass_release_ssa_names): ...new function.
10465 (pass_early_inline): Replace declaration with that of...
10466 (make_pass_early_inline): ...new function.
10467 (pass_inline_parameters): Replace declaration with that of...
10468 (make_pass_inline_parameters): ...new function.
10469 (pass_update_address_taken): Replace declaration with that of...
10470 (make_pass_update_address_taken): ...new function.
10471 (pass_convert_switch): Replace declaration with that of...
10472 (make_pass_convert_switch): ...new function.
10473 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10474 to a subclass of simple_ipa_opt_pass along with...
10475 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10476 (make_pass_ipa_tree_profile): ...new function.
10477 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10478 subclass of gimple_opt_pass along with...
10479 (pass_data_sra_early): ...new pass_data instance and...
10480 (make_pass_sra_early): ...new function.
10481 (pass_sra): Convert from a global struct to a subclass of
10482 gimple_opt_pass along with...
10483 (pass_data_sra): ...new pass_data instance and...
10484 (make_pass_sra): ...new function.
10485 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10486 gimple_opt_pass along with...
10487 (pass_data_early_ipa_sra): ...new pass_data instance and...
10488 (make_pass_early_ipa_sra): ...new function.
10489 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10490 subclass of gimple_opt_pass along with...
10491 (pass_data_ccp): ...new pass_data instance and...
10492 (make_pass_ccp): ...new function.
10493 (pass_fold_builtins): Convert from a global struct to a subclass of
10494 gimple_opt_pass along with...
10495 (pass_data_fold_builtins): ...new pass_data instance and...
10496 (make_pass_fold_builtins): ...new function.
10497 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10498 subclass of gimple_opt_pass along with...
10499 (pass_data_copy_prop): ...new pass_data instance and...
10500 (make_pass_copy_prop): ...new function.
10501 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10502 global struct to a subclass of gimple_opt_pass along with...
10503 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10504 (make_pass_rename_ssa_copies): ...new function.
10505 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10506 subclass of gimple_opt_pass along with...
10507 (pass_data_dce): ...new pass_data instance and...
10508 (make_pass_dce): ...new function.
10509 (pass_dce_loop): Convert from a global struct to a subclass of
10510 gimple_opt_pass along with...
10511 (pass_data_dce_loop): ...new pass_data instance and...
10512 (make_pass_dce_loop): ...new function.
10513 (pass_cd_dce): Convert from a global struct to a subclass of
10514 gimple_opt_pass along with...
10515 (pass_data_cd_dce): ...new pass_data instance and...
10516 (make_pass_cd_dce): ...new function.
10517 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10518 subclass of gimple_opt_pass along with...
10519 (pass_data_dominator): ...new pass_data instance and...
10520 (make_pass_dominator): ...new function.
10521 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10522 gimple_opt_pass along with...
10523 (pass_data_phi_only_cprop): ...new pass_data instance and...
10524 (make_pass_phi_only_cprop): ...new function.
10525 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10526 subclass of gimple_opt_pass along with...
10527 (pass_data_dse): ...new pass_data instance and...
10528 (make_pass_dse): ...new function.
10529 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10530 a subclass of gimple_opt_pass along with...
10531 (pass_data_forwprop): ...new pass_data instance and...
10532 (make_pass_forwprop): ...new function.
10533 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10534 struct to a subclass of gimple_opt_pass along with...
10535 (pass_data_tree_ifcombine): ...new pass_data instance and...
10536 (make_pass_tree_ifcombine): ...new function.
10537 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10538 subclass of gimple_opt_pass along with...
10539 (pass_data_ch): ...new pass_data instance and...
10540 (make_pass_ch): ...new function.
10541 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10542 subclass of gimple_opt_pass along with...
10543 (pass_data_tree_loop): ...new pass_data instance and...
10544 (make_pass_tree_loop): ...new function.
10545 (pass_tree_loop_init): Convert from a global struct to a subclass of
10546 gimple_opt_pass along with...
10547 (pass_data_tree_loop_init): ...new pass_data instance and...
10548 (make_pass_tree_loop_init): ...new function.
10549 (pass_lim): Convert from a global struct to a subclass of
10550 gimple_opt_pass along with...
10551 (pass_data_lim): ...new pass_data instance and...
10552 (make_pass_lim): ...new function.
10553 (pass_tree_unswitch): Convert from a global struct to a subclass of
10554 gimple_opt_pass along with...
10555 (pass_data_tree_unswitch): ...new pass_data instance and...
10556 (make_pass_tree_unswitch): ...new function.
10557 (pass_predcom): Convert from a global struct to a subclass of
10558 gimple_opt_pass along with...
10559 (pass_data_predcom): ...new pass_data instance and...
10560 (make_pass_predcom): ...new function.
10561 (pass_vectorize): Convert from a global struct to a subclass of
10562 gimple_opt_pass along with...
10563 (pass_data_vectorize): ...new pass_data instance and...
10564 (make_pass_vectorize): ...new function.
10565 (pass_graphite): Convert from a global struct to a subclass of
10566 gimple_opt_pass along with...
10567 (pass_data_graphite): ...new pass_data instance and...
10568 (make_pass_graphite): ...new function.
10569 (pass_graphite_transforms): Convert from a global struct to a subclass
10570 of gimple_opt_pass along with...
10571 (pass_data_graphite_transforms): ...new pass_data instance and...
10572 (make_pass_graphite_transforms): ...new function.
10573 (pass_check_data_deps): Convert from a global struct to a subclass of
10574 gimple_opt_pass along with...
10575 (pass_data_check_data_deps): ...new pass_data instance and...
10576 (make_pass_check_data_deps): ...new function.
10577 (pass_iv_canon): Convert from a global struct to a subclass of
10578 gimple_opt_pass along with...
10579 (pass_data_iv_canon): ...new pass_data instance and...
10580 (make_pass_iv_canon): ...new function.
10581 (pass_scev_cprop): Convert from a global struct to a subclass of
10582 gimple_opt_pass along with...
10583 (pass_data_scev_cprop): ...new pass_data instance and...
10584 (make_pass_scev_cprop): ...new function.
10585 (pass_record_bounds): Convert from a global struct to a subclass of
10586 gimple_opt_pass along with...
10587 (pass_data_record_bounds): ...new pass_data instance and...
10588 (make_pass_record_bounds): ...new function.
10589 (pass_complete_unroll): Convert from a global struct to a subclass of
10590 gimple_opt_pass along with...
10591 (pass_data_complete_unroll): ...new pass_data instance and...
10592 (make_pass_complete_unroll): ...new function.
10593 (pass_complete_unrolli): Convert from a global struct to a subclass of
10594 gimple_opt_pass along with...
10595 (pass_data_complete_unrolli): ...new pass_data instance and...
10596 (make_pass_complete_unrolli): ...new function.
10597 (pass_parallelize_loops): Convert from a global struct to a subclass
10598 of gimple_opt_pass along with...
10599 (pass_data_parallelize_loops): ...new pass_data instance and...
10600 (make_pass_parallelize_loops): ...new function.
10601 (pass_loop_prefetch): Convert from a global struct to a subclass of
10602 gimple_opt_pass along with...
10603 (pass_data_loop_prefetch): ...new pass_data instance and...
10604 (make_pass_loop_prefetch): ...new function.
10605 (pass_iv_optimize): Convert from a global struct to a subclass of
10606 gimple_opt_pass along with...
10607 (pass_data_iv_optimize): ...new pass_data instance and...
10608 (make_pass_iv_optimize): ...new function.
10609 (pass_tree_loop_done): Convert from a global struct to a subclass of
10610 gimple_opt_pass along with...
10611 (pass_data_tree_loop_done): ...new pass_data instance and...
10612 (make_pass_tree_loop_done): ...new function.
10613 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10614 struct to a subclass of gimple_opt_pass along with...
10615 (pass_data_cse_reciprocals): ...new pass_data instance and...
10616 (make_pass_cse_reciprocals): ...new function.
10617 (pass_cse_sincos): Convert from a global struct to a subclass of
10618 gimple_opt_pass along with...
10619 (pass_data_cse_sincos): ...new pass_data instance and...
10620 (make_pass_cse_sincos): ...new function.
10621 (pass_optimize_bswap): Convert from a global struct to a subclass of
10622 gimple_opt_pass along with...
10623 (pass_data_optimize_bswap): ...new pass_data instance and...
10624 (make_pass_optimize_bswap): ...new function.
10625 (pass_optimize_widening_mul): Convert from a global struct to a
10626 subclass of gimple_opt_pass along with...
10627 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10628 (make_pass_optimize_widening_mul): ...new function.
10629 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10630 subclass of gimple_opt_pass along with...
10631 (pass_data_phiopt): ...new pass_data instance and...
10632 (make_pass_phiopt): ...new function.
10633 (pass_cselim): Convert from a global struct to a subclass of
10634 gimple_opt_pass along with...
10635 (pass_data_cselim): ...new pass_data instance and...
10636 (make_pass_cselim): ...new function.
10637 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10638 subclass of gimple_opt_pass along with...
10639 (pass_data_phiprop): ...new pass_data instance and...
10640 (make_pass_phiprop): ...new function.
10641 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10642 subclass of gimple_opt_pass along with...
10643 (pass_data_pre): ...new pass_data instance and...
10644 (make_pass_pre): ...new function.
10645 (pass_fre): Convert from a global struct to a subclass of
10646 gimple_opt_pass along with...
10647 (pass_data_fre): ...new pass_data instance and...
10648 (make_pass_fre): ...new function.
10649 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10650 subclass of gimple_opt_pass along with...
10651 (pass_data_reassoc): ...new pass_data instance and...
10652 (make_pass_reassoc): ...new function.
10653 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10654 subclass of gimple_opt_pass along with...
10655 (pass_data_sink_code): ...new pass_data instance and...
10656 (make_pass_sink_code): ...new function.
10657 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10658 subclass of gimple_opt_pass along with...
10659 (pass_data_strlen): ...new pass_data instance and...
10660 (make_pass_strlen): ...new function.
10661 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10662 struct to a subclass of gimple_opt_pass along with...
10663 (pass_data_build_alias): ...new pass_data instance and...
10664 (make_pass_build_alias): ...new function.
10665 (pass_build_ealias): Convert from a global struct to a subclass of
10666 gimple_opt_pass along with...
10667 (pass_data_build_ealias): ...new pass_data instance and...
10668 (make_pass_build_ealias): ...new function.
10669 (pass_ipa_pta): Convert from a global struct to a subclass of
10670 simple_ipa_opt_pass along with...
10671 (pass_data_ipa_pta): ...new pass_data instance and...
10672 (make_pass_ipa_pta): ...new function.
10673 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10674 subclass of gimple_opt_pass along with...
10675 (pass_data_uncprop): ...new pass_data instance and...
10676 (make_pass_uncprop): ...new function.
10677 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10678 global struct to a subclass of gimple_opt_pass along with...
10679 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10680 (make_pass_late_warn_uninitialized): ...new function.
10681 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10682 to a subclass of gimple_opt_pass along with...
10683 (pass_data_init_datastructures): ...new pass_data instance and...
10684 (make_pass_init_datastructures): ...new function.
10685 (pass_early_warn_uninitialized): Convert from a global struct to a
10686 subclass of gimple_opt_pass along with...
10687 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10688 (make_pass_early_warn_uninitialized): ...new function.
10689 (pass_update_address_taken): Convert from a global struct to a
10690 subclass of gimple_opt_pass along with...
10691 (pass_data_update_address_taken): ...new pass_data instance and...
10692 (make_pass_update_address_taken): ...new function.
10693 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10694 struct to a subclass of gimple_opt_pass along with...
10695 (pass_data_release_ssa_names): ...new pass_data instance and...
10696 (make_pass_release_ssa_names): ...new function.
10697 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10698 subclass of gimple_opt_pass along with...
10699 (pass_data_stdarg): ...new pass_data instance and...
10700 (make_pass_stdarg): ...new function.
10701 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10702 global struct to a subclass of gimple_opt_pass along with...
10703 (pass_data_convert_switch): ...new pass_data instance and...
10704 (make_pass_convert_switch): ...new function.
10705 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10706 to a subclass of gimple_opt_pass along with...
10707 (pass_data_tail_recursion): ...new pass_data instance and...
10708 (make_pass_tail_recursion): ...new function.
10709 (pass_tail_calls): Convert from a global struct to a subclass of
10710 gimple_opt_pass along with...
10711 (pass_data_tail_calls): ...new pass_data instance and...
10712 (make_pass_tail_calls): ...new function.
10713 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10714 struct to a subclass of gimple_opt_pass along with...
10715 (pass_data_lower_vector): ...new pass_data instance and...
10716 (make_pass_lower_vector): ...new function.
10717 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10718 gimple_opt_pass along with...
10719 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10720 (make_pass_lower_vector_ssa): ...new function.
10721 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10722 to a subclass of gimple_opt_pass along with...
10723 (pass_data_slp_vectorize): ...new pass_data instance and...
10724 (make_pass_slp_vectorize): ...new function.
10725 (pass_ipa_increase_alignment): Convert from a global struct to a
10726 subclass of simple_ipa_opt_pass along with...
10727 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10728 (make_pass_ipa_increase_alignment): ...new function.
10729 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10730 gimple_opt_pass along with...
10731 (pass_data_vrp): ...new pass_data instance and...
10732 (make_pass_vrp): ...new function.
10733 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10734 subclass of simple_ipa_opt_pass along with...
10735 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10736 (make_pass_ipa_free_lang_data): ...new function.
10737 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10738 gimple_opt_pass along with...
10739 (pass_data_tsan): ...new pass_data instance and...
10740 (make_pass_tsan): ...new function.
10741 (pass_tsan_O0): Convert from a global struct to a subclass of
10742 gimple_opt_pass along with...
10743 (pass_data_tsan_O0): ...new pass_data instance and...
10744 (make_pass_tsan_O0): ...new function.
10745 * var-tracking.c (pass_variable_tracking): Convert from a global
10746 struct to a subclass of rtl_opt_pass along with...
10747 (pass_data_variable_tracking): ...new pass_data instance and...
10748 (make_pass_variable_tracking): ...new function.
10749 * web.c (pass_web): Convert from a global struct to a subclass of
10750 rtl_opt_pass along with...
10751 (pass_data_web): ...new pass_data instance and...
10752 (make_pass_web): ...new function.
10753 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10754 declaration with that of...
10755 (make_pass_mode_switch_use): ...new function.
10756 (pass_resolve_sw_modes): Replace declaration with that of...
10757 (make_pass_resolve_sw_modes): ...new function.
10758 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10759 from a global struct to a subclass of rtl_opt_pass along with...
10760 (pass_data_mode_switch_use): ...new pass_data instance and...
10761 (make_pass_mode_switch_use): ...new function.
10762 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10763 from a global struct to a subclass of rtl_opt_pass along with...
10764 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10765 (make_pass_resolve_sw_modes): ...new function.
10766 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10767 struct to a subclass of rtl_opt_pass along with...
10768 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10769 (make_pass_insert_vzeroupper): ...new function.
10770 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10771 global struct to a subclass of rtl_opt_pass along with...
10772 (pass_data_work_around_errata): ...new pass_data instance and...
10773 (make_pass_work_around_errata): ...new function.
10774 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10775 struct to a subclass of rtl_opt_pass along with...
10776 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10777 (make_pass_mips_machine_reorg2): ...new function.
10779 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10781 * passes.c (pass_manager::operator new): New.
10783 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10785 Handwritten part of conversion of passes to C++ classes.
10787 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10788 (toplev.o): Add dep on PASS_MANAGER_H.
10789 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10790 of early local pases to reflect this moving from a global to a
10791 member of gcc::pass_manager.
10792 (cgraph_add_new_function): Likewise.
10793 * lto-cgraph.c (lto_output_node): Update for conversion of
10794 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10795 * passes.c (opt_pass::clone): New.
10796 (opt_pass::gate): New.
10797 (opt_pass::execute): New.
10798 (opt_pass::opt_pass): New.
10799 (pass_manager::execute_early_local_passes): New.
10800 (pass_manager::execute_pass_mode_switching): new.
10801 (finish_optimization_passes): Convert to...
10802 (pass_manager::finish_optimization_passes): ...this.
10803 (finish_optimization_passes): Update for conversion of passes to
10805 (register_dump_files_1): Use has_gate since we cannot portably
10806 check a vtable entry against NULL.
10807 (dump_one_pass): Likewise.
10808 (ipa_write_summaries_2): Likewise.
10809 (ipa_write_optimization_summaries_1): Likewise.
10810 (ipa_read_summaries_1): Likewise.
10811 (ipa_read_optimization_summaries_1): Likewise.
10812 (execute_ipa_stmt_fixups): Likewise.
10813 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10814 pass-creation functions rather than wiring up globals, and
10815 storing the results in fields of pass_manager generated using
10816 pass-instances.def.
10817 (pass_manager::dump_profile_report): Update for conversion of
10818 passes to C++ classes.
10819 (pass_manager::execute_ipa_summary_passes): Likewise.
10820 (execute_one_ipa_transform_pass): Likewise.
10821 (execute_one_pass): Use has_gate and has_execute since we cannot
10822 portably check a vtable entry against NULL.
10823 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10824 (pass_manager): Use pass-instances.def to add fields for the
10825 various pass instances.
10826 * toplev.c (finalize): Update for move of
10827 finish_optimization_passes to a method of gcc::pass_manager.
10828 * toplev.h (finish_optimization_passes): Move to method of class
10830 * tree-pass.h (struct pass_data): New.
10831 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10832 (opt_pass::gate): Convert to virtual function.
10833 (opt_pass::~opt_pass): New.
10834 (opt_pass::clone): New.
10835 (opt_pass::execute): Convert to virtual function.
10836 (opt_pass::opt_pass): New.
10837 (opt_pass::ctxt_): new.
10838 (gimple_opt_pass): Convert to subclass of opt_pass.
10839 (gimple_opt_pass::gimple_opt_pass): New.
10840 (rtl_opt_pass): Convert to subclass of opt_pass.
10841 (rtl_opt_pass::rtl_opt_pass): New.
10842 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10843 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10844 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10845 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10846 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10847 invocation of pass_mode_switching to reflect this moving from a
10848 global to a member of gcc::pass_manager.
10849 (ix86_option_override): Rework how pass_insert_vzeroupper is
10850 added to the pass_manager to reflect autogenerated changes.
10851 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10853 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10855 PR rtl-optimization/57708
10856 * recog.c (peep2_find_free_register): Validate all regs in a
10859 2013-08-05 Jan Hubicka <jh@suse.cz>
10862 * cgraph.c (verify_cgraph_node): Accept local flags from other
10864 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10865 (function_and_variable_visibility): Likewise.
10866 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10868 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10870 * graph.c (init_graph_slim_pretty_print): Remove.
10871 (print_graph_cfg): Do not call it. Use local pretty printer.
10872 (start_graph_dump): Likewise.
10874 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10876 * gimple-pretty-print.c (buffer): Remove.
10877 (initialized): Likewise.
10878 (maybe_init_pretty_print): Likewise.
10879 (print_gimple_stmt): Do not call it. Use non-static local
10880 pretty_printer variable.
10881 (print_gimple_expr): Likewise.
10882 (print_gimple_seq): Likewise.
10883 (gimple_dump_bb): Likewise.
10885 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10887 * asan.c (asan_pp): Remove.
10888 (asan_pp_initialized): Likewise.
10889 (asan_pp_initialize): Likewise.
10890 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10891 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10892 (asan_add_global): Likewise.
10894 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10896 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10897 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10898 * pretty-print.c (pp_formatted_text_data): Likewise.
10899 (pp_indent): Rename from pp_base_indent.
10900 (pp_format): Rename from pp_base_format.
10901 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10902 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10903 (pp_flush): Rename from pp_base_flush.
10904 (pp_set_line_maximum_length): Rename from
10905 pp_base_set_line_maximum_length.
10906 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10907 (pp_set_prefix): Rename from pp_base_set_prefix.
10908 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10909 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10910 (pp_append_text): Rename from pp_base_append_text.
10911 (pp_formatted_text): Rename from pp_base_formatted_text.
10912 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10913 (pp_remaining_character_count_for_line): Rename from
10914 pp_base_remaining_character_count_for_line.
10915 (pp_newline): Rename from pp_base_newline.
10916 (pp_character): Rename from pp_base_character.
10917 (pp_string): Rename from pp_base_string.
10918 (pp_maybe_space): Rename from pp_base_maybe_space.
10919 * asan.c (asan_pp_string): Adjust.
10920 (asan_emit_stack_protection): Likewise.
10921 (asan_add_global): Likewise.
10922 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10923 * tree-mudflap.c (mf_varname_tree): Likewise.
10924 * tree-pretty-print.c (pp_tree_identifier): Rename from
10925 pp_base_tree_identifier.
10926 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10927 Declare as function.
10929 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10931 * pretty-print.h (pp_bar_bar): New.
10932 (pp_ampersand_ampersand): Likewise.
10933 (pp_less_equal): Likewise.
10934 (pp_greater_equal): Likewise.
10935 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10936 printer functions instead of pp_string or operators and punctuators.
10937 (dump_gimple_call): Likewise.
10938 (dump_gimple_omp_for): Likewise.
10939 (dump_gimple_transaction): Likewise.
10940 (dump_gimple_phi): Likewise.
10941 (pp_gimple_stmt_1): Likewise.
10942 * sched-vis.c (print_insn): Likewise.
10943 * tree-mudflap.c (mf_varname_tree): Likewise.
10944 * tree-pretty-print.c (dump_block_node): Likewise.
10945 (dump_generic_node): Likewise.
10947 2013-08-02 Jan Hubicka <jh@suse.cz>
10949 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10951 * lto-streamer-out.c (tree_is_indexable): Results decls and
10952 parm decls are not indexable.
10953 (DFS_write_tree_body): Do not follow args and results.
10954 (hash_tree): Likewise.
10955 (output_functions): Rearrange so struct function is needed
10956 only when real body is output; be able to also ouptut abstract
10957 functions; output DECL_ARGUMENTS and DECL_RESULT.
10958 (lto_output): When not in WPA, ale store abstract functions.
10959 (write_symbol): Do not care about RESULT_DECL.
10960 (output_symbol_p): Handle correctly sbtract decls.
10961 * lto-streamer-in.c (input_function): Rearrange so struct
10962 function can be NULL at entry; allow streaming of
10963 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10964 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10965 sanity check during LTO.
10966 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10967 RESULT_DECl and DECL_ARGUMENTS.
10968 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10971 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10973 * pretty-print.h (pp_underscore): New.
10975 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10976 printer functions instead of pp_character.
10977 (dump_binary_rhs): Likewise.
10978 (dump_ternary_rhs): Likewise.
10979 (dump_gimple_call_args): Likewise.
10980 (pp_points_to_solution): Likewise.
10981 (dump_gimple_call): Likewise.
10982 (dump_gimple_switch): Likewise.
10983 (dump_gimple_cond): Likewise.
10984 (dump_gimple_bind): Likewise.
10985 (dump_gimple_try): Likewise.
10986 (dump_gimple_omp_for): Likewise.
10987 (dump_gimple_omp_continue): Likewise.
10988 (dump_gimple_omp_single): Likewise.
10989 (dump_gimple_omp_sections): Likewise.
10990 (dump_gimple_omp_block): Likewise.
10991 (dump_gimple_omp_critical): Likewise.
10992 (dump_gimple_transaction): Likewise.
10993 (dump_gimple_asm): Likewise.
10994 (dump_gimple_phi): Likewise.
10995 (dump_gimple_omp_parallel): Likewise.
10996 (dump_gimple_omp_task): Likewise.
10997 (dump_gimple_omp_atomic_load): Likewise.
10998 (dump_gimple_omp_atomic_store): Likewise.
10999 (dump_gimple_mem_ops): Likewise.
11000 (pp_gimple_stmt_1): Likewise.
11001 (pp_cfg_jump): Likewise.
11002 (dump_implicit_edges): Likewise.
11003 (gimple_dump_bb_for_graph): Likewise.
11004 * graph.c (draw_cfg_node): Likewise.
11005 * langhooks.c (lhd_print_error_function): Likewise.
11006 * sched-vis.c (print_exp): Likewise.
11007 (print_value): Likewise.
11008 (print_pattern): Likewise.
11009 (print_insn): Likewise.
11010 (rtl_dump_bb_for_graph): Likewise.
11011 * tree-pretty-print.c (dump_function_declaration): Likewise.
11012 (dump_array_domain): Likewise.
11013 (dump_omp_clause): Likewise.
11014 (dump_location): Likewise.
11015 (dump_generic_node): Likewise.
11016 (print_struct_decl): Likewise.
11017 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11019 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11021 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11022 candidate table when replacing a candidate statement.
11023 (replace_rhs_if_not_dup): Likewise.
11024 (replace_one_candidate): Likewise.
11026 2013-08-02 Jan Hubicka <jh@suse.cz>
11027 Martin Liska <marxin.liska@gmail.com>
11029 * cgraphunit.c (add_new_function): Fix logic when adding from
11031 (assemble_thunk): Rename to ...
11032 (expand_thunk); .. this one; export; get it working with
11033 general functions; make produced gimple valid.
11034 * cgraph.h (expand_thunk): Declare.
11036 2013-08-02 Jan Hubicka <jh@suse.cz>
11038 * ipa-cp.c (gather_context_independent_values): Use
11039 ipa_get_param_move_cost.
11040 (get_replacement_map): Remove PARAM; move parameter folding
11042 (create_specialized_node): Update.
11043 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11044 assert that we have gimple body; update move_cost.
11045 (count_formal_params): Assert that we have gimple body.
11046 (ipa_dump_param): New function.
11047 (ipa_alloc_node_params): Break out from ...
11048 (ipa_initialize_node_params): ... here.
11049 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11050 (ipa_write_node_info): Stream move costs.
11051 (ipa_read_node_info): Read move costs.
11052 (ipa_update_after_lto_read): Do not recompute node params.
11053 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11054 (ipa_get_param): Check we are not in WPA.
11055 (ipa_get_param_move_cost): New.
11056 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11057 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11058 parm numbers to be present.
11060 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11062 PR rtl-optimization/58048
11063 * lra-constraints.c (process_alt_operands): Don't check asm
11064 operand on register.
11066 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11068 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11069 the implied StoreLoad barrier for atomic operations if before.
11071 2013-08-02 Jan Hubicka <jh@suse.cz>
11072 Martin Liska <marxin.liska@gmail.com>
11074 * cgraph.c (cgraph_function_body_availability): Do not check
11076 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11077 symtab_node_availability): Declare.
11078 * ipa.c (can_replace_by_local_alias): New.
11079 (function_and_variable_visibility): Use it.
11080 * symtab.c (symtab_for_node_and_aliases,
11081 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11083 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11085 PR rtl-optimization/57963
11086 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11087 (lra_constraints): Use them.
11089 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11091 * config/arm/types.md (define_attr "type"): Add "load_acq"
11093 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11095 (cortex_a53_store1): Likewise.
11097 2013-08-01 Jan Hubicka <jh@suse.cz>
11099 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11100 partitions are not needed.
11102 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11104 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11105 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11106 MAYBE_NON_Q_CLASS_P where appropriate.
11108 2013-08-01 Jan Hubicka <jh@suse.cz>
11110 * cgraph.h (release_function_body): Declare.
11111 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11112 of unused delcarations.
11114 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11116 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11117 RTL form when subtracting a constant.
11119 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11121 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11122 Generate canonical plus rtx with negated immediate instead of minus
11124 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11126 2013-08-01 Jan Hubicka <jh@suse.cz>
11128 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11129 (cgraph_release_function_body): Likewise.
11130 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11131 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11132 to used_as_abstract_origin.
11133 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11134 symbols used as abstract origins.
11135 * cgraphunit.c (analyze_functions): Update.
11136 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11137 used_as_abstract_origin.
11138 * tree-inline.c (tree_function_versioning): Update
11139 used_as_abstract_origin; be ready for DECL_RESULT and
11140 DECL_ARGUMENTS to be NULL.
11142 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11143 for abstract functions.
11144 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11147 2013-08-01 Jan Hubicka <jh@suse.cz>
11149 * profile.c (compute_value_histograms): Fix thinko.
11151 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11153 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11154 aarch-common-protos.h to extra_headers.
11155 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11156 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11157 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11159 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11161 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11162 Include "../arm/types.md". Define "type" attribute for all patterns.
11163 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11166 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11168 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11169 to support power8 load fusion.
11170 (fusion_gpr_mem_load): Likewise.
11172 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11174 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11175 declarations for power8 load fusion.
11176 (emit_fusion_gpr_load): Likewise.
11178 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11179 tuning for power8, turn on fusion mode by default. Turn on sign
11180 extending fusion mode if normal fusion mode is on, and we are at
11182 (fusion_gpr_load_p): New function, return true if we can fuse an
11183 addis instruction with a dependent load to a GPR.
11184 (emit_fusion_gpr_load): Emit the instructions for power8 load
11187 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11188 (VSX load fusion peepholes): New peepholes to fuse together an
11189 addi instruction with a VSX load instruction.
11191 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11192 peepholes to fuse an addis instruction with a load to a GPR base
11193 register. If we are supporting sign extending fusions, convert
11194 sign extending loads to zero extending loads and add an explicit
11197 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11199 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11200 aarch-common-protos.h to extra_headers.
11201 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11202 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11203 (arm_early_store_addr_dep): Likewise.
11204 (arm_no_early_alu_shift_dep): Likewise.
11205 (arm_no_early_alu_shift_value_dep): Likewise.
11206 (arm_no_early_mul_dep): Likewise.
11207 (arm_no_early_store_addr_dep): Likewise.
11208 (arm_mac_accumulator_is_mul_result): Likewise.
11209 (arm_mac_accumulator_is_result): Likewise.
11210 * config/arm/aarch-common.c: ... here. New file.
11211 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11213 (arm_early_store_addr_dep): Likewise.
11214 (arm_no_early_alu_shift_dep): Likewise.
11215 (arm_no_early_alu_shift_value_dep): Likewise.
11216 (arm_no_early_mul_dep): Likewise.
11217 (arm_no_early_store_addr_dep): Likewise.
11218 (arm_mac_accumulator_is_mul_result): Likewise.
11219 (arm_mac_accumulator_is_result): Likewise.
11220 * config/arm/aarch-common-protos.h: ... here. New file.
11221 * config/arm/t-arm (aarch-common.o): Define.
11223 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11225 * config/arm/arm.md: Include new file "types.md".
11226 (define_attr "type"): Move from here to ...
11227 (define_attr "mul32"): Likewise.
11228 (define_attr "mul64"): Likewise.
11229 * config/arm/types.md: ... here. New file.
11231 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11233 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11234 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11236 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11238 * gen-pass-instances.awk: Fix offset of substr().
11240 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11242 * Makefile.in (pass-instances.def): New.
11243 (passes.o): Replace dependency on passes.def with one on
11246 * gen-pass-instances.awk: New.
11248 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11249 rather than passes.def, updating local definition of NEXT_PASS
11250 macro to add an extra NUM parameter (currently unused).
11252 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11254 * Makefile.in (PASS_MANAGER_H): New.
11255 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11256 (passes.o): Likewise.
11257 (statistics.o): Likewise.
11258 (cgraphunit.o): Likewise.
11259 (context.o): Depend on PASS_MANAGER_H.
11261 * pass_manager.h: New.
11263 * cgraphunit.c (cgraph_add_new_function): Update for moves
11264 of globals to fields of pass_manager.
11265 (analyze_function): Likewise.
11266 (expand_function): Likewise.
11267 (ipa_passes): Likewise.
11268 (compile): Likewise.
11270 * context.c (context::context): New.
11271 * context.h (context::context): New.
11272 (context::get_passes): New.
11273 (context::passes_): New.
11275 * lto-cgraph.c (input_node): Update for moves of globals to
11276 fields of pass_manager.
11278 * passes.c (all_passes): Remove, in favor of a field of the
11279 same name within the new class pass_manager.
11280 (all_small_ipa_passes): Likewise.
11281 (all_lowering_passes): Likewise.
11282 (all_regular_ipa_passes): Likewise.
11283 (all_late_ipa_passes): Likewise.
11284 (all_lto_gen_passes): Likewise.
11285 (passes_by_id): Likewise.
11286 (passes_by_id_size): Likewise.
11287 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11288 the new class pass_manager.
11289 (set_pass_for_id): Convert to...
11290 (pass_manager::set_pass_for_id): ...method.
11291 (get_pass_for_id): Convert to...
11292 (pass_manager::get_pass_for_id): ...method.
11293 (register_one_dump_file): Move body of implementation into...
11294 (pass_manager::register_one_dump_file): ...here.
11295 (register_dump_files_1): Convert to...
11296 (pass_manager::register_dump_files_1): ...method.
11297 (register_dump_files): Convert to...
11298 (pass_manager::register_dump_files): ...method.
11299 (create_pass_tab): Update for moves of globals to fields of
11301 (dump_passes): Move body of implementation into...
11302 (pass_manager::dump_passes): ...here.
11303 (register_pass): Move body of implementation into...
11304 (pass_manager::register_pass): ...here.
11305 (init_optimization_passes): Convert into...
11306 (pass_manager::pass_manager): ...constructor for new
11307 pass_manager class, and initialize the pass_lists array.
11308 (check_profile_consistency): Update for moves of globals to
11309 fields of pass_manager.
11310 (dump_profile_report): Move body of implementation into...
11311 (pass_manager::dump_profile_report): ...here.
11312 (ipa_write_summaries_1): Update for moves of pass lists from
11313 being globals to fields of pass_manager.
11314 (ipa_write_optimization_summaries): Likewise.
11315 (ipa_read_summaries): Likewise.
11316 (ipa_read_optimization_summaries): Likewise.
11317 (execute_all_ipa_stmt_fixups): Likewise.
11319 * statistics.c (statistics_fini): Update for moves of globals to
11320 fields of pass_manager.
11322 * toplev.c (general_init): Replace call to
11323 init_optimization_passes with construction of the pass_manager
11326 * tree-pass.h (all_passes): Remove, in favor of a field of the
11327 same name within the new class pass_manager.
11328 (all_small_ipa_passes): Likewise.
11329 (all_lowering_passes): Likewise.
11330 (all_regular_ipa_passes): Likewise.
11331 (all_lto_gen_passes): Likewise.
11332 (all_late_ipa_passes): Likewise.
11333 (passes_by_id): Likewise.
11334 (passes_by_id_size): Likewise.
11335 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11336 the new class pass_manager.
11337 (get_pass_for_id): Remove.
11339 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11341 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11344 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11346 * arm.md (mulhi3): New expand pattern.
11348 2013-07-30 Jan Hubicka <jh@suse.cz>
11349 Martin Liska <marxin.liska@gmail.com>
11351 * profile.c (compute_value_histograms): Do not ICE when
11352 there is mismatch only on some counters.
11354 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11356 PR rtl-optimization/57637
11357 * function.c (move_insn_for_shrink_wrap): Also check the
11358 GEN set of the LIVE problem for the liveness analysis
11359 if it exists, otherwise give up.
11361 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11363 PR tree-optimization/57993
11364 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11365 replaced statement in the candidate table.
11366 (phi_add_costs): Return infinite cost when the hidden basis does
11367 not dominate all phis on which the candidate is dependent.
11368 (replace_one_candidate): Record replaced statement in the
11371 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11373 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11374 (ashlv2si3): New expander.
11375 (*ashlv2si3_i): New define_insn_and_split.
11376 * predicates.md (float_operation): Allow patterns with three
11377 basic sub-patterns.
11379 PR rtl-optimization/58021
11380 * mode-switching.c (create_pre_exit): Always split off preceding
11381 insns if we are not at the basic block head.
11383 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11385 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11386 (UCLIBC_DYNAMIC_LINKER): New macro.
11387 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11389 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11390 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11392 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11393 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11394 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11395 for SF and DF modes. Use ieee_quad_format for TF mode.
11396 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11397 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11398 (mips_option_override): Handle `-mnan=legacy'.
11399 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11400 `-mabs=2008' and `-mnan=2008'.
11401 (OPTION_DEFAULT_SPECS): Add "nan" default.
11402 (ASM_SPEC): Handle `-mnan='.
11403 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11404 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11405 comment accordingly.
11406 (neg<mode>2): Likewise.
11407 * config/mips/mips.opt (mabs, mnan): New options.
11408 * doc/install.texi (Configuration): Document `--with-nan=' option.
11409 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11411 (MIPS Options): Document them.
11412 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11413 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11414 * configure: Regenerate.
11415 * config.in: Regenerate.
11417 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11419 * config/i386/i386.md (float post-reload splitters): Do not check
11420 for subregs of SSE registers.
11422 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11423 H.J. Lu <hongjiu.lu@intel.com>
11427 * config/i386/i386.md (post-reload splitter
11428 to avoid partial SSE reg dependency stalls): New pattern.
11430 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11432 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11434 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11436 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11437 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11439 2013-07-26 Cary Coutant <ccoutant@google.com>
11441 * dwarf2out.c (die_checksum_ordered): Don't include template
11442 instantiations in signature.
11443 (is_template_parameter): New function.
11444 (is_template_instantiation): New function.
11445 (generate_skeleton_bottom_up): Don't include template instantiations
11447 (generate_skeleton): Likewise.
11448 (break_out_comdat_types): Move recursive call to break out nested
11450 (prune_unused_types_mark_generic_parms_dies): Call
11451 is_template_parameter.
11453 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11455 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11456 uses vector registers.
11457 * config/aarch64/iterators.md: Add attributes rtn and vas.
11459 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11460 Richard Earnshaw <richard.earnshaw@arm.com>
11462 * combine.c (simplify_comparison): Re-canonicalize operands
11464 * config/arm/arm.md (movcond_addsi): New splitter.
11466 2013-07-25 Sterling Augustine <saugustine@google.com>
11468 * dwarf2out.c (size_of_pubnames): Move code to...
11469 (include_pubname_in_output): ...here. New.
11470 (want_pubnames): Rearrange.
11471 (output_pubnames): Call include_pubname_in_output. Move assertion.
11473 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11475 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11477 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11480 * doc/extend.texi: Remove obsolete builtins. Fix
11481 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11483 2013-07-25 Jan Hubicka <jh@suse.cz>
11485 * cgraph.c (release_function_body): Break out from ...
11486 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11487 and DECL_ARGUMENTS.
11488 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11489 old_tree in the map.
11490 (create_specialized_node): Update.
11491 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11493 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11494 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11495 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11496 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11499 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11501 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11502 addsi3_carryin_alt2_<optab>): Correct output template.
11504 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11506 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11507 Adjust for arm_restrict_it.
11508 Remove trailing whitespace.
11510 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11512 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11513 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11515 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11517 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11519 PR rtl-optimization/57960
11520 * lra-constraints.c (process_alt_operands): Use the right mode
11521 when checking strict_low.
11523 2013-07-25 Jan Hubicka <jh@suse.cz>
11525 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11526 * cgraph.c (cgraph_remove_node): Do not release function body
11527 when in cgraph streaming.
11528 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11529 in other partitions are not considered reachable; fix handling of
11532 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11534 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11536 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11541 * config/arm/arm.md ("*sibcall_insn): Replace use of
11542 Ss with US. Adjust output for v5 and v4t.
11543 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11545 * config/arm/constraints.md ("Ss"): Rename to US.
11547 2013-07-25 Terry Guo <terry.guo@arm.com>
11549 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11550 shift_add/shift_sub0/shift_sub1 RTXs.
11552 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11553 Anton Blanchard <anton@au1.ibm.com>
11555 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11556 (altivec_vpks<VI_char>ss): Likewise.
11557 (altivec_vpks<VI_char>us): Likewise.
11558 (altivec_vpku<VI_char>us): Likewise.
11559 (altivec_vpku<VI_char>um): Likewise.
11561 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11563 Introduce context class.
11565 * Makefile.in (CONTEXT_H): New.
11566 (OBJS): Add context.o.
11567 (toplev.o): Add CONTEXT_H to dependencies.
11570 * toplev.c (general_init): Create the singleton gcc::context instance.
11576 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11578 PR rtl-optimization/57968
11579 * mode-switching.c (create_pre_exit): Allow instructions that
11580 don't set a return register to need a non-exit mode.
11582 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11583 Anton Blanchard <anton@au1.ibm.com>
11585 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11586 operands to vperm for little endian.
11587 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11588 of lvsl to create the control mask for a vperm for little endian.
11590 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11591 Anton Blanchard <anton@au1.ibm.com>
11593 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11594 two operands for little-endian.
11596 2013-07-23 Steve Ellcey <sellcey@mips.com>
11598 * config/mips/mips.c (mips_case_values_threshold): New.
11599 (TARGET_CASE_VALUES_THRESHOLD): Define.
11601 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11602 Anton Blanchard <anton@au1.ibm.com>
11604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11605 selection of field for vector splat in little endian mode.
11607 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11609 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11610 expanders to rs6000.md.
11611 (ior<mode>3): Likewise.
11612 (and<mode>3): Likewise.
11613 (one_cmpl<mode>2): Likewise.
11614 (nor<mode>3): Likewise.
11615 (andc<mode>3): Likewise.
11616 (eqv<mode>3): Likewise.
11617 (nand<mode>3): Likewise.
11618 (orc<mode>3): Likewise.
11620 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11623 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11624 to split multi-word logical operations.
11625 (rs6000_split_logical_di): Likewise.
11626 (rs6000_split_logical): Likewise.
11628 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11629 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11630 and allow TImode operations in 32-bit.
11631 (vsx_and<mode>3_64bit): Likewise.
11632 (vsx_ior<mode>3_32bit): Likewise.
11633 (vsx_ior<mode>3_64bit): Likewise.
11634 (vsx_xor<mode>3_32bit): Likewise.
11635 (vsx_xor<mode>3_64bit): Likewise.
11636 (vsx_one_cmpl<mode>2_32bit): Likewise.
11637 (vsx_one_cmpl<mode>2_64bit): Likewise.
11638 (vsx_nor<mode>3_32bit): Likewise.
11639 (vsx_nor<mode>3_64bit): Likewise.
11640 (vsx_andc<mode>3_32bit): Likewise.
11641 (vsx_andc<mode>3_64bit): Likewise.
11642 (vsx_eqv<mode>3_32bit): Likewise.
11643 (vsx_eqv<mode>3_64bit): Likewise.
11644 (vsx_nand<mode>3_32bit): Likewise.
11645 (vsx_nand<mode>3_64bit): Likewise.
11646 (vsx_orc<mode>3_32bit): Likewise.
11647 (vsx_orc<mode>3_64bit): Likewise.
11649 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11650 logical types in GPRs.
11652 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11653 logical insns to rs6000.md, and allow TImode operations in
11655 (altivec_ior<mode>3): Likewise.
11656 (altivec_xor<mode>3): Likewise.
11657 (altivec_one_cmpl<mode>2): Likewise.
11658 (altivec_nor<mode>3): Likewise.
11659 (altivec_andc<mode>3): Likewise.
11661 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11662 attributes for moving the 128-bit logical operations into
11664 (BOOL_REGS_OUTPUT): Likewise.
11665 (BOOL_REGS_OP1): Likewise.
11666 (BOOL_REGS_OP2): Likewise.
11667 (BOOL_REGS_UNARY): Likewise.
11668 (BOOL_REGS_AND_CR0): Likewise.
11669 (one_cmpl<mode>2): Add support for DI logical operations on
11670 32-bit, splitting the operations to 32-bit.
11671 (anddi3): Likewise.
11672 (iordi3): Likewise.
11673 (xordi3): Likewise.
11674 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11675 changes to combine the 32/64-bit code, allow logical operations on
11676 TI mode in 32-bit, and to use similar match_operator patterns like
11677 scalar mode uses. Combine the Altivec and VSX code for logical
11678 operations, and move it here.
11679 (ior<mode>3, 128-bit types): Likewise.
11680 (xor<mode>3, 128-bit types): Likewise.
11681 (one_cmpl<mode>3, 128-bit types): Likewise.
11682 (nor<mode>3, 128-bit types): Likewise.
11683 (andc<mode>3, 128-bit types): Likewise.
11684 (eqv<mode>3, 128-bit types): Likewise.
11685 (nand<mode>3, 128-bit types): Likewise.
11686 (orc<mode>3, 128-bit types): Likewise.
11687 (and<mode>3_internal): Likewise.
11688 (bool<mode>3_internal): Likewise.
11689 (boolc<mode>3_internal1): Likewise.
11690 (boolc<mode>3_internal2): Likewise.
11691 (boolcc<mode>3_internal1): Likewise.
11692 (boolcc<mode>3_internal2): Likewise.
11693 (eqv<mode>3_internal1): Likewise.
11694 (eqv<mode>3_internal2): Likewise.
11695 (one_cmpl1<mode>3_internal): Likewise.
11697 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11699 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11700 Rename flag_stack_usage to flag_stack_usage_info.
11702 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11704 * config/microblaze/sync.md: New file.
11705 * config/microblaze/microblaze.md: Include sync.md
11706 * config/microblaze/microblaze.c: Add print_operand 'y'.
11707 * config/microblaze/constraints.md: Add memory_contraint
11708 'Q' which is a single register.
11710 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11712 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11714 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11716 * reload.c (find_reloads): Exit loop once we find this operand
11717 cannot be reloaded somehow for this alternative.
11719 * reload.c (find_reloads): Exit loop once we find a hard register.
11721 * rtlanal.c (computed_jump_p): Exit loop once we find label
11724 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11726 * cfgloopmanip.c (remove_path): Exit loop after setting
11729 * gensupport.c (subst_dup): Avoid loop if code is not
11730 MATCH_DUP nor MATCH_OP_DUP.
11732 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11734 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11736 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11738 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11739 true for SP_REGNUM if mode == ptr_mode.
11740 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11741 with value R0_REGNUM + 31.
11743 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11745 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11746 pad pointer-typed argument downward.
11748 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11750 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11751 and __ILP32__ when the ILP32 model is in use.
11753 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11755 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11756 (aarch64_load_symref_appropriately): In the case of
11757 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11758 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11759 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11760 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11761 if 'mode' doesn't equal to 'ptr_mode'.
11762 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11763 'vcall_offset'; change to call aarch64_emit_move differently depending
11764 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11765 to calculate the upper bound of 'vcall_offset'.
11766 (aarch64_cannot_force_const_mem): Change to also return true if
11768 (aarch64_legitimize_reload_address): In the case of large
11769 displacements, add new local variable 'xmode' and an assertion
11770 based on it; change to use 'xmode' to generate the new rtx and
11772 (aarch64_asm_trampoline_template): Change to generate the template
11773 differently depending on TARGET_ILP32 or not; change to use
11774 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11775 (aarch64_trampoline_size): Removed.
11776 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11777 and replace immediate literals with it. Change to use 'ptr_mode'
11778 instead of 'DImode' and call convert_memory_address if the mode
11779 of 'fnaddr' doesn't equal to 'ptr_mode'.
11780 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11782 (aarch64_elf_asm_destructor): Likewise.
11783 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11784 on TARGET_ILP32 instead of aarch64_trampoline_size.
11785 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11786 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11787 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11788 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11789 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11790 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11791 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11792 depending on the value of 'mode'.
11793 (add_losym_<mode>): New.
11794 (ldr_got_small_<mode>): New, based on ldr_got_small.
11795 (ldr_got_small): Remove.
11796 (ldr_got_small_sidi): New.
11797 * config/aarch64/iterators.md (P): New.
11798 (PTR): Change to 'ptr_mode' in the condition.
11800 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11802 * config.gcc (aarch64*-*-*): Support --with-abi.
11803 (aarch64*-*-elf): Support --with-multilib-list.
11804 (aarch64*-*-linux*): Likewise.
11805 (supported_defaults): Add abi to aarch64*-*-*.
11806 * configure.ac: Mention AArch64 for --with-multilib-list.
11807 * configure: Re-generated.
11808 * config/aarch64/biarchilp32.h: New file.
11809 * config/aarch64/biarchlp64.h: New file.
11810 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11812 (MULTILIB_DEFAULTS): Ditto.
11813 (DRIVER_SELF_SPECS): Ditto.
11814 (ASM_SPEC): Update to also substitute -mabi.
11815 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11816 file whose name depends on -mabi= and -mbig-endian.
11817 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11819 (POINTER_SIZE): New define.
11820 (POINTERS_EXTEND_UNSIGNED): Ditto.
11821 (enum aarch64_abi_type): New enumeration tag.
11822 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11823 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11824 (TARGET_ILP32): New define.
11825 * config/aarch64/aarch64.opt (mabi): New.
11826 (aarch64_abi): New.
11827 (ilp32, lp64): New values for -mabi.
11828 * config/aarch64/t-aarch64 (comma): New define.
11829 (MULTILIB_OPTIONS): Ditto.
11830 (MULTILIB_DIRNAMES): Ditto.
11831 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11832 * doc/invoke.texi: Document -mabi for AArch64.
11834 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11836 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11838 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11839 Anton Blanchard <anton@au1.ibm.com>
11841 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11842 endianness when selecting field to splat.
11844 2013-07-22 Eric Christopher <echristo@gmail.com>
11846 * dwarf2out.c (die_odr_checksum): New function to use
11847 CHECKSUM_ macros and ULEB128 for DIE tag.
11848 (generate_type_signature): Use.
11850 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11852 * config.gcc (sparc*-*-*): Accept leon3 processor.
11853 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11854 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11855 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11856 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11857 (mfix-ut699): Adjust comment.
11858 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11859 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11860 (CPP_CPU_SPEC): Likewise.
11861 (ASM_CPU_SPEC): Likewise.
11862 * config/sparc/sparc.c (leon3_cost): New constant.
11863 (sparc_option_override): Add leon3 support.
11864 (mem_ref): New function.
11865 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11866 (sparc_do_work_around_errata): Look into the instruction in the delay
11867 slot and adjust accordingly. Add fix for the data cache nullify issues
11868 of the UT699. Change insertion position for the NOP.
11869 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11870 (leon3_load): New reservation.
11871 (leon_store): Bump latency to 2.
11872 (grfpu): New automaton.
11873 (grfpu_alu): New unit.
11874 (grfpu_ds): Likewise.
11875 (leon_fp_alu): Adjust.
11876 (leon_fp_mult): Delete.
11877 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11878 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11879 * config/sparc/sparc.md (cpu): Add leon3.
11880 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11881 (swapsi): Likewise.
11882 (atomic_test_and_set): Likewise.
11883 (ldstub): Likewise.
11885 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11887 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11888 default for R5900 targets.
11889 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11890 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11891 * config/mips/mips.c (mips_option_override): Report an error for
11892 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11893 for -march=r5900 -mhard-float.
11895 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11897 * df-problems.c (can_move_insns_across): Exit loop once we
11898 find a non-fixed, non-global register.
11900 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11903 * omega.c (omega_eliminate_red): Break after setting red_found.
11904 (omega_problem_has_red_equations): Similarly after setting found.
11905 (omega_query_variable): Similarly after setting coupled.
11907 2013-07-22 Marek Polacek <polacek@redhat.com>
11909 * gimplify.c: Don't include gimple.h twice.
11911 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11913 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11914 instead of TARGET_THUMB1.
11915 (Pz): New constraint.
11916 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11918 (compare_negsi_si): Likewise.
11919 (compare_addsi2_op0): Likewise.
11920 (compare_addsi2_op1): Likewise.
11921 (addsi3_carryin_<optab>): Likewise.
11922 (addsi3_carryin_alt2_<optab>): Likewise.
11923 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11924 for arm_restrict_it.
11925 (subsi3_carryin): Likewise.
11926 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11927 (minmax_arithsi): Disable for arm_restrict_it.
11928 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11929 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11930 (satsi_<SAT:code>_shift): Likewise.
11931 (arm_shiftsi3): Add alternative for 16-bit encoding.
11932 (arm32_movhf): Disable for arm_restrict_it.
11933 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11934 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11936 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11938 * config/arm/arm.md (attribute "insn"): Delete.
11939 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11940 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11941 (not_shiftsi): Update for attribute change.
11942 (not_shiftsi_compare0): Likewise.
11943 (not_shiftsi_compare0_scratch): Likewise.
11944 (arm_one_cmplsi2): Likewise.
11945 (thumb1_one_cmplsi2): Likewise.
11946 (notsi_compare0): Likewise.
11947 (notsi_compare0_scratch): Likewise.
11948 (thumb1_movdi_insn): Likewise.
11949 (arm_movsi_insn): Likewise.
11950 (movhi_insn_arch4): Likewise.
11951 (movhi_bytes): Likewise.
11952 (arm_movqi_insn): Likewise.
11953 (thumb1_movqi_insn): Likewise.
11954 (arm32_movhf): Likewise.
11955 (thumb1_movhf): Likewise.
11956 (arm_movsf_soft_insn): Likewise.
11957 (thumb1_movsf_insn): Likewise.
11958 (thumb_movdf_insn): Likewise.
11959 (movsicc_insn): Likewise.
11960 (movsfcc_soft_insn): Likewise.
11961 (and_scc): Likewise.
11962 (cond_move): Likewise.
11963 (if_move_not): Likewise.
11964 (if_not_move): Likewise.
11965 (if_shift_move): Likewise.
11966 (if_move_shift): Likewise.
11967 (if_shift_shift): Likewise.
11968 (if_not_arith): Likewise.
11969 (if_arith_not): Likewise.
11970 (cond_move_not): Likewise.
11971 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11972 (neon_mov<mode>): Likewise.
11973 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11974 (thumb2_movsi_vfp): Likewise.
11975 (movsf_vfp): Likewise.
11976 (thumb2_movsf_vfp): Likewise.
11977 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11979 (cortexa7_older_only): Likewise.
11980 (cortexa7_younger): Likewise.
11981 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11982 (1020alu_shift_op): Likewise.
11983 (1020alu_shift_reg_op): Likewise.
11984 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11985 (alu_shift_op): Likewise.
11986 (alu_shift_reg_op): Likewise.
11987 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11988 (11_alu_shift_op): Likewise.
11989 (11_alu_shift_reg_op): Likewise.
11990 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11991 (9_alu_shift_reg_op): Likewise.
11992 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11994 (cortex_a15_alu_shift): Likewise.
11995 (cortex_a15_alu_shift_reg): Likewise.
11996 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11997 (cortex_a5_alu_shift): Likewise.
11998 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12000 (cortex_a53_alu_shift): Likewise.
12001 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12003 (cortex_a7_alu_reg): Likewise.
12004 (cortex_a7_alu_shift): Likewise.
12005 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12006 (cortex_a8_alu_shift): Likewise.
12007 (cortex_a8_alu_shift_reg): Likewise.
12008 (cortex_a8_mov): Likewise.
12009 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12010 (cortex_a9_dp_shift): Likewise.
12011 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12012 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12013 (cortex_r4_mov): Likewise.
12014 (cortex_r4_alu_shift): Likewise.
12015 (cortex_r4_alu_shift_reg): Likewise.
12016 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12017 (526_alu_shift_op): Likewise.
12018 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12019 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12020 (626te_alu_shift_op): Likewise.
12021 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12022 (726te_alu_op): Likewise.
12023 (726te_alu_shift_op): Likewise.
12024 (726te_alu_shift_reg_op): Likewise.
12025 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12026 (mp626_alu_shift_op): Likewise.
12027 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12028 (pj4_alu_e1_conds): Likewise.
12029 (pj4_alu): Likewise.
12030 (pj4_alu_conds): Likewise.
12031 (pj4_shift): Likewise.
12032 (pj4_shift_conds): Likewise.
12033 (pj4_alu_shift): Likewise.
12034 (pj4_alu_shift_conds): Likewise.
12036 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12038 * config/arm/predicates.md (shiftable_operator_strict_it):
12040 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12041 Disable cond_exec version for arm_restrict_it.
12042 (thumb2_smaxsi3): Convert to generate cond_exec.
12043 (thumb2_sminsi3): Likewise.
12044 (thumb32_umaxsi3): Likewise.
12045 (thumb2_uminsi3): Likewise.
12046 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12047 (thumb2_neg_abssi2): Likewise.
12048 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12049 (thumb2_movsicc_insn): Adjust alternatives.
12050 (thumb2_mov_negscc): Disable for arm_restrict_it.
12051 (thumb2_mov_negscc_strict_it): New pattern.
12052 (thumb2_mov_notscc_strict_it): New pattern.
12053 (thumb2_mov_notscc): Disable for arm_restrict_it.
12054 (thumb2_ior_scc): Likewise.
12055 (thumb2_ior_scc_strict_it): New pattern.
12056 (thumb2_cond_move): Adjust for arm_restrict_it.
12057 (thumb2_cond_arith): Disable for arm_restrict_it.
12058 (thumb2_cond_arith_strict_it): New pattern.
12059 (thumb2_cond_sub): Adjust for arm_restrict_it.
12060 (thumb2_movcond): Likewise.
12061 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12062 (thumb2_zero_extendhisi2_v6): Likewise.
12063 (thumb2_zero_extendqisi2_v6): Likewise.
12064 (orsi_notsi_si): Likewise.
12065 (orsi_not_shiftsi_si): Likewise.
12067 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12069 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12070 instruction sequence is 1 byte shorter.
12072 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12074 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12075 it is not needed after split.
12077 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12080 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12081 second label for nonlocal goto receivers. Don't output pic base labels
12082 unless we're producing PIC; mark that action unreachable().
12083 (ix86_save_reg): If the function contains a nonlocal label, save the
12085 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12086 * config/darwin.c (emitted_pic_label_num): New GTY.
12087 (update_pic_label_number_if_needed): New.
12088 (machopic_output_function_base_name): Adjust for nonlocal receiver
12090 (machopic_should_output_picbase_label): New.
12091 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12092 (nonlocal_goto_receiver): New insn and split.
12094 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12096 * config/aarch64/aarch64-builtins.c
12097 (aarch64_fold_builtin): Fold abs in all modes.
12098 * config/aarch64/aarch64-simd-builtins.def
12099 (abs): Enable for all modes.
12100 * config/aarch64/arm_neon.h
12101 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12102 (vabs_f64): Add missing intrinsic.
12104 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12106 * config/aarch64/arm_neon.h (vabs_s64): New function
12108 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12111 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12112 * config/avr/avr.md (adjust_len): Add `round'.
12113 * config/avr/avr-protos.h (avr_out_round): New prototype.
12114 (avr_out_plus): Add `out_label' argument.
12115 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12116 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12117 Handle the case where `insn' is just a pattern.
12118 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12119 (avr_out_round): New function.
12120 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12122 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12124 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12125 Add check for flag_stack_usage to handle -fstack-usage support
12127 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12129 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12130 interaction for new Power8 flags and VSX.
12132 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12134 PR middle-end/57698
12135 * tree-inline.c (expand_call_inline): Emit errors during
12136 early_inlining only if optimization is not turned on.
12138 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12142 * passes.c (init_optimization_passes): Move the construction of
12143 the pass hierarchy into a new passes.def file.
12145 * Makefile.in (passes.o): Add dependency on passes.def.
12147 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12149 * passes.c (init_optimization_passes): Introduce macros for
12150 constructing the tree of passes (INSERT_PASSES_AFTER,
12151 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12152 TERMINATE_PASS_LIST).
12154 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12155 Wei Mi <wmi@google.com>
12157 PR rtl-optimization/57878
12158 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12160 (reload_pseudo_compare_func): Check nregs first for reload
12163 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12165 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12167 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12169 * read-rtl.c (validate_const_int): Once an invalid character is
12170 seen, quit the loop.
12172 * gengtype.c (write_roots): Similarly once we find the "deletable"
12173 or "if_marked" option.
12175 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12177 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12178 "xtab" and "sat". Move value "clz" from here to ...
12179 (attriubte "type"): ... here.
12180 (satsi_<SAT:code>): Delete "insn" attribute.
12181 (satsi_<SAT:code>_shift): Likewise.
12182 (arm_zero_extendqisi2addsi): Likewise.
12183 (arm_extendqisi2addsi): Likewise.
12184 (clzsi2): Update for attribute changes.
12185 (rbitsi2): Likewise.
12186 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12188 (arm_usatsihi): Likewise.
12189 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12191 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12193 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12194 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12195 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12196 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12197 in alphabetical order.
12198 (attribute "core_cycles"): Update for attribute changes.
12199 (arm_addsi3): Likewise.
12200 (addsi3_compare0): Likewise.
12201 (addsi3_compare0_scratch): Likewise.
12202 (addsi3_compare_op1): Likewise.
12203 (addsi3_compare_op2): Likewise.
12204 (compare_addsi2_op0): Likewise.
12205 (compare_addsi2_op1): Likewise.
12206 (addsi3_carryin_shift_<optab>): Likewise.
12207 (subsi3_carryin_shift): Likewise.
12208 (rsbsi3_carryin_shift): Likewise.
12209 (arm_subsi3_insn): Likewise.
12210 (subsi3_compare0): Likewise.
12211 (subsi3_compare): Likewise.
12212 (arm_andsi3_insn): Likewise.
12213 (thumb1_andsi3_insn): Likewise.
12214 (andsi3_compare0): Likewise.
12215 (andsi3_compare0_scratch): Likewise.
12216 (zeroextractsi_compare0_scratch
12217 (andsi_not_shiftsi_si): Likewise.
12218 (iorsi3_insn): Likewise.
12219 (iorsi3_compare0): Likewise.
12220 (iorsi3_compare0_scratch): Likewise.
12221 (arm_xorsi3): Likewise.
12222 (thumb1_xorsi3_insn): Likewise.
12223 (xorsi3_compare0): Likewise.
12224 (xorsi3_compare0_scratch): Likewise.
12225 (satsi_<SAT:code>_shift): Likewise.
12227 (arm_shiftsi3): Likewise.
12228 (shiftsi3_compare0): Likewise.
12229 (not_shiftsi): Likewise.
12230 (not_shiftsi_compare0): Likewise.
12231 (not_shiftsi_compare0_scratch): Likewise.
12232 (arm_one_cmplsi2): Likewise.
12233 (thumb_one_complsi2): Likewise.
12234 (notsi_compare0): Likewise.
12235 (notsi_compare0_scratch): Likewise.
12236 (thumb1_zero_extendhisi2): Likewise.
12237 (arm_zero_extendhisi2): Likewise.
12238 (arm_zero_extendhisi2_v6): Likewise.
12239 (arm_zero_extendhisi2addsi): Likewise.
12240 (thumb1_zero_extendqisi2): Likewise.
12241 (thumb1_zero_extendqisi2_v6): Likewise.
12242 (arm_zero_extendqisi2): Likewise.
12243 (arm_zero_extendqisi2_v6): Likewise.
12244 (arm_zero_extendqisi2addsi): Likewise.
12245 (thumb1_extendhisi2): Likewise.
12246 (arm_extendhisi2): Likewise.
12247 (arm_extendhisi2_v6): Likewise.
12248 (arm_extendqisi): Likewise.
12249 (arm_extendqisi_v6): Likewise.
12250 (arm_extendqisi2addsi): Likewise.
12251 (thumb1_extendqisi2): Likewise.
12252 (thumb1_movdi_insn): Likewise.
12253 (arm_movsi_insn): Likewise.
12254 (movsi_compare0): Likewise.
12255 (movhi_insn_arch4): Likewise.
12256 (movhi_bytes): Likewise.
12257 (arm_movqi_insn): Likewise.
12258 (thumb1_movqi_insn): Likewise.
12259 (arm32_movhf): Likewise.
12260 (thumb1_movhf): Likewise.
12261 (arm_movsf_soft_insn): Likewise.
12262 (thumb1_movsf_insn): Likewise.
12263 (movdf_soft_insn): Likewise.
12264 (thumb_movdf_insn): Likewise.
12265 (arm_cmpsi_insn): Likewise.
12266 (cmpsi_shiftsi): Likewise.
12267 (cmpsi_shiftsi_swp): Likewise.
12268 (arm_cmpsi_negshiftsi_si): Likewise.
12269 (movsicc_insn): Likewise.
12270 (movsfcc_soft_insn): Likewise.
12271 (arith_shiftsi): Likewise.
12272 (arith_shiftsi_compare0
12273 (arith_shiftsi_compare0_scratch
12274 (sub_shiftsi): Likewise.
12275 (sub_shiftsi_compare0
12276 (sub_shiftsi_compare0_scratch
12277 (and_scc): Likewise.
12278 (cond_move): Likewise.
12279 (if_plus_move): Likewise.
12280 (if_move_plus): Likewise.
12281 (if_move_not): Likewise.
12282 (if_not_move): Likewise.
12283 (if_shift_move): Likewise.
12284 (if_move_shift): Likewise.
12285 (if_shift_shift): Likewise.
12286 (if_not_arith): Likewise.
12287 (if_arith_not): Likewise.
12288 (cond_move_not): Likewise.
12289 (thumb1_ashlsi3): Set type attribute.
12290 (thumb1_ashrsi3): Likewise.
12291 (thumb1_lshrsi3): Likewise.
12292 (thumb1_rotrsi3): Likewise.
12293 (shiftsi3_compare0_scratch): Likewise.
12294 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12295 (neon_mov<mode>): Likewise.
12296 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12298 (thumb2_movsi_insn): Likewise.
12299 (thumb2_cmpsi_neg_shiftsi): Likewise.
12300 (thumb2_extendqisi_v6): Likewise.
12301 (thumb2_zero_extendhisi2_v6): Likewise.
12302 (thumb2_zero_extendqisi2_v6): Likewise.
12303 (thumb2_shiftsi3_short): Likewise.
12304 (thumb2_addsi3_compare0_scratch): Likewise.
12305 (orsi_not_shiftsi_si): Likewise.
12306 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12307 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12309 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12310 (1020alu_shift_op): Likewise.
12311 (1020alu_shift_reg_op): Likewise.
12312 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12313 (alu_shift_op): Likewise.
12314 (alu_shift_reg_op): Likewise.
12315 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12316 (11_alu_shift_op): Likewise.
12317 (11_alu_shift_reg_op): Likewise.
12318 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12319 (9_alu_shift_reg_op): Likewise.
12320 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12322 (cortex_a15_alu_shift): Likewise.
12323 (cortex_a15_alu_shift_reg): Likewise.
12324 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12326 (cortex_a5_alu_shift): Likewise.
12327 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12329 (cortex_a53_alu_shift): Likewise.
12330 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12332 (cortex_a7_alu_reg): Likewise.
12333 (cortex_a7_alu_shift): Likewise.
12334 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12336 (cortex_a8_alu_shift): Likewise.
12337 (cortex_a8_alu_shift_reg): Likewise.
12338 (cortex_a8_mov): Likewise.
12339 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12340 (cortex_a9_dp_shift): Likewise.
12341 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12343 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12345 (cortex_r4_mov): Likewise.
12346 (cortex_r4_alu_shift): Likewise.
12347 (cortex_r4_alu_shift_reg): Likewise.
12348 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12349 (526_alu_shift_op): Likewise.
12350 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12351 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12352 (626te_alu_shift_op): Likewise.
12353 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12354 (726te_alu_op): Likewise.
12355 (726te_alu_shift_op): Likewise.
12356 (726te_alu_shift_reg_op): Likewise.
12357 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12358 (mp626_alu_shift_op): Likewise.
12359 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12360 (pj4_alu_e1_conds): Likewise.
12361 (pj4_alu): Likewise.
12362 (pj4_alu_conds): Likewise.
12363 (pj4_shift): Likewise.
12364 (pj4_shift_conds): Likewise.
12365 (pj4_alu_shift): Likewise.
12366 (pj4_alu_shift_conds): Likewise.
12367 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12369 (cortexa7_older_only): Likewise.
12370 (cortexa7_younger): Likewise.
12372 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12374 * ipa-pure-const.c (generate_summary): Rename to...
12375 (pure_const_generate_summary): ... this.
12377 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12379 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12381 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12384 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12387 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12389 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12390 enabled without -march=zEC12.
12391 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12394 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12396 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12397 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12398 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12399 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12400 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12401 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12402 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12404 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12405 and ISA_HAS_NMADD3_NMSUB3.
12406 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12407 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12408 (nmsub4<mode>, nmsub3<mode>): Likewise.
12409 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12411 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12413 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12414 TARGET_MIPS5400 checking.
12416 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12417 Peter Bergner <bergner@vnet.ibm.com>
12419 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12420 registers in the comment.
12421 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12422 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12423 rather than FIRST_PSEUDO_REGISTERS.
12425 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12427 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12428 enable extra ISA flags with TARGET_HTM.
12430 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12432 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12435 2013-07-15 Cong Hou <congh@google.com>
12437 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12438 in compare function for sorting.
12440 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12442 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12443 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12444 * config/rs6000/rs6000.opt: Add -mhtm option.
12445 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12446 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12447 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12448 __HTM__ if the HTM instructions are available.
12449 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12450 htm_spr_reg_operand): New define_predicates.
12451 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12452 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12454 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12455 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12456 HTM builtin functions.
12457 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12458 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12459 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12460 (rs6000_builtin_mask_calculate): Likewise.
12461 (rs6000_option_override_internal): Likewise.
12462 (bdesc_htm): Add new HTM builtin support.
12463 (htm_spr_num): New function.
12464 (htm_spr_regno): Likewise.
12465 (rs6000_htm_spr_icode): Likewise.
12466 (htm_expand_builtin): Likewise.
12467 (htm_init_builtins): Likewise.
12468 (rs6000_expand_builtin): Add support for HTM builtin functions.
12469 (rs6000_init_builtins): Likewise.
12470 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12472 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12473 (TARGET_HTM, MASK_HTM): Define macros.
12474 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12475 (FIXED_REGISTERS): Likewise.
12476 (CALL_USED_REGISTERS): Likewise.
12477 (CALL_REALLY_USED_REGISTERS): Likewise.
12478 (REG_ALLOC_ORDER): Likewise.
12479 (enum reg_class): Likewise.
12480 (REG_CLASS_NAMES): Likewise.
12481 (REG_CLASS_CONTENTS): Likewise.
12482 (REGISTER_NAMES): Likewise.
12483 (ADDITIONAL_REGISTER_NAMES): Likewise.
12484 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12485 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12486 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12487 * config/rs6000/htm.md: New file.
12488 * config/rs6000/htmintrin.h: New file.
12489 * config/rs6000/htmxlintrin.h: New file.
12491 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12493 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12494 Define SYMBOL_TINY_GOT, update comment.
12495 * config/aarch64/aarch64.c
12496 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12497 (aarch64_expand_mov_immediate): Likewise.
12498 (aarch64_print_operand): Likewise.
12499 (aarch64_classify_symbol): Likewise.
12500 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12501 (ldr_got_tiny): Define.
12503 2013-07-13 Tobias Grosser <tobias@grosser.es>
12505 PR tree-optimization/54094
12506 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12507 scheduling dimension for the parallelism check from the polyhedral
12508 information in the AST.
12509 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12512 2013-07-13 Jason Merrill <jason@redhat.com>
12514 * print-tree.c (debug_vec_tree): Use debug_raw.
12515 (debug_raw (vec<tree, va_gc> &)): New.
12516 (debug_raw (vec<tree, va_gc> *)): New.
12517 * tree.h: Declare them.
12519 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12521 * ifcvt.c (ifcvt_after_combine): New static variable.
12522 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12524 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12525 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12526 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12528 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12530 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12532 2013-07-12 Michael Matz <matz@suse.de>
12534 PR middle-end/55771
12535 * convert.c (convert_to_real): Reject non-float inner types.
12537 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12539 * config/aarch64/aarch64-protos.h
12540 (aarch64_simd_immediate_valid_for_move): Remove.
12541 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12542 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12544 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12546 2013-07-11 Steve Ellcey <sellcey@mips.com>
12548 * config/mips/mips.c (mips_conditional_register_usage): Do not
12549 use t[0-7] registers in MIPS16 mode when optimizing for size.
12551 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12553 * config/i386/i386.c (dispatch_function_versions): Fix array
12554 indexing of function_version_info to match actual_versions.
12556 2013-07-11 Teresa Johnson <tejohnson@google.com>
12558 * vec.h (struct va_gc): Move release out-of-line.
12559 (va_gc::release): Call ggc_free on released vec.
12561 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12563 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12564 Require GOT register as additional operand in UNSPEC.
12565 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12566 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12567 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12568 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12569 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12570 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12571 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12573 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12576 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12577 name seen by assembler/linker rather if available.
12579 2013-07-11 Andreas Schwab <schwab@suse.de>
12581 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12583 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12585 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12587 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12589 * doc/tm.texi.in: Move hook documentation to ....
12590 * target.def: ... here.
12592 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12593 text on @hook line.
12594 * doc/tm.texi: Regenerate.
12596 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12599 * doc/invoke.texi: Document -Wconditionally-supported.
12601 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12604 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12607 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12610 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12611 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12612 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12613 Remove duplicate devices.
12614 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12615 * config/avr/t-multilib: Regenerate.
12616 * config/avr/avr-tables.opt: Regenerate.
12617 * doc/avr-mmcu.texi: Regenerate.
12619 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12622 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12624 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12626 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12627 the cost of MULT when optimizing for size.
12629 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12631 * config/cr16/cr16-protos.h: Don't include target.h.
12633 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12635 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12636 adjust register size for TDmode and TFmode for VSX registers.
12638 2013-07-08 Kai Tietz <ktietz@redhat.com>
12641 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12642 hook_bool_const_tree_true.
12644 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12646 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12647 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12648 * config/s390/s390.md: Define FPR*_REGNUM constants.
12649 Fix FPR2_REGNUM constant (18 -> 17).
12650 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12651 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12652 ("trunc<BFP:mode><DFP_ALL:mode>2")
12653 ("trunc<DFP_ALL:mode><BFP:mode>2")
12654 ("*extend<BFP:mode><DFP_ALL:mode>2")
12655 ("*extend<DFP_ALL:mode><BFP:mode>2")
12656 ("extend<BFP:mode><DFP_ALL:mode>2")
12657 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12660 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12662 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12664 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12666 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12667 and cfun_fpr_bit_p to cfun_fpr_save_p.
12668 (s390_frame_area, s390_register_info, s390_frame_info)
12669 (s390_emit_prologue, s390_emit_epilogue)
12670 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12672 * config/s390/s390.h: Define *_REGNUM macros for floating point
12675 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12677 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12679 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12681 PR rtl-optimization/57786
12682 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12683 and break out of the loop when it is set to false.
12685 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12688 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12689 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12690 (const_int 63)) 0)).
12691 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12692 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12693 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12695 PR rtl-optimization/57829
12696 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12697 mask bits outside of mode are just sign-extension from mode to HWI.
12699 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12701 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12702 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12703 adjust_address instead of change_address to keep info about alignment.
12704 (emit_strmov): Remove.
12705 (emit_memmov): New function.
12706 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12707 (expand_movmem_epilogue): Likewise and return updated rtx for
12709 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12710 destination and source.
12711 (decide_alignment): Refactor, handle vector_loop.
12712 (ix86_expand_movmem): Likewise.
12713 (ix86_expand_setmem): Likewise.
12714 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12716 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12718 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12719 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12721 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12723 * config/i386/sse.md (sse_movlhps): Change alternative 3
12724 of operand 2 to "m".
12726 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12729 * config/i386/sse.md (iptr): New mode attribute.
12730 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12731 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12732 (<sse>_vmmul<mode>3): Ditto.
12733 (<sse>_vmdiv<mode>3): Ditto.
12734 (sse_vmrcpv4sf2): Ditto.
12735 (<sse>_vmsqrt<mode>2): Ditto.
12736 (sse_vmrsqrtv4sf2): Ditto.
12737 (<sse>_vm<code><mode>3): Ditto.
12738 (avx_vmcmp<mode>3): Ditto.
12739 (<sse>_vmmaskcmp<mode>3): Ditto.
12740 (<sse>_comi): Ditto.
12741 (<sse>_ucomi): Ditto.
12742 (*xop_vmfrcz_<mode>): Ditto.
12743 (*fmai_fmadd_<mode>): Ditto.
12744 (*fmai_fmsub_<mode>): Ditto.
12745 (*fmai_fnmadd_<mode>): Ditto.
12746 (*fmai_fnmsub_<mode>): Ditto.
12747 (*fma4i_vmfmadd_<mode>): Ditto.
12748 (*fma4i_vmfmsub_<mode>): Ditto.
12749 (*fma4i_vmfnmadd_<mode>): Ditto.
12750 (*fma4i_vmfnmsub_<mode>): Ditto.
12751 (*xop_vmfrcz_<mode>): Ditto.
12752 (sse_cvtps2pi): Ditto.
12753 (sse_cvttps2pi): Ditto.
12754 (sse_cvtss2si): Ditto.
12755 (sse_cvtss2si_2): Ditto.
12756 (sse_cvtss2siq_2): Ditto.
12757 (sse_cvttss2si): Ditto.
12758 (sse_cvttss2siq): Ditto.
12759 (sse_cvtsd2si): Ditto.
12760 (sse_cvtsd2si_2): Ditto.
12761 (sse_cvtsd2siq_2): Ditto.
12762 (sse_cvttsd2si): Ditto.
12763 (sse_cvttsd2siq): Ditto.
12764 (sse_cvtsd2ss): Ditto.
12765 (sse_cvtss2sd): Ditto.
12766 (avx2_pbroadcast<mode>): Ditto.
12767 (avx2_pbroadcast<mode>_1): Ditto.
12768 (*avx_vperm_broadcast_v4sf): Ditto.
12770 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12771 (sse_movlhps): Ditto.
12772 (sse_storehps): Ditto.
12773 (sse_loadhps): Ditto.
12774 (sse_storelps): Ditto.
12775 (sse_loadlps): Ditto.
12776 (*vec_concatv4sf): Ditto.
12777 (*vec_interleave_highv2df): Ditto.
12778 (*vec_interleave_lowv2df): Ditto.
12779 (*vec_extractv2df_1_sse): Ditto.
12780 (*vec_extractv2df_0_sse): Ditto.
12781 (sse2_storelpd): Ditto.
12782 (sse2_loadlpd): Ditto.
12783 (sse2_movsd): Ditto.
12784 (*vec_concatv4si): Ditto.
12785 (vec_concatv2di): Ditto.
12787 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12788 for Intel asm dialect.
12789 (mmx_punpcklwd): Ditto.
12790 (mmx_punpckldq): Ditto.
12792 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12793 for intel assembler dialect.
12795 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12798 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12799 for BUILT_IN_C{LZ,LRSB}*.
12800 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12801 * tree-vrp.c (extract_range_basic): Handle
12802 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12803 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12804 fall thru to code calling set_value*.
12805 * builtins.c (expand_builtin): Remove *IMAX cases.
12806 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12807 if width is bigger than 2*HWI.
12809 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12811 PR rtl-optimization/55342
12812 * lra-int.h (lra_subreg_reload_pseudos): New.
12813 * lra.c: Add undoing optional reloads to the block diagram.
12814 (lra_subreg_reload_pseudos): New.
12815 (lra_optional_reload_pseudos): Change comments.
12816 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12817 lra_optional_reload_pseudos after undo transformations.
12818 * lra-assigns.c (pseudo_prefix_title): New.
12819 (lra_setup_reg_renumber): Use it.
12820 (spill_for): Ditto. Check subreg reload pseudos too.
12821 (assign_by_spills): Consider subreg reload pseudos too.
12822 * lra-constraints.c (simplify_operand_subreg): Use
12823 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12824 (curr_insn_transform): Recognize and do optional reloads.
12825 (undo_optional_reloads): New.
12826 (lra_undo_inheritance): Call undo_optional_reloads.
12828 2013-07-05 Thomas Quinot <quinot@adacore.com>
12830 * tree-complex.c (expand_complex_operations_1): Fix typo.
12832 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12834 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12835 (tune_params): New member 'const vec_costs'.
12836 * config/aarch64/aarch64.c (generic_vector_cost): New.
12837 (generic_tunings): New member 'generic_vector_cost'.
12838 (aarch64_builtin_vectorization_cost): New.
12839 (aarch64_add_stmt_cost): New.
12840 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12841 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12843 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12846 * config/i386/predicates.md (vsib_address_operand): Disallow
12847 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12849 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12851 PR middle-end/55030
12852 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12853 expand_builtin_setjmp_receiver.
12854 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12855 with NULL for the label parameter.
12856 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12857 the frame-pointer. Adjust comments.
12858 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12859 only if LABEL is non-NULL.
12861 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12863 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12864 (ARM_ABI_AAPCS64): Ditto.
12866 (ARM_DEFAULT_ABI): Ditto.
12868 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12870 * config/aarch64/aarch64-builtins.c
12871 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12872 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12874 * config/aarch64/aarch64-simd.md
12875 (aarch64_ld1<VALL:mode>): New.
12876 (aarch64_st1<VALL:mode>): Likewise.
12877 * config/aarch64/arm_neon.h
12878 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12880 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12882 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12884 (ix86_option_override_internal):Turn on all -mavx target flags by
12885 default as they are dependent on AVX anyway.
12887 2013-07-02 Cary Coutant <ccoutant@google.com>
12889 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12890 deterministic hash.
12891 (loc_checksum_ordered): Likewise.
12892 (hash_loc_operands): Remove inline keyword.
12894 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12896 PR tree-optimization/57741
12897 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12898 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12899 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12900 Allow REAL_CST step_exprs if flag_associative_math.
12901 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12903 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12905 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12907 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12909 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12911 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12913 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12915 (iorsi3_insn): Likewise.
12916 (arm_xorsi3): Likewise.
12918 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12920 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12922 (attribute "type"): ... here, and prefix with "wmmx_".
12923 (attribute "core_cycles"): Update for attribute changes.
12924 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12925 (tbcstv4hi): Likewise.
12926 (tbcstv2si): Likewise.
12927 (iwmmxt_iordi3): Likewise.
12928 (iwmmxt_xordi3): Likewise.
12929 (iwmmxt_anddi3): Likewise.
12930 (iwmmxt_nanddi3): Likewise.
12931 (iwmmxt_arm_movdi): Likewise.
12932 (iwmmxt_movsi_insn): Likewise.
12933 (mov<mode>_internal): Likewise.
12934 (and<mode>3_iwmmxt): Likewise.
12935 (ior<mode>3_iwmmxt): Likewise.
12936 (xor<mode>3_iwmmxt): Likewise.
12937 (add<mode>3_iwmmxt): Likewise.
12938 (ssaddv8qi3): Likewise.
12939 (ssaddv4hi3): Likewise.
12940 (ssaddv2si3): Likewise.
12941 (usaddv8qi3): Likewise.
12942 (usaddv4hi3): Likewise.
12943 (usaddv2si3): Likewise.
12944 (sub<mode>3_iwmmxt): Likewise.
12945 (sssubv8qi3): Likewise.
12946 (sssubv4hi3): Likewise.
12947 (sssubv2si3): Likewise.
12948 (ussubv8qi3): Likewise.
12949 (ussubv4hi3): Likewise.
12950 (ussubv2si3): Likewise.
12951 (mulv4hi3_iwmmxt): Likewise.
12952 (smulv4hi3_highpart): Likewise.
12953 (umulv4hi3_highpart): Likewise.
12954 (iwmmxt_wmacs): Likewise.
12955 (iwmmxt_wmacsz): Likewise.
12956 (iwmmxt_wmacu): Likewise.
12957 (iwmmxt_wmacuz): Likewise.
12958 (iwmmxt_clrdi): Likewise.
12959 (iwmmxt_clrv8qi): Likewise.
12960 (iwmmxt_clr4hi): Likewise.
12961 (iwmmxt_clr2si): Likewise.
12962 (iwmmxt_uavgrndv8qi3): Likewise.
12963 (iwmmxt_uavgrndv4hi3): Likewise.
12964 (iwmmxt_uavgv8qi3): Likewise.
12965 (iwmmxt_uavgv4hi3): Likewise.
12966 (iwmmxt_tinsrb): Likewise.
12967 (iwmmxt_tinsrh): Likewise.
12968 (iwmmxt_tinsrw): Likewise.
12969 (iwmmxt_textrmub): Likewise.
12970 (iwmmxt_textrmsb): Likewise.
12971 (iwmmxt_textrmuh): Likewise.
12972 (iwmmxt_textrmsh): Likewise.
12973 (iwmmxt_textrmw): Likewise.
12974 (iwmxxt_wshufh): Likewise.
12975 (eqv8qi3): Likewise.
12976 (eqv4hi3): Likewise.
12977 (eqv2si3): Likewise.
12978 (gtuv8qi3): Likewise.
12979 (gtuv4hi3): Likewise.
12980 (gtuv2si3): Likewise.
12981 (gtv8qi3): Likewise.
12982 (gtv4hi3): Likewise.
12983 (gtv2si3): Likewise.
12984 (smax<mode>3_iwmmxt): Likewise.
12985 (umax<mode>3_iwmmxt): Likewise.
12986 (smin<mode>3_iwmmxt): Likewise.
12987 (umin<mode>3_iwmmxt): Likewise.
12988 (iwmmxt_wpackhss): Likewise.
12989 (iwmmxt_wpackwss): Likewise.
12990 (iwmmxt_wpackdss): Likewise.
12991 (iwmmxt_wpackhus): Likewise.
12992 (iwmmxt_wpackwus): Likewise.
12993 (iwmmxt_wpackdus): Likewise.
12994 (iwmmxt_wunpckihb): Likewise.
12995 (iwmmxt_wunpckihh): Likewise.
12996 (iwmmxt_wunpckihw): Likewise.
12997 (iwmmxt_wunpckilb): Likewise.
12998 (iwmmxt_wunpckilh): Likewise.
12999 (iwmmxt_wunpckilw): Likewise.
13000 (iwmmxt_wunpckehub): Likewise.
13001 (iwmmxt_wunpckehuh): Likewise.
13002 (iwmmxt_wunpckehuw): Likewise.
13003 (iwmmxt_wunpckehsb): Likewise.
13004 (iwmmxt_wunpckehsh): Likewise.
13005 (iwmmxt_wunpckehsw): Likewise.
13006 (iwmmxt_wunpckelub): Likewise.
13007 (iwmmxt_wunpckeluh): Likewise.
13008 (iwmmxt_wunpckeluw): Likewise.
13009 (iwmmxt_wunpckelsb): Likewise.
13010 (iwmmxt_wunpckelsh): Likewise.
13011 (iwmmxt_wunpckelsw): Likewise.
13012 (ror<mode>3): Likewise.
13013 (ashr<mode>3_iwmmxt): Likewise.
13014 (lshr<mode>3_iwmmxt): Likewise.
13015 (ashl<mode>3_iwmmxt): Likewise.
13016 (ror<mode>3_di): Likewise.
13017 (ashr<mode>3_di): Likewise.
13018 (lshr<mode>3_di): Likewise.
13019 (ashl<mode>3_di): Likewise.
13020 (iwmmxt_wmadds): Likewise.
13021 (iwmmxt_wmaddu): Likewise.
13022 (iwmmxt_tmia): Likewise.
13023 (iwmmxt_tmiaph): Likewise.
13024 (iwmmxt_tmiabb): Likewise.
13025 (iwmmxt_tmiatb): Likewise.
13026 (iwmmxt_tmiabt): Likewise.
13027 (iwmmxt_tmiatt): Likewise.
13028 (iwmmxt_tmovmskb): Likewise.
13029 (iwmmxt_tmovmskh): Likewise.
13030 (iwmmxt_tmovmskw): Likewise.
13031 (iwmmxt_waccb): Likewise.
13032 (iwmmxt_wacch): Likewise.
13033 (iwmmxt_waccw): Likewise.
13034 (iwmmxt_waligni): Likewise.
13035 (iwmmxt_walignr): Likewise.
13036 (iwmmxt_walignr0): Likewise.
13037 (iwmmxt_walignr1): Likewise.
13038 (iwmmxt_walignr2): Likewise.
13039 (iwmmxt_walignr3): Likewise.
13040 (iwmmxt_wsadb): Likewise.
13041 (iwmmxt_wsadh): Likewise.
13042 (iwmmxt_wsadbz): Likewise.
13043 (iwmmxt_wsadhz): Likewise.
13044 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13045 (iwmmxt_wabsdiffb): Likewise.
13046 (iwmmxt_wabsdiffh): Likewise.
13047 (iwmmxt_wabsdiffw): Likewise.
13048 (iwmmxt_waddsubhx): Likewise
13049 (iwmmxt_wsubaddhx): Likewise.
13050 (addc<mode>3): Likewise.
13051 (iwmmxt_avg4): Likewise.
13052 (iwmmxt_avg4r): Likewise.
13053 (iwmmxt_wmaddsx): Likewise.
13054 (iwmmxt_wmaddux): Likewise.
13055 (iwmmxt_wmaddsn): Likewise.
13056 (iwmmxt_wmaddun): Likewise.
13057 (iwmmxt_wmulwsm): Likewise.
13058 (iwmmxt_wmulwum): Likewise.
13059 (iwmmxt_wmulsmr): Likewise.
13060 (iwmmxt_wmulumr): Likewise.
13061 (iwmmxt_wmulwsmr): Likewise.
13062 (iwmmxt_wmulwumr): Likewise.
13063 (iwmmxt_wmulwl): Likewise.
13064 (iwmmxt_wqmulm): Likewise.
13065 (iwmmxt_wqmulwm): Likewise.
13066 (iwmmxt_wqmulmr): Likewise.
13067 (iwmmxt_wqmulwmr): Likewise.
13068 (iwmmxt_waddbhusm): Likewise.
13069 (iwmmxt_waddbhusl): Likewise.
13070 (iwmmxt_wqmiabb): Likewise.
13071 (iwmmxt_wqmiabt): Likewise.
13072 (iwmmxt_wqmiatb): Likewise.
13073 (iwmmxt_wqmiatt): Likewise.
13074 (iwmmxt_wqmiabbn): Likewise.
13075 (iwmmxt_wqmiabtn): Likewise.
13076 (iwmmxt_wqmiatbn): Likewise.
13077 (iwmmxt_wqmiattn): Likewise.
13078 (iwmmxt_wmiabb): Likewise.
13079 (iwmmxt_wmiabt): Likewise.
13080 (iwmmxt_wmiatb): Likewise.
13081 (iwmmxt_wmiatt): Likewise.
13082 (iwmmxt_wmiabbn): Likewise.
13083 (iwmmxt_wmiabtn): Likewise.
13084 (iwmmxt_wmiatbn): Likewise.
13085 (iwmmxt_wmiattn): Likewise.
13086 (iwmmxt_wmiawbb): Likewise.
13087 (iwmmxt_wmiawbt): Likewise.
13088 (iwmmxt_wmiawtb): Likewise.
13089 (iwmmxt_wmiawtt): Likewise.
13090 (iwmmxt_wmiawbbn): Likewise.
13091 (iwmmxt_wmiawbtn): Likewise.
13092 (iwmmxt_wmiawtbn): Likewise.
13093 (iwmmxt_wmiawttn): Likewise.
13094 (iwmmxt_wmerge): Likewise.
13095 (iwmmxt_tandc<mode>3): Likewise.
13096 (iwmmxt_torc<mode>3): Likewise.
13097 (iwmmxt_torvsc<mode>3): Likewise.
13098 (iwmmxt_textrc<mode>3): Likewise.
13099 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13100 (wmmxt_pack): Likewise.
13101 (wmmxt_mult_c1): Likewise.
13102 (wmmxt_mult_c2): Likewise.
13103 (wmmxt_alu_c1): Likewise.
13104 (wmmxt_alu_c2): Likewise.
13105 (wmmxt_alu_c3): Likewise.
13106 (wmmxt_transfer_c1): Likewise.
13107 (wmmxt_transfer_c2): Likewise.
13108 (wmmxt_transfer_c3): Likewise.
13109 (marvell_f_iwmmxt_wstr): Likewise.
13110 (marvell_f_iwmmxt_wldr): Likewise.
13112 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13114 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13116 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13119 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13120 * lra-constraints.c (need_for_split_p): Check call used hard regs
13121 living through calls.
13123 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13124 call used regs for call insn.
13126 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13129 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13130 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13132 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13134 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13136 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13138 * lra-constraints.c (need_for_split_p): Check call used hard regs
13139 living through calls.
13141 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13144 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13145 to tie with any other modes. Eliminate Altivec vector mode tests,
13146 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13147 code, to return 0 if testing MODE2 for a condition, if we've
13148 already tested MODE1 for the same condition.
13150 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13152 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13155 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13157 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13158 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13160 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13162 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13164 (aarch64_symbolic_constant_p): Remove.
13165 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13166 static. Fix line length and white space.
13167 (aarch64_symbolic_constant_p): Remove.
13168 * config/aarch64/predicates.md (aarch64_valid_symref):
13169 Use aarch64_classify_symbol_expression.
13171 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13173 * config/arm/constraints.md (Ts): New constraint.
13174 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13176 (compare_scc): Use "Ts" constraint for operand 0.
13177 (ior_scc_scc): Likewise.
13178 (and_scc_scc): Likewise.
13179 (and_scc_scc_nodom): Likewise.
13180 (ior_scc_scc_cmp): Likewise for operand 7.
13181 (and_scc_scc_cmp): Likewise.
13182 * config/arm/thumb2.md (thumb2_movsi_insn):
13183 Add alternatives for 16-bit encodings.
13184 (thumb2_movhi_insn): Likewise.
13185 (thumb2_movsicc_insn): Likewise.
13186 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13187 (thumb2_negscc): Use "Ts" constraint.
13188 Move mvn instruction outside cond_exec block.
13189 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13190 for 16-bit encodings.
13192 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13194 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13196 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13197 (mulsi3subsi): Likewise.
13198 (mulsidi3adddi): Likewise.
13199 (mulsidi3_v6): Likewise.
13200 (umulsidi3_v6): Likewise.
13201 (umulsidi3adddi_v6): Likewise.
13202 (smulsi3_highpart_v6): Likewise.
13203 (umulsi3_highpart_v6): Likewise.
13204 (mulhisi3tb): Likewise.
13205 (mulhisi3bt): Likewise.
13206 (mulhisi3tt): Likewise.
13207 (maddhisi4): Likewise.
13208 (maddhisi4tb): Likewise.
13209 (maddhisi4tt): Likewise.
13210 (maddhidi4): Likewise.
13211 (maddhidi4tb): Likewise.
13212 (maddhidi4tt): Likewise.
13213 (zeroextractsi_compare0_scratch): Likewise.
13214 (insv_zero): Likewise.
13215 (insv_t2): Likewise.
13216 (anddi_notzesidi_di): Likewise.
13217 (anddi_notsesidi_di): Likewise.
13218 (andsi_notsi_si): Likewise.
13219 (iordi_zesidi_di): Likewise.
13220 (xordi_zesidi_di): Likewise.
13221 (andsi_iorsi3_notsi): Likewise.
13222 (smax_0): Likewise.
13223 (smax_m1): Likewise.
13224 (smin_0): Likewise.
13225 (not_shiftsi): Likewise.
13226 (unaligned_loadsi): Likewise.
13227 (unaligned_loadhis): Likewise.
13228 (unaligned_loadhiu): Likewise.
13229 (unaligned_storesi): Likewise.
13230 (unaligned_storehi): Likewise.
13231 (extv_reg): Likewise.
13232 (extzv_t2): Likewise.
13233 (divsi3): Likewise.
13234 (udivsi3): Likewise.
13235 (arm_zero_extendhisi2addsi): Likewise.
13236 (arm_zero_extendqisi2addsi): Likewise.
13237 (compareqi_eq0): Likewise.
13238 (arm_extendhisi2_v6): Likewise.
13239 (arm_extendqisi2addsi): Likewise.
13240 (arm_movt): Likewise.
13241 (thumb2_ldrd): Likewise.
13242 (thumb2_ldrd_base): Likewise.
13243 (thumb2_ldrd_base_neg): Likewise.
13244 (thumb2_strd): Likewise.
13245 (thumb2_strd_base): Likewise.
13246 (thumb2_strd_base_neg): Likewise.
13247 (arm_negsi2): Add alternative for 16-bit encoding.
13248 (arm_one_cmplsi2): Likewise.
13250 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13252 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13253 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13254 (movdfcc): Likewise.
13255 * config/arm/vfp.md (*thumb2_movsf_vfp):
13256 Disable predication for arm_restrict_it.
13257 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13258 (*thumb2_movdfcc_vfp): Likewise.
13259 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13260 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13261 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13262 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13263 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13264 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13265 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13266 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13267 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13268 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13269 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13270 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13271 Disable predication for arm_restrict_it.
13273 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13275 * config/i386/bmiintrin.h (_bextr_u32): New.
13276 (_bextr_u64): Ditto.
13278 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13280 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13281 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13282 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13283 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13284 * config/mips/n32-elf.h: ...this new file.
13286 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13289 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13290 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13292 2013-06-27 Catherine Moore <clm@codesourcery.com>
13294 * config/mips/mips-tables.opt: Regenerate.
13295 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13296 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13297 * doc/invoke.texi: Add -m14kc.
13299 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13302 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13303 constraints of operand 1 and 2.
13306 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13307 to match RTL canonicalization. Swap predicates and
13308 constraints of operand 1 and 2.
13310 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13312 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13313 Process OP_INOUT regs for splitting too.
13315 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13317 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13318 decl before the loop, initialize to NULL.
13319 (vectorizable_load): Initialize ptr_incr to NULL.
13321 2013-06-27 Martin Jambor <mjambor@suse.cz>
13324 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13325 * ipa-ref.c (ipa_maybe_record_reference): New function.
13326 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13327 * ipa-cp.c (create_specialized_node): Record potential references from
13329 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13331 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13333 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13334 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13335 'mode' to force_reg.
13336 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13337 (aarch64_expand_mov_immediate): Likewise.
13339 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13341 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13342 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13344 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13346 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13347 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13348 (struct machine_function): Add tbegin_p.
13349 (s390_canonicalize_comparison): Fold CC mode compares to
13350 conditional jump if possible.
13351 (s390_emit_jump): Return the emitted jump.
13352 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13353 Handle CCRAWmode compares.
13354 (s390_option_override): Default to -mhtm if available.
13355 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13356 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13357 FPRs instead of df_regs_ever_live_p.
13358 (s390_optimize_nonescaping_tx): New function.
13359 (s390_init_frame_layout): Extend clobbered_regs array to cover
13361 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13362 (s390_expand_tbegin): New function.
13363 (enum s390_builtin): New enum definition.
13364 (code_for_builtin): New array definition.
13365 (s390_init_builtins): New function.
13366 (s390_expand_builtin): New function.
13367 (TARGET_INIT_BUILTINS): Define.
13368 (TARGET_EXPAND_BUILTIN): Define.
13369 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13370 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13371 (s390_alc_comparison): Likewise.
13372 * config/s390/s390-modes.def: Add CCRAWmode.
13373 * config/s390/s390.h (processor_flags): Add PF_TX.
13374 (TARGET_CPU_HTM): Define macro.
13375 (TARGET_HTM): Define macro.
13376 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13377 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13378 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13379 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13380 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13381 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13382 ("*cc_to_int"): Move up.
13383 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13384 constants other than 0.
13385 ("*ccraw_to_int"): New insn and splitter definition.
13386 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13387 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13388 ("tx_assist"): New expander.
13389 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13390 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13391 * config/s390/s390.opt: Add -mhtm option.
13392 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13393 * config/s390/htmxlintrin.h: New file.
13394 * config/s390/htmintrin.h: New file.
13395 * config/s390/s390intrin.h: New file.
13396 * doc/extend.texi: Document htm builtins.
13397 * config.gcc: Add the new header files to extra_headers.
13399 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13401 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13402 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13404 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13405 Pat Haugen <pthaugen@us.ibm.com>
13406 Peter Bergner <bergner@vnet.ibm.com>
13408 * config/rs6000/power8.md: New.
13409 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13410 setting for power8 entry.
13411 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13412 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13413 test for Power4/Power5 only.
13414 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13416 (force_new_group): Adjust comment.
13417 * config/rs6000/rs6000.md: Include power8.md.
13419 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13421 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13422 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13424 (tune_params): Update comment.
13425 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13426 (arm_max_conditional_execute): New function.
13427 (thumb2_final_prescan_insn): Use max_insn_skipped and
13428 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13430 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13432 PR tree-optimization/57705
13433 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13434 SSA_NAME step, provided that it is not defined inside the loop.
13435 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13436 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13438 2013-06-25 Martin Jambor <mjambor@suse.cz>
13440 PR middle-end/57670
13441 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13442 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13444 (ipa_note_param_call): Initialize member_ptr flag.
13445 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13446 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13447 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13448 (ipa_read_indirect_edge_info): Likewise.
13450 2013-06-25 Richard Biener <rguenther@suse.de>
13452 PR middle-end/56977
13453 * passes.c (init_optimization_passes): Move pass_fold_builtins
13454 and pass_dce earlier with -Og.
13456 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13458 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13459 <BIT_FIELD_REF>: Remove trailing TAB.
13460 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13463 2013-06-24 Martin Jambor <mjambor@suse.cz>
13465 PR tree-optimization/57358
13466 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13467 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13468 (ipa_analyze_params_uses): Generate pessimistic info when true.
13470 2013-06-24 Martin Jambor <mjambor@suse.cz>
13472 PR tree-optimization/57539
13473 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13474 global.inlined_to of the new node to it. All callers changed.
13475 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13476 inlining_into, pass it to cgraph_clone_node.
13477 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13478 ipa_free_edge_args_substructures.
13479 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13480 rdesc linked list. Do not assert rdesc edges have inlined caller.
13481 Assert we have found an rdesc in the rdesc list.
13483 2013-06-24 Richard Biener <rguenther@suse.de>
13485 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13486 (pointer_set_lookup): Declare.
13487 (class pointer_map): New template class implementing a
13488 generic pointer to T map.
13489 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13490 pointer_map<T>::contains, pointer_map<T>::insert,
13491 pointer_map<T>::traverse): New functions.
13492 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13493 (pointer_set_lookup): New function.
13494 (pointer_set_contains): Use pointer_set_lookup.
13495 (pointer_set_insert): Likewise.
13496 (insert_aux): Remove.
13497 (struct pointer_map_t): Embed a pointer_set_t.
13498 (pointer_map_create): Adjust.
13499 (pointer_map_destroy): Likewise.
13500 (pointer_map_contains): Likewise.
13501 (pointer_map_insert): Likewise.
13502 (pointer_map_traverse): Likewise.
13503 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13504 pointer_map<unsigned> instead of a pointer_map_t.
13505 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13506 (streamer_tree_cache_lookup): Likewise.
13507 (streamer_tree_cache_create): Likewise.
13508 (streamer_tree_cache_delete): Likewise.
13509 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13510 pointer_map<unsigned> instead of a pointer_map_t.
13511 (lto_init_tree_ref_encoder): Adjust.
13512 (lto_destroy_tree_ref_encoder): Likewise.
13513 * lto-section-out.c (lto_output_decl_index): Likewise.
13514 (lto_record_function_out_decl_state): Likewise.
13515 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13517 2013-06-24 Richard Biener <rguenther@suse.de>
13519 PR tree-optimization/57488
13520 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13522 2013-06-24 Alan Modra <amodra@gmail.com>
13524 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13525 (gen_easy_altivec_constant): Likewise.
13526 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13527 easy_vector_constant_msb): Likewise.
13529 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13532 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13533 add missing return true.
13535 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13538 * config/sh/predicates.md (general_extend_operand): Invoke
13539 general_movsrc_operand for memory operands.
13540 (general_movsrc_operand): Allow reg+reg addressing, do not use
13541 general_operand for memory operands.
13543 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13545 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13546 when current target options does not apply.
13547 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13548 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13549 * config/i386/bmiintrin.h: Pass appropriate target
13550 attributes to header.
13551 * config/i386/mmintrin.h: Ditto.
13552 * config/i386/nmmintrin.h: Ditto.
13553 * config/i386/avx2intrin.h: Ditto.
13554 * config/i386/fxsrintrin.h: Ditto.
13555 * config/i386/tbmintrin.h: Ditto.
13556 * config/i386/xsaveintrin.h: Ditto.
13557 * config/i386/f16cintrin.h: Ditto.
13558 * config/i386/xtestintrin.h: Ditto.
13559 * config/i386/xsaveoptintrin.h: Ditto.
13560 * config/i386/bmi2intrin.h: Ditto.
13561 * config/i386/lzcntintrin.h: Ditto.
13562 * config/i386/smmintrin.h: Ditto.
13563 * config/i386/wmmintrin.h: Ditto.
13564 * config/i386/x86intrin.h: Remove all header include guards.
13565 * config/i386/prfchwintrin.h: Ditto.
13566 * config/i386/pmmintrin.h: Ditto.
13567 * config/i386/tmmintrin.h: Ditto.
13568 * config/i386/xmmintrin.h: Ditto.
13569 * config/i386/popcntintrin.h: Ditto.
13570 * config/i386/rdseedintrin.h: Ditto.
13571 * config/i386/ammintrin.h: Ditto.
13572 * config/i386/emmintrin.h: Ditto.
13573 * config/i386/immintrin.h: Remove all header include guards.
13574 * config/i386/fma4intrin.h: Ditto.
13575 * config/i386/lwpintrin.h: Ditto.
13576 * config/i386/xopintrin.h: Ditto.
13577 * config/i386/ia32intrin.h: Ditto.
13578 * config/i386/avxintrin.h: Ditto.
13579 * config/i386/rtmintrin.h: Ditto.
13580 * config/i386/fmaintrin.h: Ditto.
13581 * config/i386/mm3dnow.h: Ditto.
13583 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13585 * common/config/i386/i386-common.c: Handle LZCNT.
13587 2013-06-22 Andi Kleen <ak@linux.intel.com>
13589 * doc/extend.texi: Use __atomic_store_n instead of
13590 __atomic_store in HLE example.
13592 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13594 * config/sh/sh.c: Remove <cstdlib> workaround.
13596 2013-06-21 Andi Kleen <ak@linux.intel.com>
13598 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13600 2013-06-21 Andi Kleen <ak@linux.intel.com>
13602 * doc/extend.texi: Document that __atomic_clear and
13603 __atomic_test_and_set should only be used with bool.
13605 2013-06-20 Jan Hubicka <jh@suse.cz>
13607 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13608 types_same_for_odr.
13609 * tree.c (decls_same_for_odr): New function.
13610 (same_for_edr): New function.
13611 (types_same_for_odr): New function.
13612 (get_binfo_at_offset): Use it.
13613 * tree.h (types_same_for_odr): Declare.
13615 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13616 Jason Merrill <jason@redhat.com>
13618 * system.h: Include <cstdlib> as well as <stdlib.h>.
13620 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13623 * config/i386/i386.c (construct_container): Report error if
13624 long double is used with disabled x87 float returns.
13626 2013-06-20 Jan Hubicka <jh@suse.cz>
13628 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13630 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13632 PR rtl-optimization/57425
13633 PR rtl-optimization/57569
13634 * alias.c (write_dependence_p): Remove parameters mem_mode and
13635 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13636 Changed all callers.
13637 (canon_anti_dependence): Get comments and semantics in sync.
13638 Add parameter mem_canonicalized. Changed all callers.
13639 * rtl.h (canon_anti_dependence): Update prototype.
13641 2013-06-20 Richard Biener <rguenther@suse.de>
13643 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13644 byte case, inline streamer_read_uchar and defer section
13647 2013-06-20 Richard Biener <rguenther@suse.de>
13649 PR tree-optimization/57584
13650 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13651 SSA names into the expanded expression that take part in
13652 abnormal coalescing.
13654 2013-06-19 Sharad Singhai <singhai@google.com>
13656 * gcov.c (print_usage): Handle new option.
13657 (process_args): Ditto.
13658 (get_gcov_intermediate_filename): New function.
13659 (output_intermediate_file): New function.
13660 (output_gcov_file): New function
13661 (generate_results): Handle new option.
13662 (release_function): Relase demangled name.
13663 (read_graph_file): Handle demangled name.
13664 (output_lines): Ditto.
13665 * doc/gcov.texi: Document gcov intermediate format.
13667 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13670 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13671 (lra_emit_add): Use the functions. Add comment about Y as an
13674 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13677 * collect2.c (collect_atexit): New.
13678 (collect_exit): Delete.
13679 (main): Register collect_atexit with atexit.
13680 (collect_wait): Change collect_exit to exit.
13682 * collect2.h (collect_exit): Delete.
13683 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13685 2013-06-19 Wei Mi <wmi@google.com>
13687 PR rtl-optimization/57518
13688 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13689 if regno is used in paradoxical subreg.
13690 (update_equiv_regs): Check pdx_subregs[regno] before
13691 set a reg to be equivalent with a mem.
13693 2013-06-19 Matthias Klose <doko@ubuntu.com>
13696 * file-find.h (find_a_file): Add a mode parameter.
13697 * file-find.c (find_a_file): Likewise.
13698 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13699 with X_OK for the executables.
13700 * collect2.c (main): Call find_a_file with X_OK.
13702 2013-06-19 Steve Ellcey <sellcey@mips.com>
13705 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13706 Use NEXT_INSN instead of next_real_insn.
13708 2013-06-19 Jan Hubicka <jh@suse.cz>
13710 * cgraph.h (const_value_known_p): Replace by ...
13711 (ctor_for_folding): .. this one.
13712 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13713 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13714 * expr.c (expand_expr_real_1): Likewise.
13715 (string_constant): Likewise.
13716 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13717 * ipa.c (process_references): Likewise.
13718 (symtab_remove_unreachable_nodes): Likewise.
13719 * ipa-inline-analysis.c (param_change_prob): Likewise.
13720 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13721 (get_base_constructor): Likwise.
13722 * varpool.c (varpool_remove_node): Likewise.
13723 (varpool_remove_initializer): LIkewise.
13724 (dump_varpool_node): LIkwise.
13725 (const_value_known_p): Rewrite to ...
13726 (ctor_for_folding): ... this one.
13728 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13731 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13732 PERSONALITY in $PATH derived prefixes.
13734 2013-06-19 Jeff Law <law@redhat.com>
13736 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13739 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13740 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13743 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13745 * config/arm/vfp.md: Move VFP instruction classification documentation
13747 * config/arm/arm.md: ... here. Update instruction classification
13750 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13752 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13753 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13754 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13756 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13759 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13760 with NEXT_INSN. Use tablejump_p to check for jump table data
13763 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13766 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13767 that now in C++ the value is correct per the C++ standards.
13769 2013-06-19 Richard Biener <rguenther@suse.de>
13771 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13772 for global context.
13774 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13777 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13780 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13781 with next_active_insn.
13783 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13785 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13786 functions are inlined during failures to flag an error.
13787 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13788 in early inline pass.
13790 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13792 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13795 2013-06-18 Julian Brown <julian@codesourcery.com>
13797 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13798 Permit virtual register pre-reload if !strict.
13799 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13801 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13803 * config/arm/neon.md (movmisalign<mode>): Use
13804 neon_perm_struct_or_reg_operand instead of
13805 neon_struct_or_register_operand.
13806 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13807 neon_permissive_struct_operand instead of neon_struct_operand.
13808 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13809 neon_vector_mem_operand.
13810 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13811 neon_vector_mem_operand.
13812 (neon_permissive_struct_operand): New.
13813 (neon_struct_or_register_operand): Rename to...
13814 (neon_perm_struct_or_reg_operand): This. Adjust call to
13815 neon_vector_mem_operand.
13817 2013-06-18 Richard Biener <rguenther@suse.de>
13819 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13820 * lto-streamer.h: Include pointer-set.h.
13821 (struct lto_decl_slot): Remove.
13822 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13823 Remove next_index entry.
13824 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13825 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13826 (lto_init_tree_ref_encoder): Adjust.
13827 (lto_destroy_tree_ref_encoder): Likewise.
13828 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13829 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13830 (lto_output_decl_index): Adjust.
13831 (lto_new_out_decl_state): Likewise.
13832 (lto_record_function_out_decl_state): Likewise.
13833 * lto-streamer-out.c (copy_function): Likewise.
13835 2013-06-18 Richard Biener <rguenther@suse.de>
13837 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13838 * cgraphunit.c: Include cfgloop.h.
13839 (init_lowered_empty_function): Initialize the loop tree.
13840 (assemble_thunk): Insert new BBs into loops.
13842 2013-06-18 Richard Biener <rguenther@suse.de>
13844 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13845 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13846 the map from cache entry to cache index optional.
13847 (streamer_tree_cache_replace_tree): Adjust accordingly.
13848 (streamer_tree_cache_append): Likewise.
13849 (streamer_tree_cache_delete): Likewise.
13850 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13851 streamer cache map from cache entry to cache index.
13852 * lto-streamer-out.c (create_output_block): Adjust.
13854 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13856 * config/arm/arm.md (attribute "insn"): Move multiplication and
13857 division attributes to...
13858 (attribute "type"): ... here. Remove mult.
13859 (attribute "mul32"): New attribute.
13860 (attribute "mul64"): Add umaal.
13861 (*arm_mulsi3): Update attributes.
13862 (*arm_mulsi3_v6): Likewise.
13863 (*thumb_mulsi3): Likewise.
13864 (*thumb_mulsi3_v6): Likewise.
13865 (*mulsi3_compare0): Likewise.
13866 (*mulsi3_compare0_v6): Likewise.
13867 (*mulsi_compare0_scratch): Likewise.
13868 (*mulsi_compare0_scratch_v6): Likewise.
13869 (*mulsi3addsi): Likewise.
13870 (*mulsi3addsi_v6): Likewise.
13871 (*mulsi3addsi_compare0): Likewise.
13872 (*mulsi3addsi_compare0_v6): Likewise.
13873 (*mulsi3addsi_compare0_scratch): Likewise.
13874 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13875 (*mulsi3subsi): Likewise.
13876 (*mulsidi3adddi): Likewise.
13877 (*mulsi3addsi_v6): Likewise.
13878 (*mulsidi3adddi_v6): Likewise.
13879 (*mulsidi3_nov6): Likewise.
13880 (*mulsidi3_v6): Likewise.
13881 (*umulsidi3_nov6): Likewise.
13882 (*umulsidi3_v6): Likewise.
13883 (*umulsidi3adddi): Likewise.
13884 (*umulsidi3adddi_v6): Likewise.
13885 (*smulsi3_highpart_nov6): Likewise.
13886 (*smulsi3_highpart_v6): Likewise.
13887 (*umulsi3_highpart_nov6): Likewise.
13888 (*umulsi3_highpart_v6): Likewise.
13889 (mulhisi3): Likewise.
13890 (*mulhisi3tb): Likewise.
13891 (*mulhisi3bt): Likewise.
13892 (*mulhisi3tt): Likewise.
13893 (maddhisi4): Likewise.
13894 (*maddhisi4tb): Likewise.
13895 (*maddhisi4tt): Likewise.
13896 (maddhidi4): Likewise.
13897 (*maddhidi4tb): Likewise.
13898 (*maddhidi4tt): Likewise.
13899 (divsi3): Likewise.
13900 (udivsi3): Likewise.
13901 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13902 (thumb2_mulsi_short_compare0): Likewise.
13903 (thumb2_mulsi_short_compare0_scratch): Likewise.
13904 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13905 (1020mult2): Likewise.
13906 (1020mult3): Likewise.
13907 (1020mult4): Likewise.
13908 (1020mult5): Likewise.
13909 (1020mult6): Likewise.
13910 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13912 (cortex_a15_mult64): Likewise.
13913 (cortex_a15_sdiv): Likewise.
13914 (cortex_a15_udiv): Likewise.
13915 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13921 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13922 (pj4_ir_div): Likewise.
13923 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13924 (11_mult2): Likewise.
13925 (11_mult3): Likewise.
13926 (11_mult4): Likewise.
13927 (11_mult5): Likewise.
13928 (11_mult6): Likewise.
13929 (11_mult7): Likewise.
13930 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13931 (cortex_a8_mla): Likewise.
13932 (cortex_a8_mull): Likewise.
13933 (cortex_a8_smulwy): Likewise.
13934 (cortex_a8_smlald): Likewise.
13935 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13936 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13937 (cortex_r4_mul_3): Likewise.
13938 (cortex_r4_mla_4): Likewise.
13939 (cortex_r4_mla_3): Likewise.
13940 (cortex_r4_smlald): Likewise.
13941 (cortex_r4_mull): Likewise.
13942 (cortex_r4_sdiv): Likewise.
13943 (cortex_r4_udiv): Likewise.
13944 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13945 (cortex_a7_idiv): Likewise.
13946 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13947 (9_mult2): Likewise.
13948 (9_mult3): Likewise.
13949 (9_mult4): Likewise.
13950 (9_mult5): Likewise.
13951 (9_mult6): Likewise.
13952 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13953 (cortex_a53_sdiv): Likewise.
13954 (cortex_a53_udiv): Likewise.
13955 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13956 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13957 (mp626_mult2): Likewise.
13958 (mp626_mult3): Likewise.
13959 (mp626_mult4): Likewise.
13960 * config/arm/fa526.md (526_mult1): Update attribute change.
13961 (526_mult2): Likewise.
13962 * config/arm/arm-generic.md (mult): Update attribute change.
13963 (mult_ldsched_strongarm): Likewise.
13964 (mult_ldsched): Likewise.
13965 (multi_cycle): Likewise.
13966 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13967 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13968 (606te_mult2): Likewise.
13969 (606te_mult3): Likewise.
13970 (606te_mult4): Likewise.
13971 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13972 (cortex_a9_mac16): Likewise.
13973 (cortex_a9_multiply): Likewise.
13974 (cortex_a9_mac): Likewise.
13975 (cortex_a9_multiply_long): Likewise.
13976 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13977 (626te_mult2): Likewise.
13978 (626te_mult3): Likewise.
13979 (626te_mult4): Likewise.
13981 2013-06-18 Richard Biener <rguenther@suse.de>
13984 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13986 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13989 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13990 with next_active_insn.
13992 2013-06-18 Alan Modra <amodra@gmail.com>
13994 * config/rs6000/rs6000.h (enum data_align): New.
13995 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13996 (DATA_ABI_ALIGNMENT): Define.
13997 (CONSTANT_ALIGNMENT): Correct comment.
13998 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13999 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14001 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14003 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14004 ATTRIBUTE_UNUSED marking.
14006 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14008 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14009 alternative and update.
14010 (aarch64_dup_lanedi): Delete.
14011 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14012 * config/aarch64/aarch64-simd-builtins.def: Update.
14014 2013-06-17 Richard Biener <rguenther@suse.de>
14016 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14017 (lto_input_scc): Declare.
14018 (lto_input_tree_1): Likewise.
14019 (struct lto_stats_d): Add num_tree_bodies_output and
14020 num_pickle_refs_output.
14021 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14022 (lto_read_tree_1): Split out from ...
14023 (lto_read_tree): ... this.
14024 (lto_input_scc): New function.
14025 (lto_input_tree_1): Split out from ...
14026 (lto_input_tree): ... this. Handle LTO_tree_scc.
14027 (lto_data_in_create): Create the streamer cache without hashes.
14028 * lto-streamer-out.c (create_output_block): Create the streamer
14029 cache with hashes when not doing WPA.
14030 (lto_write_tree_1): Split out from ...
14031 (lto_write_tree): ... this.
14032 (get_symbol_initial_value): New function.
14033 (lto_output_tree_1): Split out from ...
14034 (lto_output_tree): ... this. Write trees as series of SCCs
14035 using a DFS walk via DFS_write_tree.
14036 (struct sccs, struct scc_entry): New types.
14037 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14038 (DFS_write_tree_body): New function.
14039 (DFS_write_tree): Likewise.
14040 (hash_tree): Likewise.
14041 (scc_entry_compare): Likewise.
14042 (hash_scc): Likewise.
14043 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14044 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14045 TREE_CHAIN as regular reference.
14046 (streamer_read_integer_cst): Remove.
14047 (streamer_get_pickled_tree): Adjust.
14048 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14049 of DECL_EXTERNALs in BLOCK_VARS for now.
14050 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14052 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14053 Add hash value argument and record that if hashes are recorded
14055 (streamer_tree_cache_insert_1): Adjust.
14056 (streamer_tree_cache_insert): Likewise.
14057 (streamer_tree_cache_insert_at): Rename to ...
14058 (streamer_tree_cache_replace_tree): ... this and adjust.
14059 (streamer_tree_cache_append): Adjust.
14060 (record_common_node): Likewise.
14061 (streamer_tree_cache_create): Add argument whether to
14062 record hash values together with trees.
14063 (streamer_tree_cache_delete): Adjust.
14064 * tree-streamer.h (struct streamer_tree_cache_d): Add
14066 (streamer_read_integer_cst): Remove.
14067 (streamer_tree_cache_insert): Adjust.
14068 (streamer_tree_cache_append): Likewise.
14069 (streamer_tree_cache_insert_at): Rename to ...
14070 (streamer_tree_cache_replace_tree): ... this and adjust.
14071 (streamer_tree_cache_create): Add argument whether to record hashes.
14072 (streamer_tree_cache_get): Rename to ...
14073 (streamer_tree_cache_get_tree): ... this.
14074 (streamer_tree_cache_get_hash): New function.
14075 * tree.c (cache_integer_cst): New function.
14076 * tree.h (cache_integer_cst): Declare.
14077 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14078 * lto-symtab.c (lto_varpool_replace_node): Only release
14079 DECL_INITIAL of non-prevailing decls.
14080 * varpool.c (varpool_remove_initializer): Do not release
14081 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14083 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14085 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14086 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14087 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14088 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14089 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14090 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14091 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14092 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14093 instead of TARGET_64BIT.
14094 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14095 Require ISA_HAS_<D>DIV.
14097 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14099 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14100 (mips*-*-linux*): Move default with_llsc setting to where other
14102 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14104 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14105 Likewise. Remove default with_tune setting. Move default float
14106 setting to its own block. Handle with_llsc in the same block as above.
14108 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14110 PR rtl-optimization/57425
14111 PR rtl-optimization/57569
14112 * alias.c (write_dependence_p): Add new parameters mem_mode,
14113 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14114 Changed all callers.
14115 (canon_anti_dependence): New function.
14116 * cse.c (check_dependence): Use canon_anti_dependence.
14117 * cselib.c (cselib_invalidate_mem): Likewise.
14118 * rtl.h (canon_anti_dependence): Declare.
14120 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14122 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14123 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14124 ".set mips3" for 64-bit targets.
14126 2013-06-15 Dehao Chen <dehao@google.com>
14128 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14129 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14130 (gimple_check_call_args): Likewise.
14131 * value-prof.c (check_ic_target): Likewise.
14132 * ipa-inline.c (early_inliner): Likewise.
14133 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14134 * cgraph.c (cgraph_create_edge_1): Likewise.
14135 (cgraph_make_edge_direct): Likewise.
14137 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14140 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14141 rs6000_output_move_128bit to handle emitting quad memory
14142 operations. Set attribute length to 8 bytes.
14144 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14146 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14148 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14149 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14150 (aarch64_<su>mlsl<mode>): Likewise.
14152 2013-06-14 Mike Stump <mikestump@comcast.net>
14154 * Makefile.in (TARGET_H): Add insn-codes.h.
14156 2013-06-14 Alan Modra <amodra@gmail.com>
14158 PR middle-end/57134
14159 PR middle-end/57586
14160 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14161 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14162 bitfield expansion when EXPAND_MEMORY.
14163 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14165 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14167 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14168 test for clearing quad memory on 32-bit later.
14170 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14172 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14173 (fold_negate_expr): Likewise.
14174 (fold_real_zero_addition_p): Handle vectors.
14175 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14177 2013-06-14 Alan Modra <amodra@gmail.com>
14179 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14181 2013-06-13 Jan Hubicka <jh@suse.cz>
14183 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14184 Local comdats are not externally visible.
14185 * symtab.c (dump_symtab_base): Dump externally visible.
14186 (verify_symtab_base): Verify back links in the symtab hash.
14188 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14190 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14191 CONVERT_EXPR as equal nodes.
14193 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14195 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14197 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14199 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14200 Generalize to complex and vector.
14201 * tree.c (build_all_ones_cst): New function.
14202 * tree.h (build_all_ones_cst): Declare it.
14204 2013-06-13 Alan Modra <amodra@gmail.com>
14206 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14207 * config/rs6000/rs6000.md (signbittf2): New insn.
14208 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14209 (abstf2_internal, cmptf_internal2): Likewise.
14210 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14212 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14213 Pat Haugen <pthaugen@us.ibm.com>
14214 Peter Bergner <bergner@vnet.ibm.com>
14216 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14217 power8 byte, half-word, and quad-word atomic instructions.
14218 (emit_store_conditional): Likewise.
14219 (rs6000_expand_atomic_compare_and_swap): Likewise.
14220 (rs6000_expand_atomic_op): Likewise.
14222 * config/rs6000/sync.md (larx): Add new modes for power8.
14224 (AINT): New mode iterator to include TImode as well as normal
14225 integer modes on power8.
14226 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14227 that VSX registers are not considered. Use AINT mode iterator
14228 instead of INT1 to allow inclusion of quad word atomic operations
14230 (load_locked<mode>): Likewise.
14231 (store_conditional<mode>): Likewise.
14232 (atomic_compare_and_swap<mode>): Likewise.
14233 (atomic_exchange<mode>): Likewise.
14234 (atomic_nand<mode>): Likewise.
14235 (atomic_fetch_<fetchop_name><mode>): Likewise.
14236 (atomic_nand_fetch<mode>): Likewise.
14237 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14239 (ATOMIC): On power8, add QImode, HImode modes.
14240 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14241 modes that promote to SImode.
14242 (load_lockedti): Convert TImode arguments to PTImode, so that we
14243 get a guaranteed even/odd register pair.
14244 (load_lockedpti): Likewise.
14245 (store_conditionalti): Likewise.
14246 (store_conditionalpti): Likewise.
14248 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14249 atomic load/store instructions.
14252 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14254 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14256 (insn_count): New attribute, with most cases extracted from...
14257 (length): ...here. Redefine most cases in terms of insn_count.
14258 (single_insn): Delete.
14259 (can_delay): Use insn_count to check for single instructions.
14260 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14261 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14262 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14263 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14264 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14265 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14266 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14267 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14268 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14269 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14270 rather than "length".
14271 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14272 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14273 Use "insn_count" rather than "length".
14274 * config/mips/mips-dsp.md
14275 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14276 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14279 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14281 PR tree-optimization/57361
14282 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14284 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14286 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14288 (aarch64_simd_combine<mode>): New instruction expansion.
14289 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14290 function prototype.
14291 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14292 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14294 2013-06-12 Jan Hubicka <jh@suse.cz>
14296 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14297 decl has when in streaming stage.
14298 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14299 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14301 2013-06-12 Roland Stigge <stigge@antcom.de>
14304 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14306 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14308 PR tree-optimization/57537
14309 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14310 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14312 2013-06-12 Richard Biener <rguenther@suse.de>
14314 * data-streamer.h (streamer_write_char_stream): CSE
14315 obs->current_pointer.
14316 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14317 streamer_write_char_stream manually and optimize the resulting loop.
14318 (streamer_write_hwi_stream): Likewise.
14320 2013-06-12 Jan Hubicka <jh@suse.cz>
14322 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14323 * cgraph.h (varpool_create_empty_node): Declare.
14324 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14326 * symtab.c (symtab_unregister_node): Be lax about missin entries
14328 (symtab_get_node): Update comment.
14329 * varpool.c (varpool_create_empty_node): Break out from ...
14330 (varpool_node_for_decl): ... here.
14331 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14333 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14335 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14336 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14337 part. Use straight-line flow at the end.
14338 <COMPONENT_REF>: Remove superfluous else.
14339 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14341 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14344 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14345 target hook even for !TREE_PUBLIC decls. If no resolution info
14346 is available, return false for common and external decls.
14348 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14350 * config/rl78/constraints.md (U): New constraint.
14351 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14354 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14357 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14358 to allow returning address to AT_PLATFORM name.
14360 2013-06-11 Jan Hubicka <jh@suse.cz>
14362 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14363 * cgraph.h (symtab_node_base): Add weakref flag.
14364 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14365 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14366 (output_weakrefs): Use weakref flag.
14367 * fold-const.c (simple_operand_p): Handle WEAK.
14368 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14369 * ipa.c (varpool_externally_visible_p): Drop weakref.
14370 (function_and_variable_visibility): Update comment; fix weakref
14371 sanity checks; do not clear DECL_WEAK on them.
14372 * lto-cgraph.c (lto_output_node): update.
14373 (lto_output_varpool_node): Update.
14374 (input_overwrite_node): Update.
14375 (input_node): Update.
14376 (input_varpool_node): Update.
14377 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14378 (lto_symtab_merge_symbols): Add sanity check.
14379 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14380 * passes.c (rest_of_decl_compilation): Set static flag, too.
14381 * symtab.c (dump_symtab_base): Dump weakref.
14382 (verify_symtab_base): Sanity check weakrefs.
14383 (symtab_make_decl_local): Remove duplicated code.
14384 (symtab_alias_ultimate_target): Simplify.
14385 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14387 2013-06-11 Tom de Vries <tom@codesourcery.com>
14389 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14390 sequence_vect == NULL.
14392 2013-06-11 DJ Delorie <dj@redhat.com>
14394 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14395 (rl78_unwind_word_mode): New.
14397 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14399 * final.c (debug_prefix_maps): Make static.
14401 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14403 * function.c (initial_trampoline): Remove stray copy.
14405 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14407 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14409 2013-06-11 Martin Jambor <mjambor@suse.cz>
14411 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14412 within bounds at the beginning of the function.
14414 2013-06-11 Alan Modra <amodra@gmail.com>
14416 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14418 (get_named_section): Don't NULL !DECL_P decl.
14420 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14422 * doc/invoke.texi (core-avx2): Document.
14424 (atom): Updated with MOVBE.
14426 2013-06-11 Richard Biener <rguenther@suse.de>
14428 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14430 2013-06-11 Anton Blanchard <anton@samba.org>
14432 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14433 correct shift value in little-endian mode.
14435 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14438 * varasm.c (get_variable_align): Move #endif to the right place.
14440 2013-06-10 Cary Coutant <ccoutant@google.com>
14442 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14443 for hash so that hash table traversal order is deterministic.
14445 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14446 Pat Haugen <pthaugen@us.ibm.com>
14447 Peter Bergner <bergner@vnet.ibm.com>
14449 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14450 of vectors in GPRS if they are direct moves or quad word load or
14453 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14455 (direct_move_p): Likewise.
14456 (quad_load_store_p): Likewise.
14458 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14459 classes into bins based on the physical register type.
14460 (reg_class_to_reg_type): Likewise.
14461 (IS_STD_REG_TYPE): Likewise.
14462 (IS_FP_VECT_REG_TYPE): Likewise.
14463 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14464 use direct move instructions.
14465 (reload_gpr_vsx): Likewise.
14466 (reload_vsx_gpr): Likewise.
14467 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14468 information that is a simplification of register classes. Also
14469 precalculate direct move reload helpers.
14470 (direct_move_p): New function to return true if the operation can
14471 be done as a direct move instruciton.
14472 (quad_load_store_p): New function to return true if the operation
14473 is a quad memory operation.
14474 (rs6000_legitimize_address): If quad memory, only allow register
14475 indirect for TImode addresses.
14476 (rs6000_legitimate_address_p): Likewise.
14477 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14478 (rs6000_reload_register_type): Likewise.
14479 (register_to_reg_type): Return register type.
14480 (rs6000_secondary_reload_simple_move): New helper function for
14481 secondary reload and secondary memory needed to identify anything
14482 that is a simple move, and does not need reloading.
14483 (rs6000_secondary_reload_direct_move): New helper function for
14484 secondary reload to identify cases that can be done with several
14485 instructions via the direct move instructions.
14486 (rs6000_secondary_reload_move): New helper function for secondary
14487 reload to identify moves between register types that can be done.
14488 (rs6000_secondary_reload): Add support for quad memory operations
14489 and for direct move.
14490 (rs6000_secondary_memory_needed): Likewise.
14491 (rs6000_debug_secondary_memory_needed): Change argument names.
14492 (rs6000_output_move_128bit): New function to return the move to
14493 use for 128-bit moves, including knowing about the various
14494 limitations of quad memory operations.
14496 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14497 memory operations. call rs6000_output_move_128bit for the actual
14498 instruciton(s) to generate.
14499 (vsx_movti_64bit): Likewise.
14501 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14502 (UNSPEC_P8V_MTVSRWZ): Likewise.
14503 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14504 (UNSPEC_P8V_MTVSRD): Likewise.
14505 (UNSPEC_P8V_XXPERMDI): Likewise.
14506 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14507 (UNSPEC_FUSION_GPR): Likewise.
14508 (FMOVE128_GPR): New iterator for direct move.
14509 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14510 (f32_sv): Likewise.
14511 (f32_dm): Likewise.
14512 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14513 loads and direct move instructions.
14514 (zero_extendsidi2_lfiwzx): Likewise.
14515 (extendsidi2_lfiwax): Likewise.
14516 (extendsidi2_nocell): Likewise.
14517 (floatsi<mode>2_lfiwax): Likewise.
14518 (lfiwax): Likewise.
14519 (floatunssi<mode>2_lfiwzx): Likewise.
14520 (lfiwzx): Likewise.
14521 (fix_trunc<mode>_stfiwx): Likewise.
14522 (fixuns_trunc<mode>_stfiwx): Likewise.
14523 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14524 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14525 (parity<mode>2_cmpb): Set length/type attr.
14526 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14527 for 'mr.' to fast_compare.
14528 (bpermd_<mode>): Change type attr to popcnt.
14529 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14530 (p8_mtvsrwz_1): Likewise.
14531 (p8_mtvsrwz_2): Likewise.
14532 (reload_fpr_from_gpr<mode>): Likewise.
14533 (p8_mtvsrd_1): Likewise.
14534 (p8_mtvsrd_2): Likewise.
14535 (p8_xxpermdi_<mode>): Likewise.
14536 (reload_vsx_from_gpr<mode>): Likewise.
14537 (reload_vsx_from_gprsf): Likewise.
14538 (p8_mfvsrd_3_<mode>): LIkewise.
14539 (reload_gpr_from_vsx<mode>): Likewise.
14540 (reload_gpr_from_vsxsf): Likewise.
14541 (p8_mfvsrd_4_disf): Likewise.
14542 (multi-word GPR splits): Do not split direct moves or quad memory
14545 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14547 * tree-into-ssa.c (interesting_blocks): Make static.
14549 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14552 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14553 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14554 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14555 (get_variable_align): New function.
14556 (get_variable_section, emit_bss, emit_common,
14557 assemble_variable_contents, place_block_symbol): Use
14558 get_variable_align instead of DECL_ALIGN.
14559 (assemble_noswitch_variable): Add align argument, use it
14560 instead of DECL_ALIGN.
14561 (assemble_variable): Adjust caller. Use get_variable_align
14562 instead of DECL_ALIGN.
14563 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14565 (DATA_ABI_ALIGNMENT): Define.
14566 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14567 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14568 opt is false, only return the psABI mandated alignment increase.
14569 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14570 (DATA_ABI_ALIGNMENT): ... this.
14571 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14572 (DATA_ABI_ALIGNMENT): ... this.
14573 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14574 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14575 (DATA_ABI_ALIGNMENT): ... this.
14576 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14577 * doc/tm.texi: Regenerated.
14579 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14581 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14582 cmp_code to construct REG_EQUAL note.
14584 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14587 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14588 that operands[2] doesn't overlap with operands[0].
14590 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14591 Jan Hubicka <jh@suse.cz>
14593 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14594 hack to mark symbols as used.
14596 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14598 PR rtl-optimization/57559
14599 * lra-constraints.c (process_alt_operands): Don't discourage
14600 memory with known offset for offsetable memory constraint.
14601 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14603 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14605 * varasm.c (struct oc_local_state): Reorder fields.
14606 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14607 and adjust accordingly.
14608 (output_constructor): Reorder initialization code and adjust call to
14609 output_constructor_bitfield.
14611 2013-06-07 Jan Hubicka <jh@suse.cz>
14613 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14615 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14617 * tree-object-size.c (unknown): Make const.
14619 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14621 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14622 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14623 for last alternative in the cpu_facility attribute.
14625 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14628 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14629 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14630 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14631 * config/arm/constraints.md (Dg): New constraint.
14632 * config/arm/neon.md (xordi3_neon): Remove.
14633 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14634 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14636 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14638 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14639 Clean up alternatives.
14641 2013-06-07 Alan Modra <amodra@gmail.com>
14643 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14646 2013-06-07 Alan Modra <amodra@gmail.com>
14648 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14650 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14652 * config/arm/constraints.md (Df): New constraint.
14653 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14654 Correct length attribute for last two alternatives.
14656 2013-06-07 Alan Modra <amodra@gmail.com>
14658 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14659 override user -mfp-in-toc.
14660 (offsettable_ok_by_alignment): Consider just the current access
14661 rather than the whole object, unless BLKmode. Handle
14662 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14663 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14664 for -mcmodel=medium.
14665 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14666 override user -mfp-in-toc or -msum-in-toc. Default to
14667 -mno-fp-in-toc for -mcmodel=medium.
14669 2013-06-06 DJ Delorie <dj@redhat.com>
14671 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14672 TARGET_VALID_POINTER_MODE.
14674 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14675 Pat Haugen <pthaugen@us.ibm.com>
14676 Peter Bergner <bergner@vnet.ibm.com>
14678 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14679 Document new power8 builtins.
14681 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14682 condition code register, to allow 128-bit logical operations to be
14683 done in the VSX or GPR registers.
14684 (nor<mode>3): Use the canonical form for nor.
14685 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14686 vclz*, and vpopcnt* vector instructions.
14687 (nand<mode>3): Likewise.
14688 (orc<mode>3): Likewise.
14689 (clz<mode>2): LIkewise.
14690 (popcount<mode>2): Likewise.
14692 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14693 that only the GPRs are recognized.
14695 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14696 support for new power8 builtins.
14698 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14700 (xscvdpspn): Likewise.
14706 (vpopcnt): Likewise.
14707 (vpopcntb): Likewise.
14708 (vpopcnth): Likewise.
14709 (vpopcntw): Likewise.
14710 (vpopcntd): Likewise.
14712 (vmrgew): Likewise.
14713 (vmrgow): Likewise.
14715 (eqv_v16qi3): Likewise.
14716 (eqv_v8hi3): Likewise.
14717 (eqv_v4si3): Likewise.
14718 (eqv_v2di3): Likewise.
14719 (eqv_v4sf3): Likewise.
14720 (eqv_v2df3): Likewise.
14722 (nand_v16qi3): Likewise.
14723 (nand_v8hi3): Likewise.
14724 (nand_v4si3): Likewise.
14725 (nand_v2di3): Likewise.
14726 (nand_v4sf3): Likewise.
14727 (nand_v2df3): Likewise.
14729 (orc_v16qi3): Likewise.
14730 (orc_v8hi3): Likewise.
14731 (orc_v4si3): Likewise.
14732 (orc_v2di3): Likewise.
14733 (orc_v4sf3): Likewise.
14734 (orc_v2df3): Likewise.
14736 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14737 allow power8 quad mode in 64-bit.
14738 (rs6000_builtin_vectorized_function): Add support to vectorize
14739 ISA 2.07 count leading zeros, population count builtins.
14740 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14741 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14742 (builtin_function_type): Add vgbbd builtin function which takes an
14744 (altivec_expand_vec_perm_const): Add support for new power8 merge
14747 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14748 that does not include TImdoe for use with 32-bit.
14749 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14751 (UNSPEC_VSX_CVDPSPN): Likewise.
14752 (vsx_xscvdpspn): Likewise.
14753 (vsx_xscvspdpn): Likewise.
14754 (vsx_xscvdpspn_scalar): Likewise.
14755 (vsx_xscvspdpn_directmove): Likewise.
14756 (vsx_and<mode>3): Split logical operations into 32-bit and
14757 64-bit. Add support to do logical operations on TImode as well as
14758 VSX vector types. Allow logical operations to be done in either
14759 VSX registers or in general purpose registers in 64-bit mode. Add
14760 splitters if GPRs were used. For AND, add clobber of CCmode to
14761 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14763 (vsx_and<mode>3_32bit): Likewise.
14764 (vsx_and<mode>3_64bit): Likewise.
14765 (vsx_ior<mode>3): Likewise.
14766 (vsx_ior<mode>3_32bit): Likewise.
14767 (vsx_ior<mode>3_64bit): Likewise.
14768 (vsx_xor<mode>3): Likewise.
14769 (vsx_xor<mode>3_32bit): Likewise.
14770 (vsx_xor<mode>3_64bit): Likewise.
14771 (vsx_one_cmpl<mode>2): Likewise.
14772 (vsx_one_cmpl<mode>2_32bit): Likewise.
14773 (vsx_one_cmpl<mode>2_64bit): Likewise.
14774 (vsx_nor<mode>3): Likewise.
14775 (vsx_nor<mode>3_32bit): Likewise.
14776 (vsx_nor<mode>3_64bit): Likewise.
14777 (vsx_andc<mode>3): Likewise.
14778 (vsx_andc<mode>3_32bit): Likewise.
14779 (vsx_andc<mode>3_64bit): Likewise.
14780 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14781 and xxlorc instructions.
14782 (vsx_eqv<mode>3_64bit): Likewise.
14783 (vsx_nand<mode>3_32bit): Likewise.
14784 (vsx_nand<mode>3_64bit): Likewise.
14785 (vsx_orc<mode>3_32bit): Likewise.
14786 (vsx_orc<mode>3_64bit): Likewise.
14788 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14790 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14792 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14793 (p8_vmrgow): Likewise.
14794 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14795 GPRs to be split under VSX.
14796 (p8v_clz<mode>2): Add power8 count leading zero support.
14797 (p8v_popcount<mode>2): Add power8 population count support.
14798 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14801 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14804 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14806 (vec_nand): Likewise.
14807 (vec_vclz): Likewise.
14808 (vec_vclzb): Likewise.
14809 (vec_vclzd): Likewise.
14810 (vec_vclzh): Likewise.
14811 (vec_vclzw): Likewise.
14812 (vec_vgbbd): Likewise.
14813 (vec_vmrgew): Likewise.
14814 (vec_vmrgow): Likewise.
14815 (vec_vpopcnt): Likewise.
14816 (vec_vpopcntb): Likewise.
14817 (vec_vpopcntd): Likewise.
14818 (vec_vpopcnth): Likewise.
14819 (vec_vpopcntw): Likewise.
14821 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14823 PR rtl-optimization/57468
14824 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14827 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14829 PR rtl-optimization/57459
14830 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14831 type when setting live regs.
14833 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14835 * config/s390/s390.opt (mlra): New option.
14836 * config/s390/s390.c (s390_decompose_address): Check displacement
14837 for all registers for LRA.
14838 (s390_secondary_reload): Don't used secondary reloads for LRA.
14839 (s390_lra_p): New function.
14840 (TARGET_LRA_P): Define.
14841 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14842 of attribute cpu_facility to zarch for the last alternative.
14843 (*cmpmem_short): Ditto.
14845 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14847 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14848 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14849 (arm_expand_prologue): Likewise.
14851 2013-06-06 Teresa Johnson <tejohnson@google.com>
14854 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14855 as this is now done by redirect_edge_and_branch_force.
14856 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14857 barriers, and fix interaction with splitting.
14858 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14859 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14860 reflect changes made in the routine.
14861 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14862 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14863 since this is called in cfglayout mode, and replace partition fixup
14864 with assert as that is now done by force_nonfallthru_and_redirect.
14865 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14866 already be marked with region crossing note.
14867 (insert_section_boundary_note): Make non-static, gate on flag
14868 has_bb_partition, rewrite to also check for multiple partitions.
14869 (rest_of_handle_reorder_blocks): Remove call to
14870 insert_section_boundary_note, now done later during free_cfg.
14871 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14872 * bb-reorder.h (insert_section_boundary_note): Declare.
14873 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14874 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14875 invoke insert_section_boundary_note.
14876 (try_redirect_by_replacing_jump): Remove unnecessary
14877 check for region crossing note.
14878 (fixup_partition_crossing): New function.
14879 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14880 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14881 in non-cfglayout mode.
14882 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14883 remove old code that tried to do this. Emit barrier correctly
14884 when we are in cfglayout mode.
14885 (last_bb_in_partition): New function.
14886 (rtl_split_edge): Correctly fixup partition boundaries.
14887 (commit_one_edge_insertion): Remove old code that tried to
14888 fixup region crossing edge since this is now handled in
14889 split_block, and set up insertion point correctly since
14890 block may now end in a jump.
14891 (verify_hot_cold_block_grouping): Guard against checking when not in
14892 linearized RTL mode.
14893 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14895 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14896 rtl_verify_flow_info, so not called in cfglayout mode.
14897 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14898 (fixup_reorder_chain): Remove old code that attempted to fixup region
14899 crossing note as this is now handled in force_nonfallthru_and_redirect.
14900 (duplicate_insn_chain): Don't duplicate switch section notes.
14901 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14903 * basic-block.h (emit_barrier_after_bb): Declare.
14905 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14907 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14908 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14909 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14911 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14913 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14915 * config/arm/ldmstm.md: Regenerate.
14917 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14919 * config/arm/sync.md (atomic_loaddi_1):
14920 Disable predication for arm_restrict_it.
14921 (arm_load_exclusive<mode>): Likewise.
14922 (arm_load_exclusivesi): Likewise.
14923 (arm_load_exclusivedi): Likewise.
14924 (arm_load_acquire_exclusive<mode>): Likewise.
14925 (arm_load_acquire_exclusivesi): Likewise.
14926 (arm_load_acquire_exclusivedi): Likewise.
14927 (arm_store_exclusive<mode>): Likewise.
14928 (arm_store_exclusive<mode>): Likewise.
14929 (arm_store_release_exclusivedi): Likewise.
14930 (arm_store_release_exclusive<mode>): Likewise.
14932 2013-06-06 Richard Biener <rguenther@suse.de>
14934 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14936 (lto_tag_is_tree_code_p): Adjust.
14937 (lto_tag_is_gimple_code_p): Likewise.
14938 (lto_gimple_code_to_tag): Likewise.
14939 (lto_tag_to_gimple_code): Likewise.
14940 (lto_tree_code_to_tag): Likewise.
14941 (lto_tag_to_tree_code): Likewise.
14942 * data-streamer.h (streamer_write_hwi_in_range): Use
14943 uhwi streaming to stream the normalized range.
14944 (streamer_read_hwi_in_range): Likewise.
14946 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14948 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14949 (predicable_short_it): Likewise.
14950 (predicated): Likewise.
14951 (enabled): Handle above.
14952 (define_cond_exec): Set predicated attribute to yes.
14954 2013-06-05 Mike Stump <mikestump@comcast.net>
14956 * gdbinit.in (__FUNCTION__): Add.
14958 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14960 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14961 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14963 2013-06-05 Jan Hubicka <jh@suse.cz>
14965 * varasm.c (mark_decl_referenced): Revert the removal until targets
14968 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14970 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14971 instead of mark_decl_referenced.
14973 2013-06-05 Jan Hubicka <jh@suse.cz>
14975 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14976 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14977 and symtab_used_from_object_file_p.
14978 (cgraph_make_node_local_1): Clear forced_by_abi.
14979 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14980 * cgraph.h (symtab_node_base): Add forced_by_abi.
14981 (decide_is_variable_needed): Remove.
14982 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14983 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14984 (decide_is_symbol_needed): ... this one; handle symbols in general;
14985 always analyze virtuals; honnor forced_by_abi.
14986 (cgraph_finalize_function): Update.
14987 (varpool_finalize_decl): Update.
14988 (symbol_defined_and_needed): Remove.
14989 (analyze_functions): Update.
14990 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14991 output_refs, input_overwrite_node): Handle forced_by_abi.
14992 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14993 (address_taken_from_non_vtable_p): ... this one.
14994 (comdat_can_be_unshared_p_1): New function.
14995 (cgraph_comdat_can_be_unshared_p): Rename to ...
14996 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14997 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14998 (function_and_variable_visibility): Clear forced_by_abi as needed.
14999 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15000 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15001 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15002 * varpool.c (decide_is_variable_needed): Remove.
15004 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15006 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15007 (arm_option_override): Override arm_restrict_it where appropriate.
15008 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15009 * config/arm/arm.opt (mrestrict-it): New command-line option.
15010 * doc/invoke.texi: Document -mrestrict-it.
15012 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15014 * tsan.c (tsan_atomic_table): Make const.
15016 2013-06-05 Richard Biener <rguenther@suse.de>
15018 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15019 index associated with the tree we are supposed to replace.
15020 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15021 TREE_ASM_WRITTEN as zero for everything but SSA names.
15023 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15025 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15027 2013-06-04 Jan Hubicka <jh@suse.cz>
15029 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15030 (input_node, input_varpool_node): Handle correctly external same
15032 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15033 nodes at ltrans stage.
15035 2013-06-04 Jan Hubicka <jh@suse.cz>
15037 * ipa-inline.c (update_caller_keys): Fix availability test.
15038 (update_callee_keys): Likewise.
15039 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15040 to follow ELF standard.
15042 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15044 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15045 (mips64r5900el-*-elf*): New configurations.
15046 * config/mips/mips-cpus.def (r5900): New processor.
15047 * config/mips/mips-tables.opt: Regenerate.
15048 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15049 (mips_issue_rate): Handle PROCESSOR_R5900.
15050 (mips_reorg_process_insns): Force reorder mode for the R5900.
15051 * config/mips/mips.h (TARGET_MIPS5900): Define.
15052 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15054 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15056 * config/mips/mips.md (processor): Add r5900.
15057 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15059 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15061 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15062 into function to generate MOVI instruction.
15063 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15064 (aarch64_preferred_simd_mode): Turn into wrapper.
15065 (aarch64_output_scalar_simd_mov_immediate): New function.
15066 * config/aarch64/aarch64-protos.h: Add prototype for above.
15068 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15070 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15071 element_char member.
15072 (sizetochar): Return signed char.
15073 (aarch64_simd_valid_immediate): Remove elchar and other
15074 unnecessary variables.
15075 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15076 Calculate element_char as required.
15077 * config/aarch64/aarch64-protos.h: Update and move prototype
15078 for aarch64_output_simd_mov_immediate.
15079 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15082 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15084 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15085 information completed by aarch64_simd_valid_immediate.
15086 (aarch64_legitimate_constant_p): Update arguments.
15087 (aarch64_simd_valid_immediate): Work with struct rather than many
15089 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15090 (aarch64_simd_make_constant): Update arguments.
15091 (aarch64_output_simd_mov_immediate): Work with struct rather than
15092 many pointers. Output immediate directly rather than as operand.
15093 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15095 * config/aarch64/constraints.md (Dn): Update arguments.
15097 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15099 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15101 (aarch64_simd_immediate_valid_for_move): Remove.
15102 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15103 (aarch64_simd_make_constant): Update call.
15104 (aarch64_output_simd_mov_immediate): Update call.
15105 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15107 * config/aarch64/constraints.md (Dn): Update call.
15109 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15111 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15112 return type to bool for prototype.
15113 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15114 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15115 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15117 2013-06-04 Catherine Moore <clm@codesourcery.com>
15119 * config/mips/mips.opt (meva): New.
15120 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15121 (ASM_SPEC): Handle -meva.
15122 * doc/invoke.texi (meva): Document.
15124 2013-06-04 Alan Modra <amodra@gmail.com>
15126 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15129 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15131 * rtl.def: Add extra fourth optional field to define_cond_exec.
15132 * gensupport.c (process_one_cond_exec): Process attributes from
15134 * doc/md.texi: Document fourth field in define_cond_exec.
15136 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15138 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15139 out the processing order as in store_bit_field_1.
15141 2013-06-04 Jan Hubicka <jh@suse.cz>
15143 PR middle-end/57500
15144 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15145 non-VAR_DECL node if it does not exist yet.
15147 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15149 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15150 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15151 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15152 target_cpu_default setting.
15154 2013-06-03 Teresa Johnson <tejohnson@google.com>
15156 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15157 default to -fopt-info=optimized instead of all.
15158 * doc/invoke.texi: Ditto.
15159 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15160 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15161 (execute_vect_slp): Emit BB vectorization success under
15162 MSG_OPTIMIZED_LOCATIONS.
15163 * tree-vect-slp.c (vect_slp_transform_bb): Change
15164 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15165 * tree-vect-loop.c (vect_transform_loop): Ditto.
15167 2013-06-03 Jason Merrill <jason@redhat.com>
15170 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15171 Use TARGET_EXPR for C++.
15173 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15175 PR rtl-optimization/57268
15176 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15177 if DEBUG_INSN_P (insn).
15180 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15182 PR rtl-optimization/57268
15183 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15184 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15186 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15188 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15189 (ix86_avoid_lea_for_addr): Likewise.
15190 (exact_dependency_1): Likewise.
15191 (ix86_adjust_cost): Likewise.
15192 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15194 (do_reorder_for_imul): Fix typo, formatting and
15195 !reload_completed check removed.
15196 (ix86_sched_reorder): Fix typo and formatting.
15197 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15200 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15202 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15204 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15206 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15207 <STRING_CST>: Likewise.
15208 <VECTOR_CST>: Likewise.
15210 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15211 Mikael Morin <mikael@gcc.gnu.org>
15213 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15214 * config.in: Regenerated.
15215 * configure: Regenerated.
15217 2013-06-01 Jan Hubicka <jh@suse.cz>
15219 PR middle-end/57366
15220 * cgraphunit.c (compile): When weakref is not supported,
15221 set up transparent aliases before final output pass.
15222 * varasm.c (assemble_alias): Do not try to do it here.
15224 2013-06-01 Jan Hubicka <jh@suse.cz>
15226 PR middle-end/57467
15227 * passes.c (for_per_function): Skip unanalyzed functions.
15229 2013-06-01 Jan Hubicka <jh@suse.cz>
15231 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15232 (lto_symtab_merge_symbols_1): ... this one.
15233 (lto_symtab_merge_cgraph_nodes): Rename to ...
15234 (lto_symtab_merge_symbols): ... this one; simplify.
15235 * cgraph.c (same_body_aliases_done): Rename to ...
15236 (cpp_implicit_aliases_done): ... this one.
15237 (cgraph_create_function_alias): Update.
15238 (cgraph_same_body_alias): Update.
15239 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15240 (verify_edge_corresponds_to_fndecl): Simplify.
15241 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15242 (cgraph_node): Remove same_body_alias.
15243 (varpool_node): Remove alias_of and extra_name_alias.
15244 (same_body_aliases_done): Rename to ..
15245 (cpp_implicit_aliases_done): ... this one.
15246 (symtab_alias_ultimate_target): Add default parameter.
15247 (symtab_resolve_alias): New function.
15248 (fixup_same_cpp_alias_visibility): Declare.
15249 (cgraph_function_node): Add default parameter.
15250 (cgraph_node_asm_name): Likewise.
15251 (cgraph_function_or_thunk_node): Add default parameter; do
15252 not ICE when it is NULL.
15253 (varpool_variable_node): Likewise.
15254 * tree-emutls.c (create_emultls_var): Update.
15255 (ipa_lower_emutls): Update.
15256 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15257 (cgraph_reset_node): Reset alias info.
15258 (cgraph_finalize_function): Update.
15259 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15260 (analyze_function): Simplify.
15261 (cgraph_process_same_body_aliases): Simplify.
15262 (analyze_functions): Fixup same body aliases.
15263 (handle_alias_pairs): Simplify.
15264 (assemble_thunk): Update.
15265 (assemble_thunks_and_aliases): Update.
15266 (output_weakrefs): Rewrite.
15267 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15268 (lto_output_varpool_node): Likewise.
15269 (compute_ltrans_boundary): Remve assert.
15270 (get_alias_symbol): New functoin.
15271 (input_node): Rewrite alias handling.
15272 (input_varpool_node): Likewise.
15273 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15274 * ipa.c (process_references): Handle weakrefs correctly.
15275 (symtab_remove_unreachable_nodes): Likewise.
15276 * trans-mem.c (get_cg_data): Update.
15277 (ipa_tm_create_version_alias): Update.
15278 (ipa_tm_execute): Update.
15279 * symtab.c (dump_symtab_base): Dump aliases.
15280 (verify_symtab_base): Verify aliases.
15281 (symtab_node_availability): New function.
15282 (symtab_alias_ultimate_target): Simplify.
15283 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15284 handle all the fixup cases.
15285 (symtab_resolve_alias): New function.
15286 * passes.c (ipa_write_summaries): Handle weakrefs.
15287 * varpool.c (varpool_analyze_node): Simplify.
15288 (assemble_aliases): Update.
15289 (varpool_create_variable_alias): Simplify.
15290 (varpool_extra_name_alias): Simplify.
15291 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15292 (lto_symtab_merge_symbols): ... this one.
15294 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15297 PR rtl-optimization/57268
15298 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15299 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15301 2013-06-01 Tobias Burnus <burnus@net-b.de>
15303 Partially reverted:
15304 2013-05-31 Tobias Burnus <burnus@net-b.de>
15306 PR middle-end/57073
15307 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15310 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15312 PR rtl-optimization/57268
15313 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15314 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15316 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15318 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15319 unordered comparison operators when -fno-trapping-math is in effect
15321 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15322 and implement unordered comparison operators properly on the e500.
15324 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15326 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15327 for constant scalar integers.
15328 (simplify_relational_operation_1): Likewise.
15330 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15332 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15333 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15336 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15337 Igor Zamyatin <igor.zamyatin@intel.com>
15339 Silvermont (SLM) architecture performance tuning.
15340 * config/i386/i386.h (enum ix86_tune_indices): Add
15341 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15342 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15344 * config/i386/i386.c (initial_ix86_tune_features)
15345 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15346 (ix86_lea_outperforms): Handle Silvermont tuning.
15347 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15349 (ix86_use_lea_for_mov): Likewise.
15350 (ix86_avoid_lea_for_addr): Likewise.
15351 (ix86_lea_for_add_ok): Likewise.
15352 (exact_dependency_1): New function.
15353 (exact_store_load_dependency): Likewise.
15354 (ix86_adjust_cost): Handle Silvermont tuning.
15355 (do_reoder_for_imul): Likewise.
15356 (swap_top_of_ready_list): New function.
15357 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15359 * config/i386/i386.md (peepholes that split memory operand in fp
15362 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15364 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15365 Remove un-necessary braces.
15367 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15369 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15370 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15372 2013-05-31 Tobias Burnus <burnus@net-b.de>
15374 PR middle-end/57073
15375 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15377 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15380 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15381 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15382 * config/arm/neon.md (iordi3_neon): Remove.
15383 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15384 * config/arm/predicates.md (imm_for_neon_logic_operand):
15385 Move to earlier in the file.
15386 (neon_logic_op2): Likewise.
15387 (arm_iordi_operand_neon): New predicate.
15389 2013-05-31 Richard Biener <rguenther@suse.de>
15391 PR tree-optimization/57478
15392 PR tree-optimization/57453
15393 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15396 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15398 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15399 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15401 2013-05-30 Tobias Burnus <burnus@net-b.de>
15402 Thomas Koenig <tkoenig@gcc.gnu.org>
15404 PR middle-end/57073
15405 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15406 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15408 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15410 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15412 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15414 * target.def (register_usage_leveling_p): New hook.
15415 * targhooks.c (default_register_usage_leveling_p): New.
15416 * targhooks.h (default_register_usage_leveling_p): New prototype.
15417 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15418 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15419 * doc/tm.texi: Update.
15420 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15422 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15424 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15425 (*insv_reg<mode>): New define_insn.
15427 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15429 PR rtl-optimization/57439
15430 * postreload.c (move2add_valid_value_p): Check that we have
15431 a zero subreg_regno_offset when accessing the register in
15432 the requested mode.
15434 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15435 Igor Zamyatin <igor.zamyatin@intel.com>
15437 Silvermont (SLM) architecture pipeline model, tuning and
15439 * config.gcc: Add slm config options and target.
15441 * config/i386/slm.md: New.
15443 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15445 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15447 (ix86_target_macros_internal): Likewise.
15449 * config/i386/i386.c (slm_cost): New cost.
15450 (m_SLM): New macro flag.
15451 (initial_ix86_tune_features): Set m_SLM.
15452 (x86_accumulate_outgoing_args): Likewise.
15453 (x86_arch_always_fancy_math_387): Likewise.
15454 (processor_target_table): Add slm cost.
15455 (cpu_names): Add slm cpu name.
15456 (x86_option_override_internal): Set SLM ISA.
15457 (ix86_issue_rate): New case PROCESSOR_SLM.
15458 (ia32_multipass_dfa_lookahead): Likewise.
15459 (fold_builtin_cpu): Add slm.
15461 * config/i386/i386.h (TARGET_SLM): New target macro.
15462 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15463 (processor_type): Add PROCESSOR_SLM.
15465 * config/i386/i386.md (cpu): Add new value "slm".
15466 (slm.md): Include slm.md.
15468 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15469 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15471 * config/arm/arm-protos.h: Add and update function protos.
15472 * config/arm/arm.c (use_simple_return_p): New added.
15473 (thumb2_expand_return): Check simple_return flag.
15474 * config/arm/arm.md: Add simple_return and conditional simple_return.
15475 * config/arm/iterators.md: Add iterator for return and simple_return.
15477 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15479 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15480 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15481 (arm_emit_vfp_multi_reg_pop): Likewise.
15482 (thumb2_emit_ldrd_pop): Likewise.
15483 (arm_expand_epilogue): Add misc REG_CFA notes.
15484 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15486 2013-05-29 Lawrence Crowl <crowl@google.com>
15488 * config/arm/t-arm: Update for below.
15490 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15491 Change type to hash_table. Update dependent calls and types.
15493 * config/i386/t-cygming: Update for below.
15495 * config/i386/t-interix: Update for below.
15497 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15498 Change type to hash_table. Update dependent calls and types.
15499 (i386_find_on_wrapper_list::wrappers): Likewise.
15501 * config/ia64/t-ia64: Update for below.
15503 * config/ia64/ia64.c (bundle_state_table):
15504 Change type to hash_table. Update dependent calls and types.
15506 * config/mips/mips.c (mips_reorg_process_insns::htab):
15507 Change type to hash_table. Update dependent calls and types.
15509 * config/sol2.c (solaris_comdat_htab):
15510 Change type to hash_table. Update dependent calls and types.
15512 * config/t-sol2: Update for above.
15514 2013-05-29 Teresa Johnson <tejohnson@google.com>
15516 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15517 functions are not yet marked as defined.
15519 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15520 Pat Haugen <pthaugen@us.ibm.com>
15521 Peter Bergner <bergner@vnet.ibm.com>
15523 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15527 (vrotl<mode>3): Likewise.
15528 (vashl<mode>3): Likewise.
15529 (vlshr<mode>3): Likewise.
15530 (vashr<mode>3): Likewise.
15532 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15533 support for power8 V2DI builtins.
15535 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15536 power8 V2DI builtins.
15537 (vupkhsw): Likewise.
15538 (vupklsw): Likewise.
15539 (vaddudm): Likewise.
15540 (vminsd): Likewise.
15541 (vmaxsd): Likewise.
15542 (vminud): Likewise.
15543 (vmaxud): Likewise.
15544 (vpkudum): Likewise.
15545 (vpksdss): Likewise.
15546 (vpkudus): Likewise.
15547 (vpksdus): Likewise.
15552 (vsubudm): Likewise.
15553 (vcmpequd): Likewise.
15554 (vcmpgtsd): Likewise.
15555 (vcmpgtud): Likewise.
15556 (vcmpequd_p): Likewise.
15557 (vcmpgtsd_p): Likewise.
15558 (vcmpgtud_p): Likewise.
15559 (vupkhsw): Likewise.
15560 (vupklsw): Likewise.
15561 (vaddudm): Likewise.
15562 (vmaxsd): Likewise.
15563 (vmaxud): Likewise.
15564 (vminsd): Likewise.
15565 (vminud): Likewise.
15566 (vpksdss): Likewise.
15567 (vpksdus): Likewise.
15568 (vpkudum): Likewise.
15569 (vpkudus): Likewise.
15574 (vsubudm): Likewise.
15576 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15577 support for power8 V2DI instructions.
15579 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15580 power8 V2DI instructions. Combine pack and unpack insns to use an
15581 iterator for each mode. Check whether a particular mode supports
15582 Altivec instructions instead of just checking TARGET_ALTIVEC.
15583 (UNSPEC_VPKUWUM): Likewise.
15584 (UNSPEC_VPKSHSS): Likewise.
15585 (UNSPEC_VPKSWSS): Likewise.
15586 (UNSPEC_VPKUHUS): Likewise.
15587 (UNSPEC_VPKSHUS): Likewise.
15588 (UNSPEC_VPKUWUS): Likewise.
15589 (UNSPEC_VPKSWUS): Likewise.
15590 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15591 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15592 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15593 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15594 (UNSPEC_VUPKHSB): Likewise.
15595 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15596 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15597 (UNSPEC_VUPKHSH): Likewise.
15598 (UNSPEC_VUPKLSB): Likewise.
15599 (UNSPEC_VUPKLSH): Likewise.
15601 (VI_char): Likewise.
15602 (VI_scalar): Likewise.
15603 (VI_unit): Likewise.
15605 (VP_small): Likewise.
15606 (VP_small_lc): Likewise.
15607 (VU_char): Likewise.
15608 (add<mode>3): Likewise.
15609 (altivec_vaddcuw): Likewise.
15610 (altivec_vaddu<VI_char>s): Likewise.
15611 (altivec_vadds<VI_char>s): Likewise.
15612 (sub<mode>3): Likewise.
15613 (altivec_vsubcuw): Likewise.
15614 (altivec_vsubu<VI_char>s): Likewise.
15615 (altivec_vsubs<VI_char>s): Likewise.
15616 (altivec_vavgs<VI_char>): Likewise.
15617 (altivec_vcmpbfp): Likewise.
15618 (altivec_eq<mode>): Likewise.
15619 (altivec_gt<mode>): Likewise.
15620 (altivec_gtu<mode>): Likewise.
15621 (umax<mode>3): Likewise.
15622 (smax<mode>3): Likewise.
15623 (umin<mode>3): Likewise.
15624 (smin<mode>3): Likewise.
15625 (altivec_vpkuhum): Likewise.
15626 (altivec_vpkuwum): Likewise.
15627 (altivec_vpkshss): Likewise.
15628 (altivec_vpkswss): Likewise.
15629 (altivec_vpkuhus): Likewise.
15630 (altivec_vpkshus): Likewise.
15631 (altivec_vpkuwus): Likewise.
15632 (altivec_vpkswus): Likewise.
15633 (altivec_vpks<VI_char>ss): Likewise.
15634 (altivec_vpks<VI_char>us): Likewise.
15635 (altivec_vpku<VI_char>us): Likewise.
15636 (altivec_vpku<VI_char>um): Likewise.
15637 (altivec_vrl<VI_char>): Likewise.
15638 (altivec_vsl<VI_char>): Likewise.
15639 (altivec_vsr<VI_char>): Likewise.
15640 (altivec_vsra<VI_char>): Likewise.
15641 (altivec_vsldoi_<mode>): Likewise.
15642 (altivec_vupkhsb): Likewise.
15643 (altivec_vupkhs<VU_char>): Likewise.
15644 (altivec_vupkls<VU_char>): Likewise.
15645 (altivec_vupkhsh): Likewise.
15646 (altivec_vupklsb): Likewise.
15647 (altivec_vupklsh): Likewise.
15648 (altivec_vcmpequ<VI_char>_p): Likewise.
15649 (altivec_vcmpgts<VI_char>_p): Likewise.
15650 (altivec_vcmpgtu<VI_char>_p): Likewise.
15651 (abs<mode>2): Likewise.
15652 (vec_unpacks_hi_v16qi): Likewise.
15653 (vec_unpacks_hi_v8hi): Likewise.
15654 (vec_unpacks_lo_v16qi): Likewise.
15655 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15656 (vec_unpacks_lo_v8hi): Likewise.
15657 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15658 (vec_pack_trunc_v8h): Likewise.
15659 (vec_pack_trunc_v4si): Likewise.
15660 (vec_pack_trunc_<mode>): Likewise.
15662 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15664 (vec_vmaxsd): Likewise.
15665 (vec_vmaxud): Likewise.
15666 (vec_vminsd): Likewise.
15667 (vec_vminud): Likewise.
15668 (vec_vpksdss): Likewise.
15669 (vec_vpksdus): Likewise.
15670 (vec_vpkudum): Likewise.
15671 (vec_vpkudus): Likewise.
15672 (vec_vrld): Likewise.
15673 (vec_vsld): Likewise.
15674 (vec_vsrad): Likewise.
15675 (vec_vsrd): Likewise.
15676 (vec_vsubudm): Likewise.
15677 (vec_vupkhsw): Likewise.
15678 (vec_vupklsw): Likewise.
15680 2013-05-29 Jan Hubicka <jh@suse.cz>
15682 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15683 flags; reorder rest of fields in more consistent way.
15684 (varpool_node): Remove analyzed, finalized and alias.
15685 (cgraph_ndoe): Likewise.
15686 (symtab_alias_ultimate_target): New function.
15687 (cgraph_function_node): Move offline.
15688 (cgraph_reset_node): Declare.
15689 (cgraph_comdat_can_be_unshared_p): Remove.
15690 (varpool_remove_initializer): Declare.
15691 (varpool_first_defined_variable, varpool_next_defined_variable
15692 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15693 (cgraph_function_with_gimple_body_p): Update.
15694 (varpool_all_refs_explicit_p): Update.
15695 (symtab_alias_target): New function.
15696 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15697 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15698 (cgraph_function_or_thunk_node): Simplify using
15699 symtab_alias_ultimate_target.
15700 (varpool_variable_node): Likewise.
15701 * cgraph.c (cgraph_create_function_alias): Update.
15702 (cgraph_add_thunk): Update.
15703 (cgraph_remove_node): Update.
15704 (dump_cgraph_node): Do not dump removed flags.
15705 (cgraph_function_body_availability): Update.
15706 (cgraph_propagate_frequency): Update.
15707 (verify_cgraph_node): Check sanity of local flag.
15708 (cgraph_function_node): Move here from cgraph.h; revamp for
15709 cgraph_function_or_thunk_node.
15710 * lto-symtab.c (lto_varpool_replace_node): Update.
15711 (lto_symtab_resolve_can_prevail_p): Update.
15712 (lto_symtab_merge_cgraph_nodes): Update.
15713 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15714 propagate_constants_accross_call, devirtualization_time_bonus,
15715 ipcp_propagate_stage): Update.
15716 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15717 * ipa-inline-transform.c (clone_inlined_nodes,
15718 preserve_function_body_p): Update.
15719 * ipa-reference.c (propagate): Update.
15720 (write_node_summary_p): Update.
15721 * toplev.c (wrapup_global_declaration_2): Update.
15722 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15723 (analyze_function) ... this one.
15724 (cgraph_process_new_functions): Update.
15725 (cgraph_reset_node): Export.
15726 (cgraph_finalize_function): Update.
15727 (cgraph_add_new_function): Update.
15728 (process_function_and_variable_attributes): Update.
15729 (varpool_finalize_decl): Update.
15730 (symbol_finalized): Remove.
15731 (symbol_finalized_and_needed): Rename to ...
15732 (symbol_defined_and_needed): ... update.
15733 (cgraph_analyze_functions): Update.
15734 (handle_alias_pairs): Update.
15735 (mark_functions_to_output): Update.
15736 (assemble_thunk): Update.
15737 (output_in_order): Update.
15738 (output_weakrefs): Update.
15739 (finalize_compilation_unit): Update.
15740 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15741 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15742 input_node, input_varpool_node): Update.
15743 * dbxout.c (dbxout_expand_expr): Update.
15744 * cgraphclones.c (cgraph_clone_node): Update.
15745 (cgraph_copy_node_for_versioning): Update.
15746 (cgraph_materialize_clone): Update.
15747 (cgraph_materialize_all_clones): Update.
15748 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15749 propagate_pure_const, propagate_nothrow): Update.
15750 * lto-streamer-out.c (lto_output, write_symbol): Update.
15751 * ipa-utils.c (ipa_reverse_postorder): Update.
15752 * ipa-inline.c (can_inline_edge_p): Update.
15753 (update_caller_keys, ipa_inline): Update.
15754 * dwarf2out.c (reference_to_unused,
15755 premark_types_used_by_global_vars_helper): Update.
15756 * tree-eh.c (tree_could_trap_p): Update.
15757 * ipa-split.c (consider_split, execute_split_functions): Update.
15758 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15759 has_addr_references_p): Update; move ahead in file for better
15761 (process_references): Simplify.
15762 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15763 bodies are removed.
15764 (cgraph_comdat_can_be_unshared_p): Make static.
15765 (cgraph_externally_visible_p): Update.
15766 (varpool_externally_visible_p): Update.
15767 (function_and_variable_visibility): Update.
15768 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15769 ipa_tm_mark_force_output_node): Update.
15770 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15771 estimate_edge_devirt_benefit, inline_generate_summary,
15772 inline_write_summary): Update.
15773 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15774 * ipa-prop.c (ipa_compute_jump_functions): Update.
15775 (ipa_print_node_params, ipa_prop_read_section,
15776 ipa_update_after_lto_read, read_replacements_section): Update.
15777 * varasm.c (mark_decl_referenced): Update.
15778 (assemble_alias, dump_tm_clone_pairs): Update.
15779 * tree-inline.c (copy_bb): Update.
15780 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15782 * symtab.c (dump_symtab_base): Print new flags.
15783 (verify_symtab_base): Verify new flags.
15784 (symtab_alias_ultimate_target): New function.
15785 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15786 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15788 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15790 * i386.c (ix86_get_function_versions_dispatcher,
15791 ix86_generate_version_dispatcher_body): Update.
15792 (fold_builtin_cpu): Use varpool_add_new_variable.
15793 * varpool.c (varpool_remove_initializer): Break out from ...
15794 (varpool_remove_node): ... this one.
15795 (dump_varpool_node, varpool_node_for_asm,
15796 cgraph_variable_initializer_availability, varpool_analyze_node,
15797 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15798 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15801 2013-05-29 Jan Hubicka <jh@suse.cz>
15803 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15805 2013-05-29 Easwaran Raman <eraman@google.com>
15807 PR tree-optimization/57442
15808 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15809 when control exits the main loop.
15811 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15813 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15815 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15816 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15817 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15820 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15822 PR tree-optimization/57441
15823 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15824 Don't limit size of incr_vec to number of candidates.
15826 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15828 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15829 and mips16 directories.
15830 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15831 (MULTILIB_DIRNAMES): Ditto.
15832 (MULTILIB_EXCEPTIONS): Add new exceptions.
15833 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15834 (MULTILIB_DIRNAMES): Ditto.
15835 (MULTILIB_EXCEPTIONS): Add new exceptions.
15837 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15838 Marcus Shawcroft <marcus.shawcroft@arm.com>
15840 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15841 SYMBOL_TINY_ABSOLUTE.
15842 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15843 SYMBOL_TINY_ABSOLUTE.
15844 (aarch64_expand_mov_immediate): Likewise.
15845 (aarch64_classify_symbol): Likewise.
15846 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15847 Permit SYMBOL_TINY_ABSOLUTE.
15848 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15850 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15851 Marcus Shawcroft <marcus.shawcroft@arm.com>
15853 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15854 Refactor if/switch. Replace gcc_assert with if.
15856 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15858 * config/i386/i386.c (initial_ix86_tune_features): Enable
15859 FP Reassociation for AMD bdver1 and bdver2.
15861 2013-05-29 Martin Jambor <mjambor@suse.cz>
15863 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15864 and IMAGPART_EXPR do not occur within other handled_components.
15866 2013-05-29 Richard Biener <rguenther@suse.de>
15868 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15869 access on whether the use is in the BB we currently try to
15871 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15872 vectorize to vect_bb_slp_scalar_cost.
15874 2013-05-29 Richard Biener <rguenther@suse.de>
15876 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15877 computing scalar cost offsetted by stmts that are kept live
15879 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15880 for computation of scalar cost.
15882 2013-05-28 Steve Ellcey <sellcey@mips.com>
15884 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15886 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15888 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15889 array notation built-in reduction functions.
15890 * doc/passes.texi (Passes): Added documentation about changes done
15892 * doc/invoke.texi (C Dialect Options): Added documentation about
15893 the -fcilkplus flag.
15894 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15895 (BUILTINS_DEF): Depend on cilkplus.def.
15896 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15897 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15898 * cilkplus.def: New file.
15900 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15902 PR rtl-optimization/57439
15903 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15905 2013-05-28 Easwaran Raman <eraman@google.com>
15907 PR tree-optimization/57337
15908 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15909 (find_insert_point): Correctly identify the insertion point
15910 when two statements with the same UID is compared.
15912 2013-05-28 Richard Biener <rguenther@suse.de>
15914 PR tree-optimization/56787
15915 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15916 from the list of data references.
15917 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15919 (vect_analyze_loop_operations): Likewise.
15920 (vect_transform_loop): Remove clobbers.
15922 2013-05-28 Martin Jambor <mjambor@suse.cz>
15924 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15925 and REALPART_EXPRs have scalar type.
15927 2013-05-28 Richard Biener <rguenther@suse.de>
15929 PR tree-optimization/57411
15930 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15932 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15933 virtual operand propagation.
15935 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15937 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15938 destination register for bmasksi_vis.
15939 (vector_init_bshuffle): Likewise.
15940 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15942 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15944 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15945 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15946 mode if the instruction isn't available in the original mode.
15947 * config/sparc/sparc.opt (mfix-ut699): New option.
15948 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15949 (divdf3): Turn into expander.
15950 (divdf3_nofix): New insn.
15951 (divdf3_fix): Likewise.
15952 (divsf3): Disable if -mfix-ut699.
15953 (sqrtdf2): Turn into expander.
15954 (sqrtdf2_nofix): New insn.
15955 (sqrtdf2_fix): Likewise.
15956 (sqrtsf2): Disable if -mfix-ut699.
15958 2013-05-27 Richard Biener <rguenther@suse.de>
15960 PR middle-end/57412
15961 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15962 block for the new loop.
15964 2013-05-27 Richard Biener <rguenther@suse.de>
15966 PR tree-optimization/57343
15967 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15968 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15969 (number_of_iterations_cond): Do not build the folded tree.
15971 2013-05-27 Richard Biener <rguenther@suse.de>
15974 PR middle-end/57381
15975 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15976 OEP_CONSTANT_ADDRESS_OF retained.
15978 PR tree-optimization/57417
15979 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15980 for unchanged base.
15981 (set_ssa_val_to): Compare addresses using
15982 get_addr_base_and_unit_offset.
15984 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15986 PR rtl-optimization/56833
15987 * postreload.c (move2add_record_mode): New function.
15988 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15989 (move2add_use_add2_insn): Use move2add_record_sym_value.
15990 (move2add_use_add3_insn): Likewise.
15991 (reload_cse_move2add): Use move2add_valid_value_p and
15992 move2add_record_mode. Invalidate call-clobbered and REG_INC
15993 affected regs by setting reg_mode to VOIDmode.
15994 (move2add_note_store): Don't pretend the inside of a SUBREG is
15995 the actual destination. Invalidate single/leading registers by
15996 setting reg_mode to VOIDmode.
15997 Use move2add_record_sym_value, move2add_valid_value_p and
15998 move2add_record_mode.
16000 2013-05-27 Richard Biener <rguenther@suse.de>
16002 PR tree-optimization/57396
16003 * tree-affine.c (double_int_constant_multiple_p): Properly
16004 return false for val == 0 and div != 0.
16006 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16008 * config/mips/mips.h: Use #elif in preprocessor conditions.
16010 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16013 * config/mips/constraints.md (kl): New constraint.
16014 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16015 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16016 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16017 from LO for MIPS16.
16018 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16020 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16023 * config/mips/mips.c (mips_can_inline_p): New function.
16024 (TARGET_CAN_INLINE_P): Define.
16026 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16028 * sched-int.h (ds_t, dw_t): Make unsigned int.
16029 Fix documentation that describes how all the ds_t bits are used.
16030 Reserve the last bit for delayed-branch scheduling.
16031 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16032 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16033 (gen_dep_weak_1): Remove prototype.
16034 * sched-deps.c (get_dep_weak_1): Make static.
16035 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16036 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16037 * doc/tm.texi: Regenerate.
16038 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16040 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16043 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16045 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16046 Sandra Loosemore <sandra@codesourcery.com>
16048 * config.gcc (powerpc-*): Allow native for with-cpu.
16050 2013-05-24 Jeff Law <law@redhat.com>
16052 PR tree-optimization/57124
16053 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16054 conversion feeding a condition if the range has an overflow
16055 if -fstrict-overflow. Add warnings for when we do make the
16058 2013-05-24 Dehao Chen <dehao@google.com>
16060 * tree-cfg.c (locus_discrim_map): Fix the typo.
16061 (locus_discrim_hasher): Likewise.
16062 (locus_discrim_hasher::hash): Likewise.
16063 (locus_discrim_hasher::equal): Likewise.
16065 2013-05-24 Martin Jambor <mjambor@suse.cz>
16067 PR tree-optimization/57294
16068 * cgraph.h (ipa_record_stmt_references): Declare.
16069 * cgraphbuild.c (ipa_record_stmt_references): New function.
16070 (build_cgraph_edges): Use ipa_record_stmt_references.
16071 (rebuild_cgraph_edges): Likewise.
16072 (cgraph_rebuild_references): Likewise.
16073 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16074 associated with the old statement and build references from the
16075 newly built statements.
16076 * ipa-ref.c (ipa_remove_stmt_references): New function.
16077 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16079 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16081 * lra-constraints.c (emit_spill_move): Use smaller mode for
16083 (check_and_process_move): Consider mem-reg moves for secondary
16085 (curr_insn_transform): Don't lose insns emitted before for
16086 secondary memory moves.
16087 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16088 reg set up in the current insn.
16090 2013-05-24 Dehao Chen <dehao@google.com>
16092 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16094 (locus_descrim_hasher::equal): Likewise.
16095 (build_gimple_cfg): New discriminator assignment algorithm.
16096 (make_edges): Likewise.
16097 (next_discriminator_for_locus): Likewise.
16098 (same_line_p): Likewise.
16099 (assign_discriminators): Likewise.
16100 (make_cond_expr_edges): Likewise.
16101 (make_gimple_switch_edges): Likewise.
16102 (make_goto_expr_edges): Likewise.
16103 (make_gimple_asm_edges): Likewise.
16105 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16107 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16108 X format specifier to only display bottom 16 bits.
16109 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16110 immediate to match for operand 2, since it will be masked.
16112 2013-05-24 Richard Biener <rguenther@suse.de>
16114 PR tree-optimization/57287
16115 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16116 all SSA names that occur in abnormal PHIs.
16118 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16121 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16122 that index is not negative.
16124 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16126 PR rtl-optimization/55177
16127 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16128 (simplify_byte_swapping_operation): New.
16129 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16130 (simplify_relational_operation_1): Deal with BSWAP.
16132 2013-05-23 Richard Henderson <rth@redhat.com>
16135 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16136 (ix86_reorg): Call it.
16138 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16141 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16142 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16143 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16145 2013-05-23 Christian Bruel <christian.bruel@st.com>
16148 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16150 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16151 Marcus Shawcroft <marcus.shawcroft@arm.com>
16153 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16154 * config/aarch64/constraints.md (Usa): Remove.
16155 * doc/md.texi (AArch64 Usa): Remove.
16157 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16158 Marcus Shawcroft <marcus.shawcroft@arm.com>
16160 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16161 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16162 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16163 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16165 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16167 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16168 instruction (AdvSIMD).
16169 * config/aarch64/aarch64-builtins.c
16170 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16171 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16173 2013-05-23 Martin Jambor <mjambor@suse.cz>
16175 PR middle-end/57347
16176 * tree.h (contains_bitfld_component_ref_p): Declare.
16177 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16178 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16180 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16181 not access a bit-field. Assert all final offsets are byte-aligned.
16183 2013-05-23 Richard Biener <rguenther@suse.de>
16185 PR tree-optimization/57380
16186 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16187 least one invariant or re-used load.
16188 * passes.c (init_optimization_passes): Move pass_phiprop before
16191 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16193 * config/aarch64/aarch64-simd.md
16194 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16196 2013-05-23 Richard Biener <rguenther@suse.de>
16198 PR middle-end/57381
16199 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16200 OEP_CONSTANT_ADDRESS_OF retained.
16202 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16204 PR middle-end/57344
16205 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16206 don't lower unit. Handle unit not being always BITS_PER_WORD.
16208 2013-05-23 Richard Biener <rguenther@suse.de>
16210 PR rtl-optimization/57341
16211 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16212 instead of true_dependence.
16214 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16216 * bb-reorder.c (branch_threshold): Make const.
16217 (exec_threshold): Ditto.
16219 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16220 Pat Haugen <pthaugen@us.ibm.com>
16221 Peter Bergner <bergner@vnet.ibm.com>
16223 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16224 documentation for the power8 crypto builtins.
16226 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16228 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16229 macros for defining power8 builtin functions.
16230 (BU_P8V_AV_2): Likewise.
16231 (BU_P8V_AV_P): Likewise.
16232 (BU_P8V_VSX_1): Likewise.
16233 (BU_P8V_OVERLOAD_1): Likewise.
16234 (BU_P8V_OVERLOAD_2): Likewise.
16235 (BU_CRYPTO_1): Likewise.
16236 (BU_CRYPTO_2): Likewise.
16237 (BU_CRYPTO_3): Likewise.
16238 (BU_CRYPTO_OVERLOAD_1): Likewise.
16239 (BU_CRYPTO_OVERLOAD_2): Likewise.
16240 (XSCVSPDP): Fix typo, point to the correct instruction.
16241 (VCIPHER): Add power8 crypto builtins.
16242 (VCIPHERLAST): Likewise.
16243 (VNCIPHER): Likewise.
16244 (VNCIPHERLAST): Likewise.
16245 (VPMSUMB): Likewise.
16246 (VPMSUMH): Likewise.
16247 (VPMSUMW): Likewise.
16248 (VPERMXOR_V2DI): Likewise.
16249 (VPERMXOR_V4SI: Likewise.
16250 (VPERMXOR_V8HI: Likewise.
16251 (VPERMXOR_V16QI: Likewise.
16252 (VSHASIGMAW): Likewise.
16253 (VSHASIGMAD): Likewise.
16254 (VPMSUM): Likewise.
16255 (VPERMXOR): Likewise.
16256 (VSHASIGMA): Likewise.
16258 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16259 __CRYPTO__ if the crypto instructions are available.
16260 (altivec_overloaded_builtins): Add support for overloaded power8
16263 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16264 support for power8 crypto builtins.
16265 (builtin_function_type): Likewise.
16266 (altivec_init_builtins): Add support for builtins that take vector
16267 long long (V2DI) arguments.
16269 * config/rs6000/crypto.md: New file, define power8 crypto
16272 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16273 Pat Haugen <pthaugen@us.ibm.com>
16274 Peter Bergner <bergner@vnet.ibm.com>
16276 * doc/invoke.texi (Option Summary): Add power8 options.
16277 (RS/6000 and PowerPC Options): Likewise.
16279 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16280 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16281 wm, wn, wr documentation.
16283 * config/rs6000/constraints.md (wm): New constraint for VSX
16284 registers if direct move instructions are enabled.
16285 (wn): New constraint for no registers.
16286 (wq): New constraint for quad word even GPR registers.
16287 (wr): New constraint if 64-bit instructions are enabled.
16288 (wv): New constraint if power8 vector instructions are enabled.
16289 (wQ): New constraint for quad word memory locations.
16291 * config/rs6000/predicates.md (const_0_to_15_operand): New
16292 constraint for 0..15 for crypto instructions.
16293 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16294 as GPR and floating point registers.
16295 (int_reg_operand): New predicate to match only GPR registers.
16296 (base_reg_operand): New predicate to match base registers.
16297 (quad_int_reg_operand): New predicate to match even GPR registers
16298 for quad memory operations.
16299 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16300 operations in both GPR and VSX registers.
16301 (quad_memory_operand): New predicate for quad memory operations.
16302 (reg_or_indexed_operand): New predicate for direct move support.
16304 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16305 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16306 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16307 (POWERPC_MASKS): Add power8 options.
16308 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16311 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16312 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16314 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16315 (-mpower8-fusion): New power8 options.
16316 (-mpower8-fusion-sign): Likewise.
16317 (-mpower8-vector): Likewise.
16318 (-mcrypto): Likewise.
16319 (-mdirect-move): Likewise.
16320 (-mquad-memory): Likewise.
16322 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16323 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16325 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16326 (rs6000_debug_vector_unit): Add p8_vector.
16327 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16328 definitions. Also print fusion state.
16329 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16330 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16331 (rs6000_option_override_internal): Add support for power8.
16332 (rs6000_common_init_builtins): Add debugging for skipped builtins
16333 if -mdebug=builtin.
16334 (rs6000_adjust_cost): Add power8 support.
16335 (rs6000_issue_rate): Likewise.
16336 (insn_must_be_first_in_group): Likewise.
16337 (insn_must_be_last_in_group): Likewise.
16338 (force_new_group): Likewise.
16339 (rs6000_register_move_cost): Likewise.
16340 (rs6000_opt_masks): Likewise.
16342 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16343 power8 capable assembler, default to power7 options.
16344 (TARGET_DIRECT_MOVE): Likewise.
16345 (TARGET_CRYPTO): Likewise.
16346 (TARGET_P8_VECTOR): Likewise.
16347 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16348 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16349 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16350 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16351 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16352 (TARGET_XSCVDPSPN): Likewise.
16353 (TARGET_XSCVSPDPN): Likewsie.
16354 (TARGET_SYNC_HI_QI): Likewise.
16355 (TARGET_SYNC_TI): Likewise.
16356 (MASK_CRYPTO): Likewise.
16357 (MASK_DIRECT_MOVE): Likewise.
16358 (MASK_P8_FUSION): Likewise.
16359 (MASK_P8_VECTOR): Likewise.
16360 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16361 temporary used by some of the direct move instructions to get two FP
16362 temporary registers does not force creation of a stack frame.
16363 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16364 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16365 that any VSX registers are tieable, even if they are also an
16366 Altivec vector mode.
16367 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16368 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16369 (RS6000_BTM_CRYPTO): Likewise.
16370 (RS6000_BTM_COMMON): Likewise.
16372 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16373 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16374 (enum rs6000_vector): Add power8 vector support.
16376 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16380 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16381 (any_sibcall_could_use_r3): this and handle indirect calls.
16382 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16384 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16386 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16388 2013-05-22 Richard Biener <rguenther@suse.de>
16390 PR middle-end/57349
16391 * profile.c (branch_prob): Do not split blocks that are
16392 abnormally receiving from ECF_RETURNS_TWICE functions.
16394 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16396 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16397 address mode. Move pointer mode initialization to the same place.
16399 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16401 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16402 while it has any effect.
16404 2013-05-21 Easwaran Raman <eraman@google.com>
16406 PR tree-optimization/57322
16407 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16408 UID of the statement added to the BB to be 1.
16410 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16412 PR tree-optimization/57331
16413 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16414 of conversion from pointer type to integral type with integer.
16416 2013-05-21 Martin Jambor <mjambor@suse.cz>
16419 * ipa-prop.c (ipa_read_node_info): Process param_used and
16420 controlled_uses in the same order as when writing.
16422 2013-05-21 Magnus Granberg <baldrick@free.fr>
16425 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16427 2013-05-21 Richard Biener <rguenther@suse.de>
16429 PR tree-optimization/57318
16430 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16431 estimate stmts with side-effects as likely eliminated.
16433 2013-05-21 Richard Biener <rguenther@suse.de>
16435 PR tree-optimization/57330
16436 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16437 preserve the call stmts fntype.
16439 2013-05-21 Richard Biener <rguenther@suse.de>
16441 PR tree-optimization/57303
16442 * tree-ssa-sink.c (statement_sink_location): Improve killing
16443 stmt detection and properly handle self-assignments.
16445 2013-05-21 Christian Bruel <christian.bruel@st.com>
16447 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16448 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16449 registers. Set register size out of the PARALLEL loop.
16451 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16454 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16455 (*fmasf4, *fmasf4_media): New insns.
16457 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16459 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16460 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16461 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16462 (mips_idiv_insns): Update the comments to say that the returned
16463 instruction counts are in units of BASE_INSN_LENGTH.
16464 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16465 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16466 using 2 rather than 4 as the length of indirect MIPS16 and
16467 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16468 length of a NOP. Don't divide MIPS16 lengths by 2.
16469 (mips16_split_long_branches): Assume a branch is long if the
16470 length is greater than 4 rather than 8.
16471 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16472 rather than multiplying them by 2. Multiply instruction counts
16473 by BASE_INSN_LENGTH rather than 4.
16474 (*jump_mips16, tls_get_tp_mips16_<mode>)
16475 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16477 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16479 * config/mips/mips.md (extended_mips16): Remove branch case.
16480 (length): Remove duplicated extended_mips16 test.
16482 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16484 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16486 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16488 * recog.h (Recog_data): Rename to...
16489 (recog_data_d): ...this.
16490 (recog_data): Update accordingly.
16491 * recog.c (recog_data): Likewise.
16492 * reload.c (save_recog_data): Likewise.
16493 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16494 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16496 2013-05-17 Julian Brown <julian@codesourcery.com>
16498 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16499 found in a REG_EQUAL note, invalidate it.
16501 2013-05-17 Easwaran Raman <eraman@google.com>
16503 * tree-ssa-reassoc.c (find_insert_point): New function.
16504 (insert_stmt_after): Likewise.
16505 (get_def_stmt): Likewise.
16506 (ensure_ops_are_available): Likewise.
16507 (not_dominated_by): Likewise.
16508 (rewrite_expr_tree): Do not move statements beyond what is
16509 necessary. Remove call to swap_ops_for_binary_stmt...
16510 (reassociate_bb): ... and move it here.
16511 (build_and_add_sum): Assign UIDs for new statements.
16512 (linearize_expr): Likewise.
16513 (do_reassoc): Renumber gimple statement UIDs.
16515 2013-05-17 Jan Hubicka <jh@suse.cz>
16517 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16519 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16520 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16522 (output_weakrefs): Update.
16524 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16525 Martin Jambor <mjambor@suse.cz>
16527 PR middle-end/57276
16528 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16529 value that corresponds to the given aggval is found in values vector.
16531 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16533 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16534 sse, sse2, sse3, ssse3 and sse4a flags to options.
16536 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16538 * gengtype-state.c: (s_expr_writer): New class, to handle
16539 prettifying of output layout of s-expressions.
16540 (state_writer): New class, to write out gtype.state.
16541 (state_written_type_count): Move this variable into member data of
16543 (s_expr_writer::s_expr_writer): New code: constructor for new class
16544 (state_writer::state_writer(): ditto
16545 (s_expr_writer::write_new_line): New function
16546 (s_expr_writer::write_any_indent): ditto
16547 (s_expr_writer::begin_s_expr): ditto
16548 (s_expr_writer::end_s_expr): ditto
16549 (write_state_fileloc): convert to method of state_writer...
16550 (state_writer:: write_state_fileloc): ...and use methods of
16551 s_expr_writer to write indentation into the gtype.state output file
16552 to visually represent the hierarchical structure of the list
16554 (write_state_fields): ditto, renaming to...
16555 (state_writer::write_state_fields)
16556 (write_state_a_string): ditto, renaming to...
16557 (state_writer::write_state_a_string)
16558 (write_state_string_option): ditto, renaming to...
16559 (state_writer::write_state_string_option)
16560 (write_state_type_option): ditto, renaming to...
16561 (state_writer::write_state_type_option)
16562 (write_state_nested_option): ditto, renaming to...
16563 (state_writer::write_state_nested_option)
16564 (write_state_option): ditto, renaming to...
16565 (state_writer::write_state_option)
16566 (write_state_options): ditto, renaming to...
16567 (state_writer::write_state_options)
16568 (write_state_lang_bitmap): ditto, renaming to...
16569 (state_writer::write_state_lang_bitmap)
16570 (write_state_version): ditto, renaming to...
16571 (state_writer::write_state_version)
16572 (write_state_scalar_type): ditto, renaming to...
16573 (state_writer::write_state_scalar_type)
16574 (write_state_string_type): ditto, renaming to...
16575 (state_writer::write_state_string_type)
16576 (write_state_undefined_type): ditto, renaming to...
16577 (state_writer::write_state_undefined_type)
16578 (write_state_struct_union_type): ditto, renaming to...
16579 (state_writer::write_state_struct_union_type)
16580 (write_state_struct_type): ditto, renaming to...
16581 (state_writer::write_state_struct_type)
16582 (write_state_user_struct_type): ditto, renaming to...
16583 (state_writer::write_state_user_struct_type)
16584 (write_state_lang_struct_type): ditto, renaming to...
16585 (state_writer::write_state_lang_struct_type)
16586 (write_state_param_struct_type): ditto, renaming to...
16587 (state_writer::write_state_param_struct_type)
16588 (write_state_pointer_type): ditto, renaming to...
16589 (state_writer::write_state_pointer_type)
16590 (write_state_array_type): ditto, renaming to...
16591 (state_writer::write_state_array_type)
16592 (write_state_gc_used): ditto, renaming to...
16593 (state_writer::write_state_gc_used)
16594 (write_state_common_type_content): ditto, renaming to...
16595 (state_writer::write_state_common_type_content)
16596 (write_state_type): ditto, renaming to...
16597 (state_writer::write_state_type)
16598 (write_state_pair_list): ditto, renaming to...
16599 (state_writer::write_state_pair_list)
16600 (write_state_pair): ditto, renaming to...
16601 (state_writer::write_state_pair)
16602 (write_state_typedefs): ditto, renaming to...
16603 (state_writer::write_state_typedefs)
16604 (write_state_structures): ditto, renaming to...
16605 (state_writer::write_state_structures)
16606 (write_state_param_structs): ditto, renaming to...
16607 (state_writer::write_state_param_structs)
16608 (write_state_variables): ditto, renaming to...
16609 (state_writer::write_state_variables)
16610 (write_state_srcdir): ditto, renaming to...
16611 (state_writer::write_state_srcdir)
16612 (write_state_files_list): ditto, renaming to...
16613 (state_writer::write_state_files_list)
16614 (write_state_languages): ditto, renaming to...
16615 (state_writer::write_state_languages)
16616 (write_state): create a state_writer instance and use it when
16617 writing out the state file
16619 2013-05-17 Mike Stump <mikestump@comcast.net>
16621 PR rtl-optimization/57304
16622 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16623 accessing DF_REF_REAL_LOC.
16625 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16627 PR rtl-optimization/57281
16628 PR rtl-optimization/57300
16629 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16630 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16631 what the other splitter did if the registers are dead.
16633 2013-05-17 Richard Biener <rguenther@suse.de>
16635 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16638 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16640 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16643 2013-05-17 Marek Polacek <polacek@redhat.com>
16645 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16646 length when doing non-zero store of storing '\0' to '\0'.
16648 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16650 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16651 vect_external_def oprnd1 with loop_vinfo, try to emit
16652 optional cast, negation and and stmts on the loop preheader
16653 edge instead of into the pattern def seq.
16655 PR tree-optimization/57051
16656 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16657 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16659 2013-05-16 Nick Clifton <nickc@redhat.com>
16661 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16662 (rl78_is_naked_func): New function.
16663 (rl78_expand_prologue): Skip prologue generation for naked functions.
16664 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16665 * doc/extend.texi (naked): Add RL78 to the list of processors
16666 that supports this attribute.
16668 2013-05-16 Jeff Law <law@redhat.com>
16670 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16672 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16674 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16675 cache parameters using detect_caches_amd also for CYRIX,
16676 NSC and TM2 signatures.
16678 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16679 Dzianis Kahanovich <mahatma@eu.by>
16683 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16684 VIA/Centaur processors and determine their cache parameters
16685 using detect_caches_amd.
16687 2013-05-16 Teresa Johnson <tejohnson@google.com>
16689 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16690 (rtl_verify_edges): New function.
16691 (rtl_verify_bb_insns): Ditto.
16692 (rtl_verify_bb_pointers): Ditto.
16693 (rtl_verify_bb_insn_chain): Ditto.
16694 (rtl_verify_fallthru): Ditto.
16695 (rtl_verify_bb_layout): Ditto.
16696 (rtl_verify_flow_info_1): Outline checks into new functions.
16697 (rtl_verify_flow_info): Ditto.
16699 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16701 * cfghooks.c (copy_bbs): Add update_dominance argument.
16702 * cfghooks.h (copy_bbs): Update prototype.
16703 * tree-cfg.c (gimple_duplicate_sese_region):
16704 Add update_dominance argument.
16705 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16706 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16707 gimple_duplicate_sese_region call.
16708 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16709 Update copy_bbs call.
16710 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16711 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16713 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16715 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16716 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16717 vect_recog_rotate_pattern.
16718 (vect_recog_rotate_pattern): New function.
16720 2013-05-16 Jason Merrill <jason@redhat.com>
16722 * Makefile.in (LLINKER): New variable.
16723 (mostlyclean): Remove link mutex.
16724 * configure.ac: Handle --enable-link-mutex.
16725 * lock-and-run.sh: New script.
16727 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16730 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16733 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16735 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16737 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16739 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16740 * config/arm/arm.c (next_consecutive_mem): New function.
16741 (gen_movmem_ldrd_strd): Likewise.
16742 * config/arm/arm.md (movmemqi): Update condition and code.
16743 (unaligned_loaddi, unaligned_storedi): New patterns.
16745 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16747 * config.gcc: Obsolete *-*-solaris2.9*.
16748 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16750 2013-05-16 Richard Biener <rguenther@suse.de>
16752 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16753 earlier, after GRAPHITE transforms and IV canonicalization.
16755 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16757 * omp-low.c (extract_omp_for_data): For collapsed loops,
16758 if at least one of the loops is known at compile time to
16759 iterate zero times, set count to 0.
16760 (expand_omp_regimplify_p): New function.
16761 (expand_omp_for_generic): For collapsed loops, if at least
16762 one of the loops isn't known to iterate at least once,
16763 add runtime check with setting count to 0.
16764 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16765 For unsigned types if it isn't known at compile time that
16766 the loop will iterate at least once, add runtime check to bypass
16767 the whole loop if initial condition isn't true.
16769 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16771 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16773 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16775 PR middle-end/57286
16776 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16777 transformations to avoid an infinite loop.
16779 2013-05-16 Marek Polacek <polacek@redhat.com>
16781 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16783 2013-05-15 Leif Ekblad <leif@rdos.net>
16785 * config/i386/i386.c (ix86_decompose_address): Use
16786 DEFAULT_TLS_SEG_REG to access TLS segment register.
16787 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16788 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16789 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16791 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16794 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16795 sibling calls to functions that would normally be lazily bound,
16796 unless $gp is call-clobbered.
16798 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16800 * config/i386/i386.c (ix86_option_override_internal): Update
16801 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16802 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16803 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16804 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16806 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16808 2013-05-15 Andreas Schwab <schwab@suse.de>
16810 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16811 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16814 2013-05-15 Teresa Johnson <tejohnson@google.com>
16816 * loop-unroll.c (report_unroll_peel): Check decision before
16817 emitting unroll/peel message.
16819 2013-05-15 Teresa Johnson <tejohnson@google.com>
16821 * function.h (has_bb_partition): New rtl_data flag.
16822 (bb_reorder_complete): Ditto.
16823 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16824 instead of flag_reorder_blocks_and_partition.
16825 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16826 with some enhancements.
16827 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16828 * bb-reorder.c (connect_traces): Check for has_bb_partition
16829 instead of flag_reorder_blocks_and_partition.
16830 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16831 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16832 verify_hot_cold_block_grouping.
16833 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16835 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16838 * config/arm/predicates.md (call_insn_operand): New predicate.
16839 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16840 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16841 if insn is not a tail call.
16842 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16844 * config/arm/arm.h (enum reg_class): New caller save register class.
16845 (REG_CLASS_NAMES): Likewise.
16846 (REG_CLASS_CONTENTS): Likewise.
16847 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16850 2013-05-15 Richard Biener <rguenther@suse.de>
16852 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16853 of MSG_OPTIMIZED_LOCATIONS.
16854 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16855 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16857 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16858 of MSG_OPTIMIZED_LOCATIONS.
16859 (execute_vect_slp): Likewise.
16860 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16861 (vect_create_cond_for_alias_checks): Likewise.
16862 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16863 (vect_recog_widen_mult_pattern): Likewise.
16864 (vect_recog_widen_sum_pattern): Likewise.
16865 (vect_recog_over_widening_pattern): Likewise.
16866 (vect_recog_widen_shift_pattern): Likewise.
16867 (vect_recog_vector_vector_shift_pattern): Likewise.
16868 (vect_recog_divmod_pattern): Likewise.
16869 (vect_recog_mixed_size_cond_pattern): Likewise.
16870 (vect_recog_bool_pattern): Likewise.
16871 (vect_pattern_recog_1): Likewise.
16873 2013-05-15 Martin Jambor <mjambor@suse.cz>
16875 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16876 non-functions to builtin_unreachable.
16877 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16878 correct when new direct edges were discovered.
16880 2013-05-15 Martin Jambor <mjambor@suse.cz>
16882 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16883 header, print symbol order instead of node uid, print more information
16884 about indirect edge targets.
16885 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16887 (ipa_make_edge_direct_to_target): Likewise.
16888 (remove_described_reference): Likewise.
16889 (propagate_controlled_uses): Likewise.
16890 (ipa_print_node_params): Also print symbol order.
16891 (ipcp_transform_function): Print symbol order instead of node uids.
16892 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16893 (cgraph_get_create_real_symbol_node): Likewise.
16894 * ipa-cp.c (print_lattice): Likewise.
16895 (print_all_lattices): Likewise.
16896 (determine_versionability): Likewise.
16897 (initialize_node_lattices): Likewise.
16898 (estimate_local_effects): Likewise.
16899 (update_profiling_info): Likewise.
16900 (create_specialized_node): Likewise.
16901 (perhaps_add_new_callers): Likewise.
16902 (decide_about_value): Likewise.
16903 (decide_whether_version_node): Likewise.
16904 (identify_dead_nodes): Likewise.
16905 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16906 (dump_inline_summary): Likewise.
16907 (estimate_node_size_and_time): Likewise.
16908 (inline_analyze_function): Likewise.
16909 * ipa-inline.c (report_inline_failed_reason): Likewise.
16910 (want_early_inline_function_p): Likewise.
16911 (edge_badness): Likewise.
16912 (update_edge_key): Likewise.
16913 (inline_small_functions): Likewise. Add dumping of order to two other
16915 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16916 instead of node uids.
16917 (propagate_pure_const): Likewise.
16918 (propagate_pure_const): Likewise.
16919 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16920 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16921 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16923 * tree-pretty-print.c (dump_function_header): Likewise.
16924 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16925 Print symbol order instead of node uids.
16927 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16929 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16930 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16932 2013-05-15 Richard Biener <rguenther@suse.de>
16934 PR tree-optimization/57275
16935 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16936 return value for fail to do runtime alias checks for gather loads.
16938 2013-05-15 Jan Hubicka <jh@suse.cz>
16942 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16943 weakrefs are not external.
16944 (lto_symtab_merge_decls): Fix thinko when dealing with
16945 non-lto_symtab decls.
16946 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16947 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16948 * varpool.c (dump_varpool_node): Dump more flags.
16950 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16952 * config/i386/i386.c (processor_alias_table): Add instruction
16953 FSGSBASE for AMD bdver3 architecture.
16955 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16957 * tree.c (warn_deprecated_use): Print file:line using locus color.
16958 * diagnostic.c (diagnostic_report_current_module): Print file:line
16959 and file:line:column using locus color.
16961 2013-05-14 Mike Stump <mikestump@comcast.net>
16963 * gdbinit.in: Add __null.
16965 2013-05-14 Mike Stump <mikestump@comcast.net>
16967 * recog.h: Rename struct recog_data to Recog_data.
16968 * recog.c: Likewise.
16969 * reload.c (can_reload_into): Likewise.
16970 * config/picochip/picochip.c: Likewise.
16972 2013-05-14 Mike Stump <mikestump@comcast.net>
16974 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16976 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16978 * resource.h (struct resources): Remove unch_memory member.
16979 (CLEAR_RESOURCE): Don't clear unch_memory.
16980 * resource.c (mark_referenced_resources): Don't set it.
16981 (mark_set_resources): Likewise.
16982 (mark_target_live_regs): Don't clear it.
16983 (init_resource_info): Likewise.
16984 * reorg.c (resource_conflicts_p): Don't compare it.
16985 (redundant_insn): Don't set it.
16987 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16989 * emit-rtl.c (next_label): Remove unused function.
16990 (skip_consecutive_labels, link_cc0_insns): Move to ...
16991 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16992 only place where these functions are used, and make them static.
16994 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16996 * fold-const.c (fold_negate_expr): Handle vectors.
16997 (fold_truth_not_expr): Make it static.
16998 (fold_invert_truthvalue): New static function.
16999 (invert_truthvalue_loc): Handle vectors. Do not call
17000 fold_truth_not_expr directly.
17001 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17002 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17003 (fold_comparison): Handle vector constants.
17004 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17005 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17006 * tree.h (fold_truth_not_expr): Remove declaration.
17008 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17010 * config/aarch64/aarch64-simd.md
17011 (aarch64_vcond_internal<mode>): Rename to...
17012 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17013 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17014 float modes. Clarify all iterator modes.
17015 (vcond<mode><mode>): Use new name for vcond expanders.
17016 (vcond<v_cmp_result><mode>): Likewise.
17017 (vcondu<mode><mode>: Likewise.
17018 * config/aarch64/iterators.md (VDQF_COND): New.
17020 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17023 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17024 variable for the shift amount. Check that we shift by non-negative
17027 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17030 * config/arm/arm.h (EPILOGUE_USES): Only return true
17031 for LR_REGNUM after epilogue_completed.
17033 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17035 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17036 is error_mark_node.
17038 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17041 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17042 and Solaris 11+/x86 with gld.
17043 * configure: Regenerate.
17045 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17047 * expmed.c (expand_shift_1): Canonicalize rotates by
17048 constant bitsize / 2 to bitsize - 1.
17049 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17050 case ROTATERT>: Likewise.
17053 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17055 * config/i386/i386.md (rotateinv): New code attr.
17056 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17057 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17058 roll $31, %eax, etc.
17060 2013-05-14 Richard Biener <rguenther@suse.de>
17062 PR middle-end/57235
17063 * tree-eh.c (sink_clobbers): Give up for successors with
17064 multiple predecessors and no virtual uses.
17066 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17068 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17069 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17071 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17073 PR middle-end/57251
17074 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17075 the case when both op0 and op1 have VOIDmode.
17077 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17079 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17080 in multiply-accumulate mode.
17082 2013-05-13 Guozhi Wei <carrot@google.com>
17084 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17086 2013-05-13 Kai Tietz <ktietz@redhat.com>
17089 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17090 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17091 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17092 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17093 (ix86_expand_prologue): Likewise.
17094 (ix86_expand_split_stack_prologue): Likewise.
17095 (legitimate_pic_address_disp_p): Likewise.
17096 (legitimize_pic_address): Likewise.
17097 (legitimize_tls_address): Likewise.
17098 (legitimize_pe_coff_symbol): Likewise.
17099 (output_pic_addr_const): Likewise.
17100 (construct_plt_address): Likewise.
17101 (ix86_expand_call): Likewise.
17102 (x86_output_mi_thunk): Likewise.
17103 (x86_function_profiler): Likewise.
17105 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17107 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17108 similar switch cases.
17109 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17110 (aarch64_simd_mov_to_<mode>low): Delete.
17111 (aarch64_simd_mov_to_<mode>high): Delete.
17112 (move_lo_quad_<mode>): Add w<-r alternative.
17113 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17114 (aarch64_simd_mov_from_*): Update type attribute.
17115 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17118 2013-05-13 Jan Hubicka <jh@suse.cz>
17120 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17121 * config/i386/i386.c (ix86_compute_frame_layout,
17122 ix86_expand_epilogue, emit_i387_cw_initialization,
17123 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17124 ix86_local_alignment): Fix use of size/speed predicates.
17126 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17128 PR tree-optimization/45216
17129 PR tree-optimization/57157
17130 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17131 the (-Y) & (B - 1) variant if OP is |.
17132 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17133 return shifted. Use (-op1) & (prec - 1) as other_amount
17134 instead of prec - op1.
17136 2013-05-13 Martin Jambor <mjambor@suse.cz>
17138 PR middle-end/42371
17139 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17140 (ipa_constant_data): New type.
17141 (ipa_jump_func): Use ipa_constant_data to hold information about
17142 constant jump functions.
17143 (ipa_get_jf_constant): Adjust to jump function type changes.
17144 (ipa_get_jf_constant_rdesc): New function.
17145 (ipa_param_descriptor): New field controlled_uses.
17146 (ipa_get_controlled_uses): New function.
17147 (ipa_set_controlled_uses): Likewise.
17148 * ipa-ref.h (ipa_find_reference): Declare.
17149 * ipa-prop.c (ipa_cst_ref_desc): New type.
17150 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17152 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17153 New parameter cs. Adjust all callers.
17154 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17155 (remove_described_reference): New function.
17156 (jfunc_rdesc_usable): Likewise.
17157 (try_make_edge_direct_simple_call): Decrement controlled use count,
17158 attempt to remove reference if it hits zero.
17159 (combine_controlled_uses_counters): New function.
17160 (propagate_controlled_uses): Likewise.
17161 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17162 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17163 (ipa_print_node_params): Print described use counter.
17164 (ipa_write_jump_function): Adjust to jump function type changes.
17165 (ipa_read_jump_function): New parameter CS, pass it to
17166 ipa_set_jf_constant. Adjust caller.
17167 (ipa_write_node_info): Stream controlled use count
17168 (ipa_read_node_info): Likewise.
17169 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17171 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17172 count. Remove cloning-added reference if it reaches zero.
17173 * ipa-ref.c (ipa_find_reference): New function.
17175 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17177 * config/i386/i386.c (processor_target_table): Modified default
17178 alignment values for AMD BD and BT architectures.
17180 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17182 * tree-vect-generic.c (uniform_vector_p): Move ...
17183 * tree.c (uniform_vector_p): ... here.
17184 * tree.h (uniform_vector_p): Declare it.
17185 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17188 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17190 PR tree-optimization/57230
17191 * tree-ssa-strlen.c (handle_char_store): Record length for
17192 array store from STRING_CST.
17194 PR tree-optimization/57230
17195 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17198 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17200 * config/epiphany/epiphany.c (epiphany_init): Check size of
17201 NUM_MODES_FOR_MODE_SWITCHING.
17202 (epiphany_expand_prologue):
17203 Remove CONFIG_REGNUM initial value handling code.
17204 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17205 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17206 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17207 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17208 Don't return 1 for FP_MODE_NONE.
17209 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17210 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17211 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17212 * config/epiphany/epiphany.md (save_config): New pattern.
17214 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17216 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17218 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17220 * config/i386/i386.md (memory): Handle sseishft1.
17221 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17222 (*vec_extractv2di_1): Ditto.
17224 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17226 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17229 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17231 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17232 Add mthumb/march=armv7-a multilib.
17233 Add mthumb/march=armv7-r multilib.
17234 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17236 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17238 * config/v850/t-rtems: Add more multilibs.
17240 2013-05-10 Richard Biener <rguenther@suse.de>
17242 PR tree-optimization/57214
17243 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17244 not propagate from SSA names that occur in abnormal PHI nodes.
17246 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17248 * stor-layout.c (element_precision): New function.
17249 * machmode.h (element_precision): Declare it.
17250 * tree.c (build_minus_one_cst): New function.
17251 (element_precision): Likewise.
17252 * tree.h (build_minus_one_cst): Declare new function.
17253 (element_precision): Likewise.
17254 * fold-const.c (operand_equal_p): Use element_precision.
17255 (fold_binary_loc): Handle vector types.
17256 * convert.c (convert_to_integer): Use element_precision.
17257 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17260 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17262 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17263 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17264 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17265 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17266 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17267 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17268 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17269 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17270 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17271 (Uuw8): New constraints.
17272 (Usb4): Move into alphabetical order.
17273 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17274 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17275 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17276 previously unnamed patterns.
17277 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17278 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17279 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17280 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17281 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17282 of specific lengths.
17284 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17286 * config/i386/i386.md (rotateinv): New code attr.
17287 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17288 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17289 roll $31, %eax, etc.
17291 PR tree-optimization/45216
17292 PR tree-optimization/57157
17293 * tree-ssa-forwprop.c (simplify_rotate): New function.
17294 (ssa_forward_propagate_and_combine): Call it.
17296 2013-05-10 Richard Biener <rguenther@suse.de>
17298 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17299 disable peeling when we version for aliasing.
17300 (vector_alignment_reachable_p): Honor explicit user alignment.
17301 (vect_supportable_dr_alignment): Likewise.
17302 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17303 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17304 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17305 then peeling to arrange for the cost-model check to come first.
17307 2013-05-10 Alan Modra <amodra@gmail.com>
17309 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17310 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17311 * configure: Regenerate.
17313 2013-05-10 Alan Modra <amodra@gmail.com>
17316 * varasm.c (default_elf_select_section): Move !DECL_P check..
17317 (get_named_section): ..to here before calling get_section_name.
17319 (default_section_type_flags): Add DECL_P check.
17320 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17321 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17323 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17325 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17326 When using gen_stack_adjust_str with a register offset, add a
17327 REG_FRAME_RELATED_EXPR note.
17329 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17331 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17332 (*vec_extractv4si_zext_mem): Ditto.
17333 (*vec_extractv2di): Add 0->x and x->x alternatives.
17334 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17335 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17337 2013-05-09 Jason Merrill <jason@redhat.com>
17339 N3639 C++1y VLA support
17340 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17343 * tree.c (build_constructor_va): New.
17344 * tree.h: Declare it.
17346 2013-05-09 Martin Jambor <mjambor@suse.cz>
17349 * gimple-fold.c (canonicalize_constructor_val): Call
17350 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17352 2013-05-09 Jan Hubicka <jh@suse.cz>
17353 Richard Biener <rguenther@suse.de>
17356 * symtab.c (symtab_make_decl_local): Do not add private names.
17358 2013-05-09 Jan Hubicka <jh@suse.cz>
17361 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17362 (unlink_from_assembler_name_hash): Likewise.
17363 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17364 symtab_unregister_node, symtab_initialize_asm_name_hash,
17365 change_decl_assembler_name): Update.
17367 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17369 * config/aarch64/aarch64.md: New movtf split.
17370 (*movtf_aarch64): Update.
17371 (aarch64_movdi_tilow): Handle TF modes and rename to
17372 aarch64_movdi_<mode>low.
17373 (aarch64_movdi_tihigh): Handle TF modes and rename to
17374 aarch64_movdi_<mode>high
17375 (aarch64_movtihigh_di): Handle TF modes and rename to
17376 aarch64_mov<mode>high_di
17377 (aarch64_movtilow_di): Handle TF modes and rename to
17378 aarch64_mov<mode>low_di
17379 (aarch64_movtilow_tilow): Remove spurious whitespace.
17380 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17382 (aarch64_print_operand): Update.
17384 2013-05-09 Alan Modra <amodra@gmail.com>
17386 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17388 * configure: Regenerate.
17390 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17392 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17393 splitter preparation statements.
17394 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17395 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17398 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17400 * gimple-ssa-strength-reduction.c (count_candidates): Change
17401 return value to int.
17402 (analyze_candidates_and_replace): Change type of length to int.
17404 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17406 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17407 (*vec_extract<mode>): Use VI12_128 mode iterator.
17408 (*vec_extract<mode>_mem): Ditto.
17409 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17412 2013-05-08 Diego Novillo <dnovillo@google.com>
17417 2012-08-17 Diego Novillo <dnovillo@google.com>
17420 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17421 * config.in: Regenerate.
17422 * configure: Regenerate.
17423 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17425 2013-05-08 Jan Hubicka <jh@suse.cz>
17428 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17429 * cgraph.h (symtab_node_base): Add unique_name.
17430 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17431 input_overwrite_node, input_varpool_node): Stream unique_name.
17432 * cgraphclones.c (cgraph_create_virtual_clone,
17433 cgraph_function_versioning): Set unique_name.
17434 * ipa.c (function_and_variable_visibility): Set unique_name.
17436 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17438 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17439 (alloc_cand_and_find_basis): Restrict conditional candidate
17440 processing to CAND_MULTs.
17442 2013-05-08 Jan Hubicka <jh@suse.cz>
17445 lto-symtab.c (lto_symtab_symbol_p): New function.
17446 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17447 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17448 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17449 Skip static symbols.
17451 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17453 PR tree-optimization/57200
17454 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17455 Only call inform if the preceding warning_at returns true.
17457 2013-05-07 Han Shen <shenhan@google.com>
17459 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17460 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17461 * common.opt (fstack-protector-strong): New option.
17462 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17463 * doc/invoke.texi (Optimization Options): Document
17464 "-fstack-protector-strong".
17465 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17467 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17469 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17471 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17473 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17474 it is not lhs of the insn.
17475 (setup_reg_equiv): Remove insn having equiv memory which it is not
17477 * lra-constraints.c (process_address): Try to improve generation
17478 code for address base + disp.
17479 (lra_constraints): Make correct the code for checking insn setting
17480 up backward equivalence. Remove insn only if it is in the init
17482 * lra-eliminations.c (update_reg_eliminate): Change return value.
17483 (lra_eliminate): Use the result.
17485 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17487 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17488 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17489 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17490 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17491 register target operands.
17492 (*vec_extractv8hi_sse2): New pattern.
17493 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17494 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17495 (*vec_extract<mode>_mem): New insn and split pattern.
17497 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17499 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17500 (TARGET_ASAN_SHADOW_OFFSET): Define.
17501 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17502 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17504 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17506 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17507 (incr_vec_index): Return -1 if increment not found.
17508 (create_add_on_incoming_edge): Assert if increment not found.
17509 (record_increment): Limit number of increments recorded.
17510 (all_phi_incrs_profitable): Return false if an increment not found.
17511 (replace_profitable_candidates): Don't process increments that were
17513 (analyze_candidates_and_replace): Limit size of incr_vec.
17515 2013-05-07 Richard Biener <rguenther@suse.de>
17517 * calls.c (special_function_p): setjmp-like functions are leaf.
17518 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17519 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17521 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17523 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17524 (aarch64_simd_mov<mode>): New expander.
17525 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17526 (aarch64_simd_mov_to_<mode>high): Likewise.
17527 (aarch64_simd_mov_from_<mode>low): Likewise.
17528 (aarch64_simd_mov_from_<mode>high): Likewise.
17529 (aarch64_dup_lane<mode>): Update.
17530 (aarch64_dup_lanedi): New instruction pattern.
17531 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17532 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17534 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17536 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17537 (replace_mult_candidate): Remove unnecessary argument; remove
17538 unnecessary parameter from call to introduce_cast_before_cand.
17539 (replace_unconditional_candidate): Remove unnecessary parameter
17540 from call to replace_mult_candidate.
17541 (replace_conditional_candidate): Likewise.
17542 (insert_initializers): Use make_temp_ssa_name.
17543 (introduce_cast_before_cand): Remove unnecessary argument; use
17544 make_temp_ssa_name.
17545 (replace_one_candidate): Remove unnecessary argument; remove
17546 unnecessary parameter from calls to introduce_cast_before_cand.
17547 (replace_profitable_candidates): Remove unnecessary parameters
17548 from calls to replace_one_candidate.
17550 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17552 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17553 phi def as possibly hiding a basis for a CAND_ADD whose operands
17554 have been commuted in the analysis.
17555 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17557 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17559 * config/aarch64/aarch64.md
17560 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17561 shift value between 0-4.
17563 2013-05-07 Richard Biener <rguenther@suse.de>
17565 * double-int.h (rshift): New overload.
17566 * double-int.c (rshift): New function.
17567 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17568 (create_reference_ops_from_ref): Remove.
17569 (vn_reference_insert): Use shared ops for constructing the
17570 reference and copy it.
17572 2013-05-07 Richard Biener <rguenther@suse.de>
17574 PR middle-end/57190
17575 * tree-eh.c (sink_clobbers): Properly propagate
17576 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17578 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17580 PR tree-optimization/57149
17581 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17582 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17583 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17584 uninit_undefined_value_p instead of ssa_undefined_value_p.
17587 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17588 for modifier == EXPAND_INITIALIZER.
17590 2013-05-07 Anton Blanchard <anton@samba.org>
17592 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17593 for powerpc64 little endian.
17594 * configure: Regenerate.
17596 2013-05-06 Graham Stott <grahams@btinternet.com>
17598 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17599 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17600 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17601 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17603 2013-05-06 Graham Stott <grahams@btinternet.com>
17605 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17606 codes which allow non-lvalues.
17608 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17610 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17611 components are all 1s.
17612 (integer_minus_onep): New function.
17613 * tree.h (integer_minus_onep): Declare it.
17614 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17615 integer_minus_onep instead of integer_all_onesp.
17617 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17620 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17621 variations of these patterns.
17623 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17625 * config/i386/i386.md (isa): Add x64_sse4 member.
17626 (enabled): Handle x64_sse4.
17627 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17628 instruction for 64bit SSE4_1 targets. Update insn attributes.
17629 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17630 instruction for SSE4_1 targets. Update insn attributes.
17631 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17632 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17633 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17635 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17636 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17637 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17639 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17642 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17644 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17646 * final.c (do_assembler_dialects): Don't handle curly braces and
17647 vertical bar escaped by % as dialect delimiters.
17648 (output_asm_insn): Print curly braces and vertical bar if escaped
17649 by % and ASSEMBLER_DIALECT defined.
17650 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17651 * doc/tm.texi: Regenerated.
17653 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17655 * config/mips/mips.c: Include tree-pass.h.
17656 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17657 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17658 (pass_mips_machine_reorg2): New machine specific pass.
17659 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17660 (mips_option_override): Register the new pass.
17661 * rtl.h (cleanup_barriers): Remove prototype.
17662 (dbr_schedule): Likewise.
17663 * jump.c (cleanup_barriers): Make static.
17664 * reorg.c (dbr_schedule): Likewise.
17666 2013-05-06 Richard Biener <rguenther@suse.de>
17668 PR tree-optimization/57185
17669 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17670 SSA names properly.
17672 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17675 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17676 between "first_arg" and "insn", not "last" and "insn".
17678 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17680 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17681 (find_candidates_in_block): Re-enable slsr_process_phi.
17682 (create_phi_basis): Fix double counting of candidate adjustment.
17684 2013-05-06 Richard Biener <rguenther@suse.de>
17686 PR middle-end/57147
17687 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17688 the edge is also fallthru, preserve it and just clear the
17690 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17691 also complex, preserve that and just clear the fallthru flag.
17692 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17693 update virtual operands.
17695 2013-05-06 Alan Modra <amodra@gmail.com>
17697 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17698 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17699 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17700 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17701 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17702 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17704 2013-05-06 Alan Modra <amodra@gmail.com>
17706 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17707 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17708 (DEFAULT_ASM_ENDIAN): Define.
17709 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17710 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17711 Update -K PIC clause from sysv4.h.
17712 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17713 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17715 2013-05-06 Alan Modra <amodra@gmail.com>
17717 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17718 twice for little-endian.
17719 (ashrdi3_no_power, ashrdi3): Support little-endian.
17721 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17724 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17725 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17727 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17728 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17729 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17732 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17733 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17735 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17736 * config/pa/pa.opt: Make mbig-switch a no-op.
17737 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17738 (CASE_VECTOR_MODE): Always return SImode.
17739 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17740 for the !TARGET_BIG_SWITCH case.
17741 * config/pa/pa-linux.h: Likewise.
17742 * config/pa/pa-openbsd.h: Likewise.
17743 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17744 * config/pa/pa.md (short_jump): Remove define_insn.
17745 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17746 (casesi0): Remove define_insn.
17747 (type): Remove btable_branch.
17748 (pa_combine_type): Likewise.
17749 (in_nullified_branch_delay): Likewise.
17750 (in_call_delay): Likewise.
17751 (define_delay): Likewise.
17752 (define_insn_reservation "Z3"): Likewise.
17753 (define_insn_reservation "Z4"): Likewise.
17754 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17755 (pa_adjust_insn_length): Remove adjustment for btable branches.
17756 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17759 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17761 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17762 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17763 Add m->r,x alternatives.
17764 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17765 splitters using SWI48x mode iterator.
17766 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17767 TARGET_64BIT. Add m->x alternative.
17768 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17769 Add o->x alternative. Enable for TARGET_SSE.
17770 (sse_storeq): Remove expander.
17771 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17772 with memory input operand.
17773 (*vec_extractv2di_1 splitter): New.
17774 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17775 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17777 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17779 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17780 (INT_LOWPART): Delete.
17781 (extract_MB): Adjust.
17782 (extract_ME): Adjust.
17783 (print_operand): Adjust.
17785 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17787 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17788 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17789 (reg_or_logical_cint_operand, easy_fp_constant,
17790 logical_const_operand): Delete "CONST_DOUBLE" case.
17791 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17792 "HOST_BITS_PER_WIDE_INT == 64" test.
17793 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17794 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17795 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17796 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17798 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17799 CONST_DOUBLE DImode/VOIDmode case.
17800 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17801 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17802 CONST_DOUBLE VOIDmode case.
17803 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17804 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17805 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17806 Delete CONST_DOUBLE case.
17807 (splitters for mov FMOVE64 const_double): Delete
17808 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17809 "HOST_BITS_PER_WIDE_INT >= 64" test.
17810 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17812 (mov DI const_double): Delete.
17814 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17816 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17817 on op shows all bits zero in mode of a lowpart subreg, return zero.
17819 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17822 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17823 to save TFmode registers and DImode to save TImode registers for
17824 caller save operations.
17825 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17826 mark being partially clobbered since they only use the first
17829 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17830 and TDmode only use the upper 64-bits of each VSX register.
17832 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17834 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17835 (find_candidates_in_block): Disable slsr_process_phi.
17837 2013-05-03 Guozhi Wei <carrot@google.com>
17839 * coverage.c (coverage_obj_init): Move the construction of gcov
17841 (build_init_ctor): ... here.
17843 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17845 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17846 (slsr_cand_d): Redefine def_phi.
17847 (stride_status, phi_adjust_status, count_phis_status): New enums.
17848 (find_phi_def): New.
17849 (find_basis_for_base_expr): New.
17850 (find_basis_for_candidate): Handle hidden bases.
17851 (alloc_cand_and_find_basis): Handle phi candidates.
17852 (slsr_process_phi): New.
17853 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17854 (create_mul_imm_cand): Likewise.
17855 (create_add_ssa_cand): Exclude phi base candidates.
17856 (create_add_imm_cand): Likewise.
17857 (slsr_process_cast): Likewise.
17858 (slsr_process_copy): Likewise.
17859 (find_candidates_in_block): Handle phi candidates.
17860 (dump_candidate): Likewise.
17861 (unconditional_cands): Delete.
17862 (unconditional_cands_with_known_stride_p): Delete.
17863 (phi_dependent_cand_p): New.
17864 (cand_increment): Handle phi-dependent candidates.
17865 (replace_dependent): Delete.
17866 (replace_mult_candidate): New.
17867 (replace_unconditional_candidate): New.
17868 (incr_vec_index): Move to avoid forward reference.
17869 (create_add_on_incoming_edge): New.
17870 (create_phi_basis): New.
17871 (replace_dependents): Delete.
17872 (replace_conditional_candidate): New.
17873 (phi_add_costs): New.
17874 (replace_uncond_cands_and_profitable_phis): New.
17875 (record_increment): Handle phi adjustments.
17876 (record_phi_increments): New.
17877 (record_increments): Handle phi adjustments.
17878 (phi_incr_cost): New.
17879 (lowest_cost_path): Handle phis.
17880 (total_savings): Likewise.
17881 (analyze_increments): Likewise.
17882 (ncd_with_phi): New.
17883 (ncd_of_cand_and_phis): New.
17884 (nearest_common_dominator_for_cands): Handle phi increments.
17885 (all_phi_incrs_profitable): New.
17886 (replace_profitable_candidates): Handle phi-dependent candidates.
17887 (analyze_candidates_and_replace): Likewise.
17889 2013-05-03 Teresa Johnson <tejohnson@google.com>
17892 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17893 do not exceed REG_BR_PROB_BASE.
17895 2013-05-03 Jeff Law <law@redhat.com>
17897 PR tree-optimization/57144
17898 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17899 operand of the condition will bit into the new type when eliminating
17900 a cast feeding a condition.
17902 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17904 PR rtl-optimization/57130
17905 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17906 of COMPARE as in_code to the recursive call if needed.
17908 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17910 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17911 (enabled): Handle new members.
17912 * config/i386/sse.md (*vec_concatv2si): Merge from
17913 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17914 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17916 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17918 PR tree-optimization/57027
17919 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17920 for fnms opportunity, check we got the prerequisite kind
17921 of tree / gimple before using accessor functions.
17923 2013-05-03 Richard Biener <rguenther@suse.de>
17925 * double-int.h (lshift): New overload without precision
17926 and arith argument.
17927 (operator *=, operator +=, operator -=): Move ...
17928 * double-int.c (operator *=, operator +=, operator -=): ... here
17929 and implement more efficiently.
17930 (mul_double_with_sign): Remove.
17931 (lshift_double): Adjust to take unsinged shift argument, push
17932 dispatching code to callers.
17933 (mul_double_wide_with_sign): Add early out for callers that
17934 are not interested in high parts or overflow.
17935 (lshift): New function.
17936 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17937 dispatch code here.
17938 (lrotate, rrotate): Use logical shifts.
17939 * expr.c (get_inner_reference): Use lshift.
17940 * fixed-value.c (do_fixed_divide): Likewise.
17941 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17942 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17943 (indirect_refs_may_alias_p): Likewise.
17944 (stmt_kills_ref_p_1): Likewise.
17946 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17948 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17950 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17952 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17953 scalar form of FABD instruction.
17955 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17957 * lra-constraints.c (process_alt_operands): Add checking alt
17958 number to choose the best alternative.
17960 2013-05-02 Richard Biener <rguenther@suse.de>
17962 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17963 bitmap and its handling.
17964 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17966 2013-05-02 Richard Biener <rguenther@suse.de>
17968 PR middle-end/57140
17969 * tree-inline.c (copy_loops): Properly handle removed loops.
17970 (copy_cfg_body): Mark destination loops for fixup if source
17971 loops needed fixup.
17973 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17976 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17977 generating simple_return for naked functions.
17979 2013-05-02 Martin Jambor <mjambor@suse.cz>
17981 PR middle-end/56988
17982 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17983 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17985 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17986 ipa_agg_replacement_value structures.
17987 (known_aggs_to_agg_replacement_list): Likewise.
17988 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17989 (read_agg_replacement_chain): Likewise.
17990 (ipcp_transform_function): Also check that by_ref flags match.
17992 2013-05-02 Richard Biener <rguenther@suse.de>
17994 * graphds.h (struct graph): Add obstack member.
17995 * graphds.c (new_graph): Initialize obstack and allocate
17997 (add_edge): Allocate edge from the obstack.
17998 (free_graph): Free the obstack instead of all edges and vertices.
18000 2013-05-02 Teresa Johnson <tejohnson@google.com>
18002 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18004 * cfg.c (update_bb_profile_for_threading): Ditto.
18005 * tree-inline.c (copy_bb): Ditto.
18006 (copy_edges_for_bb): Ditto.
18007 (initialize_cfun): Ditto.
18008 (copy_cfg_body): Ditto.
18009 (expand_call_inline): Ditto.
18010 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18011 (estimate_node_size_and_time): Ditto.
18012 (inline_merge_summary): Ditto.
18013 * cgraphclones.c (cgraph_clone_edge): Ditto.
18014 (cgraph_clone_node): Ditto.
18015 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18016 (compute_trg_info): Ditto.
18018 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18020 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18021 S reg when fp attribute set.
18022 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18024 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18026 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18028 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18029 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18030 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18032 2013-05-02 Richard Biener <rguenther@suse.de>
18034 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18035 (struct instantiate_cache_entry): New type.
18036 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18037 (struct instantiate_cache_type): New type.
18038 (set_instantiated_value, get_instantiated_value): Remove.
18039 (get_instantiated_value_entry): New function.
18040 (instantiate_scev_name): Use the new cache and adjust.
18041 (instantiate_scev_poly): Adjust.
18042 (instantiate_scev_binary): Likewise.
18043 (instantiate_array_ref): Likewise.
18044 (instantiate_scev_convert): Likewise.
18045 (instantiate_scev_not): Likewise.
18046 (instantiate_scev_3): Likewise.
18047 (instantiate_scev_2): Likewise.
18048 (instantiate_scev_r): Likewise.
18049 (instantiate_scev): Likewise.
18050 (resolve_mixers): Likewise.
18052 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18055 * lra-constraints.c (best_small_class_operands_num): Remove.
18056 (process_alt_operands): Remove small_class_operands_num. Take
18057 small classes operands into losers and only if the operand is not
18058 matched. Modify debugging output.
18059 (curr_insn_transform): Remove best_small_class_operands_num.
18062 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18064 * config/aarch64/aarch64-builtins.c
18065 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18066 * config/aarch64/aarch64-simd-builtins.def
18067 (reduc_splus_): Add new modes.
18068 (reduc_uplus_): New.
18069 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18070 (reduc_uplus_v4sf): Likewise.
18071 (reduc_splus_v4sf): Likewise.
18072 (aarch64_addv<mode>): Likewise.
18073 (reduc_uplus_<mode>): Likewise.
18074 (reduc_splus_<mode>): Likewise.
18075 (aarch64_addvv2di): Likewise.
18076 (reduc_uplus_v2di): Likewise.
18077 (reduc_splus_v2di): Likewise.
18078 (aarch64_addvv2si): Likewise.
18079 (reduc_uplus_v2si): Likewise.
18080 (reduc_splus_v2si): Likewise.
18081 (reduc_<sur>plus_<mode>): New.
18082 (reduc_<sur>plus_v2di): Likewise.
18083 (reduc_<sur>plus_v2si): Likewise.
18084 (reduc_<sur>plus_v4sf): Likewise.
18085 (aarch64_addpv4sf): Likewise.
18086 * config/aarch64/arm_neon.h
18087 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18088 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18089 add UNSPEC_SADDV, UNSPEC_UADDV.
18091 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18093 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18095 * config/aarch64/arm_neon.h
18096 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18098 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18100 * config/aarch64/aarch64-builtins
18101 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18103 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18105 * config/aarch64/aarch64-simd-builtins.def
18106 (reduc_smax_): New.
18107 (reduc_smin_): Likewise.
18108 (reduc_umax_): Likewise.
18109 (reduc_umin_): Likewise.
18110 (reduc_smax_nan_): Likewise.
18111 (reduc_smin_nan_): Likewise.
18114 (smax): Update for V2SF, V4SF and V2DF modes.
18117 (smin_nan): Likewise.
18118 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18119 (<su><maxmin><mode>3): ...This, refactor.
18120 (s<maxmin><mode>3): New.
18121 (<maxmin_uns><mode>3): Likewise.
18122 (reduc_<maxmin_uns>_<mode>): Refactor.
18123 (reduc_<maxmin_uns>_v4sf): Likewise.
18124 (reduc_<maxmin_uns>_v2si): Likewise.
18125 (aarch64_<fmaxmin><mode>: Remove.
18126 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18128 (vmin<q>_f<32,64>): Likewise.
18129 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18131 (su): Add mappings for smax, smin, umax, umin.
18133 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18134 (FMAXMIN): Rename as...
18135 (FMAXMIN_UNS): ...This.
18137 (fmaxminv): Likewise.
18138 (fmaxmin): Likewise.
18140 (maxmin_uns_op): Likewise.
18142 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18144 * config/aarch64/arm_neon.h
18145 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18146 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18147 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18149 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18151 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18152 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18154 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18156 * config/aarch64/aarch64-simd.md
18157 (vcond<mode>_internal): Handle special cases for constant masks.
18158 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18159 (vcondu<mode><mode>): Likewise.
18160 (vcond<v_cmp_result><mode>): New.
18162 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18164 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18165 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18166 * config/aarch64/aarch64-simd-builtins.def
18167 (cmeq): Update to BUILTIN_VALLDI.
18172 * config/aarch64/arm_neon.h
18173 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18174 to builtins or C as appropriate.
18176 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18178 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18180 (cmhi): Rename to...
18182 * config/aarch64/aarch64-simd.md
18183 (simd_mode): Add SF.
18184 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18185 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18186 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18187 (cstore<mode>_neg): ...This.
18188 * config/aarch64/iterators.md
18190 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18191 (COMPARISONS): New.
18192 (UCOMPARISONS): Likewise.
18193 (optab): Add missing comparisons.
18199 (VCMP_S): Likewise.
18200 (VCMP_U): Likewise.
18201 (V_cmp_result): Add DF, SF modes.
18202 (v_cmp_result): Likewise.
18204 (vmtype): Likewise.
18205 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18207 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18209 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18210 define_insn to define_insn_and_split.
18211 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18212 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18213 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18214 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18215 (thumb2_negscc): Likewise.
18217 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18219 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18221 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18223 * config/arm/thumb2.md: Remove trailing whitespaces.
18225 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18227 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18228 Use gen_int_mode rather than GEN_INT.
18230 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18232 * value-prof.c (stream_in_histogram_value): Remove the strayed
18235 2013-04-30 Richard Biener <rguenther@suse.de>
18237 PR middle-end/57122
18238 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18240 2013-04-30 Richard Biener <rguenther@suse.de>
18242 PR middle-end/57107
18243 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18245 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18247 PR rtl-optimization/56957
18248 PR rtl-optimization/57105
18249 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18250 variable. Use just INSN_UID for determining whether an insn
18251 should be only disconnected from the insn stream.
18252 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18254 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18256 PR tree-optimization/57104
18257 * tsan.c (instrument_expr): Don't instrument accesses to
18258 DECL_HARD_REGISTER VAR_DECLs.
18260 2013-04-30 Richard Biener <rguenther@suse.de>
18262 * function.h (loops_for_fn): New inline function.
18263 (set_loops_for_fn): Likewise.
18264 * cfgloop.h (place_new_loop): Add struct function parameter.
18265 (get_loop): Likewise.
18266 (get_loops): Likewise.
18267 (number_of_loops): Likewise.
18268 (fel_next): Adjust.
18269 (fel_init): Likewise.
18270 * cfg.c (get_loop_copy): Adjust.
18271 * cfgloop.c (flow_loops_dump): Likewise.
18272 (record_loop_exits): Likewise.
18273 (verify_loop_structure): Likewise.
18274 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18275 (estimate_reg_pressure_cost): Likewise.
18276 (mark_loop_exit_edges): Likewise.
18277 * cfgloopmanip.c (place_new_loop): Likewise.
18278 (add_loop): Likewise.
18279 (duplicate_loop): Likewise.
18280 * graph.c (draw_cfg_nodes): Likewise.
18281 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18282 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18283 (extract_affine_chrec): Likewise.
18284 (build_scop_iteration_domain): Likewise.
18285 * graphite.c (graphite_initialize): Likewise.
18286 * ira-build.c (create_loop_tree_nodes): Likewise.
18287 (more_one_region_p): Likewise.
18288 (rebuild_regno_allocno_maps): Likewise.
18289 (mark_loops_for_removal): Likewise.
18290 (mark_all_loops_for_removal): Likewise.
18291 (remove_unnecessary_regions): Likewise.
18292 (ira_build): Likewise.
18293 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18294 * loop-init.c (fix_loop_structure): Likewise.
18295 (gate_rtl_move_loop_invariants): Likewise.
18296 (gate_rtl_unswitch): Likewise.
18297 (gate_rtl_unroll_and_peel_loops): Likewise.
18298 (rtl_doloop): Likewise.
18299 * lto-streamer-in.c (input_cfg): Likewise.
18300 * lto-streamer-out.c (output_cfg): Likewise.
18301 * modulo-sched.c (sms_schedule): Likewise.
18302 * predict.c (tree_estimate_probability): Likewise.
18303 (tree_estimate_probability_driver): Likewise.
18304 (estimate_loops): Likewise.
18305 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18306 (move_sese_region_to_fn): Likewise.
18307 (debug_loop_num): Likewise.
18308 * tree-chrec.c (chrec_evaluate): Likewise.
18309 (hide_evolution_in_other_loops_than_loop): Likewise.
18310 (chrec_component_in_loop_num): Likewise.
18311 (reset_evolution_in_loop): Likewise.
18312 (evolution_function_is_invariant_rec_p): Likewise.
18313 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18314 * tree-inline.c (copy_loops): Likewise.
18315 (copy_cfg_body): Likewise.
18316 (tree_function_versioning): Likewise.
18317 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18318 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18320 (add_to_evolution_1): Likewise.
18321 (scev_const_prop): Likewise.
18322 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18323 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18324 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18325 (tree_ssa_lim_initialize): Likewise.
18326 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18327 (verify_loop_closed_ssa): Likewise.
18328 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18329 (tree_ssa_loop_im): Likewise.
18330 (tree_ssa_loop_unswitch): Likewise.
18331 (tree_vectorize): Likewise.
18332 (check_data_deps): Likewise.
18333 (tree_ssa_loop_ivcanon): Likewise.
18334 (tree_ssa_loop_bounds): Likewise.
18335 (tree_complete_unroll): Likewise.
18336 (tree_complete_unroll_inner): Likewise.
18337 (tree_parallelize_loops): Likewise.
18338 (tree_ssa_loop_prefetch): Likewise.
18339 (tree_ssa_loop_ivopts): Likewise.
18340 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18341 * tree-vectorizer.c (vectorize_loops): Likewise.
18343 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18345 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18346 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18347 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18348 with EABI_LINK_SPEC.
18350 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18353 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18356 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18359 * lra-constraints.c (process_alt_operands): Discourage a bit more
18360 using memory for pseudos. Print cost dump for alternatives.
18361 Modify cost values for conflicts with early clobbers.
18362 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18364 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18367 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18369 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18371 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18372 from/to S register.
18373 (movdi_aarch64): Support LDR/STR from/to D register.
18375 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18377 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18378 or higher by default.
18380 2013-04-29 Richard Biener <rguenther@suse.de>
18382 PR middle-end/57075
18383 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18384 even if not adding abnormal edges for calls that can make
18387 2013-04-29 Richard Biener <rguenther@suse.de>
18389 PR middle-end/57103
18390 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18392 (move_stmt_r): Remove redundant checking.
18394 2013-04-29 Teresa Johnson <tejohnson@google.com>
18397 * basic-block.h (apply_scale): New function.
18398 (apply_probability): Use apply_scale.
18399 * gimple-streamer-in.c (input_bb): Ditto.
18400 * lto-streamer-in.c (input_cfg): Ditto.
18401 * lto-cgraph.c (merge_profile_summaries): Ditto.
18402 * tree-optimize.c (execute_fixup_cfg): Ditto.
18403 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18404 (copy_edges_for_bb): Ditto.
18405 (copy_cfg_body): Ditto.
18407 2013-04-29 Tom de Vries <tom@codesourcery.com>
18409 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18410 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18411 (tail_merge_optimize): Handle current_loops == NULL.
18413 2013-04-26 Jeff Law <law@redhat.com>
18415 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18416 (simplify_cond_using_ranges): Generalize code to simplify
18417 COND_EXPRs where one argument is a constant and the other
18418 is an SSA_NAME created by an integral type conversion.
18420 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18422 * config/arm/arm.md (store_minmaxsi): Use only when
18423 optimize_insn_for_size_p.
18425 2013-04-29 Christian Bruel <christian.bruel@st.com>
18428 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18430 2013-04-29 Richard Biener <rguenther@suse.de>
18432 PR middle-end/57089
18433 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18434 loop tree make sure to schedule a fixup for the child as well.
18435 (expand_omp_for_generic): Properly add loops.
18436 (expand_omp_for_static_nochunk): Likewise.
18437 (expand_omp_for_static_chunk): Likewise.
18438 (expand_omp_for): For the degenerate case fixup loops.
18439 (expand_omp_sections): Fix default bb placement in loops.
18440 (expand_omp_atomic_pipeline): Properly add loops.
18442 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18444 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18446 2013-04-29 Tom de Vries <tom@codesourcery.com>
18448 * tree-ssa-tail-merge.c: Update header comment.
18450 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18452 * config/aarch64/arm_neon.h
18453 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18454 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18455 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18456 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18457 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18458 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18459 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18460 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18462 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18464 * config/aarch64/aarch64-simd.md
18465 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18466 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18467 fix_trunc, fixuns_trunc.
18468 (ftrunc<VDQF:mode>2): New.
18469 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18470 (fix_trunc_optab): New.
18472 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18474 * config/aarch64/aarch64-builtins.c
18475 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18476 iceilf, lround, iroundf.
18478 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18481 * config/i386/i386.h (enum ix86_tune_indices)
18482 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18483 New, split from X86_TUNE_INTER_UNIT_MOVES.
18484 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18485 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18486 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18487 (TARGET_INTER_UNIT_MOVES): Remove.
18488 * config/i386/i386.c (initial_ix86_tune_features): Update.
18489 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18490 (ix86_expand_convert_uns_didf_sse): Use
18491 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18492 (ix86_expand_vector_init_one_nonzero): Ditto.
18493 (ix86_expand_vector_init_interleave): Ditto.
18494 (inline_secondary_memory_needed): Return true for moves from SSE class
18495 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18496 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18497 * config/i386/constraints.md (Yi, Ym): Depend on
18498 TARGET_INTER_UNIT_MOVES_TO_VEC.
18499 (Yj, Yn): New constraints.
18500 * config/i386/i386.md (*movdi_internal): Change constraints of
18501 operand 1 from Yi to Yj and from Ym to Yn.
18502 (*movsi_internal): Ditto.
18503 (*movdf_internal): Ditto.
18504 (*movsf_internal): Ditto.
18505 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18506 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18507 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18508 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18509 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18510 * config/i386/sse.md (movdi_to_sse): Ditto.
18511 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18512 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18513 TARGET_INTER_UNIT_MOVES.
18514 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18515 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18516 instead of TARGET_INTER_UNIT_MOVES.
18517 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18518 operand 1 from Yi to Yj and from Ym to Yn.
18520 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18522 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18523 (float_truncate_hi_): Likewise.
18524 (float_extend_lo_): Likewise.
18525 (float_truncate_lo_): Likewise.
18526 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18527 (aarch64_float_extend_lo_v2df): Likewise.
18528 (vec_unpacks_hi_v4sf): Likewise.
18529 (aarch64_float_truncate_lo_v2sf): Likewise.
18530 (aarch64_float_truncate_hi_v4sf): Likewise.
18531 (vec_pack_trunc_v2df): Likewise.
18532 (vec_pack_trunc_df): Likewise.
18534 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18536 * config/aarch64/aarch64-builtins.c
18537 (aarch64_fold_builtin): Fold float conversions.
18538 * config/aarch64/aarch64-simd-builtins.def
18539 (floatv2si, floatv4si, floatv2di): New.
18540 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18541 * config/aarch64/aarch64-simd.md
18542 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18543 * config/aarch64/iterators.md (FLOATUORS): New.
18544 (optab): Add float, floatuns.
18545 (su_optab): Likewise.
18547 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18549 * config/aarch64/aarch64-builtins.c
18550 (aarch64_builtin_vectorized_function): Use new names for
18552 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18553 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18554 (fcvtzu): Split as...
18555 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18556 (fcvtas): Split as...
18557 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18558 (fcvtau): Split as...
18559 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18560 (fcvtps): Split as...
18561 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18562 (fcvtpu): Split as...
18563 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18564 (fcvtms): Split as...
18565 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18566 (fcvtmu): Split as...
18567 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18568 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18569 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18570 (lfrintnusf, lfrintnudf): Likewise.
18571 * config/aarch64/aarch64-simd.md
18572 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18574 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18575 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18576 (fcvt_pattern): Likewise.
18578 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18580 * config/aarch64/aarch64-simd.md
18581 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18582 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18584 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18586 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18587 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18588 (vrnd<a,m,n,p>_f32): Implement using builtins.
18589 (vrnd<i,x><q>_f<32, 64>): New.
18591 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18593 * config/aarch64/aarch64-builtins.c
18594 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18595 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18596 (frintz): Rename to...
18598 (frintp): Rename to...
18600 (frintm): Rename to...
18602 (frinti): Rename to...
18603 (nearbyint): ...this.
18604 (frintx): Rename to...
18606 (frinta): Rename to...
18608 * config/aarch64/aarch64-simd.md
18609 (aarch64_frint<frint_suffix><mode>): Delete.
18610 (<frint_pattern><mode>2): Convert to insn.
18611 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18612 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18613 (frint_pattern): Likewise.
18614 (frint_suffix): Likewise.
18616 2013-04-29 Richard Biener <rguenther@suse.de>
18618 PR tree-optimization/57081
18619 * loop-init.c: Include tree-flow.h.
18620 (loop_optimizer_finalize): Free number of iteration estimates.
18621 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18623 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18625 PR tree-optimization/57083
18626 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18627 non-singleton shift count range, zero extend low_bound for uns case.
18629 * config/i386/predicates.md (general_vector_operand): New predicate.
18630 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18631 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18632 if they aren't nonimmediate operands. If their original values
18633 satisfy const_vector_equal_evenodd_p, don't shift them.
18634 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18635 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18636 if not nonimmediate_operand.
18637 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18638 instead of register_operand.
18639 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18641 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18643 * stor-layout.c (finalize_size_functions): Allocate a structure and
18644 reset cfun before dumping the functions.
18646 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18648 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18651 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18652 use xop_pmacsdqh if uns_p.
18653 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18654 the immediate rotate count.
18656 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18658 * rtl.h (struct rtx_def): Add comment for field jump.
18659 (LRA_SUBREG_P): New macro.
18660 * recog.c (register_operand): Check LRA_SUBREG_P.
18661 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18663 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18664 (lra_final_code_change): Skip subreg change for operators.
18665 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18666 if there are no operand changes.
18667 * lra-constraints.c (curr_insn_set): New.
18668 (match_reload): Set LRA_SUBREG_P.
18669 (emit_spill_move): Ditto.
18670 (check_and_process_move): Use curr_insn_set. Process only single
18671 set insns. Don't initialize sec_mem_p and change_p.
18672 (simplify_operand_subreg): Use LRA_SUBREG_P.
18673 (reg_in_class_p): New function.
18674 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18675 of #ifdef. Add code to remove cycling.
18676 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18677 non-null disp. Reload inner instead of disp when base and index
18678 are null. Try to put lo_sum into register.
18679 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18680 (check_and_process_move): Move code for move cost check to
18681 simple_move_p. Remove equiv_substitution.
18682 (simple_move_p): New function.
18683 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18684 curr_insn_set. Call check_and_process_move only for single set
18685 insns. Use the new function. Move call of check_and_process_move
18686 after operand equiv substitution and address process.
18688 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18691 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18692 with nonlocal goto receivers or returns twice calls, ignore
18693 unininitialized values from abnormal edges to nl goto receiver
18694 or returns twice call.
18696 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18698 PR tree-optimization/57051
18699 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18700 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18703 2013-04-26 Richard Biener <rguenther@suse.de>
18705 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18706 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18707 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18708 (fixup_loop_arrays_after_move): New function.
18709 (move_sese_region_to_fn): Properly outline the loop tree parts
18710 of the SESE region.
18712 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18714 * config/i386/i386.md (type, unit): Fix long lines.
18716 2013-04-26 Richard Biener <rguenther@suse.de>
18718 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18719 (lto-streamer-out.o): Likewise.
18720 * cfgloop.c (init_loops_structure): Export, add struct function
18721 argument and adjust.
18722 (flow_loops_find): Adjust.
18723 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18724 (init_loops_structure): Declare.
18725 * lto-streamer-in.c: Include cfgloop.h.
18726 (input_cfg): Input the loop tree.
18727 * lto-streamer-out.c: Include cfgloop.h.
18728 (output_cfg): Output the loop tree.
18729 (output_struct_function_base): Do not drop PROP_loops.
18731 2013-03-26 Richard Biener <rguenther@suse.de>
18733 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18734 (pass_build_cfg): Provide PROP_loops.
18735 (move_sese_region_to_fn): Remove loops that are outlined into fn
18737 * tree-inline.c: Include cfgloop.h.
18738 (initialize_cfun): Do not drop PROP_loops.
18739 (copy_loops): New function.
18740 (copy_cfg_body): Copy loop structure.
18741 (tree_function_versioning): Initialize destination loop tree.
18742 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18743 (pass_parallelize_loops): Do IL verification.
18744 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18745 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18746 the CFG make sure we fixup loops as well.
18747 * tree-ssa-tail-merge.c: Include cfgloop.h.
18748 (replace_block_by): When merging loop latches mark loops for fixup.
18749 * lto-streamer-out.c (output_struct_function_base): Drop
18750 PROP_loops for now.
18751 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18752 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18753 * ipa-split.c: Include cfgloop.h.
18754 (split_function): Add the new return block to the loop tree root.
18755 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18756 whether we have removed the forwarder block.
18757 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18758 * cfgloop.h (place_new_loop): Declare.
18759 * cfgloopmanip.c (place_new_loop): Export.
18760 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18761 (tree-switch-conversion.o): Likewise.
18762 (tree-complex.o): Likewise.
18763 (tree-inline.o): Likewise.
18764 (tree-ssa-tailmerge.o): Likewise.
18765 (ipa-split.o): Likewise.
18766 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18767 (tree-ssa-copy.o): Likewise.
18768 * tree-switch-conversion.c: Include cfgloop.h
18769 (process_switch): If we emit a bit-test cascade, schedule loops
18771 * tree-complex.c: Include cfgloop.h.
18772 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18773 * asan.c: Include cfgloop.h.
18774 (create_cond_insert_point): Properly add new basic-blocks to
18775 loops, schedule loop fixup.
18776 * cfgloop.c (verify_loop_structure): Check that looks are not
18778 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18780 (expand_omp_for_generic): Likewise.
18781 (expand_omp_sections): Likewise.
18782 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18783 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18784 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18785 is initialized, not when loops are present.
18786 * tree-parloops.c (parallelize_loops): Remove checking here.
18787 * passes.c (init_optimization_passes): Schedule a copy-propagation
18788 pass before complete unrolling of inner loops.
18790 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18792 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18793 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18794 (colorize_init): Add argument to _WIN32 version.
18795 * toplev.c: Include diagnostic-color.h.
18796 (process_options): Default to -fdiagnostics-color=auto if
18797 GCC_COLORS env var is in the environment.
18798 * common.opt (fdiagnostics-color=): Add Var and Init.
18799 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18800 env var is in the environment, the default is auto rather than never.
18802 * diagnostic.h (file_name_as_prefix): Add context argument.
18803 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18804 the string as locus.
18805 * langhooks.c (lhd_print_error_function): Adjust caller.
18807 2013-04-25 Lawrence Crowl <crowl@google.com>
18809 * var-tracking.c (shared_hash_def::htab):
18810 Change type to hash_table. Update dependent calls and types.
18812 2013-04-25 Lawrence Crowl <crowl@google.com>
18814 * Makefile.in: Update as needed below.
18816 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18817 Move declaration to after the type's method definitons.
18819 * attribs.c (htab_t scoped_attributes::attribute_hash):
18820 Change type to hash_table. Update dependent calls and types.
18822 * bitmap.c (htab_t bitmap_desc_hash):
18823 Change type to hash_table. Update dependent calls and types.
18825 * cselib.c (htab_t cselib_hash_table):
18826 Change type to hash_table. Update dependent calls and types.
18828 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18829 (hash_string_slot_node): Move implementation into lto-streamer.h
18830 struct string_slot_hasher.
18831 (eq_string_slot_node): Likewise.
18833 * data-streamer-out.c: Update output_block::string_hash_table
18834 dependent calls and types.
18836 * dwarf2cfi.c (htab_t trace_index):
18837 Change type to hash_table. Update dependent calls and types.
18839 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18840 Change type to hash_table. Update dependent calls and types.
18841 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18842 (htab_t optimize_external_refs::map): Likewise.
18843 (htab_t output_comp_unit::extern_map): Likewise.
18844 (htab_t output_comdat_type_unit::extern_map): Likewise.
18845 (htab_t output_macinfo::macinfo_htab): Likewise.
18846 (htab_t optimize_location_lists::htab): Likewise.
18847 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18849 * except.c (htab_t ehspec_hash_type):
18850 Change type to hash_table. Update dependent calls and types.
18851 (assign_filter_values::ttypes): Likewise.
18852 (assign_filter_values::ehspec): Likewise.
18853 (sjlj_assign_call_site_values::ar_hash): Likewise.
18854 (convert_to_eh_region_ranges::ar_hash): Likewise.
18856 * gcse.c (htab_t pre_ldst_table):
18857 Change type to hash_table. Update dependent calls and types.
18859 * ggc-common.c (htab_t saving_htab):
18860 Change type to hash_table. Update dependent calls and types.
18861 (htab_t loc_hash): Likewise.
18862 (htab_t ptr_hash): Likewise.
18863 (call_count): Rename ggc_call_count.
18864 (call_alloc): Rename ggc_call_alloc.
18865 (loc_descriptor): Rename make_loc_descriptor.
18866 (add_statistics): Rename ggc_add_statistics.
18868 * ggc-common.c (saving_htab):
18869 Change type to hash_table. Update dependent calls and types.
18871 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18872 (push_gimplify_context): Likewise.
18873 (pop_gimplify_context): Likewise.
18874 (struct gimple_temp_hash_elt): Added.
18875 (struct gimplify_hasher): Likewise.
18876 (struct gimplify_ctx.temp_htab):
18877 Change type to hash_table. Update dependent calls and types.
18879 * gimple-fold.c: Include gimplify-ctx.h.
18881 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18882 Change type to hash_table. Update dependent calls and types.
18883 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18884 avoid potential global name collision.
18886 * gimplify.c: Include gimplify-ctx.h.
18887 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18888 (htab_t gimplify_ctx::temp_htab):
18889 Update dependent calls and types for new type hash_table.
18890 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18891 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18893 * gimplify-ctx.h: New.
18894 (struct gimple_temp_hash_elt): Move from gimplify.c.
18895 (class gimplify_hasher): New.
18896 (struct gimplify_ctx): Move from gimple.h.
18897 (htab_t gimplify_ctx::temp_htab):
18898 Change type to hash_table. Update dependent calls and types.
18900 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18901 (htab_t ivs_params::newivs_index):
18902 Change type to hash_table. Update dependent calls and types.
18903 (htab_t ivs_params::params_index): Likewise.
18904 (htab_t print_generated_program::params_index): Likewise.
18905 (htab_t gloog::newivs_index): Likewise.
18906 (htab_t gloog::params_index): Likewise.
18908 * graphite.c: Include graphite-htab.h.
18909 4htab_t graphite_transform_loops::bb_pbb_mapping):
18910 Change type to hash_table. Update dependent calls and types.
18912 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18913 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18914 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18916 * graphite-dependences.c: Include graphite-htab.h.
18917 (loop_is_parallel_p): Change hash table type of parameter.
18919 * graphite-htab.h: New.
18920 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18921 (extern find_pbb_via_hash): Move from graphite-poly.h.
18922 (extern loop_is_parallel_p): Move from graphite-poly.h.
18923 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18925 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18926 (extern loop_is_parallel_p): Move to graphite-htab.h.
18927 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18929 * haifa-sched.c (htab_t delay_htab):
18930 Change type to hash_table. Update dependent calls and types.
18931 (htab_t delay_htab_i2): Likewise.
18933 * ira-color.c (htab_t allocno_hard_regs_htab):
18934 Change type to hash_table. Update dependent calls and types.
18936 * ira-costs.c (htab_t cost_classes_htab):
18937 Change type to hash_table. Update dependent calls and types.
18939 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18940 Change type to hash_table. Update dependent calls and types.
18942 * loop-iv.c (htab_t bivs):
18943 Change type to hash_table. Update dependent calls and types.
18945 * loop-unroll.c (htab_t opt_info::insns_to_split):
18946 Change type to hash_table. Update dependent calls and types.
18947 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18949 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18950 (struct string_slot_hasher): New.
18951 (htab_t output_block::string_hash_table):
18952 Change type to hash_table. Update dependent calls and types.
18954 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18955 (htab_t file_name_hash_table):
18956 Change type to hash_table. Update dependent calls and types.
18958 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18961 * lto-streamer.c (htab_t tree_htab):
18962 Change type to hash_table. Update dependent calls and types.
18964 * omp-low.c: Include gimplify-ctx.h.
18966 * passes.c (htab_t name_to_pass_map):
18967 Change type to hash_table. Update dependent calls and types.
18968 (pass_traverse): Rename to passes_pass_traverse.
18970 * plugin.c (htab_t event_tab):
18971 Change type to hash_table. Update dependent calls and types.
18973 * postreload-gcse.c (htab_t expr_table):
18974 Change type to hash_table. Update dependent calls and types.
18975 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18977 * sese.c (debug_rename_map_1): Make extern.
18978 (htab_t copy_bb_and_scalar_dependences::rename_map):
18979 Change type to hash_table. Update dependent calls and types.
18981 * sese.h (extern debug_rename_map): Move to .c file.
18983 * store-motion.c (htab_t store_motion_mems_table):
18984 Change type to hash_table. Update dependent calls and types.
18986 * trans-mem.c (htab_t tm_new_mem_hash):
18987 Change type to hash_table. Update dependent calls and types.
18989 * tree-browser.c (htab_t TB_up_ht):
18990 Change type to hash_table. Update dependent calls and types.
18992 * tree-cfg.c (htab_t discriminator_per_locus):
18993 Change type to hash_table. Update dependent calls and types.
18995 * tree-complex.c: Include tree-hasher.h
18996 (htab_t complex_variable_components):
18997 Change type to hash_table. Update dependent calls and types.
18999 * tree-eh.c (htab_t finally_tree):
19000 Change type to hash_table. Update dependent calls and types.
19002 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19003 struct int_tree_hasher.
19004 (extern int_tree_map_eq): Likewise.
19005 (uid_decl_map_hash): Removed.
19006 (extern decl_tree_map_eq): Likewise.
19008 * tree-hasher.h: New.
19009 (struct int_tree_hasher): New.
19010 (typedef int_tree_htab_type): New.
19012 * tree-inline.c: Include gimplify-ctx.h.
19014 * tree-mudflap.c: Include gimplify-ctx.h.
19016 * tree-parloops.c: Include tree-hasher.h.
19017 (htab_t eliminate_local_variables_stmt::decl_address):
19018 Change type to hash_table. Update dependent calls and types.
19019 (htab_t separate_decls_in_region::decl_copies): Likewise.
19021 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19022 Change type to hash_table. Update dependent calls and types.
19024 * tree-sra.c (candidates):
19025 Change type to hash_table. Update dependent calls and types.
19027 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19029 (int_tree_map_hash): Likewise.
19031 * tree-ssa-dom.c (htab_t avail_exprs):
19032 Change type to hash_table. Update dependent calls and types.
19034 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19035 Change type to hash_table. Update dependent calls and types.
19037 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19038 Change type to hash_table. Update dependent calls and types.
19040 * tree-ssa-phiopt.c (seen_ssa_names):
19041 Change type to hash_table. Update dependent calls and types.
19043 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19044 Change type to hash_table. Update dependent calls and types.
19046 * tree-ssa-uncprop.c (equiv):
19047 Change type to hash_table. Update dependent calls and types.
19049 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19051 PR rtl-optimization/57003
19052 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19053 call note_stores with kill_clobbered_value callback again after
19054 killing regs_invalidated_by_call.
19056 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19058 * config/aarch64/aarch64-simd.md
19059 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19060 (aarch64_simd_bsl<mode>): Likewise.
19061 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19063 2013-04-25 Marek Polacek <polacek@redhat.com>
19065 PR tree-optimization/57066
19066 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19068 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19070 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19072 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19074 * config/aarch64/aarch64-builtins.c
19075 (aarch64_fold_builtin): New.
19076 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19077 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19078 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19079 * config/aarch64/arm_neon.h
19080 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19082 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19083 Tejas Belagod <tejas.belagod@arm.com>
19085 * config/aarch64/aarch64-builtins.c
19086 (aarch64_gimple_fold_builtin): New.
19087 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19088 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19089 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19090 (addvv4sf): Update.
19091 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19093 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19095 * config/aarch64/aarch64.md
19096 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19098 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19100 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19101 (*ngcsi_uxtw): New pattern.
19103 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19104 Julian Brown <julian@codesourcery.com>
19106 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19107 (TB_DREG): Add T_V4HF.
19108 (v4hf_UP): New macro.
19109 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19110 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19111 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19113 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19114 (arm_vector_mode_supported_p): Handle V4HF.
19115 (arm_mangle_map): Handle V4HFmode.
19116 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19117 * config/arm/arm_neon_builtins.def: Add entries for
19118 vcvtv4hfv4sf, vcvtv4sfv4hf.
19119 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19120 (neon_vcvtv4hfv4sf): Likewise.
19121 * config/arm/neon-gen.ml: Handle half-precision floating point
19123 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19124 * config/arm/arm_neon.h: Regenerate.
19125 * config/arm/neon.ml (type elts): Add F16.
19126 (type vectype): Add T_float16x4, T_floatHF.
19127 (type vecmode): Add V4HF.
19128 (type features): Add Requires_FP_bit feature.
19129 (elt_width): Handle F16.
19130 (elt_class): Likewise.
19131 (elt_of_class_width): Likewise.
19132 (mode_of_elt): Refactor.
19133 (type_for_elt): Handle F16, fix error messages.
19134 (vectype_size): Handle T_float16x4.
19135 (vcvt_sh): New function.
19136 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19137 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19138 (string_of_mode): Handle V4HF.
19139 * doc/arm-neon-intrinsics.texi: Regenerate.
19141 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19143 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19144 format specifier in 'X' case.
19146 2013-04-25 Alan Modra <amodra@gmail.com>
19149 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19150 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19151 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19152 Repeat for many other rotate/shift and mask patterns using subregs.
19153 Name lshiftrt insns.
19154 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19155 on WORDS_BIG_ENDIAN.
19157 2013-04-25 Alan Modra <amodra@gmail.com>
19159 * config.gcc: Support little-endian powerpc-linux targets.
19160 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19161 (LINK_OS_LINUX_SPEC): Define.
19162 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19163 Preserve MASK_LITTLE_ENDIAN.
19164 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19165 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19166 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19167 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19168 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19169 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19170 Correct fp word order for little-endian. Don't shift toc entries
19171 smaller than a word for little-endian.
19172 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19173 (bswapdi2 splits): Correct low-part subreg for little-endian.
19174 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19175 low/high where such is correct only for be.
19176 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19177 little-endian for -mcall-aixdesc.
19179 2013-04-25 Alan Modra <amodra@gmail.com>
19181 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19182 replace_equiv_address_nv.
19184 2013-04-25 Alan Modra <amodra@gmail.com>
19186 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19188 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19191 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19192 * rtl.h (struct rtx_def): ...
19194 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19196 PR rtl-optimizations/57046
19197 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19198 for multi-reg splits.
19200 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19202 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19204 2013-04-24 Sterling Augustine <saugustine@google.com>
19206 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19207 (comp_dir_string, debug_str_dwo_section): New.
19208 (DEBUG_STR_DWO_SECTION): Rename to ...
19209 (DEBUG_DWO_STR_SECTION): ... this.
19210 (DEBUG_NORM_STR_SECTION): Delete.
19211 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19212 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19213 (find_AT_string): Move most logic to ...
19214 (find_AT_string_in_table): ... here. New.
19215 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19216 add_skeleton_AT_string. Delete logic.
19217 (output_skeleton_debug_sections): Remove call to
19218 add_top_level_skeleton_die_attrs.
19219 (add_comp_dir_attribute): Move logic to comp_dir_string.
19220 (dwarf2out_init): Initialize debug_str_dwo_section.
19221 (output_indirect_string): Call find_string_form.
19222 (output_indirect_strings): Rewrite.
19223 (prune_unused_types): Empty skeleton_debug_str_hash.
19224 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19225 (dwarf2out_finish): Call output_indirect_strings.
19227 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19229 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19231 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19233 * rtl.h (struct rtx_def): Add comment for field jump.
19234 (LRA_SUBREG_P): New macro.
19235 * recog.c (register_operand): Check LRA_SUBREG_P.
19236 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19238 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19239 (lra_final_code_change): Skip subreg change for operators.
19240 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19241 if there are no operand changes.
19242 * lra-constraints.c (curr_insn_set): New.
19243 (match_reload): Set LRA_SUBREG_P.
19244 (emit_spill_move): Ditto.
19245 (check_and_process_move): Use curr_insn_set. Process only single
19246 set insns. Don't initialize sec_mem_p and change_p.
19247 (simplify_operand_subreg): Use LRA_SUBREG_P.
19248 (reg_in_class_p): New function.
19249 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19250 of #ifdef. Add code to remove cycling.
19251 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19252 non-null disp. Reload inner instead of disp when base and index
19253 are null. Try to put lo_sum into register.
19254 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19255 (check_and_process_move): Move code for move cost check to
19256 simple_move_p. Remove equiv_substitution.
19257 (simple_move_p): New function.
19258 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19259 curr_insn_set. Call check_and_process_move only for single set
19260 insns. Use the new function. Move call of check_and_process_move
19261 after operand equiv substitution and address process.
19263 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19265 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19266 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19267 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19269 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19271 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19273 2013-04-24 Marek Polacek <polacek@redhat.com>
19275 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19276 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19277 (select_loops_exit_conditions): Likewise.
19278 (number_of_iterations_for_all_loops): Likewise.
19279 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19280 (scev_analysis): Likewise.
19282 2013-04-02 Catherine Moore <clm@codesourcery.com>
19283 Chao-ying Fu <fu@mips.com>
19285 * config/mips/micromips.md (jraddiusp): New pattern.
19286 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19287 instruction if possible.
19289 2013-04-24 Alan Modra <amodra@gmail.com>
19291 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19293 2013-04-24 Julian Brown <julian@codesourcery.com>
19294 Chung-Lin Tang <cltang@codesourcery.com>
19296 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19297 dependency behavior in enumeration type DIE generation. Add TODO note
19298 to comments about future DW_FORM_sdata/udata re-work of related code.
19300 2013-04-23 Lawrence Crowl <crowl@google.com>
19302 * Makefile.in: Update as needed below.
19304 * hash-table.h (class hash_table):
19305 Correct many methods with parameter types compare_type to the correct
19306 value_type. (Correct code was unlikely to notice the change.)
19307 (hash_table::elements_with_deleted) New.
19308 (class hashtable::iterator): New.
19309 (hashtable::begin()): New.
19310 (hashtable::end()): New.
19311 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19313 * statistics.c (statistics_hashes):
19314 Change type to hash_table. Update dependent calls and types.
19316 * tree-into-ssa.c (var_infos):
19317 Change type to hash_table. Update dependent calls and types.
19319 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19320 Change type to hash_table. Update dependent calls and types.
19322 * tree-ssa-loop-im.c (struct mem_ref.refs):
19323 Change type to hash_table. Update dependent calls and types.
19325 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19326 Change type to hash_table. Update dependent calls and types.
19328 * tree-ssa-sccvn.c (vn_tables_s::nary):
19329 Change type to hash_table. Update dependent calls and types.
19330 (vn_tables_s::phis): Likewise.
19331 (vn_tables_s::references): Likewise.
19333 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19334 (vn_reference_eq): Update parameter and return types.
19336 * tree-ssa-structalias.c (pointer_equiv_class_table):
19337 Change type to hash_table. Update dependent calls and types.
19338 (location_equiv_class_table): Likewise.
19340 * tree-vect-data-refs.c: Consequential changes for making
19341 peeling a hash_table.
19343 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19344 (destroy_loop_vec_info): Dependent hash_table update.
19346 * tree-vectorizer.h (peeling_htab):
19347 Change type to hash_table. Update dependent calls and types.
19349 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19351 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19352 to check the register content is equal or not.
19353 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19354 to assign register content record.
19355 * lra-eliminations.c (update_reg_eliminate): Use
19356 lra_update_reg_val_offset to update register content offset.
19357 * lra-int.h (struct lra_reg): Add offset member.
19358 (lra_reg_val_equal_p): New static inline function.
19359 (lra_update_reg_val_offset): New static inline function.
19360 (lra_assign_reg_val): New static inline function.
19361 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19362 to assign register content record.
19363 (initialize_lra_reg_info_element): Initial offset to zero.
19365 2013-04-23 Catherine Moore <clm@codesourcery.com>
19367 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19368 operands. Record compression.
19370 2013-04-23 Xinliang David Li <davidxl@google.com>
19372 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19374 2013-04-23 Richard Biener <rguenther@suse.de>
19376 PR middle-end/57036
19377 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19378 parameter, only add abnormal goto edges from the copied body
19379 if the call could perform abnormal gotos.
19380 (copy_cfg_body): Adjust.
19382 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19384 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19386 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19388 * coretypes.h (gimple_stmt_iterator): Add struct to make
19391 2013-04-23 Richard Biener <rguenther@suse.de>
19393 PR tree-optimization/57026
19394 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19395 from SSA names occuring in abnormal PHI nodes.
19397 2013-04-22 Andi Kleen <ak@linux.intel.com>
19399 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19401 2013-04-22 Andi Kleen <ak@linux.intel.com>
19403 * lto/lto.c (print_lto_report_1): Declare early.
19404 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19406 2013-04-22 Andi Kleen <ak@linux.intel.com>
19408 * common.opt (-flto-report-wpa): Add.
19409 * doc/invoke.texi (-flto-report-wpa): Add.
19410 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19413 2013-04-22 Xinliang David Li <davidxl@google.com>
19415 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19416 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19417 * Makefile.in: New dependency
19419 David Daney <ddaney.cavm@gmail.com>
19421 * configure.ac (gcc_cv_as_micromips_support): Use the
19422 --fatal-warnings option.
19423 * configure: Regenerate.
19425 2013-04-22 Marek Polacek <polacek@redhat.com>
19428 * tsan.c (instrument_expr): Don't instrument expression
19429 in case its size is zero.
19431 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19435 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19437 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19439 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19441 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19442 (gimple_stmt_iterator): New typedef.
19443 * gimple.h (gimple_stmt_iterator): Rename to...
19444 (gimple_stmt_iterator_d): ... This.
19445 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19446 trees be valid for GIMPLE and GENERIC.
19447 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19448 * gimple-fold.c (gimple_fold_call): Call target hook
19449 gimple_fold_builtin.
19450 * hooks.c (hook_bool_gsiptr_false): New.
19451 * hooks.h (hook_bool_gsiptr_false): New.
19452 * target.def (fold_stmt): New.
19453 * doc/tm.texi: Regenerate.
19455 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19458 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19459 a set sp if no stack realignment.
19461 2013-04-22 Nick Clifton <nickc@redhat.com>
19463 * config.gcc (tilegx-linux): Extend extra_objs rather than
19465 (tilepro-linux): Likewise.
19467 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19469 * config/aarch64/aarch64-builtins.c
19471 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19472 (VAR<1-12>): Add MAP parameter.
19473 (BUILTIN_*): Likewise.
19474 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19475 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19476 (aarch64_ushl_n<mode>): Likewise.
19477 (aarch64_sshr_n<mode>): Likewise.
19478 (aarch64_ushr_n<mode>): Likewise.
19479 (aarch64_<maxmin><mode>): Likewise.
19480 (aarch64_sqrt<mode>): Likewise.
19481 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19482 (vshr<q>_n_*): Likewise.
19484 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19486 * config/aarch64/aarch64-builtins.c
19487 (aarch64_simd_builtin_type_mode): Handle SF types.
19489 (BUILTIN_GPF): Define.
19490 (aarch64_init_simd_builtins): Handle SF types.
19491 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19492 (frecps): Likewise.
19493 (frecpx): Likewise.
19494 * config/aarch64/aarch64-simd.md
19495 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19496 (aarch64_frecpe<mode>): New.
19497 (aarch64_frecps<mode>): Likewise.
19498 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19499 (v8type): Add frecp<esx>.
19500 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19501 (aarch64_frecps<mode>): Likewise.
19502 * config/aarch64/iterators.md (FRECP): New.
19503 (frecp_suffix): Likewise.
19504 * config/aarch64/arm_neon.h
19505 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19507 2013-04-22 Christian Bruel <christian.bruel@st.com>
19510 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19511 (REG_CLASS_NAMES): Idem.
19512 (REG_CLASS_CONTENTS): Idem.
19513 (REGCLASS_HAS_FP_REG): Idem.
19514 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19515 (sh_conditional_register_usage): Idem.
19517 2013-04-21 Jeff Law <law@redhat.com>
19519 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19520 (ssa_forward_propagate_and_combine): Use it.
19522 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19524 * lra.c: Update the flow chart diagram.
19526 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19528 PR rtl-optimization/56847
19529 * lra-constraints.c (process_alt_operands): Discourage alternative
19530 with non-matche doffsettable memory constraint fro memory with
19533 2013-04-19 Richard Biener <rguenther@suse.de>
19535 PR tree-optimization/56982
19536 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19538 * gimplify.c (gimplify_call_expr): Notice special calls.
19539 (gimplify_modify_expr): Likewise.
19540 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19541 abnormal control flow receivers.
19542 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19543 in the same way as cfun->has_nonlocal_labels.
19544 (gimple_purge_dead_abnormal_call_edges): Likewise.
19545 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19546 receivers start a basic-block.
19548 2013-04-19 Richard Biener <rguenther@suse.de>
19550 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19552 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19553 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19554 (SLP_TREE_LOAD_PERMUTATION): Add.
19555 (vect_transform_slp_perm_load): Adjust prototype.
19556 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19557 (vect_free_slp_instance): Likewise.
19558 (vect_create_new_slp_node): Likewise.
19559 (vect_supported_slp_permutation_p): Remove.
19560 (vect_slp_rearrange_stmts): Adjust.
19561 (vect_supported_load_permutation_p): Likewise. Inline
19562 vect_supported_slp_permutation_p here.
19563 (vect_analyze_slp_instance): Compute load permutations per
19564 slp node instead of per instance.
19565 (vect_get_slp_defs): Adjust.
19566 (vect_transform_slp_perm_load): Likewise.
19567 (vect_schedule_slp_instance): Remove redundant code.
19568 (vect_schedule_slp): Remove hack for PR56270, add it ...
19569 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19570 CSE loads for SLP. Adjust.
19572 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19574 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19575 spelling in two comments.
19577 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19580 * config/arm/arm.c (load_multiple_sequence): Require SP
19581 as base register for loads if SP is in the register list.
19583 2013-04-19 Martin Jambor <mjambor@suse.cz>
19585 PR tree-optimization/56718
19586 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19587 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19588 and made public. Adjusted all callers.
19589 (ipa_intraprocedural_devirtualization): New function.
19590 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19591 (ipa_intraprocedural_devirtualization): Likewise.
19592 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19594 2013-04-19 Richard Biener <rguenther@suse.de>
19596 PR tree-optimization/57000
19597 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19599 2013-04-19 Terry Guo <terry.guo@arm.com>
19601 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19603 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19604 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19605 (cortex_m4_fmacs): Use new reservations.
19606 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19608 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19610 PR rtl-optimization/56999
19611 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19613 (lra_coalesce): Remove split_origin_bitmap and related code.
19614 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19615 ranges if necessary.
19617 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19619 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19621 (ix86_expand_call): Remove clobbered_registers array and use
19622 x86_64_ms_sysv_extra_clobbered_registers instead.
19623 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19625 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19627 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19628 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19629 (*call_value_rex64_ms_sysv): Ditto.
19631 2013-04-18 Cary Coutant <ccoutant@google.com>
19633 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19634 parent instead of die_mark.
19636 2013-04-18 Diego Novillo <dnovillo@google.com>
19638 * gimple.c (create_gimple_tmp): New.
19639 (get_expr_type): New.
19640 (build_assign): New.
19641 (build_type_cast): New.
19642 * gimple.h (enum ssa_mode): Define.
19643 (gimple_seq_set_location): New.
19644 * asan.c (build_check_stmt): Change some gimple_build_* calls
19645 to use build_assign and build_type_cast.
19647 2013-04-18 Richard Biener <rguenther@suse.de>
19649 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19650 handle negative step. Remove redundant checks.
19651 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19652 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19653 for negative step and grouped loads fail to vectorize.
19655 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19657 * emit-rtl.c (reset_insn_used_flags): New function.
19658 (reset_all_used_flags): Use it.
19659 (verify_insn_sharing): New function.
19660 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19662 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19664 PR tree-optimization/56984
19665 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19666 and (x >> M) >= N don't register any assertion if N << M is the
19669 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19671 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19672 do not delete_insn insns that have not yet been emitted, only
19673 unlink them with remove_insn.
19674 * df-scan.c (df_insn_delete): Revert r197492.
19676 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19678 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19679 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19681 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19683 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19684 define_insn_and_split.
19685 (and_scc,ior_scc,negscc): Likewise.
19686 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19688 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19690 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19691 can benefit from using a sequence of LDRD instructions in epilogue
19692 instead of a single LDM instruction.
19694 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19697 * doc/extend.texi: Fix typo.
19699 2013-04-17 Richard Biener <rguenther@suse.de>
19701 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19702 (vect_build_slp_tree): ... here.
19703 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19704 match. Remove special-casing of mismatched complex loads.
19705 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19706 re-try the match with swapped commutative operands.
19707 (vect_supported_load_permutation_p): Remove special-casing of
19708 mismatched complex loads.
19709 (vect_analyze_slp_instance): Adjust.
19711 2013-04-17 Richard Biener <rguenther@suse.de>
19713 PR rtl-optimization/56921
19714 * cfgloop.h (struct loop): Add simple_loop_desc member.
19715 (struct niter_desc): Mark with GTY(()).
19716 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19717 * loop-iv.c (get_simple_loop_desc): Likewise.
19718 (free_simple_loop_desc): Likewise.
19721 2013-04-16 Richard Biener <rguenther@suse.de>
19723 PR rtl-optimization/56921
19724 * loop-init.c (pass_rtl_move_loop_invariants): Add
19725 TODO_do_not_ggc_collect to todo_flags_finish.
19726 (pass_rtl_unswitch): Same.
19727 (pass_rtl_unroll_and_peel_loops): Same.
19728 (pass_rtl_doloop): Same.
19730 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19732 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19733 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19734 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19736 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19737 * tree-streamer.c (record_common_node): Adjust reference in comment.
19739 2013-04-17 Terry Guo <terry.guo@arm.com>
19741 * config/arm/cortex-m4.md: Add a new bypass.
19743 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19745 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19747 (*subs_<optab><mode>_multp2): New pattern.
19748 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19749 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19751 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19753 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19754 (*subs_mul_imm_<mode>): New pattern.
19756 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19759 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19760 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19761 (vsx_movti_32bit): Change j->wa to O->wa.
19763 2013-04-16 Richard Biener <rguenther@suse.de>
19765 PR rtl-optimization/56921
19766 * loop-init.c (pass_rtl_move_loop_invariants): Add
19767 TODO_do_not_ggc_collect to todo_flags_finish.
19768 (pass_rtl_unswitch): Same.
19769 (pass_rtl_unroll_and_peel_loops): Same.
19770 (pass_rtl_doloop): Same.
19772 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19774 * config/arm/arm.c (emit_multi_reg_push): New declaration
19775 for an existing function.
19776 (arm_emit_strd_push): New function.
19777 (arm_expand_prologue): Used here.
19778 (arm_emit_ldrd_pop): New function.
19779 (arm_expand_epilogue): Used here.
19780 (arm_get_frame_offsets): Update condition.
19781 (arm_emit_multi_reg_pop): Add a special case for load of a single
19782 register with writeback.
19784 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19786 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19789 2013-04-16 Richard Biener <rguenther@suse.de>
19791 PR tree-optimization/56756
19792 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19793 (first_mem_ref_loc): New.
19794 (execute_sm): Place the load temporarily before a previous
19795 access instead of in the latch edge to ensure its SSA dependencies
19796 are defined at points dominating the load.
19798 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19800 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19801 correct fix by moving header and footer insn to the footer of
19802 the merged basic block. Clear BB_END of the merged-away block.
19804 PR middle-end/43631
19805 * emit-rtl.c (make_note_raw): New function.
19806 (link_insn_into_chain): New static inline function.
19807 (add_insn): Use it.
19808 (add_insn_before, add_insn_after): Factor insn chain linking code...
19809 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19810 using link_insn_into_chain.
19811 (note_outside_basic_block_p): New helper function for emit_note_after
19812 and emit_note_before.
19813 (emit_note_after): Use nobb variant of add_insn_after if the note
19814 should not be contained in a basic block.
19815 (emit_note_before): Use nobb variant of add_insn_before if the note
19816 should not be contained in a basic block.
19817 (emit_note_copy): Use make_note_raw.
19818 (emit_note): Likewise.
19819 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19820 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19821 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19822 the moved barrier the tail of the basic block it follows.
19823 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19825 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19827 PR tree-optimization/56962
19828 * gimple-ssa-strength-reduction.c (record_increment): Only set
19829 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19830 either rhs1 or rhs2 is equal to c->base_expr.
19832 2013-04-15 Richard Biener <rguenther@suse.de>
19834 PR tree-optimization/56933
19835 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19837 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19838 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19839 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19840 dependence check ...
19841 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19843 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19844 GROUP_READ_WRITE_DEPENDENCE.
19846 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19848 * emit-rtl.c (reset_all_used_flags): New function.
19849 (verify_rtl_sharing): Call reset_all_used_flags before and after
19850 performing the checks.
19852 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19854 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19855 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19856 * config/arm/constraints.md (De): New constraint.
19857 * config/arm/neon.md (anddi3_neon): Delete.
19858 (neon_vand<mode>): Expand to standard anddi3 pattern.
19859 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19860 Move earlier in the file.
19861 (neon_inv_logic_op2): Likewise.
19862 (arm_anddi_operand_neon): New predicate.
19864 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19866 * configure.ac (gcc_cv_ld_as_needed): Set
19867 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19868 Use -z ignore, -z record on *-*-solaris2*.
19869 (HAVE_LD_AS_NEEDED): Update comment.
19870 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19871 * configure: Regenerate.
19872 * config.in: Regenerate.
19873 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19874 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19875 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19876 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19877 equivalents. Fix markup.
19878 * doc/tm.texi: Regenerate.
19880 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19882 * config/i386/i386.opt: New option mstack-protector-guard=.
19883 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19884 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19885 TARGET_SSP_TLS_GUARD.
19886 * config/i386/i386.c (ix86_option_override_internal): Set
19887 ix86_stack_protector_guard.
19888 * config/i386/i386.md (stack_protect_set): Enable for
19889 TARGET_SSP_TLS_GUARD only.
19890 (stack_protect_set_<mode>): Ditto.
19891 (stack_protect_test): Ditto.
19892 (stack_protect_test_<mode>): Ditto.
19893 * doc/invoke.texi (i386 Option): Document.
19895 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19898 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19899 (S_MODES): Set H_MODE bit.
19900 (SF_MODES): Set only S_MODE and SF_MODE bits.
19901 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19902 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19903 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19904 <MODE_FLOAT>: Likewise.
19906 2013-04-15 Joey Ye <joey.ye@arm.com>
19908 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19910 2013-04-15 Joey Ye <joey.ye@arm.com>
19912 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19914 (thumb_far_jump_used_p): Count instruction size and set
19917 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19919 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19920 * resource.c (mark_target_live_regs): Reformat conditional block.
19922 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19924 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19925 notes, they are emitted only just before final.
19926 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19928 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19930 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19931 * cfgrtl.c (delete_insn): Call it here instead.
19932 * lra-spills.c (lra_final_code_change): Use delete_insn.
19933 * haifa-sched.c (sched_remove_insn): Likewise.
19934 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19935 returning to the nop pool.
19936 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19937 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19939 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19941 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19942 * doc/tm.texi: Regenerated.
19944 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19946 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19949 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19951 * df-core.c (df_find_def): Compare register numbers.
19952 (df_find_use): Likewise.
19954 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19957 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19958 lra_in_progress for return.
19960 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19962 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19963 define_insn into define_insn_and_split and emit movsicc patterns.
19965 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19967 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19969 2013-04-12 Richard Biener <rguenther@suse.de>
19971 * tree-pass.h (TODO_do_not_ggc_collect): New.
19972 * passes.c (execute_one_ipa_transform_pass): Honor
19973 TODO_do_not_ggc_collect.
19974 (execute_one_pass): Likewise.
19977 2013-04-10 Richard Biener <rguenther@suse.de>
19979 * passes.c (init_optimization_passes): Remove reload pass.
19980 * ira.c (do_reload): Merge into ...
19982 (rest_of_handle_reload): Remove.
19983 (pass_reload): Likewise.
19984 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19985 of reload for vzeroupper pass placement.
19987 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19989 PR tree-optimization/56918
19990 PR tree-optimization/56920
19991 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19992 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19993 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19994 use wide_mul_with_sign method.
19996 2013-04-12 Richard Biener <rguenther@suse.de>
19998 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19999 not be considered a gimple constant.
20001 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20003 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20004 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20007 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20008 Jakub Jelinek <jakub@redhat.com>
20010 * opts.c: Include diagnostic-color.h.
20011 (common_handle_option): Handle OPT_fdiagnostics_color_.
20012 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20013 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20014 (diagnostic-color.o): New.
20015 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20016 (diagnostic_color_rule): New enum.
20017 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20018 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20019 the location string.
20020 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20021 either NULL, or color kind.
20022 * diagnostic-color.c: New file.
20023 * diagnostic-color.h: New file.
20024 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20026 * doc/invoke.texi (-fdiagnostics-color): Document.
20027 * pretty-print.h (pp_show_color): Define.
20028 (struct pretty_print_info): Add show_color field.
20029 * diagnostic.c: Include diagnostic-color.h.
20030 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20031 macros. Colorize error:, warning: etc. strings and also the location
20033 (diagnostic_show_locus): Colorize the caret line.
20034 * pretty-print.c: Include diagnostic-color.h.
20035 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20036 inside of %< %> quotes or quoted through q format modifier.
20038 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20040 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20042 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20044 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20045 code in CC_NZ mode.
20046 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20049 2013-04-11 Marek Polacek <polacek@redhat.com>
20051 PR tree-optimization/48184
20052 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20054 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20056 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20057 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20058 (skip_simple_arithmetic): Tidy up.
20059 * tree.h (skip_simple_constant_arithmetic): Declare.
20061 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20063 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20065 2013-04-11 Richard Biener <rguenther@suse.de>
20067 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20068 generate vector constants.
20070 2013-04-11 Richard Biener <rguenther@suse.de>
20072 PR tree-optimization/56878
20073 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20074 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20076 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20077 Prefer to align the DR with the most invariant base address.
20079 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20081 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20084 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20086 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20087 floating-point vector comparisons against 0.
20089 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20091 PR tree-optimization/56899
20092 * fold-const.c (extract_muldiv_1): Apply distributive law
20093 only if TYPE_OVERFLOW_WRAPS (ctype).
20095 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20098 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20099 loading has side effect.
20101 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20103 * config/sparc/sparc.c: Include tree-pass.h.
20104 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20105 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20106 head of file. Change return type. Split off gate function.
20107 (sparc_gate_work_around_errata): New function.
20108 (pass_work_around_errata): New pass definition.
20109 (insert_pass_work_around_errata) New pass insert definition to
20110 insert pass_work_around_errata just after delayed-branch scheduling.
20111 (sparc_option_override): Insert the pass.
20112 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20114 2013-04-10 David S. Miller <davem@davemloft.net>
20116 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20117 or -mcpu=hypersparc.
20119 * target.def (cstore_mode): New hook.
20120 * target.h: Include insn-codes.h
20121 * targhooks.c: Likewise.
20122 (default_cstore_mode): New function.
20123 * targhooks.h: Declare it.
20124 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20125 * doc/tm.texi: Rebuild.
20126 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20127 target hook, rather than inspecting the insn_data.
20128 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20129 (TARGET_CSTORE_MODE): Redefine.
20130 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20132 * config/sparc/predicates.md (cstore_result_operand): New special
20134 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20135 Use it for operand 0.
20136 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20137 (*snesi_special): Likewise.
20138 (*snesi_zero): Likewise.
20139 (*seqsi_zero): Likewise.
20140 (*sltu_insn): Likewise.
20141 (*sgeu_insn): Likewise.
20142 (*seqdi_special): Make operand 0 and comparison operation be of
20144 (*snedi_special): Likewise.
20145 (*snedi_special_vis3): Likewise.
20146 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20147 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20148 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20149 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20150 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20151 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20152 (*sltu_extend_sp64): Likewise.
20153 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20154 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20155 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20156 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20157 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20159 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20161 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20162 (aarch64_start_file): Use the new function.
20164 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20165 Jason Merrill <jason@redhat.com>
20167 * common.opt: Add -gdwarf.
20168 * opts.c (common_handle_option): Handle it.
20169 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20171 2013-04-10 Richard Biener <rguenther@suse.de>
20173 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20174 (execute_one_ipa_transform_pass): But unconditionally here.
20175 (execute_one_pass): And here.
20176 (init_optimization_passes): Remove reload pass.
20177 * tree-pass.h (TODO_ggc_collect): Remove.
20178 (pass_reload): Likewise.
20179 * ira.c (do_reload): Merge into ...
20181 (rest_of_handle_reload): Remove.
20182 (pass_reload): Likewise.
20183 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20184 of reload for vzeroupper pass placement.
20185 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20186 and todo_flags_finish of all passes.
20188 2013-04-10 Richard Biener <rguenther@suse.de>
20190 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20191 first_const_oprnd field, rename first_def_type to first_op_type.
20192 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20193 (vect_get_and_check_slp_defs): Always use the type of the
20194 operand. Allow mixed vect_external_def, vect_constant_def types.
20195 (vect_get_constant_vectors): Handle mixed vect_external_def,
20196 vect_constant_def types.
20198 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20200 PR tree-optimization/55524
20201 * tree-ssa-math-opts.c
20202 (convert_mult_to_fma): Don't use an fms construct
20203 when we don't have an fms operation, but fnma, and it looks
20204 likely that we'll be able to use the latter.
20206 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20208 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20210 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20211 inline fail caused by overwritable functions.
20213 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20215 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20216 unnecessary bits in the constant power of two case.
20218 2013-04-10 Richard Biener <rguenther@suse.de>
20220 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20221 broken code swapping operands.
20222 (vect_build_slp_tree): Do not compute load permutations here.
20223 (vect_analyze_slp_instance): Compute load permutations here,
20224 after building the SLP tree.
20226 2013-04-09 Christian Bruel <christian.bruel@st.com>
20228 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20229 of next/prev_real_insn.
20231 2013-04-09 Jan Hubicka <jh@suse.cz>
20233 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20234 Drop aliased parameter.
20235 (function_and_variable_visibility): Do not handle alias pairs.
20236 * cgraph.c (varpool_externally_visible_p): Update prototype.
20237 * varpool.c (varpool_add_new_variable): Update.
20239 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20241 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20243 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20245 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20247 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20249 2013-04-09 Marek Polacek <polacek@redhat.com>
20251 PR tree-optimization/48762
20252 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20254 2013-04-09 Richard Biener <rguenther@suse.de>
20256 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20258 (vect_build_slp_tree): Likewise.
20259 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20260 calculating the cost of a SLP instance.
20261 (vect_analyze_slp_instance): Use it from here, after building
20264 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20266 PR middle-end/56883
20267 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20268 expand_omp_for_static_chunk): Use simple_p = true in
20269 force_gimple_operand_gsi calls when assigning to addressable decls.
20271 2013-04-09 Jeff Law <law@redhat.com>
20273 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20274 when the boolean was created by converting a wider object which
20275 had a boolean range.
20277 2013-04-09 Richard Biener <rguenther@suse.de>
20279 * tree-vectorizer.h (slp_void_p): Remove.
20280 (slp_tree): Typedef before _slp_tree declaration.
20281 (struct _slp_tree): Use a vector of slp_tree as children.
20282 (vect_get_place_in_interleaving_chain): Remove.
20283 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20285 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20287 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20288 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20289 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20290 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20291 Use slp_node instead of slp_void_p and adjust.
20293 2013-04-09 Richard Biener <rguenther@suse.de>
20295 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20296 work that is not necessary.
20298 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20300 PR tree-optimization/56854
20301 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20302 forward into clobber stmts if it would change MEM_REF lhs into
20305 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20307 * tree.c (type_hash_lookup, type_hash_add): Make static.
20308 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20310 2013-04-09 Richard Biener <rguenther@suse.de>
20312 * tree.h (unsave_expr_now): Remove.
20313 * tree-inline.c (mark_local_for_remap_r): Remove.
20314 (unsave_expr_1): Likewise.
20315 (unsave_r): Likewise.
20316 (unsave_expr_now): Likewise.
20317 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20318 (propagate_tree_value): Likewise.
20320 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20322 * doc/rtl.texi (sequence): Rewrite documentation to match the
20323 current use of SEQUENCE rtl objects.
20324 * rtl.def (SEQUENCE): Likewise.
20326 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20327 Update documentation.
20328 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20329 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20331 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20333 2013-04-08 Teresa Johnson <tejohnson@google.com>
20335 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20336 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20338 (estimate_edge_size_and_time): Add comment to suggest using rounding
20340 (estimate_node_size_and_time): Ditto.
20341 (remap_edge_change_prob): Use helper rounding divide methods.
20342 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20343 (gimple_mod_pow2_value_transform): Ditto.
20344 (gimple_mod_subtract_transform): Ditto.
20345 (gimple_ic_transform): Ditto.
20346 (gimple_stringops_transform): Ditto.
20347 * stmt.c (conditional_probability): Ditto.
20348 (emit_case_dispatch_table): Ditto.
20349 * lto-cgraph.c (merge_profile_summaries): Ditto.
20350 * tree-optimize.c (execute_fixup_cfg): Ditto.
20351 * cfgcleanup.c (try_forward_edges): Ditto.
20352 * cfgloopmanip.c (scale_loop_profile): Ditto.
20354 (duplicate_loop_to_header_edge): Ditto.
20355 (lv_adjust_loop_entry_edge): Ditto.
20356 * tree-vect-loop.c (vect_transform_loop): Ditto.
20357 * profile.c (compute_branch_probabilities): Ditto.
20358 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20359 * lto-streamer-in.c (input_cfg): Ditto.
20360 * gimple-streamer-in.c (input_bb): Ditto.
20361 * ipa-cp.c (update_profiling_info): Ditto.
20362 (update_specialized_profile): Ditto.
20363 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20364 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20366 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20367 (compute_trg_info): Ditto.
20368 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20369 (purge_dead_edges): Ditto.
20370 * loop-unswitch.c (unswitch_loop): Ditto.
20371 * cgraphclones.c (cgraph_clone_edge): Ditto.
20372 (cgraph_clone_node): Ditto.
20373 * tree-inline.c (copy_bb): Ditto.
20374 (copy_edges_for_bb): Ditto.
20375 (initialize_cfun): Ditto.
20376 (copy_cfg_body): Ditto.
20377 (expand_call_inline): Ditto.
20379 2013-04-08 Kai Tietz <ktietz@redhat.com>
20381 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20382 TARGET_CYGWIN64 by TARGET_64BIT.
20384 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20386 * config/epiphany/epiphany.md (GPR_1): New constant.
20387 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20388 * config/epiphany/epiphany.c (gen_compare_reg):
20389 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20390 is already in place.
20391 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20392 Don't require being called during rtl expansion; If y operlaps r0,
20394 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20395 (epiphany_expand_epilogue): Likewise.
20397 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20398 Don't use CC_FPmode for ORDERED / UNORDERED.
20399 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20401 * config/epiphany/constraints.md (CnL): New constraint.
20402 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20403 * config/epiphany/predicates.md (add_operand): Allow 1024.
20405 * config/epiphany/epiphany.md (logical_op): New code iterator.
20406 (op_mnc): New code attribute.
20407 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20408 (mov_f+1, mov_f+2): New peephole2 patterns.
20410 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20411 (cstoresi4): Also allow re-use of zero result when doing a NE
20412 comparison to a non-zero operand.
20413 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20415 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20416 Use gen_addsi3_i / gen_subsi3_i.
20418 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20422 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20423 contain anything but clobbers, at most one __builtin_stack_restore,
20424 optionally debug stmts and final resx, and if it has at least one
20425 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20426 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20427 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20428 which isn't defaut definition, remove them.
20429 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20430 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20431 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20432 with MEM_REF LHS with SSA_NAME address.
20434 2013-04-08 Jeff Law <law@redhat.com>
20436 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20438 2013-04-08 Richard Biener <rguenther@suse.de>
20440 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20442 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20443 determined vector type.
20444 (vect_analyze_data_refs): Likewise.
20445 (vect_get_new_vect_var): Adjust.
20446 (vect_create_destination_var): Preserve SSA name versions.
20447 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20448 not dump anything here.
20450 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20452 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20453 Add member lr_slot_known.
20454 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20456 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20457 Remove code that sets lr_slot_offset according to what a previous
20458 version of epiphany_emit_save_restore used to do.
20459 (epiphany_emit_save_restore): When doing an lr save or restore,
20460 set/verify lr_slot_known and lr_slot_offset.
20462 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20465 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20468 2013-04-08 Richard Biener <rguenther@suse.de>
20470 * alias.c (find_base_term): Fix thinko in previous change.
20472 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20474 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20475 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20476 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20477 if possible to compute val.
20478 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20479 For QImode integers don't require anything about precision. Use
20480 const_with_all_bytes_same to find out if the constant doesn't have
20481 repeated bytes in it.
20483 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20485 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20488 2013-04-08 Marek Polacek <polacek@redhat.com>
20490 PR rtl-optimization/48182
20491 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20494 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20497 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20498 nuses, make sure we have a label.
20500 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20503 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20504 (rs6000_emit_swdiv_low_precision): Remove.
20505 (rs6000_emit_swdiv): Rewrite to handle between one and four
20506 iterations of Newton-Raphson generally; modify required number of
20507 iterations for some cases.
20508 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20510 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20512 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20513 set-but-unused variable.
20515 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20516 basic blocks of released function bodies garbage-collectable.
20518 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20519 (struct rtl_opt_pass): Add TODO_df_finish.
20521 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20523 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20525 * config/arm/constraints.md (q): New constraint.
20526 * config/arm/ldrdstrd.md: New file.
20527 * config/arm/arm.md (ldrdstrd.md) New include.
20528 (arm_movdi): Use "q" instead of "r" constraint
20529 for double-word memory access.
20530 (movdf_soft_insn): Likewise.
20531 * config/arm/vfp.md (movdi_vfp): Likewise.
20532 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20533 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20534 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20535 (mem_ok_for_ldrd_strd): Likewise.
20536 (output_move_double): Update assertion.
20538 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20540 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20542 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20544 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20545 define_insn_and_split.
20546 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20548 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20550 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20551 define_insn_and_split.
20552 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20553 (shiftsi3_compare): New pattern.
20554 (rrx): New pattern.
20555 * config/arm/unspecs.md (UNSPEC_RRX): New.
20557 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20559 * config/arm/arm.md (negdi_extendsidi): New pattern.
20560 (negdi_zero_extendsidi): Likewise.
20562 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20564 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20565 define_insn_and_split.
20566 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20567 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20569 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20571 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20572 define_insn_and_split.
20573 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20574 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20576 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20578 * config/arm/arm.md (subsi3_carryin): New pattern.
20579 (subsi3_carryin_const): Likewise.
20580 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20581 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20583 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20585 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20587 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20589 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20590 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20592 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20594 * config/arm/arm.c (arm_expand_builtin): Change fcode
20595 type to unsigned int.
20597 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20599 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20601 2013-04-04 Ian Lance Taylor <iant@google.com>
20603 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20606 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20608 PR middle-end/56729
20609 * df-scan.c (df_insn_delete): Disable failing assert.
20611 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20613 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20614 New function prototype.
20615 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20616 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20617 (arm_builtin_vectorized_function): New function.
20619 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20621 * config/arm/arm_neon_builtins.def: New file.
20622 * config/arm/arm.c (neon_builtin_data): Move contents to
20623 arm_neon_builtins.def.
20624 (enum arm_builtins): Include neon builtin definitions.
20625 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20626 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20628 2013-04-04 Marek Polacek <polacek@redhat.com>
20630 PR tree-optimization/48186
20631 * predict.c (maybe_hot_frequency_p): Return false if
20632 HOT_BB_FREQUENCY_FRACTION is 0.
20633 (cgraph_maybe_hot_edge_p): Likewise.
20635 2013-04-04 Richard Biener <rguenther@suse.de>
20637 PR tree-optimization/56826
20638 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20641 2013-04-04 Richard Biener <rguenther@suse.de>
20643 PR tree-optimization/56213
20644 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20645 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20647 2013-04-04 Richard Biener <rguenther@suse.de>
20649 PR tree-optimization/56837
20650 * tree-loop-distribution.c (classify_partition): For non-zero
20651 values require that the value has the same precision as its
20652 mode to be useful as memset value.
20654 2013-04-03 Nick Clifton <nickc@redhat.com>
20656 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20657 (fmssf4): Use fmsf.s on E3V5 architectures.
20658 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20659 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20661 2013-04-03 Jeff Law <law@redhat.com>
20663 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20664 (lra-eliminations.o): Likewise.
20666 2013-04-03 Teresa Johnson <tejohnson@google.com>
20668 * gcov-io.c (compute_working_sets): Moved most of body of old
20669 compute_working_sets here from profile.c.
20670 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20671 (gcov_working_set_t): Moved typedef here from basic-block.h
20672 (compute_working_set): Declare.
20673 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20674 (get_working_sets): Renamed from compute_working_set,
20675 replace most of body with call to new compute_working_sets.
20676 (get_exec_counts): Replace call to compute_working_sets
20677 to get_working_sets.
20678 * profile.h (get_working_sets): Renamed from compute_working_set.
20679 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20680 to get_working_sets.
20681 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20682 * gcov-dump.c (dump_working_sets): New function.
20684 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20686 * hwint.c (sext_hwi, zext_hwi): New functions.
20687 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20688 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20689 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20690 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20691 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20692 (sext_hwi, zext_hwi): New functions.
20694 2013-04-03 Jeff Law <law@redhat.com>
20696 PR tree-optimization/56799
20697 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20698 back test for widening conversion erroneously dropped in prior change.
20700 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20703 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20704 instead of next_real_insn.
20706 2013-04-03 Marek Polacek <polacek@redhat.com>
20709 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20711 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20714 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20716 (thumb1_output_casesi): Likewise.
20717 (thumb2_output_casesi): Likewise.
20719 2013-04-03 Richard Biener <rguenther@suse.de>
20721 PR tree-optimization/56817
20722 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20724 (tree_unroll_loops_completely_1): ... new function to manually
20725 walk the loop tree, properly defering outer loops of unrolled
20726 loops to later iterations.
20728 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20730 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20731 (vectorizable_load): Likewise.
20732 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20733 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20735 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20737 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20740 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20742 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20744 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20746 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20748 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20750 PR tree-optimization/56790
20751 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20754 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20756 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20758 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20759 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20762 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20765 * tree.h (force_folding_builtin_constant_p): New decl.
20766 * builtins.c (force_folding_builtin_constant_p): New variable.
20767 (fold_builtin_constant_p): Fold immediately also if
20768 force_folding_builtin_constant_p.
20770 2013-04-03 Richard Biener <rguenther@suse.de>
20772 PR tree-optimization/56812
20773 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20774 DRs of the same interleaving chain are independent.
20776 2013-04-02 Jason Merrill <jason@redhat.com>
20778 * gdbinit.in (pbb): Use debug fn.
20780 2013-04-02 Lawrence Crowl <crowl@google.com>
20782 * sese.h (struct ivtype_map_elt_s): Remove unused.
20783 (extern debug_ivtype_map): Remove unused.
20784 (extern eq_ivtype_map_elts): Remove unused.
20785 * sese.c (debug_ivtype_map): Removed unused.
20786 (debug_ivtype_map_1): Removed unused.
20787 (debug_ivtype_elt): Remove unused.
20788 (eq_ivtype_map_elts): Remove unused.
20791 2013-04-02 Kai Tietz <ktietz@redhat.com>
20794 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20795 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20796 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20798 (legitimize_pe_coff_symbol): Likewise.
20799 (is_imported_p): New helper-function.
20800 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20801 for Windows x64 targets.
20802 (ix86_expand_prologue): Optimize for pe-coff targets.
20803 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20804 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20805 medium/large code-model.
20806 (legitimize_pic_address): Likewise.
20807 (legitimize_tls_address): Likewise.
20808 (ix86_expand_call): Likewise.
20809 (x86_output_mi_thunk): Likewise.
20810 (get_dllimport_decl): Add new beimport argument.
20811 (construct_plt_address): Don't assert for x64 pe-coff targets.
20812 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20814 (SYMBOL_FLAG_STUBVAR): New macro.
20815 (SYMBOL_REF_STUBVAR_P): Likewise.
20816 * config/i386/winnt.c (stub_list): New structure.
20817 (stub_head): New local variable.
20818 (i386_pe_record_stub): New function.
20819 (i386_pe_file_end): Emit refptr-stubs.
20821 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20823 PR rtl-optimization/56745
20824 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20825 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20828 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20829 and both of them are MEM_REFs, just compare first argument for
20830 equality and attempt to deal even with differing offsets.
20833 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20834 of gimple_clobber_p to be MEM_REF.
20835 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20836 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20837 after gimplification.
20838 * asan.c (get_mem_ref_of_assignment): Don't instrument
20839 gimple_clobber_p stmts.
20840 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20841 gimple_clobber_p stmt if they have MEM_REF lhs and
20842 are dead because of another gimple_clobber_p stmt.
20843 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20844 gimple_clobber_p stmts like debug stmts.
20845 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20846 that refer to unused VAR_DECLs or uninitialized values.
20847 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20848 gimple_clobber_p stmts if they refer to removed parameters.
20849 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20852 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20854 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20855 using SWI48 mode attribute.
20857 2013-04-02 Wei Mi <wmi@google.com>
20859 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20860 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20861 *<rotate_insn><mode>3_mask in i386.md.
20863 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20865 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20867 2013-04-02 Richard Biener <rguenther@suse.de>
20869 PR tree-optimization/56778
20870 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20871 Runtime alias tests are not supported for gather loads.
20872 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20873 stmts referenced from SSA operands before updating SSA form.
20875 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20876 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20878 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20879 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20880 * config/arm/cortex-a53.md: New file.
20881 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20882 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20883 * config/arm/arm.c (arm_issue_rate): Likewise.
20884 * config/arm/arm-tune.md: Regenerate
20885 * config/arm/arm-tables.opt: Regenerate.
20886 * config/arm/arm-cores.def: Add cortex-a53.
20888 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20890 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20893 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20895 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20896 scalar load/store operations using B/H registers.
20897 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20899 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20901 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20903 * config/aarch64/aarch64.c
20904 (aarch64_simd_scalar_immediate_valid_for_move): New.
20905 * config/aarch64/aarch64-protos.h
20906 (aarch64_simd_scalar_immediate_valid_for_move): New.
20907 * config/aarch64/constraints.md (Dh, Dq): New.
20908 * config/aarch64/iterators.md (hq): New.
20910 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20912 * reorg.c (get_branch_condition): Deal with conditional returns.
20913 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20915 2013-04-01 Wei Mi <wmi@google.com>
20917 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20918 Truncate operand 2 using %b asm operand modifier.
20919 (*<shift_insn><mode>3_mask): Ditto.
20920 (*<rotate_insn><mode>3_mask): Ditto.
20922 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20924 PR middle-end/56798
20925 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20927 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20929 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20931 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20933 2013-03-30 Lawrence Crowl <crowl@google.com>
20935 * dse.c (clear_alias_sets): Remove never set.
20936 (disqualified_clear_alias_sets): Remove never set.
20937 (clear_alias_mode_pool): Remove never set.
20938 (dse_step0): Remove condition that is never true.
20939 (canon_address): Remove condition that is never true.
20940 (dse_step7): Remove condition that is never true.
20941 (rest_of_handle_dse): Remove condition that is never true.
20942 (rest_of_handle_dse::did_global): Remove never read from above.
20943 (dse_step2_spill): Remove never called from above.
20944 (dse_step5_spill): Remove never called from above.
20946 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20948 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20949 JUMP_TABLE_DATA changes.
20950 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20951 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20952 (Insns) <jump_table_data>: New entry.
20953 * doc/tm.texi: Regenerate.
20955 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20957 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20958 for table jump at the end of a basic block using tablejump_p.
20959 * targhooks.c (default_invalid_within_doloop): Likewise.
20960 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20961 target hook implementation that is identical to the default hook.
20962 (rs6000_invalid_within_doloop): Remove.
20964 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20965 unused variable from tablejump_p call.
20967 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20968 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20969 (INSN_DELETED_P): Likewise.
20970 (emit_jump_table_data): New prototype.
20971 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20972 after 4th as unused.
20973 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20974 * sched-vis.c (print_insn): Likewise.
20975 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20976 insn for compatibility with back ends that use next_active_insn to
20977 identify jump table data.
20978 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20979 (remove_insn): Likewise.
20980 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20982 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20983 (emit_jump_table_data): New function.
20985 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20986 basic block, a JUMP_TABLE_DATA never is.
20987 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20988 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20989 off from code handling real insns.
20990 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20991 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20992 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20993 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20994 is not a NONDEBUG_INSN_P.
20995 * ira-costs.c (scan_one_insn): Likewise.
20996 * jump.c (mark_all_labels): Likewise.
20997 (mark_jump_label_1): Likewise.
20998 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20999 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21001 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21002 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21003 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21004 JUMP_TABLE_DATA_P insns.
21005 (calculate_elim_costs_all_insns): Likewise.
21006 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21007 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21008 (delete_output_reload): Code style fixups.
21009 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21010 insn flags on this non-insn.
21011 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21012 as scheduling barriers, for pre-change compatibility.
21013 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21014 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21016 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21017 redundant JUMP_TABLE_DATA_P test.
21018 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21019 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21020 (frv_for_each_packet): Likewise.
21021 * config/i386/i386.c (min_insn_size): Likewise.
21022 (ix86_avoid_jump_mispredicts): Likewise.
21023 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21024 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21025 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21026 (mips16_insn_length): Robustify.
21027 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21028 (mips16_split_long_branches): Likewise.
21029 * config/pa/pa.c (pa_combine_instructions): Likewise.
21030 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21031 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21032 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21033 as contributing to pool range lengths.
21034 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21035 Remove redundant JUMP_TABLE_DATA_P test.
21036 (sh_loop_align): Likewise.
21037 (split_branches): Likewise.
21038 (sh_insn_length_adjustment): Likewise.
21039 * config/spu/spu.c (get_branch_target): Likewise.
21041 2013-03-29 Jan Hubicka <jh@suse.cz>
21043 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21044 gcov streaming; stream hot bb threshold to ltrans.
21045 * predict.c (get_hot_bb_threshold): Break out from ....
21046 (maybe_hot_count_p): ... here.
21047 (set_hot_bb_threshold): New function.
21048 * lto-section-in.c (lto_section_name): Add profile.
21049 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21050 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21051 and data-streamer.h
21052 (histogram_entry): New structure.
21053 (histogram, histogram_pool): New global vars.
21054 (histogram_hash): New structure.
21055 (histogram_hash::hash): New method.
21056 (histogram_hash::equal): Likewise.
21057 (account_time_size): New function.
21058 (cmp_counts): New function.
21059 (dump_histogram): New function.
21060 (ipa_profile_generate_summary): New function.
21061 (ipa_profile_write_summary): New function.
21062 (ipa_profile_read_summary): New function.
21063 (ipa_profile): Decide on threshold.
21064 (pass_ipa_profile): Add ipa_profile_write_summary and
21065 ipa_profile_read_summary.
21066 * Makefile.in (ipa.o): Update dependencies.
21067 * lto-streamer.h (LTO_section_ipa_profile): New section.
21069 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21071 * tree.h (VAR_P): New.
21073 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21076 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21078 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21080 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21081 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21082 (control_flow_insn_p): Likewise.
21083 * cfgrtl.c (duplicate_insn_chain): Likewise.
21084 * final.c (get_attr_length_1): Likewise.
21085 (shorten_branches): Likewise.
21086 (final_scan_insn): Likewise.
21087 * function.c (instantiate_virtual_regs): Likewise.
21088 * gcse.c (insert_insn_end_basic_block): Likewise.
21089 * ira-costs.c (scan_one_insn): Likewise.
21090 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21091 * lra.c (check_rtl): Likewise.
21092 * reload1.c (elimination_costs_in_insn): Likewise.
21093 * reorg.c (follow_jumps): Likewise.
21095 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21096 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21097 (thumb_far_jump_used_p): Likewise.
21098 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21099 (workaround_speculation): Likewise.
21100 (add_sched_insns_for_speculation): Likewise.
21101 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21102 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21103 (frv_for_each_packet): Likewise.
21104 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21105 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21106 (final_emit_insn_group_barriers): Likewise.
21107 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21108 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21109 (mips16_insn_length): Likewise.
21110 * config/pa/pa.c (pa_reorg): Likewise.
21111 (pa_combine_instructions): Likewise.
21112 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21113 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21114 (sh_reorg): Likewise.
21115 (split_branches): Likewise.
21116 * config/spu/spu.c (get_branch_target): Likewise.
21118 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21121 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21123 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21124 Fix declaration name.
21126 2013-03-28 Lawrence Crowl <crowl@google.com>
21128 * graphds.h (struct graph.indicies): Remove unused.
21129 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21130 (SCOP_ORIGINAL_PDDRS): Remove unused.
21131 * sese.h (extern insert_loop_close_phis): Removed unused.
21132 (extern insert_guard_phis): Removed unused.
21133 (extern ivtype_map_elt_info): Removed unused.
21134 (new_ivtype_map_elt): Removed unused.
21135 * sese.c (ivtype_map_elt_info): Removed unused.
21137 2013-03-28 Lawrence Crowl <crowl@google.com>
21139 * Makefile.in: Add several missing include dependences.
21141 (test-dump.o): New. This object is not added to any executable,
21142 but is present for ad-hoc testing.
21144 (debug (const bitmap_head_def &)): New.
21145 (debug (const bitmap_head_def *)): New.
21147 (extern debug (const bitmap_head_def &)): New.
21148 (extern debug (const bitmap_head_def *)): New.
21150 (debug (edge_def &)): New.
21151 (debug (edge_def *)): New.
21153 (debug (basic_block_def &)): New.
21154 (debug (basic_block_def *)): New.
21156 (dump_node (const_tree, int, FILE *)): Correct source file.
21158 (debug (die_struct &)): New.
21159 (debug (die_struct *)): New.
21161 (extern debug (die_struct &)): New.
21162 (extern debug (die_struct *)): New.
21163 * gimple-pretty-print.c
21164 (debug (gimple_statement_d &)): New.
21165 (debug (gimple_statement_d *)): New.
21166 * gimple-pretty-print.h
21167 (extern debug (gimple_statement_d &)): New.
21168 (extern debug (gimple_statement_d *)): New.
21170 (debug (ira_allocno_copy &)): New.
21171 (debug (ira_allocno_copy *)): New.
21172 (debug (ira_allocno &)): New.
21173 (debug (ira_allocno *)): New.
21175 (extern debug (ira_allocno_copy &)): New.
21176 (extern debug (ira_allocno_copy *)): New.
21177 (extern debug (ira_allocno &)): New.
21178 (extern debug (ira_allocno *)): New.
21180 (debug (live_range &)): New.
21181 (debug (live_range *)): New.
21183 (debug (lra_live_range &)): New.
21184 (debug (lra_live_range *)): New.
21186 (debug (lra_live_range &)): New.
21187 (debug (lra_live_range *)): New.
21189 (debug (omega_pb_d &)): New.
21190 (debug (omega_pb_d *)): New.
21192 (extern debug (omega_pb_d &)): New.
21193 (extern debug (omega_pb_d *)): New.
21195 (debug (const rtx_def &)): New.
21196 (debug (const rtx_def *)): New.
21198 (debug_tree (tree): Move within file.
21199 (debug_raw (const tree_node &)): New.
21200 (debug_raw (const tree_node *)): New.
21201 (dump_tree_via_hooks (const tree_node *, int)): New.
21202 (debug (const tree_node &)): New.
21203 (debug (const tree_node *)): New.
21204 (debug_verbose (const tree_node &)): New.
21205 (debug_verbose (const tree_node *)): New.
21206 (debug_head (const tree_node &)): New.
21207 (debug_head (const tree_node *)): New.
21208 (debug_body (const tree_node &)): New.
21209 (debug_body (const tree_node *)): New.
21210 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21211 (debug (vec<tree, va_gc> &)): New.
21212 (debug (vec<tree, va_gc> *)): New.
21214 (extern debug (const rtx_def &)): New.
21215 (extern debug (const rtx_def *)): New.
21217 (debug_raw (simple_bitmap_def &)): New.
21218 (debug_raw (simple_bitmap_def *)): New.
21219 (debug (simple_bitmap_def &)): New.
21220 (debug (simple_bitmap_def *)): New.
21222 (extern debug (simple_bitmap_def &)): New.
21223 (extern debug (simple_bitmap_def *)): New.
21224 (extern debug_raw (simple_bitmap_def &)): New.
21225 (extern debug_raw (simple_bitmap_def *)): New.
21227 (debug (vinsn_def &)): New.
21228 (debug (vinsn_def *)): New.
21229 (debug_verbose (vinsn_def &)): New.
21230 (debug_verbose (vinsn_def *)): New.
21231 (debug (expr_def &)): New.
21232 (debug (expr_def *)): New.
21233 (debug_verbose (expr_def &)): New.
21234 (debug_verbose (expr_def *)): New.
21235 (debug (vec<rtx> &)): New.
21236 (debug (vec<rtx> *)): New.
21238 (extern debug (vinsn_def &)): New.
21239 (extern debug (vinsn_def *)): New.
21240 (extern debug_verbose (vinsn_def &)): New.
21241 (extern debug_verbose (vinsn_def *)): New.
21242 (extern debug (expr_def &)): New.
21243 (extern debug (expr_def *)): New.
21244 (extern debug_verbose (expr_def &)): New.
21245 (extern debug_verbose (expr_def *)): New.
21246 (extern debug (vec<rtx> &)): New.
21247 (extern debug (vec<rtx> *)): New.
21249 (_list_iter_cond_expr): Make inline instead of static.
21251 (debug (sreal &)): New.
21252 (debug (sreal *)): New.
21254 (extern debug (sreal &)): New.
21255 (extern debug (sreal *)): New.
21257 (extern debug_raw (const tree_node &)): New.
21258 (extern debug_raw (const tree_node *)): New.
21259 (extern debug (const tree_node &)): New.
21260 (extern debug (const tree_node *)): New.
21261 (extern debug_verbose (const tree_node &)): New.
21262 (extern debug_verbose (const tree_node *)): New.
21263 (extern debug_head (const tree_node &)): New.
21264 (extern debug_head (const tree_node *)): New.
21265 (extern debug_body (const tree_node &)): New.
21266 (extern debug_body (const tree_node *)): New.
21267 (extern debug (vec<tree, va_gc> &)): New.
21268 (extern debug (vec<tree, va_gc> *)): New.
21270 (debug (struct loop &)): New.
21271 (debug (struct loop *)): New.
21272 (debug_verbose (struct loop &)): New.
21273 (debug_verbose (struct loop *)): New.
21274 * tree-dump.c: Add header dependence.
21276 (extern debug (struct loop &)): New.
21277 (extern debug (struct loop *)): New.
21278 (extern debug_verbose (struct loop &)): New.
21279 (extern debug_verbose (struct loop *)): New.
21281 (debug (data_reference &)): New.
21282 (debug (data_reference *)): New.
21283 (debug (vec<data_reference_p> &)): New.
21284 (debug (vec<data_reference_p> *)): New.
21285 (debug (vec<ddr_p> &)): New.
21286 (debug (vec<ddr_p> *)): New.
21288 (extern debug (data_reference &)): New.
21289 (extern debug (data_reference *)): New.
21290 (extern debug (vec<data_reference_p> &)): New.
21291 (extern debug (vec<data_reference_p> *)): New.
21292 (extern debug (vec<ddr_p> &)): New.
21293 (extern debug (vec<ddr_p> *)): New.
21295 (debug (pt_solution &)): New.
21296 (debug (pt_solution *)): New.
21298 (extern debug (pt_solution &)): New.
21299 (extern debug (pt_solution *)): New.
21301 (debug (_var_map &)): New.
21302 (debug (_var_map *)): New.
21303 (debug (tree_live_info_d &)): New.
21304 (debug (tree_live_info_d *)): New.
21306 (extern debug (_var_map &)): New.
21307 (extern debug (_var_map *)): New.
21308 (extern debug (tree_live_info_d &)): New.
21309 (extern debug (tree_live_info_d *)): New.
21311 2013-03-28 Jan Hubicka <jh@suse.cz>
21313 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21315 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21317 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21318 record only when desired or required.
21320 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21322 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21323 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21325 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21327 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21328 (*andsi3_compare0_uxtw): New pattern.
21329 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21330 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21332 2013-03-28 Jan Hubicka <jh@suse.cz>
21334 * data-streamer-in.c (streamer_read_gcov_count): New function.
21335 * gimple-streamer-out.c: Include value-prof.h.
21336 (output_gimple_stmt): Output histogram.
21337 (output_bb): Use streamer_write_gcov_count.
21338 * value-prof.c: Include data-streamer.h
21339 (dump_histogram_value): Add HIST_TYPE_MAX.
21340 (stream_out_histogram_value): New function.
21341 (stream_in_histogram_value): New function.
21342 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21343 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21344 * data-streamer-out.c (streamer_write_gcov_count): New function.
21345 (streamer_write_gcov_count_stream): New function.
21346 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21347 (lto_output_node): Likewise.
21348 (input_node, input_edge): Likewise.
21349 * lto-streamer-out.c (output_cfg): Update streaming.
21350 * lto-streamer-in.c (input_cfg): Likewise.
21351 * data-streamer.h (streamer_write_gcov_count,
21352 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21353 * gimple-streamer-in.c: Include value-prof.h
21354 (input_gimple_stmt): Input histograms.
21355 (input_bb): Update profile streaming.
21357 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21359 * genmodes.c (emit_max_int): New function.
21360 (emit_insn_modes_h): Added call to emit_max_function.
21361 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21363 * machmode.def: Fixed comment.
21365 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21367 * combine.c (try_combine): Removed useless assert.
21368 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21370 2013-03-28 Marek Polacek <polacek@redhat.com>
21371 Richard Biener <rguenther@suse.de>
21373 PR tree-optimization/56695
21374 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21375 build signed result of a vector comparison.
21376 * tree-cfg.c (verify_gimple_comparison): Check that a result
21377 of a vector comparison has signed type.
21379 2013-03-28 Richard Biener <rguenther@suse.de>
21381 PR tree-optimization/37021
21382 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21383 do not restrict gaps between groups.
21384 * tree-vect-stmts.c (vectorizable_load): Properly account for
21385 a gap between groups.
21387 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21389 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21390 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21393 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21395 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21396 * doc/extend.texi (Named Address Spaces): Ditto.
21397 (Variable Attributes): Ditto.
21399 2013-03-27 Kai Tietz <ktietz@redhat.com>
21401 * config.build: Add support for cygwin x64 target.
21402 * config.gcc: Likewise.
21403 * config.host: Likewise.
21404 * configure.ac: Likewise
21405 * configure: Regenerated.
21407 2013-03-27 Kai Tietz <ktietz@redhat.com>
21409 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21410 * config/i386/t-cygwin-w64: New file.
21411 * config/i386/cygwin-w64.h: New file.
21412 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21413 and add support for x64-cygwin target.
21414 (CPP_SPEC): Likewise.
21415 (CXX_WRAP_SPEC_LIST): Undefine before define.
21416 (LIBGCJ_SONAME): Use 15 as version.
21418 2013-03-27 Richard Biener <rguenther@suse.de>
21420 PR tree-optimization/56716
21421 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21422 dumping for ref nodes.
21424 2013-03-27 Martin Jambor <mjambor@suse.cz>
21426 PR tree-optimization/55334
21427 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21428 restricted pointers to arrays.
21430 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21432 * Makefile.in (.SUFFIXES): Add .cc.
21433 (.c.o): Apply same recipe for implicit rule .cc.o.
21435 2013-03-27 Richard Biener <rguenther@suse.de>
21437 PR tree-optimization/37021
21438 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21439 REALPART/IMAGPART_EXPRs around the supported refs.
21440 * tree-ssa-structalias.c (find_func_aliases): Assume that
21441 floating-point values are not used to transfer pointers.
21443 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21445 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21446 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21447 * doc/tm.texi: Regenerate.
21448 * targhooks.h (default_has_ifunc_p): New.
21449 * targhooks.c (default_has_ifunc_p): Ditto.
21450 * config/linux-protos.h: New file.
21451 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21452 hook for linux which disables support of indirect functions in android.
21453 * config/linux-android.c: New file.
21454 * config/t-linux-android.c: Ditto.
21455 * config.gcc: Added new object file linux-android.o.
21456 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21457 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21458 * varasm.c (do_assemble_alias): Likewise.
21459 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21460 doesn't support indirect functions.
21461 * configure: Regenerate.
21463 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21466 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21467 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21470 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21472 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21474 2013-03-27 Terry Guo <terry.guo@arm.com>
21476 * config/arm/arm-cores.def: Added core cortex-r7.
21477 * config/arm/arm-tune.md: Regenerated.
21478 * config/arm/arm-tables.opt: Regenerated.
21479 * doc/invoke.texi: Added entry for core cortex-r7.
21481 2013-03-27 Walter Lee <walt@tilera.com>
21483 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21484 double-decrement of next_scratch_regno.
21486 2013-03-27 Walter Lee <walt@tilera.com>
21488 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21490 (insn_v1mulus): Ditto.
21491 (insn_v2muls): Ditto.
21493 2013-03-27 Walter Lee <walt@tilera.com>
21495 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21496 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21498 2013-03-27 Walter Lee <walt@tilera.com>
21500 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21501 (*sibcall_value): Ditto.
21503 2013-03-27 Walter Lee <walt@tilera.com>
21505 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21506 (insn_mnz_v8qi): ... this ...
21507 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21509 (insn_v<n>mnz): Replaced by ...
21510 (insn_v1mnz): ... this ...
21511 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21513 (insn_mz_<mode>): Replaced by ...
21514 (insn_mz_v8qi): ... this ...
21515 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21517 (insn_v<n>mz): Replaced by ...
21518 (insn_v1mz): ... this ...
21519 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21522 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21524 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21526 2013-03-26 Roland McGrath <mcgrathr@google.com>
21528 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21529 than fprintf with a non-constant, non-format string.
21531 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21533 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21534 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21535 operand 0 predicate.
21536 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21537 attribute. Use general_x64nomem_operand as operand 1 predicate.
21538 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21539 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21540 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21541 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21542 (mov<mode>_insv_1): Remove expander. Merge insn with
21543 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21544 Use general_x64nomem_operand as operand 1 predicate.
21545 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21546 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21547 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21548 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21549 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21550 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21551 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21552 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21553 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21554 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21555 (general_x64nomem_operand): Ditto.
21557 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21559 * config/rtems.opt: Add -pthread option.
21561 2013-03-26 Richard Biener <rguenther@suse.de>
21563 * alias.c (find_base_term): Avoid redundant and not used recursion.
21564 (base_alias_check): Get the initial base term from the caller.
21565 (true_dependence_1): Compute and pass base terms to base_alias_check.
21566 (write_dependence_p): Likewise.
21567 (may_alias_p): Likewise.
21569 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21571 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21572 PC-relative load in SI modes and above only.
21574 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21576 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21577 * config/arm/iwmmxt.md (WCGR0): Update.
21578 (WCGR1, WCGR2, WCGR3): Likewise.
21580 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21582 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21583 Use x64 and nox64 isa attributes.
21585 2013-03-26 Richard Biener <rguenther@suse.de>
21587 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21588 alignment computations and rely on get_object_alignment_1
21589 for the !TYPE_P case.
21590 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21592 2013-03-26 Walter Lee <walt@tilera.com>
21594 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21595 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21597 2013-03-25 Jeff Law <law@redhat.com>
21599 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21600 check for INTEGRAL_TYPE_P that was missing due to checking in
21601 wrong version of prior patch.
21603 2013-03-25 Walter Lee <walt@tilera.com>
21605 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21606 TILEGX_INSN_SHUFFLEBYTES1.
21607 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21609 (tilegx_builtins): Ditto.
21610 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21612 2013-03-25 Walter Lee <walt@tilera.com>
21614 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21615 (floatunssisf2): New pattern.
21616 (floatsidf2): New pattern.
21617 (floatunssidf2): New pattern.
21619 2013-03-25 Walter Lee <walt@tilera.com>
21621 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21622 tests for constraint J, K, N, P.
21624 2013-03-25 Walter Lee <walt@tilera.com>
21626 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21627 Use indirect/pcrel encoding.
21628 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21631 2013-03-25 Steve Ellcey <sellcey@mips.com>
21633 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21634 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21635 * config/mips/mips.c (mips_option_override): Set IMADD default.
21636 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21637 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21638 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21639 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21640 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21642 2013-03-25 Jeff Law <law@redhat.com>
21644 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21645 slightly to avoid creating and folding useless trees. Simplify
21646 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21648 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21650 * config/i386/i386.md (*zero_extendsidi2): Merge with
21651 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21652 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21653 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21654 targets. Clarify comment.
21656 2013-03-25 Martin Jambor <mjambor@suse.cz>
21658 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21659 pass-through jump functions differently.
21660 (ipa_read_jump_function): Likewise. Also use setter functions to set
21663 2013-03-25 Martin Jambor <mjambor@suse.cz>
21665 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21666 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21668 (ipa_get_indirect_edge_target): New function.
21669 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21670 ipa_get_indirect_edge_target. Update all callers.
21671 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21672 ipa_get_indirect_edge_target_1 instead of calling
21673 ipa_get_indirect_edge_target.
21674 (create_specialized_node): Pass aggvlas to
21675 ipcp_discover_new_direct_edges.
21677 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21679 * config/arm/arm.md (f_sels, f_seld): New types.
21680 (*cmov<mode>): New pattern.
21681 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21684 2013-03-25 Kai Tietz <ktietz@redhat.com>
21686 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21687 POSIX-printf for mingw-hosted builds.
21689 2013-03-25 Richard Biener <rguenther@suse.de>
21691 PR middle-end/56694
21692 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21693 must-not-throw stmt location.
21695 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21697 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21698 Emit load-acquire versions when acq is true.
21699 (arm_emit_store_exclusive): Add rel parameter.
21700 Emit store-release versions when rel is true.
21701 (arm_split_compare_and_swap): Use acquire-release instructions
21703 of barriers when appropriate.
21704 (arm_split_atomic_op): Likewise.
21705 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21706 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21707 (VUNSPEC_SLX): Likewise.
21708 (VUNSPEC_LDA): Likewise.
21709 (VUNSPEC_STL): Likewise.
21710 * config/arm/sync.md (atomic_load<mode>): New pattern.
21711 (atomic_store<mode>): Likewise.
21712 (arm_load_acquire_exclusive<mode>): Likewise.
21713 (arm_load_acquire_exclusivesi): Likewise.
21714 (arm_load_acquire_exclusivedi): Likewise.
21715 (arm_store_release_exclusive<mode>): Likewise.
21717 2013-03-25 Catherine Moore <clm@codesourcery.com>
21719 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21720 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21721 * config/mip/predicates.md (lwsp_swsp_operand,
21722 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21723 sb16_operand, db4_operand, db7_operand, ib3_operand,
21724 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21725 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21726 andi16_operand): New predicates.
21727 * config/mips/mips.md (compression): New attribute.
21728 (enabled): New attribute.
21729 (length): Consider compression in computing length.
21730 (shift_compression): New code attribute.
21731 (*add<mode>3): New operands. Record compression.
21732 (sub<mode>3): Likewise.
21733 (one_cmpl<mode>2): Likewise.
21734 (*and<mode>3): Likewise.
21735 (*ior<mode>3): Likewise.
21736 (unnamed pattern for xor): Likewise.
21737 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21738 (*<optab><mode>3): Likewise.
21739 (*mov<mode>_internal: Likewise.
21740 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21741 (mips_unsigned_immediate_p): New.
21742 (umips_lwsp_swsp_address_p): New.
21743 (m16_based_address_p): New.
21744 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21745 (mips_unsigned_immediate_p): New prototype.
21746 (lwsp_swsp_address_p): New prototype.
21747 (m16_based_address_p): New prototype.
21748 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21749 (mips_signed_immediate_p): New function.
21750 (m16_based_address_p): New function.
21751 (lwsp_swsp_address_p): New function.
21752 (mips_print_operand_punctuation): Recognize short delay slot insns
21753 for microMIPS.add<mode>3"
21755 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21758 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21759 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21761 2013-03-25 Richard Biener <rguenther@suse.de>
21763 PR tree-optimization/56689
21764 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21767 2013-03-25 Richard Biener <rguenther@suse.de>
21769 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21771 (memory_references): Likewise.
21772 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21773 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21774 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21775 (gather_mem_refs_in_loops): Fold into ...
21776 (analyze_memory_references): ... this. Move initialization
21777 to tree_ssa_lim_initialize.
21778 (fill_always_executed_in): Rename to ...
21779 (fill_always_executed_in_1): ... this.
21780 (fill_always_executed_in): Move contains_call computation to
21781 this new function from ...
21782 (tree_ssa_lim_initialize): ... here.
21783 (tree_ssa_lim): Call fill_always_executed_in.
21785 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21787 * postreload.c (reload_combine): Fix code detecting returns.
21789 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21791 * function.c (emit_use_return_register_into_block): On cc0 targets,
21792 do not emit the sequence between cc0 setter and user.
21794 2013-03-25 Kai Tietz <ktietz@redhat.com>
21796 * config/i386/predicates.md (local_symbolic_operand): Interpret
21797 dll-imported symbols as none-local.
21799 2013-03-25 Richard Biener <rguenther@suse.de>
21801 * tree-ssa-loop-im.c (struct depend): Remove.
21802 (struct lim_aux_data): Make depends a vec of gimples.
21803 (free_lim_aux_data): Adjust.
21804 (add_dependency): Likewise.
21805 (set_level): Likewise.
21807 2013-03-25 Richard Biener <rguenther@suse.de>
21809 PR middle-end/56434
21810 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21811 the pointer returned by calls with ECF_MALLOC set.
21813 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21815 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21817 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21819 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21820 using MMXMODE mode iterator.
21821 (*move<mode>_internal): Merge with *movv2sf_internal and
21822 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21824 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21826 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21827 (record_last_mem_set_info): Likewise.
21829 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21830 of XNEWVEC followed by memset.
21831 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21833 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21835 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21836 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21837 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21838 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21839 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21840 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21841 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21842 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21843 BARRIER_P instead of GET_CODE.
21845 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21847 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21848 inaccuracy in the probing code.
21850 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21851 (ctrapdi4): Likewise.
21853 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21855 * calls.c (expand_call): Add missing guard to code handling return
21856 of non-BLKmode structures in MSB.
21857 * function.c (expand_function_end): Likewise.
21859 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21861 * combine.c (try_combine): Adjust comment. Do not add the set of
21862 insn #0 if the destination indirectly is set or dies in insn #2.
21863 Tidy up code to distribute a new note.
21865 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21867 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21868 also for alternatives 16 and 17.
21870 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21872 * config/i386/sse.md (*mov<mode>_internal): Merge with
21873 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21874 Emit insn template depending on type attribute. Use
21875 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21876 movd instead of movq mnemonic for interunit moves. Rewrite mode
21877 attribute calculation. Remove unit attribute calculation.
21878 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21879 Set prefix_data16 attribute for DImode ssemov types.
21880 Use Ym instead of y for SSE-MMX conversion alternatives.
21881 Reorder operand constraints.
21883 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21885 * df.h (df_insn_delete): Adjust prototype.
21886 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21887 and let it decide whether mark the basic block dirty.
21888 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21889 * df-scan.c (df_insn_info_delete): New helper function, split
21890 off from df_insn_delete.
21891 (df_scan_free_bb_info): Use it.
21892 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21894 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21895 that the insn is actually an insn and it has a non-NULL basic block.
21896 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21898 2013-03-22 Richard Biener <rguenther@suse.de>
21900 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21902 (mem_ref_alloc): Do not allocate them.
21903 (refs_independent_p): Do not query or maintain a cache.
21905 2013-03-22 Richard Biener <rguenther@suse.de>
21907 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21908 (gather_mem_refs_in_loops): Do not compute it.
21909 (analyze_memory_references): Do not allocate it.
21910 (tree_ssa_lim_finalize): Do not free it.
21911 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21913 2013-03-22 Richard Biener <rguenther@suse.de>
21915 * is-a.h (as_a): Use gcc_checking_assert.
21917 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21919 * config/aarch64/aarch64.c (aarch64_print_operand): New
21920 format specifier for printing a constant in hex.
21921 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21922 format specifier for printing second operand.
21924 2013-03-22 Richard Biener <rguenther@suse.de>
21926 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21928 (gather_mem_refs_in_loops): Perform store accumulation here.
21929 (create_vop_ref_mapping_loop): Remove.
21930 (create_vop_ref_mapping): Likewise.
21931 (analyze_memory_references): Initialize refs_stored_in_loop.
21932 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21933 (record_indep_loop): Remove.
21934 (record_dep_loop): New function.
21935 (ref_indep_loop_p_1): Adjust to only walk over references
21936 in the loop, not its subloops.
21937 (ref_indep_loop_p): Rename to ...
21938 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21939 maintaining a more fine-grained cache.
21940 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21941 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21943 2013-03-22 Richard Biener <rguenther@suse.de>
21945 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21946 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21947 aggregate mem_ref_loc.
21948 (free_mem_ref_locs): Inline into ...
21949 (memref_free): ... this and adjust.
21950 (mem_ref_alloc): Adjust.
21951 (mem_ref_locs_alloc): Remove.
21952 (record_mem_ref_loc): Adjust.
21953 (get_all_locs_in_loop): Rewrite into ...
21954 (for_all_locs_in_loop): ... this iterator.
21955 (rewrite_mem_ref_loc): New functor.
21956 (rewrite_mem_refs): Use for_all_locs_in_loop.
21957 (sm_set_flag_if_changed): New functor.
21958 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21959 (ref_always_accessed): New functor.
21960 (ref_always_accessed_p): Use for_all_locs_in_loop.
21962 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21964 * tree-pass.h (PROP_gimple_lvec): New.
21965 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21966 (init_optimization_passes): Move pass_lower_vector.
21967 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21969 (pass_lower_vector): Provide PROP_gimple_lvec.
21970 (pass_lower_vector_ssa): Likewise.
21971 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21973 2013-03-21 Mark Wielaard <mjw@redhat.com>
21975 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21977 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21979 * config/i386/i386.md (*movdi_internal): Disparage slightly
21980 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21981 conversion alternatives.
21983 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21985 PR middle-end/48087
21986 * diagnostic.def (DK_WERROR): New kind.
21987 * diagnostic.h (werrorcount): Define.
21988 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21989 promoted to DK_ERROR, increment DK_WERROR counter instead of
21991 * toplev.c (toplev_main): Call print_ignored_options even if
21992 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21993 even if just werrorcount is non-zero.
21996 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21998 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21999 (string_cst_pool_decl): New function.
22000 (optimize_one_addr_into_implicit_ptr): New function.
22001 (resolve_addr_in_expr): Optimize DWARF location expression
22002 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22003 which doesn't live in memory, but has DW_AT_location or
22004 DW_AT_const_value, or refers to a string literal, into
22005 DW_OP_GNU_implicit_pointer.
22006 (optimize_location_into_implicit_ptr): New function.
22007 (resolve_addr): If removing DW_AT_location of a variable because
22008 it was DW_OP_addr of address of the variable, but the variable doesn't
22009 live in memory, try to emit const value attribute for the initializer.
22011 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22013 * tree.h (VECTOR_TYPE_P): New macro.
22014 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22015 TYPE_MODE): Use it.
22016 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22017 VEC_COND_EXPR cannot be lvalues.
22018 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22020 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22022 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22023 Restrict the transformation to equal modes.
22025 2013-03-21 Richard Biener <rguenther@suse.de>
22027 PR tree-optimization/39326
22028 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22029 (MEM_ANALYZABLE): Adjust.
22030 (record_mem_ref_loc): Move bitmap ops ...
22031 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22032 unanalyzable refs, do not record locations for it.
22033 (analyze_memory_references): Allocate ref zero as shared
22035 (refs_independent_p): Do not test for unanalyzed mems here.
22036 (ref_indep_loop_p_1): Special-case disambiguation against
22037 the unanalyzed ref.
22038 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22040 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22042 * config/arm/arm-protos.h (tune_params): Add
22043 prefer_neon_for_64bits field.
22044 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22045 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22046 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22047 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22048 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22049 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22050 (arm_option_override): Handle -mneon-for-64bits new option.
22051 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22052 (prefer_neon_for_64bits): Declare new variable.
22053 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22054 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22055 (arch_enabled): Handle new arch types. Remove support for onlya8
22057 (one_cmpldi2): Use new arch names.
22058 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22059 * config/arm/arm.opt (mneon-for-64bits): Add option.
22060 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22061 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22062 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22064 * doc/invoke.texi (-mneon-for-64bits): Document.
22066 2013-03-21 Richard Biener <rguenther@suse.de>
22068 PR tree-optimization/39326
22069 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22070 (sort_bbs_in_loop_postorder_cmp): New function.
22071 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22073 2013-03-21 Richard Biener <rguenther@suse.de>
22075 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22076 (vect_insert_into_interleaving_chain): Likewise.
22077 (vect_drs_dependent_in_basic_block): Inline ...
22078 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22080 (vect_analyze_data_ref_dependence): ... here. Simplify.
22081 (vect_check_interleaving): Simplify.
22082 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22083 (vect_slp_analyze_data_ref_dependences): ... this new function.
22084 (dr_group_sort_cmp): New function.
22085 (vect_analyze_data_ref_accesses): Compute data-reference groups
22086 here instead of in vect_analyze_data_ref_dependence. Use
22087 a more efficient algorithm.
22088 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22089 vect_slp_analyze_data_ref_dependences. Call
22090 vect_analyze_data_ref_accesses earlier.
22091 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22092 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22093 (vect_slp_analyze_data_ref_dependences): New prototype.
22095 2013-03-21 Richard Biener <rguenther@suse.de>
22097 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22098 ref is stored in the loop.
22099 (find_refs_for_sm): Walk only over all stores.
22100 (store_motion_loop): Allocate from lim_bitmap_obstack.
22101 (store_motion): Likewise.
22103 2013-03-21 Richard Biener <rguenther@suse.de>
22105 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22106 Update virtual SSA form.
22108 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22110 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22111 * configure: Regenerate.
22112 * config.in: Regenerate.
22113 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22114 if !HAVE_LD_EH_FRAME_CIEV3.
22116 2013-03-21 Richard Biener <rguenther@suse.de>
22118 * tree-cfg.c (verify_expr_no_block): New function.
22119 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22120 nor DECL_VALUE_EXPR have locations with associated blocks.
22121 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22122 (clear_unused_block_pointer): Remove code dealing with
22123 blocks in DECL_DEBUG_EXPR locations.
22125 2013-03-21 Richard Biener <rguenther@suse.de>
22127 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22128 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22129 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22130 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22131 instead of DECL_DEBUG_EXPR_IS_FROM.
22132 * gimplify.c (gimplify_modify_expr): Likewise.
22133 * tree-cfg.c (verify_expr_location_1): Likewise.
22134 * tree-complex.c (create_one_component_var): Likewise.
22135 * tree-sra.c (create_access_replacement): Likewise.
22136 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22137 (clear_unused_block_pointer): Likewise.
22138 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22139 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22140 * var-tracking.c (var_debug_decl): Likewise.
22141 (track_expr_p): Likewise.
22142 * tree-inline.c (add_local_variables): Likewise. Set
22143 DECL_HAS_DEBUG_EXPR_P after copying it.
22144 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22145 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22147 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22150 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22151 * configure: Regenerate.
22152 * config.in: Regenerate.
22153 * config/i386/i386.md (*movdf_internal): Use
22154 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22155 movd instead of movq mnemonic for interunit moves.
22156 (*movdi_internal): Ditto.
22158 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22160 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22161 (abd<mode>_3): New pattern.
22162 (aba<mode>_3): New pattern.
22163 (fabd<mode>_3): New pattern.
22165 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22167 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22168 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22169 occurrence of REGISTER_PREFIX as its empty string.
22171 2013-03-20 Jeff Law <law@redhat.com>
22173 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22174 addititional equivalences for equality comparisons between an SSA_NAME
22175 and a constant where the SSA_NAME was set from a widening conversion.
22177 2013-03-20 Walter Lee <walt@tilera.com>
22179 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22181 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22183 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22184 depending on type attribute.
22185 (*movti_internal): Ditto.
22186 (*movtf_internal): Ditto.
22187 (*movxf_internal): Ditto.
22188 (*movdf_internal): Ditto.
22189 (*movsf_internal): Ditto.
22191 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22193 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22194 maybe_vex for sselog1 and ssemov types.
22195 (*movdi_internal): Reorder operand constraints.
22196 (*movsi_internal): Ditto. Set prefix attribute to
22197 maybe_vex for sselog1 and ssemov types.
22198 (*movtf_internal): Set prefix attribute to maybe_vex
22199 for sselog1 and ssemov types.
22200 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22201 DImode ssemov types. Reorder operand constraints.
22202 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22203 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22204 attribute for SImode ssemov types. Reorder operand constraints.
22206 2013-03-20 Martin Jambor <mjambor@suse.cz>
22208 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22209 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22211 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22213 * config/rs6000/predicates.md (indexed_address, update_address_mem
22214 update_indexed_address_mem): New predicates.
22215 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22216 attribute for load/store instructions.
22217 * config/rs6000/dfp.md (movsd_store): Likewise.
22218 (movsd_load): Likewise.
22219 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22220 (unnamed HI->DI extend define_insn): Likewise.
22221 (unnamed SI->DI extend define_insn): Likewise.
22222 (unnamed QI->SI extend define_insn): Likewise.
22223 (unnamed QI->HI extend define_insn): Likewise.
22224 (unnamed HI->SI extend define_insn): Likewise.
22225 (unnamed HI->SI extend define_insn): Likewise.
22226 (extendsfdf2_fpr): Likewise.
22227 (movsi_internal1): Likewise.
22228 (movsi_internal1_single): Likewise.
22229 (movhi_internal): Likewise.
22230 (movqi_internal): Likewise.
22231 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22232 attribute for load/store instructions.
22233 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22235 (mov<mode>_softfloat): Likewise.
22236 (mov<mode>_hardfloat32): Likewise.
22237 (mov<mode>_hardfloat64): Likewise.
22238 (mov<mode>_softfloat64): Likewise.
22239 (movdi_internal32): Likewise.
22240 (movdi_internal64): Likewise.
22241 (probe_stack_<mode>): Likewise.
22243 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22245 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22246 floating point, and decimal floating point to reload iterator.
22248 * config/rs6000/constraints.md (wl constraint): New constraints to
22249 return FLOAT_REGS if certain options are used to reduce the number
22250 of separate patterns that exist in the file.
22251 (wx constraint): Likewise.
22252 (wz constraint): Likewise.
22254 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22255 -mdebug=reg, print wg, wl, wx, and wz constraints.
22256 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22257 Initialize the reload functions for 64-bit binary/decimal floating
22259 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22260 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22261 create the buffer on the stack to overcome not having a 32-bit
22263 (rs6000_emit_move): Likewise.
22264 (rs6000_secondary_memory_needed_rtx): Likewise.
22265 (rs6000_alloc_sdmode_stack_slot): Likewise.
22266 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22267 via xxlxor, just like DFmode 0.0.
22269 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22270 define as 1 if we are running on a power7 or newer.
22271 (enum r6000_reg_class_enum): Add new constraints.
22273 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22274 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22275 with other moves by using conditional constraits (wg). Use LFIWZX
22276 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22277 (movsd splitter): Likewise.
22278 (movsd_hardfloat): Likewise.
22279 (movsd_softfloat): Likewise.
22281 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22282 binary and decimal floating point moves.
22283 (fmove_ok): New attributes to combine binary and decimal floating
22284 point moves, and to combine power6x (mfpgpr) moves along normal
22286 (real_value_to_target): Likewise.
22287 (f32_lr): Likewise.
22288 (f32_lm): Likewise.
22289 (f32_li): Likewise.
22290 (f32_sr): Likewise.
22291 (f32_sm): Likewise.
22292 (f32_si): Likewise.
22293 (movsf): Combine binary and decimal floating point moves. Combine
22294 power6x (mfpgpr) moves with other moves by using conditional
22295 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22296 (mov<mode> for SFmode/SDmode); Likewise.
22297 (SFmode/SDmode splitters): Likewise.
22298 (movsf_hardfloat): Likewise.
22299 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22300 (movsf_softfloat): Likewise.
22301 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22303 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22304 wx and wz constraints.
22306 * config/rs6000/constraints.md (wg constraint): New constraint to
22307 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22309 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22312 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22313 -mdebug=reg, print wg, wl, wx, and wz constraints.
22314 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22315 Initialize the reload functions for 64-bit binary/decimal floating
22317 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22318 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22319 create the buffer on the stack to overcome not having a 32-bit
22321 (rs6000_emit_move): Likewise.
22322 (rs6000_secondary_memory_needed_rtx): Likewise.
22323 (rs6000_alloc_sdmode_stack_slot): Likewise.
22324 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22325 via xxlxor, just like DFmode 0.0.
22327 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22328 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22329 with other moves by using conditional constraits (wg). Use LFIWZX
22330 and STFIWX for loading SDmode on power7.
22331 (movdd splitters): Likewise.
22332 (movdd_hardfloat32): Likewise.
22333 (movdd_softfloat32): Likewise.
22334 (movdd_hardfloat64_mfpgpr): Likewise.
22335 (movdd_hardfloat64): Likewise.
22336 (movdd_softfloat64): Likewise.
22338 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22339 64-bit binary and decimal floating point moves.
22340 (FMOVE64X): Likewise.
22341 (movdf): Combine 64-bit binary and decimal floating point moves.
22342 Combine power6x (mfpgpr) moves with other moves by using
22343 conditional constraits (wg).
22344 (mov<mode> for DFmode/DDmode): Likewise.
22345 (DFmode/DDmode splitters): Likewise.
22346 (movdf_hardfloat32): Likewise.
22347 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22348 (movdf_softfloat32): Likewise.
22349 (movdf_hardfloat64_mfpgpr): Likewise.
22350 (movdf_hardfloat64): Likewise.
22351 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22352 (movdf_softfloat64): Likewise.
22353 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22354 (reload_<mode>_load): Move to later in the file so they aren't in
22355 the middle of the floating point move insns.
22356 (reload_<mode>_store): Likewise.
22358 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22361 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22362 constraint if -mdebug=reg.
22363 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22364 Enable using dd reload support if needed.
22366 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22367 binary and decimal floating point moves in rs6000.md.
22368 (movtd_internal): Likewise.
22370 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22371 decimal floating point moves.
22373 (movtf_internal): Likewise.
22374 (mov<mode>_internal, TDmode/TFmode): Likewise.
22375 (movtf_softfloat): Likewise.
22376 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22378 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22379 movdi_internal64, using wg constraint for move direct operations.
22380 (movdi_internal64): Likewise.
22382 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22383 MODES_TIEABLE_P for selected modes. Print the numerical value of
22384 the various virtual registers. Use GPR/FPR first/last values,
22385 instead of hard coding the register numbers. Print which modes
22386 have reload functions registered.
22387 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22388 settings before/after setting cpu, target and subtarget settings.
22389 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22390 and for secondary reload failures in rs6000_secondary_reload_inner.
22391 (rs6000_secondary_reload_fail): Likewise.
22392 (rs6000_secondary_reload_inner): Likewise.
22394 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22395 macros for first/last GPR and FPR registers.
22396 (LAST_GPR_REGNO): Likewise.
22397 (FIRST_FPR_REGNO): Likewise.
22398 (LAST_FPR_REGNO): Likewise.
22400 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22401 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22402 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22403 (vcond<mode><mode>): Likewise.
22404 (vcondu<mode><mode>): Likewise.
22405 (vector_gtu<mode>): Likewise.
22406 (vector_gte<mode>): Likewise.
22407 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22408 to prevent the compiler from converting DImode operations to TImode.
22409 (ior<mode>3): Likewise.
22410 (and<mode>3): Likewise.
22411 (one_cmpl<mode>2): Likewise.
22412 (nor<mode>3): Likewise.
22413 (andc<mode>3): Likewise.
22415 * config/rs6000/constraints.md (wt constraint): New constraint
22416 that returns VSX_REGS if TImode is allowed in VSX registers.
22418 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22419 constant under VSX.
22421 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22422 similar to TImode, but it is restricted to being in the GPRs.
22424 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22425 TImode to occupy a single VSX register.
22427 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22428 -mvsx-timode for power7/power8.
22429 (power7 cpu): Likewise.
22430 (power8 cpu): Likewise.
22432 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22433 sure that TFmode/TDmode take up two registers if they are ever
22434 allowed in the upper VSX registers.
22435 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22437 (rs6000_init_hard_regno_mode_ok): Likewise.
22438 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22439 constraint. Print if LRA is turned on.
22440 (rs6000_option_override_internal): Give an error if -mvsx-timode
22441 and VSX is not enabled.
22442 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22443 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22444 to reg+offset addressing. Use PTImode when checking offset
22445 addresses for validity.
22446 (reg_offset_addressing_ok_p): Likewise.
22447 (rs6000_legitimate_offset_address_p): Likewise.
22448 (rs6000_legitimize_address): Likewise.
22449 (rs6000_legitimize_reload_address): Likewise.
22450 (rs6000_legitimate_address_p): Likewise.
22451 (rs6000_eliminate_indexed_memrefs): Likewise.
22452 (rs6000_emit_move): Likewise.
22453 (rs6000_secondary_reload): Likewise.
22454 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22455 reloads to fpr registers to continue to use reg+offset addressing,
22456 but 64-bit reloads to altivec registers need reg+reg addressing.
22457 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22458 it. Treat LO_SUM like a PLUS operation.
22459 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22460 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22461 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22462 registers to share a register with a smaller sized type, since VSX
22463 puts scalars in the upper 64-bits.
22464 (print_operand): Add support for PTImode.
22465 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22466 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22467 registers, but don't have arithmetic support.
22468 (rs6000_memory_move_cost): Add test for VSX.
22469 (rs6000_opt_masks): Add -mvsx-timode.
22471 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22474 (VSr): Use wt constraint for TImode.
22475 (VSv): Drop TImode support.
22476 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22477 (vsx_movti_64bit): Likewise.
22478 (vsx_movti_32bit): Likewise.
22479 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22480 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22481 one '?' on the appropriate output constraint. Do not allow TImode
22482 logical operations on 32-bit systems.
22483 (vsx_ior<mode>3): Likewise.
22484 (vsx_xor<mode>3): Likewise.
22485 (vsx_one_cmpl<mode>2): Likewise.
22486 (vsx_nor<mode>3): Likewise.
22487 (vsx_andc<mode>3): Likewise.
22488 (vsx_concat_<mode>): Likewise.
22489 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22491 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22492 OPTION_MASK_VSX_TIMODE.
22493 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22494 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22496 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22497 (TI2 iterator): New iterator for TImode, PTImode.
22498 (wd mode attribute): Add values for vector types.
22499 (movti_string): Replace TI move operations with operations for TImode
22500 and PTImode. Add support for TImode being allowed in VSX registers.
22501 (mov<mode>_string, TImode/PTImode): Likewise.
22502 (movti_ppc64): Likewise.
22503 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22504 (TI mode splitters): Likewise.
22506 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22509 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22511 PR tree-optimization/56355
22512 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22513 Also handle integers with undefined overflow.
22515 2013-03-20 Catherine Moore <clm@codesourcery.com>
22516 Maciej W. Rozycki <macro@codesourcery.com>
22517 Tom de Vries <tom@codesourcery.com>
22518 Nathan Sidwell <nathan@codesourcery.com>
22519 Iain Sandoe <iain@codesourcery.com>
22520 Nathan Froyd <froydnj@codesourcery.com>
22521 Chao-ying Fu <fu@mips.com>
22523 * doc/extend.texi: (micromips, nomicromips, nocompression):
22524 Document new function attributes.
22525 * doc/invoke.texi (minterlink-compressed, mmicromips,
22526 m14k, m14ke, m14kec): Document new options.
22527 (minterlink-mips16): Update documentation.
22528 * doc/md.texi (ZC, ZD): Document new constraints.
22529 * configure.ac (gcc_cv_as_micromips): Check if linker
22530 supports the .set micromips directive.
22531 * configure: Regenerate.
22532 * config.in: Regenerate.
22533 * config/mips/mips-tables.opt: Regenerate.
22534 * config/mips/micromips.md: New file.
22535 * constraints.md (ZC, ZD): New constraints.
22536 * config/mips/predicates.md (movep_src_register): New predicate.
22537 (movep_src_operand): New predicate.
22538 (non_volatile_mem_operand): New predicate.
22539 * config/mips/mips.md (multimem): New type.
22540 (length): Differentiate between 17-bit and 18-bit branch offsets.
22541 (MOVEP1, MOVEP2): New mode iterator.
22542 (mov_<load>l): Use ZC constraint.
22543 (mov_<load>r): Likewise.
22544 (mov_<store>l): Likewise.
22545 (mov_<store>r): Likewise.
22546 (*branch_equality<mode>_inverted): Add microMIPS support.
22547 (*branch_equality<mode>): Likewise.
22548 (*jump_absolute): Likewise.
22549 (indirect_jump_<mode>): Likewise.
22550 (tablejump_<mode>): Likewise.
22551 (<optab>_internal): Likewise.
22552 (sibcall_internal): Likewise.
22553 (sibcall_value_internal): Likewise.
22554 (prefetch): Use constraint ZD.
22555 * config/mips/mips.opt (minterlink-compressed): New option.
22556 (minterlink-mips16): Now an alias for minterlink-compressed.
22557 (mmicromips): New option.
22558 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22559 (compare_and_swap_12): Likewise.
22560 (sync_add<mode>): Likewise.
22561 (sync_<optab>_12): Likewise.
22562 (sync_old_<optab>_12): Likewise.
22563 (sync_new_<optab>_12): Likewise.
22564 (sync_nand_12): Likewise.
22565 (sync_old_nand_12): Likewise.
22566 (sync_new_nand_12): Likewise.
22567 (sync_sub<mode>): Likewise.
22568 (sync_old_add<mode>): Likewise.
22569 (sync_old_sub<mode>): Likewise.
22570 (sync_new_add<mode>): Likewise.
22571 (sync_new_sub<mode>): Likewise.
22572 (sync_<optab><mode>): Likewise.
22573 (sync_old_<optab><mode>): Likewise.
22574 (sync_new_<optab><mode>): Likewise.
22575 (sync_nand<mode>): Likewise.
22576 (sync_old_nand<mode>): Likewise.
22577 (sync_new_nand<mode>): Likewise.
22578 (sync_lock_test_and_set<mode>): Likewise.
22579 (test_and_set_12): Likewise.
22580 (atomic_compare_and_swap<mode>): Likewise.
22581 (atomic_exchange<mode>_llsc): Likewise.
22582 (atomic_fetch_add<mode>_llsc): Likewise.
22583 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22584 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22585 (umips_save_restore_pattern_p): Likewise.
22586 (umips_load_store_pair_p): Likewise.
22587 (umips_output_load_store_pair): Likewise.
22588 (umips_movep_target_p): Likewise.
22589 (umips_12bit_offset_address_p): Likewise.
22590 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22591 (mips_base_mips16): Rename this...
22592 (mips_base_compression_flags): ...to this. Update all uses.
22593 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22594 (mips_mips16_decl_p): Delete.
22595 (mips_nomips16_decl_p): Delete.
22596 (mips_get_compress_on_flags): New function.
22597 (mips_get_compress_off_flags): New function.
22598 (mips_get_compress_mode): New function.
22599 (mips_get_compress_on_name): New function.
22600 (mips_get_compress_off_name): New function.
22601 (mips_insert_attributes): Support multiple compression types.
22602 (mips_merge_decl_attributes): Likewise.
22603 (umips_12bit_offset_address_p): New function.
22604 (mips_start_function_definition): Emit .set micromips directive.
22605 (mips_call_may_need_jalx_p): New function.
22606 (mips_function_ok_for_sibcall): Add microMIPS support.
22607 (mips_print_operand_punctuation): Support short delay slots and
22609 (umips_swm_mask, umips_swm_encoding): New.
22610 (umips_build_save_restore): New function.
22611 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22612 (was_mips16_p): Remove.
22613 (old_compression_mode): New.
22614 (mips_set_compression_mode): New function.
22615 (mips_set_current_function): Add microMIPS support.
22616 (mips_option_override): Likewise.
22617 (umips_save_restore_pattern_p): New function.
22618 (umips_output_save_restore): New function.
22619 (umips_load_store_pair_p_1): New function.
22620 (umips_load_store_pair_p): New function.
22621 (umips_output_load_store_pair_1): New function.
22622 (umips_output_load_store_pair): New function.
22623 (umips_movep_target_p) New function.
22624 (mips_prepare_pch_save): Add microMIPS support.
22625 * config/mips/mips.h (TARGET_COMPRESSION): New.
22626 (TARGET_CPU_CPP_BUILTINS): Update macro
22627 to use new compression flags and to support microMIPS.
22628 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22629 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22630 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22631 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22632 (ASM_SPEC): Support mmicromips and mno-micromips.
22633 (M16STORE_REG_P): New macro.
22634 (MIPS_CALL): Support TARGET_MICROMIPS.
22635 (MICROMIPS_J): New macro.
22636 (mips_base_mips16): Rename this...
22637 (mips_base_compression_flags): ...to this.
22638 (UMIPS_12BIT_OFFSET_P): New macro.
22639 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22640 (MULTILIB_DIRNAMES): Likewise.
22641 2013-03-20 Richard Biener <rguenther@suse.de>
22643 PR tree-optimization/56661
22644 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22645 the result does not have to be distinct.
22647 2013-03-20 Richard Biener <rguenther@suse.de>
22649 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22652 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22653 Steven Bosscher <steven@gcc.gnu.org>
22655 PR rtl-optimization/56605
22656 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22658 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22661 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22662 that require movd instead of movq.
22664 2013-03-20 Richard Biener <rguenther@suse.de>
22666 * tree-ssa-structalias.c (struct variable_info): Add pointer
22667 to the first field of an aggregate with sub-vars. Make
22668 this and the pointer to the next subfield its ID.
22669 (vi_next): New function.
22670 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22671 storedanything_id, integer_id): Increment by one.
22672 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22673 get_call_clobber_vi): Adjust.
22674 (solution_set_expand): Simplify and speedup.
22675 (solution_set_add): Inline into ...
22676 (set_union_with_increment): ... this. Adjust accordingly.
22677 (do_sd_constraint): Likewise.
22678 (do_ds_constraint): Likewise.
22679 (do_complex_constraint): Simplify.
22680 (build_pred_graph): Adjust.
22681 (solve_graph): Likewise. Simplify and speedup.
22682 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22683 get_constraint_for_component_ref, get_constraint_for_1,
22684 first_vi_for_offset, first_or_preceding_vi_for_offset,
22685 create_function_info_for, create_variable_info_for_1,
22686 create_variable_info_for, intra_create_variable_infos): Adjust.
22687 (init_base_vars): Push NULL for ID zero.
22688 (compute_points_to_sets): Adjust.
22690 2013-03-20 Richard Biener <rguenther@suse.de>
22692 * cfgloop.c (verify_loop_structure): Streamline and avoid
22693 ICEing on corrupt loop tree.
22694 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22697 2013-03-20 Richard Biener <rguenther@suse.de>
22699 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22700 check whether an SSA update is needed.
22702 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22704 * config/mips/constraints.md (T): Rename to...
22708 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22709 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22711 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22713 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22714 (*subsi3_carryin_uxtw): Likewise.
22716 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22718 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22719 (*rorsi3_insn_uxtw): Likewise.
22721 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22723 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22724 (*extrsi5_insn_uxtw): Likewise.
22726 2013-03-19 Richard Biener <rguenther@suse.de>
22728 PR tree-optimization/56273
22729 * passes.c (init_optimization_passes): Move second VRP after DOM.
22731 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22733 * config/i386/i386.md (*movti_internal): Merge from
22734 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22735 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22736 nox64 isa attributes.
22738 2013-03-18 Richard Biener <rguenther@suse.de>
22740 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22742 (merge_node_constraints): Likewise.
22743 (build_succ_graph): Likewise.
22744 (valid_graph_edge): Inline into single caller.
22745 (unify_nodes): Likewise. Use bitmap_set_bit return value
22747 (scc_visit): Fix formatting and variable use.
22748 (do_sd_constraint): Use gcc_checking_assert.
22749 (do_ds_constraint): Likewise.
22750 (do_complex_constraint): Likewise.
22751 (condense_visit): Likewise. Cleanup.
22752 (dump_pred_graph): New function.
22753 (perform_var_substitution): Dump the pred-graph before
22754 variable substitution.
22755 (find_equivalent_node): Use gcc_checking_assert.
22756 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22758 2013-03-18 Richard Biener <rguenther@suse.de>
22760 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22761 Remove cond_expr_stmt_list argument and do not gimplify the
22763 (vect_loop_versioning): Adjust.
22764 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22765 Cleanup to use less temporaries.
22766 (vect_create_data_ref_ptr): Cleanup.
22768 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22770 PR tree-optimization/56635
22771 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22772 require types_compatible_p types.
22774 2013-03-18 Nick Clifton <nickc@redhat.com>
22776 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22777 spurious backslash.
22779 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22780 Add missing line to comment describing function.
22782 2013-03-18 Richard Biener <rguenther@suse.de>
22784 PR tree-optimization/56210
22785 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22786 Handle string / character search functions.
22787 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22789 2013-03-18 Richard Biener <rguenther@suse.de>
22791 PR middle-end/56483
22792 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22793 and implement properly.
22794 * gimple.h (gimple_cond_single_var_p): Remove.
22796 2013-03-18 Richard Biener <rguenther@suse.de>
22798 * tree-data-ref.h (find_data_references_in_loop): Declare.
22799 * tree-data-ref.c (get_references_in_stmt): Use a stack
22800 vector pre-allocated in the callers.
22801 (find_data_references_in_stmt): Adjust.
22802 (graphite_find_data_references_in_stmt): Likewise.
22803 (create_rdg_vertices): Likewise.
22804 (find_data_references_in_loop): Export.
22805 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22806 Compute dependences here...
22807 (vect_analyze_data_refs): ...not here. When we encounter
22808 a non-vectorizable data reference in basic-block vectorization
22809 truncate the data reference vector. Do not bother to
22810 fixup data-dependence information for gather loads.
22811 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22812 of data references, as reported.
22814 2013-03-18 Richard Biener <rguenther@suse.de>
22816 PR tree-optimization/3713
22817 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22818 has_constants and expr.
22819 (stmt_has_constants): Properly valueize SSA names when deciding
22820 whether the stmt has constants.
22822 2013-03-18 Richard Biener <rguenther@suse.de>
22824 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22825 whole function when there is nothing to do.
22826 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22827 * tree-vectorizer.c (vectorize_loops): Update virtual and
22828 loop-closed SSA once.
22829 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22831 2013-03-18 Richard Biener <rguenther@suse.de>
22833 PR middle-end/56113
22834 * domwalk.c (bb_postorder): New global static.
22835 (cmp_bb_postorder): New function.
22836 (walk_dominator_tree): Replace scheme imposing an order for
22837 visiting dominator sons by one sorting them at the time they
22838 are pushed on the stack.
22840 2013-03-18 Richard Biener <rguenther@suse.de>
22842 PR tree-optimization/39326
22843 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22844 (struct mem_ref): Replace mem member with ao_ref typed member.
22845 (MEM_ANALYZABLE): Adjust.
22846 (memref_eq): Likewise.
22847 (mem_ref_alloc): Likewise.
22848 (gather_mem_refs_stmt): Likewise.
22849 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22850 (execute_sm_if_changed_flag_set): Adjust.
22851 (execute_sm): Likewise.
22852 (ref_always_accessed_p): Likewise.
22853 (refs_independent_p): Likewise.
22854 (can_sm_ref_p): Likewise.
22856 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22859 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22860 return 1 even for !unsignedp.
22862 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22864 * config/i386/i386.md (isa): Add x64 and nox64.
22865 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22866 (*pushtf): Enable *roF alternative for x64 isa only.
22867 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22868 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22869 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22870 (*movtf_internal): Merge from *movtf_internal_rex64 and
22871 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22872 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22873 nox64 isa attributes.
22874 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22875 nox64 isa attributes.
22876 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22878 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22880 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22882 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22885 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22886 class names. Remove trailing comma after "ALL_REGS".
22888 2013-03-16 Jan Hubicka <jh@suse.cz>
22890 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22891 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22892 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22893 of cgraph_get_create_node.
22894 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22896 2013-03-16 Jason Merrill <jason@redhat.com>
22899 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22900 with DW_AT_default_value.
22902 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22904 * BASE-VER: Set to 4.9.0.
22906 2013-03-14 Andi Kleen <ak@linux.intel.com>
22909 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22910 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22911 Document _x* TSX intrinsics.
22913 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22914 David Holsgrove <david.holsgrove@xilinx.com>
22916 * configure.ac: Add MicroBlaze TLS support detection.
22917 * configure: Regenerate.
22918 * config/microblaze/microblaze-protos.h
22919 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22920 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22921 * config/microblaze/microblaze.c (microblaze_address_type): Add
22922 ADDRESS_TLS and tls_reloc address types.
22923 (microblaze_address_info): Add tls_reloc.
22924 (TARGET_HAVE_TLS): Define.
22925 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22926 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22927 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22928 load_tls_operand, microblaze_call_tls_get_addr,
22929 microblaze_legitimize_tls_address): New functions.
22930 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22931 (get_base_reg): Use microblaze_tls_symbol_p.
22932 (microblaze_classify_address): Handle TLS.
22933 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22934 label_mentioned_p and microblaze_tls_referenced_p.
22935 (microblaze_legitimize_address): Handle TLS.
22936 (microblaze_address_insns): Handle ADDRESS_TLS.
22937 (pic_address_needs_scratch): Handle TLS.
22938 (print_operand_address): Handle TLS.
22939 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22940 (microblaze_expand_move): Handle TLS.
22941 (microblaze_legitimate_constant_p): Check
22942 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22943 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22944 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22945 (PIC_OFFSET_TABLE_REGNUM): Set.
22946 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22947 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22948 (addsi3, movsi_internal2, movdf_internal): Update constraints
22949 * config/microblaze/predicates.md (arith_plus_operand): Define
22950 (move_operand): Redefine as move_src_operand,
22951 check microblaze_tls_referenced_p.
22953 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22955 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22956 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22958 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22960 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22963 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22965 PR tree-optimization/53265
22966 * common.opt (Waggressive-loop-optimizations): New option.
22967 * tree-ssa-loop-niter.c: Include tree-pass.h.
22968 (do_warn_aggressive_loop_optimizations): New function.
22969 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22970 if number_of_latch_executions returned constant.
22971 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22972 early. If number_of_latch_executions returned constant, set
22973 nb_iterations_upper_bound back to it.
22974 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22976 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22977 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22979 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22980 (MULTILIB_OSDIRNAMES): Set.
22981 * genmultilib: If defaultosdirname doesn't start with :: , set
22982 defaultosdirname2 instead, clear it and emit two . multilib_raw
22983 entries instead of just one.
22985 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22987 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22988 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22989 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22990 (SUBTARGET_OVERRIDE_OPTIONS): New.
22992 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22995 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22996 (musermode): Convert to Var(TARGET_USERMODE).
22997 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22998 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22999 * config/sh/sh.c (sh_option_override): Use
23000 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23001 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23003 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23005 (udivsi3_i4_single, divsi3_i4_single): Use
23006 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23008 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23010 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23013 2013-03-13 Richard Biener <rguenther@suse.de>
23015 PR tree-optimization/56608
23016 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23017 calls when vectorizing basic-blocks.
23019 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23022 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23025 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23027 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23029 2013-03-11 Jan Hubicka <jh@suse.cz>
23032 * lto-streamer-out.c (output_symbol_p): Skip references from
23033 constructors of external variables.
23035 2013-03-11 Jan Hubicka <jh@suse.cz>
23037 PR middle-end/56571
23038 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23040 * emit-rtl.c (verify_rtx_sharing): Likewise.
23041 (copy_insn_1): Likewise.
23042 * rtl.c (copy_rtx): Likewise.
23044 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23047 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23048 output_operand_lossage message.
23050 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23053 * arm.c (shift_op): Validate RTL pattern on the fly.
23054 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23057 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23060 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23061 new call patterns that clobber %fr12.
23062 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23063 split and postreload patterns.
23064 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23065 registers %fr12 and %fr12R as call used.
23067 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23069 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23070 (canon_address, record_store, replace_read, check_mem_read_rtx,
23071 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23072 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23073 rest_of_handle_dse): Likewise.
23075 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23077 PR middle-end/56524
23078 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23080 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23081 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23082 (save_optabs_if_changed): Replace with...
23083 (init_tree_optimization_optabs): ...this.
23084 * optabs.c (save_optabs_if_changed): Rename to...
23085 (init_tree_optimization_optabs): ...this. Take the optimization node
23086 as argument. Do nothing if the base optabs are already correct.
23087 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23088 to recompute optabs.
23089 * function.h (function): Remove optabs field.
23090 * function.c (invoke_set_current_function_hook): Call
23091 init_tree_optimization_optabs. Use the result to initialize
23094 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23096 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23097 if GTMA_HAS_NO_INSTRUMENTATION.
23098 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23099 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23100 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23101 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23102 GTMA_HAS_NO_INSTRUMENTATION.
23104 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23106 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23109 2013-03-08 Marek Polacek <polacek@redhat.com>
23110 Jakub Jelinek <jakub@redhat.com>
23112 PR tree-optimization/56478
23113 * predict.c (is_comparison_with_loop_invariant_p): Change the
23114 type of loop_step to tree.
23115 (predict_loops): Adjust.
23116 (predict_iv_comparison): Perform the computations on double_ints.
23118 2013-03-08 Richard Biener <rguenther@suse.de>
23120 PR tree-optimization/56570
23121 * tree-cfg.c (verify_expr_location_1): Verify locations for
23123 * tree-sra.c (create_access_replacement): Strip locations
23124 from DECL_DEBUG_EXPRs.
23126 2013-03-08 Richard Biener <rguenther@suse.de>
23128 * tree-inline.c (expand_call_inline): Do not associate
23129 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23130 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23132 2013-03-08 Richard Biener <rguenther@suse.de>
23134 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23135 or block changes with -Og. Fix for location / block encoding
23136 changes and PHI arguments with locations.
23138 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23140 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23142 (struct output_info): Likewise.
23143 (register_overhead): Remove bad gcc_assert.
23144 (bitmap_find_bit): If there is only a single bitmap element, do not
23145 count a miss as a search.
23146 (print_statistics): Update for counter type changes.
23147 (dump_bitmap_statistics): Likewise. Print headers such that they
23148 are properly lined up with the printed counters.
23150 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23152 PR tree-optimization/56559
23153 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23154 check that it has only a single use.
23156 2013-03-07 Richard Biener <rguenther@suse.de>
23158 * doc/invoke.texi (fwhole-program): Discourage use in combination
23161 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23163 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23165 PR tree-optimization/56539
23166 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23167 instead of GSI_CONTINUE_LINKING as last argument to
23168 force_gimple_operand_gsi. Adjust function comment.
23170 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23173 PR middle-end/56548
23174 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23175 promoted mode, convert the result back to the original mode.
23177 2013-03-06 Richard Biener <rguenther@suse.de>
23179 PR middle-end/56294
23180 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23181 (insert_updated_phi_nodes_compare_uids): New function.
23182 (update_ssa): Sort symbols_to_rename after UID before
23183 traversing it to insert PHI nodes.
23185 2013-03-06 Richard Biener <rguenther@suse.de>
23187 PR middle-end/50494
23188 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23189 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23192 2013-02-13 Richard Biener <rguenther@suse.de>
23195 * varasm.c (output_constant_def_1): Get the decl representing
23196 the constant as argument.
23197 (output_constant_def): Wrap output_constant_def_1.
23198 (make_decl_rtl): Use output_constant_def_1 with the decl
23199 representing the constant.
23200 (build_constant_desc): Optionally re-use a decl already
23201 representing the constant.
23202 (tree_output_constant_def): Adjust.
23204 2013-03-06 Joey Ye <joey.ye@arm.com>
23207 * gcc.c (convert_white_space): New function.
23208 (main): Handles white space in function name.
23210 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23213 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23214 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23215 to SH_DIV_CALL_TABLE for TARGET_SH2.
23216 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23218 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23219 call-table options.
23221 2013-03-05 Sterling Augustine <saugustine@google.com>
23222 Cary Coutant <ccoutant@google.com>
23225 * dwarf2out.c (resolve_addr): Don't call
23226 remove_loc_list_addr_table_entries a second time for the same
23229 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23232 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23233 (avoid_complex_debug_insns): New function.
23234 (expand_debug_locations): Call it.
23236 PR rtl-optimization/56484
23237 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23238 lifetimes of hard registers on small register class machines.
23240 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23242 * config/microblaze/microblaze-protos.h: Rename
23243 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23244 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23246 (microblaze_fast_interrupt_function_p): New function.
23247 (microblaze_is_interrupt_handler): Rename to
23248 microblaze_is_interrupt_variant and add fast_interrupt check.
23249 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23250 (save_restore_insns): Likewise.
23251 (compute_frame_size): Likewise.
23252 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23253 (microblaze_globalize_label): Likewise.
23254 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23255 * config/microblaze/microblaze.md: Use wrapper
23256 microblaze_is_interrupt_variant.
23258 2013-03-05 Kai Tietz <ktietz@redhat.com>
23260 * sdbout.c (sdbout_one_type): Switch to current function's section
23261 supporting cold/hot.
23263 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23265 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23268 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23270 PR middle-end/56461
23271 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23272 if VALGRIND_GET_VBITS is defined, temporarily make object
23273 memory all defined, and restore previous valgrind addressability
23274 and definability afterwards. Free this_object at the end.
23276 PR middle-end/56461
23277 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23278 right before calling lra_create_live_ranges, also call it
23279 when clearing live_p. Only call lra_clear_live_ranges
23280 at the end if live_p.
23282 PR middle-end/56461
23283 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23285 2013-03-05 Richard Biener <rguenther@suse.de>
23287 PR tree-optimization/56521
23288 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23291 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23294 * except.h (remove_unreachable_eh_regions): New prototype.
23295 * except.c (remove_eh_handler_splicer): New function, split out
23296 of remove_eh_handler.
23297 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23298 warning about running it on many EH regions one at a time.
23299 (remove_unreachable_eh_regions_worker): New function, walk the
23300 EH tree in depth-first order and remove non-marked regions.
23301 (remove_unreachable_eh_regions): New function.
23302 * tree-eh.c (mark_reachable_handlers): New function, split out
23303 from remove_unreachable_handlers.
23304 (remove_unreachable_handlers): Use mark_reachable_handlers and
23305 remove_unreachable_eh_regions.
23306 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23307 and remove_unreachable_eh_regions.
23309 2013-03-05 Richard Biener <rguenther@suse.de>
23311 PR middle-end/56525
23312 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23313 not freeing them until the end.
23315 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23317 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23319 2013-03-05 Richard Biener <rguenther@suse.de>
23321 PR tree-optimization/56270
23322 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23323 of loads after scheduling an SLP instance.
23325 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23327 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23329 (check_gcc_parallelize): Run guality.exp as a separate job from
23330 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23331 struct-layout-1.exp with stackalign.exp.
23333 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23335 PR middle-end/56461
23336 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23337 load_index sbitmap even if some bit in it isn't set.
23339 PR middle-end/56461
23340 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23341 (discover_iteration_bound_by_body_walk): Change queues to
23342 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23343 spelling in comment. Call safe_push on queues[bound_index] directly.
23344 Release queues[queue_index] in every iteration unconditionally.
23345 Release bounds vector.
23347 PR middle-end/56461
23348 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23349 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23350 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23353 2013-03-05 Richard Biener <rguenther@suse.de>
23356 * tree-inline.c (remap_blocks_to_null): New function.
23357 (expand_call_inline): When expanding a call stmt without
23358 an associated block inline remap all callee blocks to NULL.
23360 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23362 PR rtl-optimization/56494
23363 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23364 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23365 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23367 PR middle-end/56461
23368 * sel-sched-ir.c (free_sched_pools): Release
23369 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23370 if succs_info_pool.max_top isn't -1.
23373 * opts.c (opts_obstack, opts_concat): Moved to...
23374 * opts-common.c (opts_obstack, opts_concat): ... here.
23376 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23378 PR middle-end/56461
23379 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23381 2013-03-04 Martin Jambor <mjambor@suse.cz>
23383 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23384 all appropriate places.
23386 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23388 PR tree-optimization/56424
23389 * ipa-split.c (split_function): Do not set the RSO flag if result is
23390 not by reference and its type is a register type.
23392 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23394 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23395 (microblaze_legitimate_pic_operand): Likewise
23396 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23397 new function microblaze_legitimate_pic_operand
23398 * config/microblaze/microblaze-protos.h
23399 (microblaze_legitimate_pic_operand): Declare.
23401 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23403 * config/microblaze/predicates.md (call_insn_simple_operand):
23404 New predicate for supported rtx code types.
23405 * config/microblaze/microblaze.md (call_internal1): Use
23406 call_insn_simple_operand predicate.
23408 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23410 PR middle-end/56461
23411 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23412 partitions.ordered_remove.
23414 PR middle-end/56461
23415 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23416 vec_oprnds0.create (1) for modifier == NONE.
23418 PR middle-end/56461
23419 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23420 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23421 vec_oprnds1 right before pushing anything to it for
23424 PR middle-end/56461
23425 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23426 set nbbs to 0 instead of having separate code path.
23427 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23428 instead of false as last argument if returning NULL.
23430 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23432 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23433 the attribute is now called "target" instead of "option".
23434 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23435 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23436 attribute/pragma name for TARGET_OPTION_VALID_P and
23437 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23438 * doc/tm.texi: Regenerated.
23440 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23442 * config/microblaze/microblaze.c:
23443 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23444 * config/microblaze/microblaze.h: Add -mxl-reorder to
23446 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23447 instructions emitted if TARGET_REORDER.
23448 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23449 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23452 2013-03-01 Xinliang David Li <davidxl@google.com>
23454 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23457 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23459 PR middle-end/56461
23460 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23461 vector even when returning true. Fix up function comment formatting.
23463 PR middle-end/56461
23464 * ira-build.c (ira_loop_nodes_count): New variable.
23465 (create_loop_tree_nodes): Initialize it.
23466 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23468 PR middle-end/56461
23469 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23470 method on dr_chain and result_chain.
23471 * tree-vect-stmts.c (vectorizable_store): Only call
23472 result_chain.create if j == 0.
23474 PR middle-end/56461
23475 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23476 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23477 before overwriting it.
23479 2013-03-01 Tobias Burnus <burnus@net-b.de>
23481 * doc/extended.texi (C Extensions): Change order in @menu
23483 (Other MIPS Built-in Functions): Move last MIPS entry before
23484 "picoChip Built-in Functions".
23485 (SH Built-in Functions): Move after RX Built-in Functions.
23486 * doc/gcc.texi (Introduction): Change order in @menu
23488 * doc/md.texi (Constraints): Ditto.
23489 * gty.texi (Type Information): Ditto.
23490 (User-provided marking routines for template types): Make
23492 * doc/invoke.texi (AArch64 Options): Move before
23493 "Adapteva Epiphany Options".
23495 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23496 Jakub Jelinek <jakub@redhat.com>
23499 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23500 no_address_safety_analysis attribute.
23501 * doc/extend.texi (no_address_safety_attribute): Rename to
23502 no_sanitize_address attribute, mention no_address_safety_analysis
23503 attribute as deprecated alias.
23505 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23507 PR middle-end/56461
23508 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23509 type to vec<vec<tree> > *.
23510 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23511 to be vec<tree> instead of vec<tree> *, set vec_defs
23512 to vNULL and call vec_defs.create (number_of_vects), adjust other
23514 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23515 vectorizable_condition): Adjust vect_get_slp_defs callers.
23517 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23519 * config/aarch64/aarch64.c
23520 (aarch64_float_const_representable): Remove unused variable.
23522 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23524 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23526 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23528 * config/aarch64/aarch64-builtins.c
23529 (aarch64_init_simd_builtins): Make static.
23531 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23533 * config/aarch64/aarch64.c
23534 (aarch64_simd_make_constant): Make static.
23536 2013-02-28 Martin Jambor <mjambor@suse.cz>
23538 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23539 with no initialization to the RHS of debug statements.
23541 2013-02-28 Martin Jambor <mjambor@suse.cz>
23543 PR tree-optimization/56294
23544 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23546 (get_access_replacement): Do not call create_access_replacement.
23547 Assert a replacement exists.
23548 (get_repl_default_def_ssa_name): Create the replacement declaration
23551 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23553 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23554 final_end_function.
23556 2013-02-28 Marek Polacek <polacek@redhat.com>
23558 PR rtl-optimization/56466
23559 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23560 if we're changing a loop.
23561 (peel_loops_completely): Likewise.
23563 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23566 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23568 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23571 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23572 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23573 INTX_FTYPE_FX, FX_FTYPE_INTX.
23574 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23576 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23578 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23579 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23580 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23581 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23582 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23583 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23584 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23585 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23586 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23587 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23588 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23589 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23590 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23591 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23592 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23593 (avrxmega6): Increase max flash segments from 5 to 6.
23594 * config/avr/t-multilib: Regenerate.
23595 * config/avr/avr-tables.opt: Regenerate.
23596 * doc/avr-mmcu.texi: Regenerate.
23598 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23600 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23601 (avr_device_to_arch): Rename to avr_device_to_ld.
23602 (avr_device_to_as): New prototype.
23603 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23604 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23605 * config/avr/driver-avr.c (avr_device_to_as): New.
23606 (avr_device_to_arch): Rename to avr_device_to_ld.
23608 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23610 PR middle-end/56461
23611 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23612 method on dr_chain and result_chain.
23614 PR middle-end/56461
23615 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23616 pointer_set_destroy on not_executed_last_iteration.
23618 PR middle-end/56461
23619 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23621 PR middle-end/56461
23622 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23623 FOR_EACH_DEFINED_FUNCTION when freeing state.
23625 PR middle-end/56461
23626 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23628 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23631 PR middle-end/56461
23632 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23633 known_aggs[i].items and release known_aggs vector.
23635 PR middle-end/56461
23636 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23638 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23640 * config/microblaze/microblaze.c (microblaze_emit_compare):
23641 Use xor for EQ/NE comparisions.
23642 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23643 (cbranchsf4): Adjust operator to comparison_operator.
23645 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23647 PR middle-end/56461
23648 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23650 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23651 vec_safe_push, always update *slot.
23652 (redirect_edge_var_map_clear): Use vec_free.
23653 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23654 (free_var_map_entry): Use vec_free.
23655 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23656 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23658 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23660 PR middle-end/45472
23661 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23662 when the may_trap_p bit of the exprs being merged differs.
23663 Reorder tests for speculativeness in the logical and operator.
23665 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23667 * incpath.c (add_standard_paths): Use reconcat instead of concat
23668 where appropriate and avoid leaking memory.
23670 * opts.h: Include obstack.h.
23671 (opts_concat): New prototype.
23672 (opts_obstack): New declaration.
23673 * opts.c (opts_concat): New function.
23674 (opts_obstack): New variable.
23675 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23676 (finish_options): Use opts_concat instead of concat
23677 and XOBNEWVEC instead of XNEWVEC.
23678 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23679 generate_option): Likewise.
23680 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23681 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23684 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23685 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23687 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23689 PR middle-end/56461
23690 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23692 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23694 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23695 (arm_block_move_unaligned_straight): Likewise.
23696 (arm_adjust_block_mem): Likewise.
23698 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23701 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23702 temp, cond and label.
23703 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23706 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23707 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23710 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23711 of prologue/epilogue functions.
23714 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23715 Remove unused variable cfa_offset.
23716 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23719 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23720 type promotion to unsigned.
23723 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23724 for HOST_WIDE_INT of 32 bit / same size as int.
23725 (arm_block_move_unaligned_straight): Likewise.
23726 (arm_adjust_block_mem): Likewise.
23729 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23732 2013-02-26 Marek Polacek <polacek@redhat.com>
23734 PR tree-optimization/56426
23735 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23737 2013-02-26 Richard Biener <rguenther@suse.de>
23740 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23741 unused variable loops.
23743 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23745 PR tree-optimization/56448
23746 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23747 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23748 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23749 later operands of the references, or even first operand for
23750 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23752 PR tree-optimization/56443
23753 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23754 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23755 to type_for_mode langhook.
23757 2013-02-25 Matt Turner <mattst88@gmail.com>
23759 * doc/invoke.texi: Document r4700.
23761 2013-02-25 Richard Biener <rguenther@suse.de>
23763 PR tree-optimization/56175
23764 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23766 (simplify_bitwise_binary): ... here. Also guard the conversion
23767 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23769 2013-02-25 Catherine Moore <clm@codesourcery.com>
23772 2013-02-24 Catherine Moore <clm@codesourcery.com>
23773 Maciej W. Rozycki <macro@codesourcery.com>
23774 Tom de Vries <tom@codesourcery.com>
23775 Nathan Sidwell <nathan@codesourcery.com>
23776 Iain Sandoe <iain@codesourcery.com>
23777 Nathan Froyd <froydnj@codesourcery.com>
23778 Chao-ying Fu <fu@mips.com>
23780 * doc/extend.texi: (micromips, nomicromips, nocompression):
23781 Document new function attributes.
23782 * doc/invoke.texi (minterlink-compressed, mmicromips,
23783 m14k, m14ke, m14kec): Document new options.
23784 (minterlink-mips16): Update documentation.
23785 * doc/md.texi (ZC, ZD): Document new constraints.
23786 * configure.ac (gcc_cv_as_micromips): Check if linker
23787 supports the .set micromips directive.
23788 * configure: Regenerate.
23789 * config.in: Regenerate.
23790 * config/mips/mips-tables.opt: Regenerate.
23791 * config/mips/micromips.md: New file.
23792 * constraints.md (ZC, AD): New constraints.
23793 * config/mips/predicates.md (movep_src_register): New predicate.
23794 (movep_src_operand): New predicate.
23795 (non_volatile_mem_operand): New predicate.
23796 * config/mips/mips.md (multimem): New type.
23797 (length): Differentiate between 17-bit and 18-bit branch offsets.
23798 (MOVEP1, MOVEP2): New mode iterator.
23799 (mov_<load>l): Use ZC constraint.
23800 (mov_<load>r): Likewise.
23801 (mov_<store>l): Likewise.
23802 (mov_<store>r): Likewise.
23803 (*branch_equality<mode>_inverted): Add microMIPS support.
23804 (*branch_equality<mode>): Likewise.
23805 (*jump_absolute): Likewise.
23806 (indirect_jump_<mode>): Likewise.
23807 (tablejump_<mode>): Likewise.
23808 (<optab>_internal): Likewise.
23809 (sibcall_internal): Likewise.
23810 (sibcall_value_internal): Likewise.
23811 (prefetch): Use constraint ZD.
23812 * config/mips/mips.opt (minterlink-compressed): New option.
23813 (minterlink-mips16): Now an alias for minterlink-compressed.
23814 (mmicromips): New option.
23815 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23816 (compare_and_swap_12): Likewise.
23817 (sync_add<mode>): Likewise.
23818 (sync_<optab>_12): Likewise.
23819 (sync_old_<optab>_12): Likewise.
23820 (sync_new_<optab>_12): Likewise.
23821 (sync_nand_12): Likewise.
23822 (sync_old_nand_12): Likewise.
23823 (sync_new_nand_12): Likewise.
23824 (sync_sub<mode>): Likewise.
23825 (sync_old_add<mode>): Likewise.
23826 (sync_old_sub<mode>): Likewise.
23827 (sync_new_add<mode>): Likewise.
23828 (sync_new_sub<mode>): Likewise.
23829 (sync_<optab><mode>): Likewise.
23830 (sync_old_<optab><mode>): Likewise.
23831 (sync_new_<optab><mode>): Likewise.
23832 (sync_nand<mode>): Likewise.
23833 (sync_old_nand<mode>): Likewise.
23834 (sync_new_nand<mode>): Likewise.
23835 (sync_lock_test_and_set<mode>): Likewise.
23836 (test_and_set_12): Likewise.
23837 (atomic_compare_and_swap<mode>): Likewise.
23838 (atomic_exchange<mode>_llsc): Likewise.
23839 (atomic_fetch_add<mode>_llsc): Likewise.
23840 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23841 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23842 (umips_save_restore_pattern_p): Likewise.
23843 (umips_load_store_pair_p): Likewise.
23844 (umips_output_load_store_pair): Likewise.
23845 (umips_movep_target_p): Likewise.
23846 (umips_12bit_offset_address_p): Likewise.
23847 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23848 (mips_base_mips16): Rename this...
23849 (mips_base_compression_flags): ...to this. Update all uses.
23850 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23851 (mips_mips16_decl_p): Delete.
23852 (mips_nomips16_decl_p): Delete.
23853 (mips_get_compress_on_flags): New function.
23854 (mips_get_compress_off_flags): New function.
23855 (mips_get_compress_mode): New function.
23856 (mips_get_compress_on_name): New function.
23857 (mips_get_compress_off_name): New function.
23858 (mips_insert_attributes): Support multiple compression types.
23859 (mips_merge_decl_attributes): Likewise.
23860 (umips_12bit_offset_address_p): New function.
23861 (mips_start_function_definition): Emit .set micromips directive.
23862 (mips_call_may_need_jalx_p): New function.
23863 (mips_function_ok_for_sibcall): Add microMIPS support.
23864 (mips_print_operand_punctuation): Support short delay slots and
23866 (umips_swm_mask, umips_swm_encoding): New.
23867 (umips_build_save_restore): New function.
23868 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23869 (was_mips16_p): Remove.
23870 (old_compression_mode): New.
23871 (mips_set_compression_mode): New function.
23872 (mips_set_current_function): Add microMIPS support.
23873 (mips_option_override): Likewise.
23874 (umips_save_restore_pattern_p): New function.
23875 (umips_output_save_restore): New function.
23876 (umips_load_store_pair_p_1): New function.
23877 (umips_load_store_pair_p): New function.
23878 (umips_output_load_store_pair_1): New function.
23879 (umips_output_load_store_pair): New function.
23880 (umips_movep_target_p) New function.
23881 (mips_prepare_pch_save): Add microMIPS support.
23882 * config/mips/mips.h (TARGET_COMPRESSION): New.
23883 (TARGET_CPU_CPP_BUILTINS): Update macro
23884 to use new compression flags and to support microMIPS.
23885 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23886 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23887 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23888 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23889 (ASM_SPEC): Support mmicromips and mno-micromips.
23890 (M16STORE_REG_P): New macro.
23891 (MIPS_CALL): Support TARGET_MICROMIPS.
23892 (MICROMIPS_J): New macro.
23893 (mips_base_mips16): Rename this...
23894 (mips_base_compression_flags): ...to this.
23895 (UMIPS_12BIT_OFFSET_P): New macro.
23896 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23897 (MULTILIB_DIRNAMES): Likewise.
23899 2013-02-25 Tom de Vries <tom@codesourcery.com>
23901 PR rtl-optimization/56131
23902 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23903 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23904 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23906 2013-02-25 Tobias Burnus <burnus@net-b.de>
23908 * doc/invoke.texi (-fsanitize=): Move from optimization
23909 to debugging options.
23911 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23913 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23915 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23916 Alexander Monakov <amonakov@ispras.ru>
23918 PR middle-end/56077
23919 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23920 flush pending lists also on non-jumps. Adjust comment.
23922 2013-02-24 Catherine Moore <clm@codesourcery.com>
23923 Maciej W. Rozycki <macro@codesourcery.com>
23924 Tom de Vries <tom@codesourcery.com>
23925 Nathan Sidwell <nathan@codesourcery.com>
23926 Iain Sandoe <iain@codesourcery.com>
23927 Nathan Froyd <froydnj@codesourcery.com>
23928 Chao-ying Fu <fu@mips.com>
23930 * doc/extend.texi: (micromips, nomicromips, nocompression):
23931 Document new function attributes.
23932 * doc/invoke.texi (minterlink-compressed, mmicromips,
23933 m14k, m14ke, m14kec): Document new options.
23934 (minterlink-mips16): Update documentation.
23935 * doc/md.texi (ZC, ZD): Document new constraints.
23936 * configure.ac (gcc_cv_as_micromips): Check if linker
23937 supports the .set micromips directive.
23938 * configure: Regenerate.
23939 * config.in: Regenerate.
23940 * config/mips/mips-tables.opt: Regenerate.
23941 * config/mips/micromips.md: New file.
23942 * constraints.md (ZC, AD): New constraints.
23943 * config/mips/predicates.md (movep_src_register): New predicate.
23944 (movep_src_operand): New predicate.
23945 (non_volatile_mem_operand): New predicate.
23946 * config/mips/mips.md (multimem): New type.
23947 (length): Differentiate between 17-bit and 18-bit branch offsets.
23948 (MOVEP1, MOVEP2): New mode iterator.
23949 (mov_<load>l): Use ZC constraint.
23950 (mov_<load>r): Likewise.
23951 (mov_<store>l): Likewise.
23952 (mov_<store>r): Likewise.
23953 (*branch_equality<mode>_inverted): Add microMIPS support.
23954 (*branch_equality<mode>): Likewise.
23955 (*jump_absolute): Likewise.
23956 (indirect_jump_<mode>): Likewise.
23957 (tablejump_<mode>): Likewise.
23958 (<optab>_internal): Likewise.
23959 (sibcall_internal): Likewise.
23960 (sibcall_value_internal): Likewise.
23961 (prefetch): Use constraint ZD.
23962 * config/mips/mips.opt (minterlink-compressed): New option.
23963 (minterlink-mips16): Now an alias for minterlink-compressed.
23964 (mmicromips): New option.
23965 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23966 (compare_and_swap_12): Likewise.
23967 (sync_add<mode>): Likewise.
23968 (sync_<optab>_12): Likewise.
23969 (sync_old_<optab>_12): Likewise.
23970 (sync_new_<optab>_12): Likewise.
23971 (sync_nand_12): Likewise.
23972 (sync_old_nand_12): Likewise.
23973 (sync_new_nand_12): Likewise.
23974 (sync_sub<mode>): Likewise.
23975 (sync_old_add<mode>): Likewise.
23976 (sync_old_sub<mode>): Likewise.
23977 (sync_new_add<mode>): Likewise.
23978 (sync_new_sub<mode>): Likewise.
23979 (sync_<optab><mode>): Likewise.
23980 (sync_old_<optab><mode>): Likewise.
23981 (sync_new_<optab><mode>): Likewise.
23982 (sync_nand<mode>): Likewise.
23983 (sync_old_nand<mode>): Likewise.
23984 (sync_new_nand<mode>): Likewise.
23985 (sync_lock_test_and_set<mode>): Likewise.
23986 (test_and_set_12): Likewise.
23987 (atomic_compare_and_swap<mode>): Likewise.
23988 (atomic_exchange<mode>_llsc): Likewise.
23989 (atomic_fetch_add<mode>_llsc): Likewise.
23990 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23991 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23992 (umips_save_restore_pattern_p): Likewise.
23993 (umips_load_store_pair_p): Likewise.
23994 (umips_output_load_store_pair): Likewise.
23995 (umips_movep_target_p): Likewise.
23996 (umips_12bit_offset_address_p): Likewise.
23997 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23998 (mips_base_mips16): Rename this...
23999 (mips_base_compression_flags): ...to this. Update all uses.
24000 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24001 (mips_mips16_decl_p): Delete.
24002 (mips_nomips16_decl_p): Delete.
24003 (mips_get_compress_on_flags): New function.
24004 (mips_get_compress_off_flags): New function.
24005 (mips_get_compress_mode): New function.
24006 (mips_get_compress_on_name): New function.
24007 (mips_get_compress_off_name): New function.
24008 (mips_insert_attributes): Support multiple compression types.
24009 (mips_merge_decl_attributes): Likewise.
24010 (umips_12bit_offset_address_p): New function.
24011 (mips_start_function_definition): Emit .set micromips directive.
24012 (mips_call_may_need_jalx_p): New function.
24013 (mips_function_ok_for_sibcall): Add microMIPS support.
24014 (mips_print_operand_punctuation): Support short delay slots and
24016 (umips_swm_mask, umips_swm_encoding): New.
24017 (umips_build_save_restore): New function.
24018 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24019 (was_mips16_p): Remove.
24020 (old_compression_mode): New.
24021 (mips_set_compression_mode): New function.
24022 (mips_set_current_function): Add microMIPS support.
24023 (mips_option_override): Likewise.
24024 (umips_save_restore_pattern_p): New function.
24025 (umips_output_save_restore): New function.
24026 (umips_load_store_pair_p_1): New function.
24027 (umips_load_store_pair_p): New function.
24028 (umips_output_load_store_pair_1): New function.
24029 (umips_output_load_store_pair): New function.
24030 (umips_movep_target_p) New function.
24031 (mips_prepare_pch_save): Add microMIPS support.
24032 * config/mips/mips.h (TARGET_COMPRESSION): New.
24033 (TARGET_CPU_CPP_BUILTINS): Update macro
24034 to use new compression flags and to support microMIPS.
24035 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24036 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24037 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24038 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24039 (ASM_SPEC): Support mmicromips and mno-micromips.
24040 (M16STORE_REG_P): New macro.
24041 (MIPS_CALL): Support TARGET_MICROMIPS.
24042 (MICROMIPS_J): New macro.
24043 (mips_base_mips16): Rename this...
24044 (mips_base_compression_flags): ...to this.
24045 (UMIPS_12BIT_OFFSET_P): New macro.
24046 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24047 (MULTILIB_DIRNAMES): Likewise.
24049 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24052 * target-globals.c (save_target_globals): For init_reg_sets and
24053 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24055 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24057 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24058 * config/aarch64/t-aarch64
24059 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24061 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24063 PR inline-asm/56148
24064 * lra-constraints.c (process_alt_operands): Reload operand
24065 conflicting with earlier clobber only if no more other conflicting
24068 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24071 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24072 if not linking a shared library.
24074 2013-02-22 Seth LaForge <sethml@google.com>
24076 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24078 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24080 * config/arm/arm.md (split for extendsidi): Update condition.
24081 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24082 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24083 (qhs_zextenddi_cstr): Likewise.
24085 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24087 PR middle-end/56420
24088 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24089 avoid signed wrapping.
24090 (expand_mult): Handle properly multiplication by
24091 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24092 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24093 in the compiler if coeff is HOST_WIDE_INT_MIN.
24094 (expand_divmod): Don't make ext_op1 static, change it's type to
24095 uhwi. Avoid undefined behavior in -INTVAL (op1).
24097 PR rtl-optimization/50339
24098 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24100 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24101 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24102 into splitting_ashiftrt field.
24103 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24105 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24108 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24110 PR middle-end/56108
24111 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24112 are sure to go irrevocable.
24114 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24116 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24117 scalars are valid operands.
24119 2013-02-21 Martin Jambor <mjambor@suse.cz>
24121 PR tree-optimization/56310
24122 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24123 only matching indices and non-negative final offsets.
24124 (intersect_aggregates_with_edge): Pass src_idx to
24125 agg_replacements_to_vector. Pass src_idx insstead of index to
24126 intersect_with_agg_replacements.
24128 2013-02-21 Martin Jambor <mjambor@suse.cz>
24130 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24131 instead of hard-wired defaults.
24133 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24135 * doc/invoke.texi (MIPS Options): Update documentation of the
24136 floating-point multiply-accumulate instruction restrictions.
24138 2013-02-21 Kostya Serebryany <kcc@google.com>
24140 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24141 asan_shadow_offset on x86_64 linux.
24143 2013-02-21 Richard Biener <rguenther@suse.de>
24145 PR tree-optimization/56415
24147 2013-02-11 Richard Biener <rguenther@suse.de>
24149 PR tree-optimization/56273
24150 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24153 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24156 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24159 PR inline-asm/56405
24160 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24161 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24163 2013-02-20 Jan Hubicka <jh@suse.cz>
24165 PR tree-optimization/56265
24166 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24167 when target is referenced for first time.
24169 2013-02-20 Richard Biener <rguenther@suse.de>
24171 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24172 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24173 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24174 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24175 not return anything.
24176 (rename_ssa_copies): Do not remove unused locals.
24177 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24178 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24179 * passes.c (execute_function_todo): Do not schedule unused locals
24180 removal if cleanup_tree_cfg did something.
24181 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24182 about the number of removed locals.
24184 2013-02-20 Richard Biener <rguenther@suse.de>
24186 PR tree-optimization/56398
24187 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24189 2013-02-20 Martin Jambor <mjambor@suse.cz>
24191 PR tree-optimization/55334
24192 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24193 restricted pointers to arrays.
24195 2013-02-20 Richard Biener <rguenther@suse.de>
24196 Jakub Jelinek <jakub@redhat.com>
24198 PR tree-optimization/56396
24199 * tree-ssa-ccp.c (n_const_val): New static variable.
24200 (get_value): Return NULL for SSA names we don't have a lattice
24202 (ccp_initialize): Initialize n_const_val.
24203 * tree-ssa-copy.c (n_copy_of): New static variable.
24204 (init_copy_prop): Initialize n_copy_of.
24205 (get_value): Return NULL_TREE for SSA names we don't have a
24208 2013-02-20 Martin Jambor <mjambor@suse.cz>
24210 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24212 2013-02-20 Richard Biener <rguenther@suse.de>
24214 * genpreds.c (write_lookup_constraint): Do not compare first
24215 letter of the constraint again.
24217 2013-02-20 Richard Biener <rguenther@suse.de>
24219 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24221 (get_use_iv_cost): Terminate hashtable walk when coming across
24224 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24226 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24227 reassociation for avx2 targets.
24229 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24231 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24232 Add version check for v8.10.a to enable microblaze_has_clz
24233 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24234 version and TARGET_PATTERN_COMPARE check
24235 * config/microblaze/microblaze.md: New clzsi2 instruction
24237 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24239 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24240 function before branching.
24242 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24244 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24246 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24248 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24250 PR middle-end/55889
24251 * sel-sched.c: Include ira.h.
24252 (implicit_clobber_conflict_p): New function.
24253 (moveup_expr): Use it.
24254 * Makefile.in (sel-sched.o): Depend on ira.h.
24256 2013-02-19 Richard Biener <rguenther@suse.de>
24258 PR tree-optimization/56384
24259 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24260 (vn_hash_type): Split out from ...
24261 (vn_hash_constant_with_type): ... here.
24262 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24263 (vn_phi_eq): Compare types from vn_phi_s structure.
24264 (vn_phi_lookup): Populate vn_phi_s type.
24265 (vn_phi_insert): Likewise.
24267 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24269 PR tree-optimization/56350
24270 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24271 if haven't found reduction or nested cycle operand, rather than
24272 asserting we must find it.
24274 PR tree-optimization/56381
24275 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24278 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24279 Jakub Jelinek <jakub@redhat.com>
24282 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24283 (swap_optab_enable): Same.
24284 (init_all_optabs): Use argument instead of global.
24285 * tree.h (struct tree_optimization_option): New field target_optabs.
24286 * expr.h (init_all_optabs): Add argument to prototype.
24287 (TREE_OPTIMIZATION_OPTABS): New.
24288 (save_optabs_if_changed): Protoize.
24289 * optabs.h: Declare this_fn_optabs.
24290 * optabs.c (save_optabs_if_changed): New.
24291 Declare this_fn_optabs.
24292 (init_optabs): Add argument to init_all_optabs() call.
24293 * function.c (invoke_set_current_function_hook): Handle per
24295 * function.h (struct function): New field optabs.
24296 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24297 optimization_current_node has changed.
24298 * target-globals.h (save_target_globals_default_opts): Protoize.
24299 * target-globals.c (save_target_globals_default_opts): New.
24301 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24304 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24305 registers %fr12 and %fr12R as call used.
24308 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24309 and HImode, require all displacements to be an integer multiple of
24311 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24312 only allow QImode and HImode when reload is in progress and strict is
24313 true. Likewise for symbolic addresses. Use base14_operand to check
24314 displacements in REG+BASE addresses.
24316 2013-02-18 Richard Biener <rguenther@suse.de>
24318 PR tree-optimization/56366
24319 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24320 handle sign-conversion of outer-loop initial induction value.
24322 2013-02-18 Richard Biener <rguenther@suse.de>
24324 PR middle-end/56349
24325 * cfghooks.c (merge_blocks): If we merge a latch into another
24326 block adjust references to it.
24327 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24328 (verify_loop_structure): Verify that a recorded latch is in fact
24331 2013-02-18 Richard Biener <rguenther@suse.de>
24333 PR tree-optimization/56321
24334 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24335 order SSA name release and virtual operand unlinking.
24337 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24339 * config/microblaze/microblaze.md (save_stack_block): Define.
24340 (restore_stack_block): Likewise.
24342 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24344 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24345 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24346 * config/microblaze/microblaze.c (microblaze_option_override):
24347 Bail out early for PIC modes when target does not support PIC.
24349 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24351 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24352 Replace with a microblaze version.
24353 (microblaze_trampoline_init): Adapt for microblaze.
24354 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24357 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24358 Dodji Seketeli <dodji@redhat.com>
24361 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24362 (instrument_mem_region_access): Do not forget to always put
24363 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24364 0) statement, even for cases where either 'base' or 'base + len'
24365 are not instrumented -- because they have been previously
24366 instrumented. Simplify the logic by putting all the statements
24367 instrument 'base + len' inside a sequence, and then insert that
24368 sequence right before the current insertion point. Then, to
24369 instrument 'base + len', just get an iterator on that statement.
24370 And do not forget to update the pointer to iterator the function
24371 received as argument.
24373 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24375 PR rtl-optimization/56348
24376 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24378 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24380 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24381 (clean_graph_dump_file): Pass base to start_graph_dump.
24383 2013-02-14 Richard Henderson <rth@redhat.com>
24386 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24388 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24390 * collect2-aix.h: Define F_LOADONLY.
24392 2013-02-14 Richard Biener <rguenther@suse.de>
24395 * varasm.c (output_constant_def_1): Get the decl representing
24396 the constant as argument.
24397 (output_constant_def): Wrap output_constant_def_1.
24398 (make_decl_rtl): Use output_constant_def_1 with the decl
24399 representing the constant.
24400 (build_constant_desc): Optionally re-use a decl already
24401 representing the constant.
24402 (tree_output_constant_def): Adjust.
24404 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24407 * asan.c (instrument_builtin_call): Really put the length of the
24408 second source argument into src1_len.
24410 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24412 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24413 argument. If it is false, don't create edge from then_bb to
24415 (insert_if_then_before_iter): Pass true to it.
24416 (build_check_stmt): Pass false to it.
24417 (transform_statements): Flush hash table only on extended basic
24418 block boundaries, rather than at the beginning of every bb.
24419 Don't flush hash table on nonfreeing_call_p calls.
24420 * tree-flow.h (nonfreeing_call_p): New prototype.
24421 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24423 2013-02-13 David S. Miller <davem@davemloft.net>
24425 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24427 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24430 * ira.c (max_regno_before_ira): Move from ...
24432 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24435 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24437 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24439 2013-02-13 Richard Biener <rguenther@suse.de>
24442 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24443 globals in MEM_REFs.
24445 2013-02-13 Richard Biener <rguenther@suse.de>
24447 * loop-init.c (loop_optimizer_init): Clear loop state when
24448 re-initializing preserved loops.
24449 * loop-unswitch.c (unswitch_single_loop): Return whether
24450 we unswitched the loop. Do not verify loop state here.
24451 (unswitch_loops): When we unswitched a loop discover new loops.
24453 2013-02-13 Kostya Serebryany <kcc@google.com>
24455 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24457 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24459 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24461 Avoid instrumenting duplicated memory access in the same basic block
24462 * Makefile.in (asan.o): Add new dependency on hash-table.h
24463 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24464 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24465 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24466 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24467 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24468 (get_mem_ref_of_assignment): New functions.
24469 (get_mem_refs_of_builtin_call): Extract from
24470 instrument_builtin_call and tweak a little bit to make it fit with
24472 (instrument_builtin_call): Use the new
24473 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24474 of is_gimple_builtin_call.
24475 (instrument_derefs, instrument_mem_region_access): Insert the
24476 instrumented memory reference into the hash table.
24477 (maybe_instrument_assignment): Renamed instrument_assignment into
24478 this, and change it to advance the iterator when instrumentation
24479 actually happened and return true in that case. This makes it
24480 homogeneous with maybe_instrument_assignment, and thus give a
24481 chance to callers to be more 'regular'.
24482 (transform_statements): Clear the memory reference hash table
24483 whenever we enter a new BB, when we cross a function call, or when
24484 we are done transforming statements. Use
24485 maybe_instrument_assignment instead of instrumentation. No more
24486 need to special case maybe_instrument_assignment and advance the
24487 iterator after calling it; it's now handled just like
24488 maybe_instrument_call. Update comment.
24490 2013-02-13 Richard Biener <rguenther@suse.de>
24492 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24493 Fix loop discovery code.
24495 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24497 PR inline-asm/56148
24498 * lra-constraints.c (process_alt_operands): Match early clobber
24499 operand with itself. Check conflicts with earlyclobber only if
24500 the operand is not reloaded. Prefer to reload conflicting operand
24501 if earlyclobber and matching operands are the same.
24503 2013-02-12 Richard Biener <rguenther@suse.de>
24506 * lto-streamer-out.c (write_symbol): Do not output symbols
24507 for hard register variables.
24509 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24512 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24513 (umulsidi3_insn, mulsidi3_insn): New insns.
24515 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24517 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24518 (struct tune_params): Add vec_costs field.
24519 * config/arm/arm.c (arm_builtin_vectorization_cost)
24520 (arm_add_stmt_cost): New functions.
24521 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24522 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24523 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24524 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24525 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24526 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24527 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24529 2013-02-12 Richard Biener <rguenther@suse.de>
24532 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24533 decls again if possible.
24535 2013-02-12 Richard Biener <rguenther@suse.de>
24537 PR middle-end/56288
24538 * tree-ssa.c (verify_ssa_name): Fix check, move
24539 SSA_NAME_IN_FREE_LIST check up.
24541 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24542 Steven Bosscher <steven@gcc.gnu.org>
24544 PR rtl-optimization/56151
24545 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24546 equal to op0 or op1, and last_insn pattern is CODE operation
24547 with MEM dest and one of the operands matches that MEM.
24549 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24551 * doc/extend.texi: Document Function Multiversioning and "default"
24552 parameter string to target attribute.
24553 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24554 target attribute parameter is "default".
24555 (ix86_compare_version_priority): Remove checks for target attribute.
24556 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24557 Remove check for target attribute equal to NULL. Add assert.
24558 (ix86_generate_version_dispatcher_body): Change error to sorry.
24560 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24561 Jack Howarth <howarth@bromo.med.uc.edu>
24562 Patrick Marlier <patrick.marlier@gmail.com>
24565 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24566 define ENDFILE_SPEC as TM_DESTRUCTOR.
24567 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24569 2013-02-11 Alexander Potapenko <glider@google.com>
24570 Jack Howarth <howarth@bromo.med.uc.edu>
24571 Jakub Jelinek <jakub@redhat.com>
24574 * config/darwin.c (cdtor_record): Rename ctor_record.
24575 (sort_cdtor_records): Rename sort_ctor_records.
24576 (finalize_dtors): New routine to sort destructors by
24577 priority before use in assemble_integer.
24578 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24580 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24582 PR rtl-optimization/56275
24583 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24584 offset is non-negative and less than cmode size before
24585 calling simplify_subreg.
24587 2013-02-11 Richard Biener <rguenther@suse.de>
24589 PR tree-optimization/56264
24590 * cfgloop.h (fix_loop_structure): Adjust prototype.
24591 * loop-init.c (fix_loop_structure): Return the number of
24592 newly discovered loops.
24593 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24594 are discovered, do a full loop-closed SSA rewrite.
24596 2013-02-11 Richard Biener <rguenther@suse.de>
24598 PR tree-optimization/56273
24599 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24601 (check_array_ref): Fix missing newline in dumps.
24602 (search_for_addr_array): Likewise.
24604 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24606 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24608 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24611 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24613 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24615 PR rtl-optimization/56246
24616 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24618 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24619 constraints are satisfied.
24621 2013-02-08 Jeff Law <law@redhat.com>
24624 * emit-rtl.c (reg_is_parm_p): New function.
24625 * regs.h (reg_is_parm_p): New prototype.
24626 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24627 callee-clobbered registers.
24629 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24632 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24633 If there is no implicit builtin declaration, just return NULL.
24635 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24637 * config/i386/sse.md (FMAMODEM): New mode iterator.
24638 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24639 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24641 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24643 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24644 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24645 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24647 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24649 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24650 (microblaze*-*-elf): Likewise.
24651 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24653 * config/microblaze/microblaze-c.c: Add builtin defines for
24654 _LITTLE_ENDIAN and _BIG_ENDIAN.
24655 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24656 add to TARGET_DEFAULT flags.
24657 Expand ASM_SPEC and LINK_SPEC.
24658 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24659 * config/microblaze/microblaze.md: Update extendsidi2 and
24660 movdi_internal instructions to use low-order / high-order reg
24662 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24663 options and inversemask / mask of LITTLE_ENDIAN.
24664 * config/microblaze/t-microblaze: Expand multilib options to
24665 include mlittle-endian (le) and update exceptions patterns.
24667 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24669 PR rtl-optimization/56195
24670 * lra-constraints.c (get_reload_reg): Don't reuse regs
24671 if they have smaller mode than requested, if they have
24672 wider mode than requested, try to return a SUBREG.
24674 PR tree-optimization/56250
24675 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24676 if type is unsigned and code isn't MULT_EXPR.
24678 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24680 PR tree-optimization/56064
24681 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24682 bits according to mode.
24683 * fixed-value.h (fixed_from_double_int)
24684 (const_fixed_from_double_int): Adjust comments.
24686 2013-02-08 Richard Biener <rguenther@suse.de>
24689 * lto-streamer.h (struct data_in): Remove current_file, current_line
24690 and current_col members.
24691 * lto-streamer-out.c (lto_output_location): Stream changed bits
24692 en-block for efficiency.
24693 * lto-streamer-in.c (clear_line_info): Remove.
24694 (lto_input_location): Cache current file, line and column
24695 globally via local statics. Read changed bits en-block.
24696 (input_function): Do not call clear_line_info.
24697 (lto_read_body): Likewise.
24698 (lto_input_toplevel_asms): Likewise.
24700 2013-02-08 Michael Matz <matz@suse.de>
24702 PR tree-optimization/52448
24703 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24704 (nt_call_phase): New static.
24705 (add_or_mark_expr): Only mark accesses with newer phase than any
24707 (nonfreeing_call_p): New.
24708 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24709 (nt_fini_block): Keep blocks marked as visited.
24710 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24712 2013-02-08 Richard Biener <rguenther@suse.de>
24714 * ira.c (ira): Free broken dominator information.
24716 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24718 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24720 2013-02-08 Marek Polacek <polacek@redhat.com>
24722 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24724 2013-02-08 Richard Biener <rguenther@suse.de>
24726 PR middle-end/56181
24727 * cfgloop.h (flow_loops_find): Adjust.
24728 (bb_loop_header_p): Declare.
24729 * cfgloop.c (bb_loop_header_p): New function split out from ...
24730 (flow_loops_find): ... here. Adjust function signature,
24731 support incremental loop structure update.
24732 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24733 * cfgloopmanip.c (fix_loop_structure): Move ...
24734 * loop-init.c (fix_loop_structure): ... here.
24735 (apply_loop_flags): Split out from ...
24736 (loop_optimizer_init): ... here.
24737 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24738 in incremental mode, only remove dead loops here.
24740 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24743 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24744 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24745 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24746 (*round<mode>3.libgcc): New insns for fixed-modes.
24747 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24748 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24749 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24750 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24751 implementations. Define to __builtin_avr_absFX,
24752 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24753 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24754 __builtin_avr_countlsFX, respectively.
24755 * config/avr/avr-c.c (target.h): Include it.
24756 (enum avr_builtin_id): New enum.
24757 (avr_resolve_overloaded_builtin): New static function.
24758 (avr_register_target_pragmas): Use it to set
24759 targetm.resolve_overloaded_builtin.
24760 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24761 tree nodes used by DEF_BUILTIN.
24762 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24763 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24764 <AVR_BUILTIN_xxBITS>: Same.
24766 2013-02-08 Richard Biener <rguenther@suse.de>
24768 * cfgloop.c (verify_loop_structure): Properly handle
24769 a loop exiting to another loop header.
24770 * ira-int.h (ira_loops): Remove.
24771 * ira.c (ira_loops): Remove.
24772 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24773 (do_reload): Use loop_optimizer_finalize.
24774 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24775 number_of_loops to access the loop tree.
24776 (more_one_region_p): Likewise.
24777 (finish_loop_tree_nodes): Likewise.
24778 (rebuild_regno_allocno_maps): Likewise.
24779 (mark_loops_for_removal): Likewise.
24780 (mark_all_loops_for_removal): Likewise.
24781 (remove_unnecessary_regions): Likewise.
24782 (ira_build): Likewise.
24783 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24785 2013-02-08 Richard Biener <rguenther@suse.de>
24787 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24788 * ipa-pure-const.c (analyze_function): Avoid calling
24789 mark_irreducible_loops twice.
24790 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24792 2013-02-07 David S. Miller <davem@davemloft.net>
24794 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24796 * var-tracking.c (vt_add_function_parameter): Test the presence of
24797 HAVE_window_save properly and do not remap argument registers when
24798 we have a leaf function.
24800 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24803 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24805 * config/i386/i386.c (ix86_print_operand): Ditto.
24807 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24809 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24811 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24813 PR rtl-optimization/56225
24814 * lra-constraints.c (process_alt_operands): Check that reload hard
24815 reg can hold value for strict_low_part.
24817 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24820 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24821 dwarf2out_end_function.
24822 (in_first_function_p, maybe_at_text_label_p,
24823 first_loclabel_num_not_at_text_label): New variables.
24824 (dwarf2out_var_location): In the first function find out
24825 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24826 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24829 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24831 PR rtl-optimization/56178
24832 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24833 SUBREG of a register. Tidy up related block of code.
24834 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24835 note if the source is a register or a SUBREG of a register.
24837 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24840 * config/rs6000/rs6000.md (ptrm): New mode attr.
24841 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24842 call_value_indirect_aix<pttrsize>,
24843 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24846 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24848 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24849 if -bnortl. Convert to strcmp and strncmp.
24851 2013-02-07 Alan Modra <amodra@gmail.com>
24854 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24855 addresses won't wrap when offsetting.
24856 (rs6000_secondary_reload): Provide secondary reloads needed for
24857 wrapping LO_SUM addresses.
24859 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24861 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24862 MACH, just __MACH__.
24864 2013-02-06 Richard Biener <rguenther@suse.de>
24866 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24867 instead of calling fix_loop_structure.
24869 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24871 PR middle-end/56217
24872 * omp-low.c (use_pointer_for_field): Return false if
24873 lower_send_shared_vars doesn't generate any copy-out code.
24875 2013-02-06 Tom de Vries <tom@codesourcery.com>
24877 PR rtl-optimization/56131
24878 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24879 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24880 of the label is NULL. Add comment.
24882 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24884 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24887 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24888 (STATIC_LIBTSAN_LIBS): Likewise.
24889 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24890 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24891 is defined, don't add anything else beyond that.
24892 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24893 (LINK_COMMAND_SPEC): Use them.
24895 PR tree-optimization/56205
24896 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24897 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24898 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24900 2013-02-05 Richard Biener <rguenther@suse.de>
24902 PR tree-optimization/53342
24903 PR tree-optimization/53185
24904 * tree-vectorizer.h (vect_check_strided_load): Remove.
24905 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24906 not disallow peeling for vectorized strided loads.
24907 (vect_check_strided_load): Make static and simplify.
24908 (vect_analyze_data_refs): Adjust.
24909 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24910 correctly when vectorizing strided loads.
24912 2013-02-05 Richard Biener <rguenther@suse.de>
24914 * doc/install.texi: Refer to ISL, not PPL.
24916 2013-02-05 Jan Hubicka <jh@suse.cz>
24918 PR tree-optimization/55789
24919 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24921 2013-02-05 Jan Hubicka <jh@suse.cz>
24923 PR tree-optimization/55789
24924 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24925 the dead call anyway.
24927 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24930 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24932 2013-02-04 Alexander Potapenko <glider@google.com>
24933 Jack Howarth <howarth@bromo.med.uc.edu>
24934 Jakub Jelinek <jakub@redhat.com>
24937 * config/darwin.c (sort_ctor_records): Stabilized qsort
24938 on constructor priority by using original position.
24939 (finalize_ctors): New routine to sort constructors by
24940 priority before use in assemble_integer.
24941 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24943 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24946 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24947 about visibility on artificial decls.
24948 * config/sol2.c (solaris_assemble_visibility): Likewise.
24950 2013-02-04 Kai Tietz <ktietz@redhat.com>
24953 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24954 argument and improve checking of return-argument types for 16-byte
24956 (ix86_function_value_1): Add additional valtype argument on call
24957 of function_value_64.
24958 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24959 handling infunction_value_64 function.
24961 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24963 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24965 2013-02-04 Richard Biener <rguenther@suse.de>
24967 PR tree-optimization/56188
24968 * tree-ssa-structalias.c (label_visit): Consider case with
24969 initially non-empty points-to set.
24970 (perform_var_substitution): Dump node mapping and clean up.
24972 2013-02-04 Richard Guenther <rguenther@suse.de>
24975 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24976 node prevail as last resort.
24977 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24978 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24980 2013-02-04 Richard Biener <rguenther@suse.de>
24982 PR tree-optimization/56113
24983 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24985 (equiv_class_lookup_or_add): ... this.
24986 (label_visit): Adjust and fix error in previous patch.
24987 (perform_var_substitution): Adjust.
24989 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24991 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24992 * config/sh/sh4-300.md: Likewise.
24993 * config/sh/sh4a.md: Likewise.
24994 * config/sh/constraints.md: Likewise.
24995 * config/sh/sh.md: Likewise.
24996 * config/sh/netbsd-elf.h: Likewise.
24997 * config/sh/predicates.md: Likewise.
24998 * config/sh/sh-protos.h: Likewise.
24999 * config/sh/ushmedia.h: Likewise.
25000 * config/sh/linux.h: Likewise.
25001 * config/sh/sh.c: Likewise.
25002 * config/sh/superh.h: Likewise.
25003 * config/sh/elf.h: Likewise.
25004 * config/sh/sh4.md: Likewise.
25005 * config/sh/sh.h: Likewise.
25007 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25009 * config/pa/constraints.md: Adjust unused letters. Change "T"
25010 constraint to match_test floating_point_store_memory_operand().
25011 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25012 (base14_operand): New.
25013 (floating_point_store_memory_operand): New.
25014 (integer_store_memory_operand): Revise to use base14_operand and
25015 reg_plus_base_memory_operand.
25016 (move_dest_operand): Allow symbolic_memory_operands.
25017 (symbolic_memory_operand): Check for LO_SOM.
25018 (symbolic_operand): Change default case to break.
25019 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25020 CONST_DOUBLE values to be reloaded by putting them into memory when
25021 the destination is a floating point register.
25022 (movdf): Remove code to handle CONST_DOUBLE.
25024 (reload_indf_r1): New.
25025 (reload_insf_r1): New.
25026 Consistently use "Q" and "T" constraints with integer and floating
25027 point move instructions, respectively.
25028 (movdi): Remove FAIL.
25029 Change predicate for source operand unamed DImode move from
25030 general_operand to move_src_operand.
25031 (umulsidi3): Change predicate for destination operand to
25033 Likewise for similar unamed patterns.
25034 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25035 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25036 (hppa_legitimize_address): Simplify mask calculation.
25037 (pa_emit_move_sequence): Revised handling of secondary reloads from
25038 REG+D addresses for floating point loads and stores. Directly handle
25039 loading CONST0_RTX (mode) to a floating point register.
25040 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25041 to floating point registers. Don't restrict secondary reloads to
25042 floating point registers to integer modes. Revise some comments and
25044 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25045 (pa_legitimate_address_p): New.
25046 (pa_legitimize_reload_address): New.
25047 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25048 (STRICT_REG_OK_FOR_BASE_P): New.
25049 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25050 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25052 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25053 Andrew Dixie <andrewd@gentrack.com>
25055 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25058 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25060 * expmed.c (extract_bit_field_1): Pass the full width of the
25061 structure to get_best_reg_extraction_insn.
25063 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25066 * configure.ac (use_cxa_atexit): Add AIX.
25067 * configure: Regenerate.
25069 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25071 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25074 * final.c (need_profile_function): New variable.
25075 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25076 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25077 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25078 notes, targetm.asm_out.function_prologue doesn't emit anything,
25079 HAVE_prologue and profiler should be emitted before prologue,
25080 set need_profile_function instead of emitting it.
25081 (final_scan_insn): If need_profile_function, emit
25082 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25083 NOTE_INSN_FUNCTION_BEG note.
25085 2013-02-01 Richard Henderson <rth@redhat.com>
25087 * config/rs6000/rs6000.md (smulditi3): New.
25090 * config/alpha/alpha.md (umulditi3): New.
25092 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25094 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25095 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25097 2013-02-01 Richard Biener <rguenther@suse.de>
25099 PR tree-optimization/56113
25100 * tree-ssa-structalias.c (label_visit): Reduce work for
25101 single-predecessor nodes.
25103 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25105 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25106 range isn't testing for zero.
25108 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25110 PR middle-end/56113
25111 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25113 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25114 Nick Clifton <nickc@redhat.com>
25116 * config/v850/constraints.md (Q): Define as a memory constraint.
25117 * config/v850/predicates.md (label_ref_operand): New predicate.
25118 (e3v5_shift_operand): New predicate.
25119 (ior_operator): New predicate.
25120 * config/v850/t-v850: Add e3v5 multilib.
25121 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25122 (v850_gen_movdi): Prototype.
25123 * config/v850/v850.c: Add support for e3v5 architecture.
25124 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25126 (construct_save_jarl): Add e3v5 long JARL support.
25127 (v850_adjust_insn_length): New function. Adjust length of call
25128 insns when using e3v5 instructions.
25129 (v850_gen_movdi): New function: Generate instructions to move a
25131 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25132 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25133 (TARGET_USE_FPU): Enable for e3v5.
25134 (CONST_OK_FOR_W): New macro.
25135 (ADJUST_INSN_LENGTH): Define.
25136 * config/v850/v850.md (UNSPEC_LOOP): Define.
25137 (attr cpu): Add v850e3v5.
25138 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25139 (movdi): New pattern.
25140 (movdi_internal): New pattern.
25141 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25142 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25143 (cstoresf4): Likewise.
25144 (cstoredf4): Likewise.
25145 (insv): New pattern.
25146 (rotlso3_a): New pattern.
25147 (rotlsi3_b): New pattern
25148 (rotlsi3_v850e3v5): New pattern.
25149 (doloop_begin): New pattern.
25150 (fix_loop_counter): New pattern.
25151 (doloop_end): New pattern.
25152 (branch_normal): Add e3v5 long branch support.
25153 (branch_invert): Likewise.
25154 (branch_z_normal): Likewise.
25155 (branch_z_invert): Likewise.
25156 (branch_nz_normal): Likewise.
25157 (branch_nz_invert): Likewise.
25158 (call_internal_short): Add e3v5 register-indirect JARL support.
25159 (call_internal_long): Likewise.
25160 (call_value_internal_short): Likewise.
25161 (call_value_internal_long): Likewise.
25162 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25163 (mloop): New option.
25164 * config.gcc: Add support for configuring v840e3v5 target.
25165 * doc/invoke.texi: Document new v850 specific command line options.
25167 2013-01-31 Paul Koning <ni1d@arrl.net>
25171 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25172 children if parent is a class.
25173 (prune_unused_types_prune): Don't add DW_AT_declaration.
25175 2013-01-31 Richard Biener <rguenther@suse.de>
25177 PR tree-optimization/56157
25178 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25179 match up operand with SLP child.
25181 2013-01-31 Jason Merrill <jason@redhat.com>
25184 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25185 parameters the first time.
25186 (gen_scheduled_generic_parms_dies): Check completeness here.
25188 2013-01-31 Richard Biener <rguenther@suse.de>
25190 PR middle-end/53073
25191 * common.opt (faggressive-loop-optimizations): New flag,
25192 enabled by default.
25193 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25194 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25195 infer_loop_bounds_from_undefined by it.
25197 2013-01-31 Richard Biener <rguenther@suse.de>
25199 PR tree-optimization/56150
25200 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25201 visit virtual operands.
25202 (find_uses_to_rename_bb): Likewise.
25204 2013-01-31 Richard Biener <rguenther@suse.de>
25206 PR tree-optimization/56150
25207 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25208 mixed store non-store stmts.
25210 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25213 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25214 LIBASAN_EARLY_SPEC is defined.
25215 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25216 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25218 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25221 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25222 invalid args instead of ICEing on it.
25223 (ix86_valid_target_attribute_tree): Return error_mark_node if
25224 ix86_valid_target_attribute_inner_p failed.
25225 (ix86_valid_target_attribute_p): Return false only if
25226 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25227 target("default") attribute.
25228 (sorted_attr_string): Change argument from const char * to tree,
25229 merge in all target attribute arguments rather than just one.
25230 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25231 instead of free. Avoid using strcat.
25232 (ix86_mangle_function_version_assembler_name): Mangle
25233 target("default") as if no target attribute is present. Adjust
25234 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25235 instead of xmalloc and XDELETEVEC instead of free.
25236 (ix86_function_versions): Don't return true if one of the decls
25237 doesn't have target attribute. If they don't and one of the decls
25238 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25239 sorted_attr_string caller. Use XDELETEVEC instead of free.
25240 (ix86_supports_function_versions): Remove.
25241 (make_name): Fix up formatting.
25242 (make_dispatcher_decl): Remove resolver_name and its initialization.
25243 Avoid leaking memory.
25244 (is_function_default_version): Return true if there is
25245 target("default") attribute rather than no target attribute at all.
25246 (make_resolver_func): Avoid leaking memory.
25247 (ix86_generate_version_dispatcher_body): Likewise.
25248 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25249 * target.def (supports_function_versions): Remove.
25250 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25251 * doc/tm.texi: Regenerated.
25253 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25255 PR rtl-optimization/56144
25256 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25257 for values with side effects.
25259 2013-01-30 Richard Biener <rguenther@suse.de>
25261 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25262 (sparseset_pop): Likewise.
25263 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25264 to be able to use quick_push in the worker loop.
25266 2013-01-30 Marek Polacek <polacek@redhat.com>
25268 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25270 2013-01-30 Richard Biener <rguenther@suse.de>
25273 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25275 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25277 PR tree-optimization/56064
25278 * fixed-value.c (fixed_from_double_int): New function.
25279 * fixed-value.h (fixed_from_double_int): New prototype.
25280 (const_fixed_from_double_int): New static inline function.
25281 * fold-const.c (native_interpret_fixed): New static function.
25282 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25283 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25284 (native_encode_fixed): New static function.
25285 (native_encode_expr) <FIXED_CST>: Use it.
25286 (native_interpret_int): Move double_int worker code to...
25287 * double-int.c (double_int::from_buffer): ...this new static method.
25288 * double-int.h (double_int::from_buffer): Prototype it.
25290 2013-01-30 Richard Biener <rguenther@suse.de>
25292 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25293 New pointer-map and obstack.
25294 (init_alias_vars): Allocate pointer-map and obstack.
25295 (delete_points_to_sets): Free them.
25296 (find_what_var_points_to): Cache result.
25297 (find_what_p_points_to): Adjust for changed interface of
25298 find_what_var_points_to.
25299 (compute_points_to_sets): Likewise.
25300 (ipa_pta_execute): Likewise.
25302 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25304 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25305 * configure: Regenerate.
25306 * config.in: Regenerate.
25307 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25308 #nobits/#progbits if supported.
25310 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25313 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25314 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25315 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25317 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25319 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25320 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25322 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25324 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25325 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25327 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25329 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25331 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25332 * config/arm/cortex-a7.md: New bypasses using
25333 arm_mac_accumulator_is_result.
25335 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25337 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25338 (cortex_a7_neon_mla): Likewise.
25339 (cortex_a7_fpfmad): New reservation.
25340 (cortex_a7_fpmacs): Use ffmas and update required units.
25341 (cortex_a7_fpmuld): Update required units and latency.
25342 (cortex_a7_fpmacd): Likewise.
25343 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25344 (cortex_a7_neon). Likewise.
25345 (bypass) Update participating units.
25347 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25349 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25350 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25352 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25353 (vfp_fmul): Use ffmad.
25354 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25355 (cortex_r4_fmacd): Use ffmad.
25356 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25357 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25358 (cortex_a9_fmacd): Use ffmad.
25359 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25360 (cortex_a8_vfp_macd): Use ffmad.
25361 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25362 (cortex_a5_fpmacd): Use ffmad.
25363 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25364 (cortex_a15_vfp_macd): Use ffmad.
25365 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25367 2013-01-29 Jason Merrill <jason@redhat.com>
25370 * varasm.c (default_assemble_visibility): Don't warn about
25371 visibility on artificial decls.
25373 2013-01-29 Richard Biener <rguenther@suse.de>
25375 PR tree-optimization/56113
25376 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25378 (label_visit): Free duplicate bitmaps and record the leader instead.
25379 (perform_var_substitution): Adjust.
25381 2013-01-29 Richard Biener <rguenther@suse.de>
25383 PR tree-optimization/55270
25384 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25385 the CFG, schedule loops for fixup.
25387 2013-01-29 Nick Clifton <nickc@redhat.com>
25389 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25392 2013-01-28 Leif Ekblad <leif@rdos.net>
25394 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25395 * config/i386/i386.h (TARGET_RDOS): New macro.
25396 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25397 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25398 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25399 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25400 DEFAULT_LARGE_SECTION_THRESHOLD.
25401 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25402 * config/i386/rdos.h: New file.
25403 * config/i386/rdos64.h: New file.
25405 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25408 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25411 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25413 PR rtl-optimization/56117
25414 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25415 call cselib_lookup_from_insn on the MEM before calling
25416 add_insn_mem_dependence.
25418 2013-01-28 Richard Biener <rguenther@suse.de>
25420 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25421 to a stmt that didn't have one.
25422 (copy_phis_for_bb): Likewise for PHI arguments.
25423 (copy_debug_stmt): Likewise for debug stmts.
25425 2013-01-28 Richard Biener <rguenther@suse.de>
25427 PR tree-optimization/56034
25428 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25429 (partition_builtin_p): Adjust.
25430 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25431 it is the last partition.
25432 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25433 up the vertex for the definition.
25434 (classify_partition): Classify whether a partition is a
25435 PKIND_REDUCTION, thus has uses outside of the loop.
25436 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25437 Merge all PKIND_REDUCTION partitions into the last partition.
25438 (tree_loop_distribution): Seed partitions from reductions as well.
25440 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25442 PR tree-optimization/56125
25443 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25444 pow(x,c) into sqrt(x) * powi(x, n/2) or
25445 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25446 optimizing for size.
25447 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25448 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25451 PR tree-optimization/56094
25452 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25453 to UNKNOWN_LOCATION while gimplifying expr.
25455 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25458 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25459 operand 0 in movabs insn template for -masm=intel asm alternative.
25460 (*movabs<mode>_2): Ditto for operand 1.
25462 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25465 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25468 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25470 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25473 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25475 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25476 Undef to avoid warning.
25478 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25480 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25481 * configure: Regenerate.
25483 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25485 PR tree-optimization/56098
25486 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25487 for stmts with volatile ops.
25488 (cond_store_replacement): Don't optimize if assign has volatile ops.
25489 (cond_if_else_store_replacement_1): Don't optimize if either
25490 then_assign or else_assign have volatile ops.
25491 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25494 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25496 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25498 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25500 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25501 missing ':' in asm example.
25503 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25505 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25506 entries into lane and laneq entries.
25507 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25508 Remove AdvSIMD scalar modes.
25509 (aarch64_sq<r>dmulh_laneq<mode>): New.
25510 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25512 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25513 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25514 * config/aarch64/iterators.md (VCOND): New.
25517 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25520 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25521 Add NULL LIBNAME argument to existing definitions.
25522 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25523 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25524 * config/avr/avr.c (DEF_BUILTIN): Same.
25525 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25526 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25527 implementation is available (DECL_ASSEMBLER_NAME is set).
25528 (avr_fold_absfx): New static function.
25529 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25530 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25531 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25532 AVR_BUILTIN_ABSLLK.
25533 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25534 (abshk, absk, abslk, absllk): Provide as static inline functions.
25536 2013-01-25 Marek Polacek <polacek@redhat.com>
25538 PR tree-optimization/56035
25539 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25541 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25543 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25544 (*movtf_internal_rex64): Add (!o,C) alternative
25545 (*movxf_internal_rex64): Ditto.
25546 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25548 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25550 * doc/invoke.texi: fix typo.
25551 * doc/objc.texi: fix typo.
25553 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25555 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25556 for the first two alternatives.
25558 2013-01-24 Diego Novillo <dnovillo@google.com>
25560 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25561 (ggc-zone.o): Remove.
25562 * configure.ac: Remove option --with-gc.
25563 * configure: Re-generate.
25564 * doc/install.texi: Remove documentation for --with-gc.
25565 * gengtype.c (write_enum_defn): Remove. Update all users.
25566 (write_Types_process_field): Remove generation of gt_e_* argument.
25567 (output_type_enum): Remove. Update all users.
25568 (write_enum_defn): Remove. Update all users.
25569 (enum alloc_zone): Remove. Update all users.
25570 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25571 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25572 Update all callers.
25573 (struct ptr_data): Remove field TYPE. Update all users.
25574 (gt_pch_note_object): Remove argument TYPE. Update all users.
25575 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25577 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25578 (struct alloc_zone): Remove.
25579 (ggc_internal_alloc_zone_stat): Remove.
25580 (ggc_internal_cleared_alloc_zone_stat): Remove.
25581 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25582 (ggc_pch_count_object): Remove last argument. Update all users.
25583 (ggc_pch_alloc_object): Remove last argument. Update all users.
25584 (struct alloc_zone): Remove.
25585 * ggc-zone.c: Remove.
25586 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25587 (struct alloc_zone): Remove.
25588 (ggc_alloc_typed_stat): Remove.
25589 (ggc_alloc_typed): Remove.
25590 (ggc_splay_alloc): Remove first argument.
25591 (rtl_zone): Remove. Update all users.
25592 (tree_zone): Remove. Update all users.
25593 (tree_id_zone): Remove. Update all users.
25594 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25595 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25596 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25597 * tree-ssanames.c: Remove references to zone allocator in comments.
25599 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25601 * config/avr/avr.c (avr_out_fract): Make register numbers that
25602 might be outside of source operand signed.
25604 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25606 * config/i386/constraints.md (Yf): New constraint.
25607 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25608 of f constraint to conditionaly disable x87 register preferences.
25609 (*movdf_internal): Ditto.
25610 (*movsf_internal): Ditto.
25612 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25614 PR inline-asm/55934
25615 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25616 that have operands with impossible constraints.
25617 Add a FIXME for a speed-up opportunity.
25618 * lra-constraints.c (process_alt_operands): Verify that a class
25619 selected from constraints on asms is valid for the operand mode.
25620 (curr_insn_transform): Remove incorrect comment.
25622 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25624 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25625 TOC operand is a valid symbol ref in the constant pool.
25627 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25629 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25631 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25634 * config/avr/stdfix.h: New file.
25635 * t-avr (stdfix-gcc.h): New rule to build it.
25636 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25638 2013-01-23 Kostya Serebryany <kcc@google.com>
25640 * config/darwin.h: remove dependency on
25641 CoreFoundation (asan on Mac OS).
25643 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25646 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25647 instead of cmpdi_operand for first comparison operand.
25648 Don't assert that comparison operands aren't both constants.
25650 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25652 * doc/install.texi (Downloading the Source): Update references to
25653 downloading separate components.
25655 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25657 * doc/extend.texi (__int128): Improve grammar.
25659 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25662 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25663 alternative to (o,r).
25664 (*movdi_internal_rex64): Remove (!o,n) alternative.
25665 (DImode immediate->memory splitter): Remove.
25666 (DImode immediate->memory peephole2): Remove.
25667 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25668 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25669 alternative to (!o,*r).
25670 (*movtf_internal_sse): New pattern.
25671 (*movxf_internal_rex64): New pattern.
25672 (*movxf_internal): Disable for TARGET_64BIT.
25673 (*movdf_internal_rex64): Remove (!o,F) alternative.
25675 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25677 PR middle-end/56074
25678 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25679 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25680 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25681 stmt locations where LOCATION_LOCUS of the stmt location is
25682 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25685 * config/i386/i386.md (UNSPEC_STOS): New.
25686 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25687 *strsetqi_1): Add UNSPEC_STOS.
25689 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25692 * doc/invoke.texi: Remove left over -Wsynth example.
25694 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25696 PR tree-optimization/56051
25697 * fold-const.c (fold_binary_loc): Don't fold
25698 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25699 a narrowing conversion, or widening conversion from signed
25702 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25704 PR rtl-optimization/56023
25705 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25706 dependent on debug instruction.
25708 2013-01-21 Martin Jambor <mjambor@suse.cz>
25710 PR middle-end/56022
25711 * function.c (allocate_struct_function): Call
25712 invoke_set_current_function_hook earlier.
25714 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25716 * reload1.c (init_reload): Only initialize reload_obstack
25717 during the first call.
25719 2013-01-21 Marek Polacek <polacek@redhat.com>
25721 * cfgloop.c (verify_loop_structure): Fix up grammar.
25723 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25725 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25726 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25728 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25731 * config/arm/marvell-pj4.md: Update copyright year.
25732 Fix up use of alu to alu_reg and simple_alu_imm.
25734 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25736 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25738 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25741 * lra-constraints.c (curr_insn_transform): Don't reuse original
25742 insn for secondary memory move when memory mode should be different.
25744 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25746 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25747 atomic_storedi_1): New patterns.
25749 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25751 btver2 pipeline descriptions.
25752 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25754 * config/i386/i386.md (btver2_decode): New type attributes.
25755 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25757 * config/i386/btver2.md: New file describing btver2 pipelines.
25759 2013-01-19 Andrew Pinski <apinski@cavium.com>
25761 PR tree-optimization/52631
25762 * tree-ssa-sccvn (visit_use): Before looking up the original
25763 statement, try looking up the simplified expression.
25765 2013-01-19 Anthony Green <green@moxielogic.com>
25767 * config/moxie/moxie.c (moxie_expand_prologue): Set
25768 current_function_static_stack_size.
25770 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25772 PR tree-optimization/56029
25773 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25774 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25776 2013-01-18 Sharad Singhai <singhai@google.com>
25778 PR tree-optimization/55995
25779 * dumpfile.c (dump_loc): Print location only if available.
25780 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25782 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25785 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25786 secondary memory move.
25787 (inherit_reload_reg): Use rclass instead of cl for
25788 check_secondary_memory_needed_p.
25790 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25792 PR middle-end/56015
25793 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25794 the case where writing real complex part of target modifies op1.
25796 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25798 * config/aarch64/aarch64-simd.md
25799 (aarch64_vcond_internal<mode>): Handle unordered cases.
25800 * config/aarch64/iterators.md (v_cmp_result): New.
25802 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25803 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25805 * config/arm/marvell-pj4.md: New file.
25806 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25807 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25808 (generic_vfp): Likewise.
25809 * config/arm/arm-cores.def: Add marvell-pj4.
25810 * config/arm/arm-tune.md: Regenerate.
25811 * config/arm/arm-tables.opt: Regenerate.
25812 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25813 * doc/invoke.texi: Document marvell-pj4.
25815 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25817 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25819 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25824 * var-tracking.c (negative_power_of_two_p): New.
25825 (global_get_addr_cache, local_get_addr_cache): New.
25826 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25827 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25829 (vt_stack_offset_p): Remove.
25830 (vt_canon_true_dep): Always canonicalize loc's address.
25831 (clobber_overlapping_mems): Make sure we have a MEM.
25832 (local_get_addr_clear_given_value): New.
25833 (val_reset): Clear local cached entries.
25834 (compute_bb_dataflow): Create and release the local cache.
25835 Disable duplicate MEMs clobbering.
25836 (emit_notes_in_bb): Clobber MEMs likewise.
25837 (vt_emit_notes): Create and release the local cache.
25838 (vt_initialize, vt_finalize): Create and release the global
25839 cache, respectively.
25840 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25842 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25844 PR libmudflap/53359
25845 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25846 not found in the symtab.
25848 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25851 PR rtl-optimization/55547
25852 PR rtl-optimization/53827
25855 * alias.c (offset_overlap_p): New, factored out of...
25856 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25857 the conservative special case for symbolic constants. Don't
25858 adjust zero sizes on alignment.
25860 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25862 PR rtl-optimization/52573
25863 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25864 REG_UNUSED for the same register.
25866 2013-01-17 Richard Biener <rguenther@suse.de>
25867 Marek Polacek <polacek@redhat.com>
25869 PR rtl-optimization/55833
25870 * loop-unswitch.c (unswitch_loops): Move loop verification...
25871 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25872 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25873 Set it to true when we're removing a loop from hierarchy tree in
25874 an irreducible region.
25875 (fix_bb_placements): Adjust caller.
25876 (fix_loop_placements): Likewise.
25878 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25880 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25881 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25882 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25883 Remove ID. Adjust comments.
25884 * config/avr/avr-c.c (avr_builtin_name): Remove.
25885 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25886 * config/avr/avr.c (avr_tolower): New static function.
25887 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25888 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25889 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25892 2013-01-17 Jan Hubicka <jh@suse.cz>
25894 PR tree-optimization/55273
25895 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25897 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25900 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25901 store through atomic_store<mode>_1.
25902 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25904 2013-01-17 Martin Jambor <mjambor@suse.cz>
25906 PR tree-optimizations/55264
25907 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25908 for virtual methods.
25909 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25910 virtual methods before inlining is over.
25911 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25913 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25916 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25918 PR rtl-optimization/56005
25919 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25920 pending reads for prefetch.
25922 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25924 * config/aarch64/aarch64.md
25925 (*cstoresi_neg_uxtw): New pattern.
25926 (*cmovsi_insn_uxtw): New pattern.
25927 (*<optab>si3_uxtw): New pattern.
25928 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25929 (*<optab>si3_insn_uxtw): New pattern.
25930 (*bswapsi2_uxtw): New pattern.
25932 2013-01-16 Richard Biener <rguenther@suse.de>
25934 * tree-inline.c (tree_function_versioning): Remove set but
25935 never used variable.
25937 2013-01-16 Richard Biener <rguenther@suse.de>
25939 PR tree-optimization/55964
25940 * tree-flow.h (rename_variables_in_loop): Remove.
25941 (rename_variables_in_bb): Likewise.
25942 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25943 (copy_loop_before): Adjust and delete update-ssa status.
25944 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25945 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25946 (rename_variables_in_loop): Remove.
25947 (slpeel_update_phis_for_duplicate_loop): Likewise.
25948 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25949 use available cfg machinery instead of duplicating it.
25950 Update PHI nodes and perform poor-mans SSA update here.
25951 (slpeel_tree_peel_loop_to_edge): Adjust.
25953 2013-01-16 Richard Biener <rguenther@suse.de>
25955 PR tree-optimization/54767
25956 PR tree-optimization/53465
25957 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25958 (vrp_visit_phi_node): For PHI arguments coming via backedges
25959 drop all symbolical range information.
25960 (execute_vrp): Compute backedges.
25962 2013-01-16 Richard Biener <rguenther@suse.de>
25964 * doc/install.texi: Update CLooG and ISL requirements to
25967 2013-01-16 Christian Bruel <christian.bruel@st.com>
25970 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25971 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25972 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25974 2013-01-16 DJ Delorie <dj@redhat.com>
25976 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25977 (UNSPECV_SP_SWITCH_E): New.
25978 (sp_switch_1): Change to an unspec.
25979 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25982 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25984 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25985 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25986 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25987 (expand_mem_thread_fence): Ditto.
25988 (expand_mem_signal_fence): Ditto.
25989 (expand_atomic_load): Ditto.
25990 (expand_atomic_store): Ditto.
25992 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25994 PR rtl-optimization/55547
25995 PR rtl-optimization/53827
25998 * alias.c (memrefs_conflict_p): Set sizes to negative after
26001 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26004 * function.c (thread_prologue_and_epilogue_insns): Always
26005 add crtl->drap_reg to set_up_by_prologue.set, even if
26006 stack_realign_drap is false.
26008 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26010 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26011 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26012 *call): Fix indention.
26014 2013-01-15 Tom de Vries <tom@codesourcery.com>
26017 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26020 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26022 PR rtl-optimization/55153
26023 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26025 2013-01-15 Martin Jambor <mjambor@suse.cz>
26027 PR tree-optimization/55920
26028 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26029 accesses as grp_to_be_debug_replaced.
26031 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26033 PR tree-optimization/55920
26034 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26035 there is non-useless type conversion needed from debug rhs to lhs,
26036 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26038 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26039 Mikael Pettersson <mikpe@it.uu.se>
26042 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26044 (ASM_OUTPUT_CASE_LABEL): Remove.
26045 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26046 * final.c (shorten_branches): Update alignment of labels before
26047 jump tables if CASE_VECTOR_SHORTEN_MODE.
26049 2013-01-15 Richard Biener <rguenther@suse.de>
26052 * system.h: Do not include gmp.h for building host tools.
26054 2013-01-15 Richard Biener <rguenther@suse.de>
26056 PR middle-end/55882
26057 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26058 account for bitpos when computing alignment.
26060 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26062 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26063 (ix86_target_macros_internal): Likewise.
26065 * config/i386/i386.c (m_CORE2I7): Removed.
26066 (m_CORE_HASWELL): New macro.
26067 (m_CORE_ALL): Likewise.
26068 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26069 (initial_ix86_arch_features): Likewise.
26070 (processor_target_table): Initializations for Core avx2.
26071 (cpu_names): New names "core-avx2".
26072 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26073 PROCESSOR_CORE_HASWELL.
26074 (ix86_issue_rate): New case.
26075 (ia32_multipass_dfa_lookahead): Likewise.
26076 (ix86_sched_init_global): Likewise.
26078 * config/i386/i386.h (TARGET_HASWELL): New macro.
26079 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26080 (processor_type): New PROCESSOR_HASWELL.
26082 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26084 PR tree-optimization/55955
26085 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26086 *SHIFT_EXPR and *ROTATE_EXPR codes.
26088 PR tree-optimization/48766
26089 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26090 -ftrapv disable -fwrapv.
26092 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26095 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26096 etc. to 1 and not to __flash.
26097 Use LL suffix for __INT24_MAX__ with -mint8.
26098 Use ULL suffix for __UINT24_MAX__ with -mint8.
26100 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26102 * config/avr/avr-arch.h
26103 (struct base_arch_s): Use typedef avr_arch_t instead.
26104 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26105 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26106 * config/avr/avr.c: Same.
26107 * config/avr/avr-devices.c: Same.
26108 * config/avr/driver-avr.c: Same.
26109 * config/avr/gen-avr-mmcu-texi.c: Same.
26110 * config/avr/avr-mcus.def: Adjust comment.
26112 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26114 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26115 * config/aarch64/iterators.md (VALLDI): New.
26117 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26118 Andi Kleen <ak@linux.intel.com>
26121 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26122 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26125 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26127 * config/avr/avr-stdint.h: Remove trailing blanks.
26128 * config/avr/avr-log.h: Same.
26129 * config/avr/avr-arch.h: Same.
26130 * config/avr/avr-devices.c: Same.
26131 * config/avr/avr-dimode.md: Same.
26132 * config/avr/predicates.md: Same.
26133 * config/avr/avr-c.c: Same. And fix typo.
26135 * config/avr/avr-protos.h: Same. And:
26136 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26137 (init_cumulative_args): Rename to avr_init_cumulative_args.
26138 (expand_prologue): Rename to avr_expand_prologue.
26139 (expand_epilogue): Rename to avr_expand_epilogue.
26140 (adjust_insn_length): Rename to avr_adjust_insn_length.
26141 (notice_update_cc): Rename to avr_notice_update_cc.
26142 (final_prescan_insn): Rename to avr_final_prescan_insn.
26143 * config/avr/avr.c: Same.
26144 * config/avr/avr.h: Same.
26145 * config/avr/avr.md: Remove trailing blanks.
26146 (prologue): Use avr_expand_prologue.
26147 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26149 2013-01-14 Richard Biener <rguenther@suse.de>
26151 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26152 verify_location, collect_subblocks): New functions.
26153 (verify_gimple_in_cfg): Verify that locations only reference
26154 BLOCKs in the functions BLOCK tree.
26156 2013-01-14 Richard Biener <rguenther@suse.de>
26158 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26160 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26162 (insert_out_of_ssa_copy_on_edge): Likewise.
26163 (rewrite_close_phi_out_of_ssa): Likewise.
26164 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26166 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26167 propagated constants.
26168 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26171 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26173 * config/avr/avr-modes.def: Add GPL copyright notice.
26175 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26177 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26178 MEMMODEL_MASK to determine memory model.
26179 (atomic_store<mode>): Ditto from operands[2].
26180 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26182 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26185 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26186 (fold_gimple_assign): Don't call unshare_expr here.
26187 (fold_ctor_reference): Call unshare_expr.
26189 2013-01-13 Terry Guo <terry.guo@arm.com>
26191 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26192 * doc/fragments.texi: Document MULTILIB_REUSE.
26193 * gcc.c (multilib_reuse): New internal spec.
26194 (set_multilib_dir): Also search multilib from multilib_reuse.
26195 * genmultilib (tmpmultilib3): Refactor code.
26196 (tmpmultilib4): Ditto.
26197 (multilib_reuse): New multilib argument.
26199 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26201 * Makefile.in: Update copyright.
26203 2013-01-12 Tom de Vries <tom@codesourcery.com>
26205 PR middle-end/55890
26206 * calls.c (expand_call): Check if arg_nr is valid.
26208 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26210 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26211 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26212 documentation. Add missing '__' in front of
26213 __builtin_ia32_packssdw256.
26215 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26218 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26219 NO_REGS for larl operands.
26220 (s390_reload_larl_operand): Use s390_load_address instead of
26223 2013-01-11 Richard Biener <rguenther@suse.de>
26225 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26226 (verify_node_sharing): ... here.
26227 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26229 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26231 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26232 Substitute TREECHECKING.
26233 * configure: Regenerate.
26234 * Makefile.in (TREECHECKING): New.
26236 2013-01-11 Richard Guenther <rguenther@suse.de>
26238 PR tree-optimization/44061
26239 * tree-vrp.c (extract_range_basic): Compute zero as
26240 value-range for __builtin_constant_p of function parameters.
26242 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26244 Update copyright years.
26246 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26248 PR rtl-optimization/55672
26249 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26250 const to be eliminable.
26252 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26254 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26255 * configure: Regenerate.
26257 2013-01-10 Richard Biener <rguenther@suse.de>
26259 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26261 2013-01-10 Richard Biener <rguenther@suse.de>
26264 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26265 locations for virtual PHI arguments.
26266 (rewrite_update_phi_arguments): Likewise.
26268 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26270 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26273 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26275 PR tree-optimization/55921
26276 * tree-complex.c (expand_complex_asm): New function.
26277 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26279 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26282 * config/s390/s390.c (s390_symref_operand_p)
26283 (s390_loadrelative_operand_p): Merge the two functions.
26284 (s390_check_qrst_address, print_operand_address): Add parameters
26285 to s390_loadrelative_operand_p invokation.
26286 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26287 (s390_reload_larl_operand, s390_secondary_reload): Use
26288 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26289 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26291 2013-01-09 Mike Stump <mikestump@comcast.net>
26293 * dse.c (record_store): Remove unnecessary assert.
26295 2013-01-09 Jan Hubicka <jh@suse.cz>
26297 PR tree-optimization/55569
26298 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26299 * cfgloop.h (scale_loop_profile): Likewise.
26301 2013-01-09 Jan Hubicka <jh@suse.cz>
26304 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26306 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26308 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26310 PR middle-end/55114
26311 * expr.h (maybe_emit_group_store): Declare.
26312 * expr.c (maybe_emit_group_store): New function.
26313 * builtins.c (expand_builtin_int_roundingfn): Call it.
26314 (expand_builtin_int_roundingfn_2): Likewise.
26316 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26318 PR rtl-optimization/55829
26319 * lra-constraints.c (match_reload): Add code for absent output.
26320 (curr_insn_transform): Add code for reloads of matched inputs
26323 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26325 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26326 attribute of movddup insn to DF.
26327 (*vec_interleave_lowv2df): Ditto.
26328 (vec_dupv2df): Ditto.
26330 2013-01-09 Jan Hubicka <jh@suse.cz>
26332 PR tree-optimiation/55875
26333 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26334 EVERY_ITERATION parameter.
26335 (number_of_iterations_exit): Check if exit is executed every iteration.
26336 (idx_infer_loop_bounds): Similarly here.
26337 (n_of_executions_at_most): Simplify
26338 to only test for cases where statement is dominated by the
26339 particular bound; handle correctly the "postdominance" test.
26340 (scev_probably_wraps_p): Use max loop iterations info
26341 as a global bound first.
26343 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26344 Nick Clifton <nickc@redhat.com>
26346 * config/v850/v850.md (cbranchsf4): New pattern.
26347 (cstoresf4): New pattern.
26348 (cbranchdf4): New pattern.
26349 (cstoredf4): New pattern.
26350 (movsicc): Disallow floating point comparisons.
26351 (cmpsf_le_insn): Fix order of operators.
26352 (cmpsf_lt_insn): Likewise.
26353 (cmpsf_eq_insn): Likewise.
26354 (cmpdf_le_insn): Likewise.
26355 (cmpdf_lt_insn): Likewise.
26356 (cmpdf_eq_insn): Likewise.
26357 (cmpsf_ge_insn): Use LE comparison.
26358 (cmpdf_ge_insn): Likewise.
26359 (cmpsf_gt_insn): Use LT comparison.
26360 (cmpdf_gt_insn): Likewise.
26361 (cmpsf_ne_insn): Delete pattern.
26362 (cmpdf_ne_insn): Delete pattern.
26363 * config/v850/v850.c (v850_gen_float_compare): Use
26364 gen_cmpdf_eq_insn for NE comparison.
26365 (v850_float_z_comparison_operator)
26366 (v850_float_nz_comparison_operator): Move from here ...
26367 * config/v850/predicates.md: ... to here. Move GT and GE
26368 comparisons into v850_float_z_comparison_operator.
26369 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26371 (v850_float_nz_comparison_operator): Likewise.
26373 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26375 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26376 with calls to gen_insvsi/gen_insvdi.
26378 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26380 * config/i386/i386.c (initial_ix86_tune_features): Set up
26381 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26383 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26384 Jakub Jelinek <jakub@redhat.com>
26386 PR tree-optimization/48189
26387 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26388 If nitercst is 0, don't predict the exit edge.
26390 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26392 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26393 in asm_fprintf with reg_names.
26394 (aarch64_print_operand_address): Likewise.
26395 (aarch64_return_addr): Likewise.
26396 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26398 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26400 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26401 (INT_U6_BITS): Likewise.
26402 * config/pa/predicates.md (uint6_operand): New predicate.
26403 (shift5_operand, shift6_operand): Likewise.
26404 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26406 (lshrdi3): Use shift6_operand.
26407 (shrpsi4, shrpdi4): New insn patterns.
26408 (extzv): Delete expander.
26409 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26410 predicates in unamed zero extract patterns. Tighten common constraint.
26411 (extv): Delete expander.
26412 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26413 predicates in unamed sign extract patterns. Tighten common constraint.
26414 (insv): Delete expander.
26415 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26416 predicates in unamed insert patterns. Tighten common constraint.
26417 Change uint32_operand predicate to uint6_operand predicate in unamed
26418 DImode pattern to insert constant values of type 1...1xxxx.
26420 2013-01-04 Jan Hubicka <jh@suse.cz>
26422 PR tree-optimization/55823
26423 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26426 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26427 Uros Bizjak <ubizjak@gmail.com>
26429 PR rtl-optimization/55845
26430 * df-problems.c (can_move_insns_across): Stop scanning at
26431 volatile_insn_p source instruction or give up if
26432 across_from .. across_to range contains any volatile_insn_p
26435 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26437 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26438 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26440 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26441 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26443 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26446 * asan.c (asan_clear_shadow): New function.
26447 (asan_emit_stack_protection): Use it.
26449 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26451 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26452 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26453 with tab instead of space.
26455 2013-01-08 Nick Clifton <nickc@redhat.com>
26457 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26458 register bank 0 at the start of an interrupt handler.
26459 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26462 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26464 * config/aarch64/aarch64-simd.md
26465 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26466 (aarch64_simd_bsl): Likewise.
26467 (aarch64_vcond_internal<mode>): Likewise.
26468 (vcond<mode><mode>): Likewise.
26469 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26470 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26472 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26474 * config/aarch64/aarch64-builtins.c
26475 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26477 2013-01-08 Martin Jambor <mjambor@suse.cz>
26480 * tree-sra.c (analyze_access_subtree): Return true also after
26481 potentially creating a debug-only replacement.
26483 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26485 PR middle-end/55890
26486 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26488 PR tree-optimization/54120
26489 * tree-vrp.c (range_fits_type_p): Don't allow
26490 src_precision < precision from signed vr to unsigned_p
26491 if vr->min or vr->max is negative.
26492 (simplify_float_conversion_using_ranges): Test can_float_p
26493 against CODE_FOR_nothing.
26495 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26496 Richard Biener <rguenther@suse.de>
26498 PR middle-end/55851
26499 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26500 types instead of just INTEGER_TYPE types.
26502 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26504 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26507 2013-01-07 Steve Ellcey <sellcey@mips.com>
26510 * config/mips/mips.opt: Change mad to mmad to match documentation.
26512 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26515 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26516 .progmemx.data now.
26518 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26521 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26522 (avr_addrspace_t): Add .section_name field.
26523 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26525 (avr_addrspace): Same. Initialize .section_name. Remove last
26526 NULL entry. Put __memx into .progmemx.data.
26527 (progmem_section_prefix): Remove.
26528 (avr_asm_init_sections): No need to initialize progmem_section.
26529 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26530 section name prefix.
26531 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26532 retrieve the progmem section.
26533 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26534 boundary to run over avr_addrspace[].
26535 (avr_register_target_pragmas): Ditto.
26537 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26539 * varasm.c (output_constant_def_contents): For asan_protect_global
26540 protected strings, adjust DECL_ALIGN if needed, before testing for
26542 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26543 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26545 (output_object_block): For asan protected decls, emit asan padding
26546 after their contents.
26547 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26548 (asan_finish_file): Test it here instead.
26550 2013-01-07 Nick Clifton <nickc@redhat.com>
26551 Matthias Klose <doko@debian.org>
26552 Doug Kwan <dougkwan@google.com>
26553 H.J. Lu <hongjiu.lu@intel.com>
26556 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26558 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26560 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26562 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26564 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26566 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26569 * doc/install.texi (Cross-Compiler-Specific Options): Document
26572 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26574 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26575 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26576 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26577 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26578 vqmovun_high_s64): Fix source operand number and update copyright.
26580 2013-01-07 Richard Biener <rguenther@suse.de>
26582 PR middle-end/55890
26583 * gimple.h (gimple_call_builtin_p): New overload.
26584 * gimple.c (validate_call): New function.
26585 (gimple_call_builtin_p): Likewise.
26586 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26587 Use gimple_call_builtin_p.
26588 (find_func_clobbers): Likewise.
26589 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26590 (strlen_optimize_stmt): Likewise.
26592 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26594 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26595 (vld1q_dup_*): Likewise.
26596 (vld1_*): Likewise.
26597 (vld1q_*): Likewise.
26598 (vld1_lane_*): Likewise.
26599 (vld1q_lane_*): Likewise.
26601 2013-01-07 Richard Biener <rguenther@suse.de>
26603 * lto-streamer.h (LTO_minor_version): Bump to 2.
26605 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26607 * config/aarch64/aarch64-protos.h
26608 (aarch64_const_double_zero_rtx_p): Rename to...
26609 (aarch64_float_const_zero_rtx_p): ...this.
26610 (aarch64_float_const_representable_p): New.
26611 (aarch64_output_simd_mov_immediate): Likewise.
26612 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26613 move immediate case.
26614 * config/aarch64/aarch64.c
26615 (aarch64_const_double_zero_rtx_p): Rename to...
26616 (aarch64_float_const_zero_rtx_p): ...this.
26617 (aarch64_print_operand): Allow printing of new constants.
26618 (aarch64_valid_floating_const): New.
26619 (aarch64_legitimate_constant_p): Check for valid floating-point
26621 (aarch64_simd_valid_immediate): Likewise.
26622 (aarch64_vect_float_const_representable_p): New.
26623 (aarch64_float_const_representable_p): Likewise.
26624 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26625 (aarch64_output_simd_mov_immediate): New.
26626 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26627 (*movdf_aarch64): Likewise.
26628 * config/aarch64/constraints.md (Ufc): New.
26629 (Y): call aarch64_float_const_zero_rtx.
26630 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26632 2013-01-07 Richard Biener <rguenther@suse.de>
26634 PR tree-optimization/55888
26635 PR tree-optimization/55862
26636 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26637 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26638 not if it is contained therein.
26640 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26642 * config/avr/t-avr: Typo.
26644 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26647 * config/avr/t-avr: Don't automatically rebuild
26648 $(srcdir)/config/avr/t-multilib
26649 $(srcdir)/config/avr/avr-tables.opt
26650 $(srcdir)/doc/avr-mmcu.texi
26651 (avr-mcus): New phony target to build them on request.
26652 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26653 * avr/avr-mcus.def: Adjust comments.
26655 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26657 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26659 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26661 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26663 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26665 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26667 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26669 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26670 to generate profiling.
26671 * config/rs6000/aix64.h (LIB_SPEC): Same.
26673 2013-01-04 Andrew Pinski <apinski@cavium.com>
26675 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26677 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26679 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26681 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26683 (ix86_expand_move): Ditto.
26684 (ix86_zero_extend_to_Pmode): Ditto.
26685 (ix86_expand_call): Ditto.
26686 (ix86_expand_special_args_builtin): Ditto.
26687 (ix86_expand_builtin): Ditto.
26689 2013-01-04 Richard Biener <rguenther@suse.de>
26691 PR tree-optimization/55862
26692 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26693 translating them through PHI nodes.
26695 2013-01-04 Martin Jambor <mjambor@suse.cz>
26697 PR tree-optimization/55755
26698 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26699 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26701 2013-01-04 Marek Polacek <polacek@redhat.com>
26703 PR middle-end/55859
26704 * opts.c (default_options_optimization): Clarify error message.
26706 2013-01-04 Richard Biener <rguenther@suse.de>
26708 PR middle-end/55863
26709 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26712 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26715 * config/pa/pa.md (movsi): Revert previous change.
26716 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26719 2013-01-03 Richard Henderson <rth@redhat.com>
26721 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26722 after eliminating TLS symbols.
26724 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26727 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26728 * graphite-poly.c (debug_gmp_value): Likewise.
26730 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26733 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26734 selected code model, define __code_mode_small__, __code_model_medium__,
26735 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26736 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26737 xchg temporary register with %k. Declare temporary register as
26739 [__x86_64__]: For medium and large code models, preserve %rbx register.
26741 2013-01-03 Richard Biener <rguenther@suse.de>
26743 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26744 (dump_subscript): Adjust.
26745 (finalize_ddr_dependent): Do not dump redundant info.
26746 (analyze_siv_subscript): Adjust.
26747 (subscript_dependence_tester): Likewise.
26748 (compute_affine_dependence): Likewise.
26750 2013-01-03 Richard Biener <rguenther@suse.de>
26753 2013-01-03 Richard Biener <rguenther@suse.de>
26755 PR tree-optimization/55857
26756 * tree-vect-stmts.c (vectorizable_load): Do not setup
26757 re-alignment for invariant loads.
26759 2013-01-02 Richard Biener <rguenther@suse.de>
26761 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26762 invariant load do not generate a vector load from the scalar location.
26764 2013-01-03 Richard Biener <rguenther@suse.de>
26766 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26767 for not vectorizing.
26768 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26769 not build INDIRECT_REFs, call get_name once only.
26770 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26771 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26773 2013-01-03 Richard Biener <rguenther@suse.de>
26775 PR tree-optimization/55857
26776 * tree-vect-stmts.c (vectorizable_load): Do not setup
26777 re-alignment for invariant loads.
26779 2013-01-03 Richard Biener <rguenther@suse.de>
26782 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26783 prefer a built-in decl.
26785 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26787 * gcc.c (process_command): Update copyright notice dates.
26788 * gcov.c (print_version): Likewise.
26789 * gcov-dump.c (print_version): Likewise.
26791 PR rtl-optimization/55838
26792 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26793 iv0.step, iv1.step and step.
26795 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26796 Marc Glisse <marc.glisse@inria.fr>
26798 PR tree-optimization/55832
26799 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26800 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26801 integer_{one,zero}_node.
26803 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26806 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26807 * var-tracking.c (reverse_op): Don't add reverse ops to
26808 VALUEs that have already
26809 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26811 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26813 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26815 2013-01-02 Teresa Johnson <tejohnson@google.com>
26817 * dumpfile.c (dump_loc): Print filename with location.
26818 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26819 new location_t parameter to emit complete unroll message with
26820 new dump framework.
26821 (canonicalize_loop_induction_variables): Compute loops location
26822 and pass to try_unroll_loop_completely.
26823 * loop-unroll.c (report_unroll_peel): New function.
26824 (peel_loops_completely): Use new dump format with location
26825 for main dumpfile message, and invoke report_unroll_peel on success.
26826 (decide_unrolling_and_peeling): Ditto.
26827 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26828 by report_unroll_peel.
26829 (decide_peel_completely): Ditto.
26830 (decide_unroll_constant_iterations): Ditto.
26831 (decide_unroll_runtime_iterations): Ditto.
26832 (decide_peel_simple): Ditto.
26833 (decide_unroll_stupid): Ditto.
26834 * cfgloop.c (get_loop_location): New function.
26835 * cfgloop.h (get_loop_location): Declare.
26837 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26839 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26842 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26844 PR middle-end/55198
26845 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26846 BLKmode objects when EXPAND_MEMORY is specified.
26848 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26850 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26852 (fold_builtin_cpu): Do not share cpu model decls across statements.
26854 2013-01-02 Jason Merrill <jason@redhat.com>
26857 * tree.c (build_array_type_1): Revert earlier change.
26859 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26861 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26863 * config/aarch64/aarch64-tune.md: Re-generate.
26865 2013-01-02 Richard Biener <rguenther@suse.de>
26867 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26868 invariant load do not generate a vector load from the scalar location.
26870 2013-01-02 Richard Biener <rguenther@suse.de>
26873 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26874 * configure: Regenerate.
26876 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26878 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26879 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26880 (expand_builtin_int_roundingfn_2): Keep the original target around
26881 for the fallback case.
26883 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26885 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26886 to be clear for sign changes.
26888 2013-01-01 Jan Hubicka <jh@suse.cz>
26890 * ipa-inline-analysis.c: Fix formatting.
26892 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26894 PR tree-optimization/55831
26895 * tree-vect-loop.c (get_initial_def_for_induction): Use
26896 gsi_after_labels instead of gsi_start_bb.
26898 Copyright (C) 2013 Free Software Foundation, Inc.
26900 Copying and distribution of this file, with or without modification,
26901 are permitted in any medium without royalty provided the copyright
26902 notice and this notice are preserved.