1 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3 PR tree-optimization/58697
4 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5 estimated_loop_iterations_int.
6 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7 (get_max_loop_iterations_int): New. HWINT version of
8 get_max_loop_iterations.
9 * cfgloop.h: Add prototypes.
10 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
11 * loop-unroll.c (decide_peel_once_rolling): Call
12 get_estimated_loop_iterations_int.
13 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
14 * tree-ssa-loop-niter.h: Tweak prototypes.
16 2013-10-16 David Malcolm <dmalcolm@redhat.com>
18 * gengtype-parse.c (struct_field_seq): Ignore access-control
19 keywords ("public:" etc).
21 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
23 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
24 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
26 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
28 * config/arm/arm.opt (mlra): New option.
29 * config/arm/arm.c (arm_lra_p): New function.
30 (TARGET_LRA_P): Define.
32 2013-10-16 Paulo Matos <pmatos@broadcom.com>
34 * tree-core.h (tree_code_name): Remove.
35 * tree.h (get_tree_code_name): New prototype.
36 * tree.c (tree_code_name): Make static.
37 (get_tree_code_name): New function.
38 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
39 tree_class_check_failed, tree_range_check_failed,
40 tree_not_class_check_failed, omp_clause_check_failed,
41 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
42 wrapper get_tree_code_name instead of calling tree_code_name directly.
43 * tree-vrp.c (dump_asserts_for): Likewise.
44 * tree-dump.c (dequeue_and_dump): Likewise.
45 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
46 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
47 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
48 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
49 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
50 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
51 dump_gimple_omp_for): Likewise.
52 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
53 * tree-ssa-pre.c (print_pre_expr): Likewise.
54 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
55 * print-tree.c (print_node_brief, print_node): Likewise.
56 * gimple.c (gimple_check_failed): Likewise.
57 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
58 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
59 * config/mep/mep.c (mep_validate_vliw): Likewise.
60 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
61 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
63 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
65 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
68 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
70 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
71 (MULTILIB_MATCHES): Add multilib for -march=v8.
73 2013-10-15 Sriraman Tallam <tmsriram@google.com>
76 * optc-save-gen.awk: Add extra parameter to the save and restore
78 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
79 * tree.c (build_optimization_node): New parameter. Add extra parameter
80 to call to cl_optimization_save.
81 (build_target_option_node): New parameter. Add extra parameter
82 to call to cl_target_option_save.
83 * tree.h (build_optimization_node): New parameter.
84 (build_target_option_node): New parameter.
85 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
86 build_optimization_node and build_target_option_node.
87 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
88 (handle_pragma_push_options): Ditto.
89 * toplev.c (process_options): Ditto.
90 * opts.c (init_options_struct): Check for opts_set non-null.
91 * target.def (target_option.save): New parameter.
92 (target_option.restore): New parameter.
94 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
95 (ix86_pragma_target_parse): Ditto.
96 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
98 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
99 to build_optimization_node and build_target_option_node.
100 (rs6000_valid_attribute_p): Ditto.
101 (rs6000_pragma_target_parse): Ditto.
102 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
104 * config/i386/i386.h:
105 TARGET_64BIT_P: New Macro
106 TARGET_MMX_P: New Macro.
107 TARGET_3DNOW_P: New Macro.
108 TARGET_3DNOW_A_P: New Macro.
109 TARGET_SSE_P: New Macro.
110 TARGET_SSE2_P: New Macro.
111 TARGET_SSE3_P: New Macro.
112 TARGET_SSSE3_P: New Macro.
113 TARGET_SSE4_1_P: New Macro.
114 TARGET_SSE4_2_P: New Macro.
115 TARGET_AVX_P: New Macro.
116 TARGET_AVX2_P: New Macro.
117 TARGET_AVX512F_P: New Macro.
118 TARGET_AVX512PF_P: New Macro.
119 TARGET_AVX512ER_P: New Macro.
120 TARGET_AVX512CD_P: New Macro.
121 TARGET_FMA_P: New Macro.
122 TARGET_SSE4A_P: New Macro.
123 TARGET_FMA4_P: New Macro.
124 TARGET_XOP_P: New Macro.
125 TARGET_LWP_P: New Macro.
126 TARGET_ABM_P: New Macro.
127 TARGET_BMI_P: New Macro.
128 TARGET_BMI2_P: New Macro.
129 TARGET_LZCNT_P: New Macro.
130 TARGET_TBM_P: New Macro.
131 TARGET_POPCNT_P: New Macro.
132 TARGET_SAHF_P: New Macro.
133 TARGET_MOVBE_P: New Macro.
134 TARGET_CRC32_P: New Macro.
135 TARGET_AES_P: New Macro.
136 TARGET_PCLMUL_P: New Macro.
137 TARGET_CMPXCHG16B_P: New Macro.
138 TARGET_FSGSBASE_P: New Macro.
139 TARGET_RDRND_P: New Macro.
140 TARGET_F16C_P: New Macro.
141 TARGET_RTM_P: New Macro.
142 TARGET_HLE_P: New Macro.
143 TARGET_RDSEED_P: New Macro.
144 TARGET_PRFCHW_P: New Macro.
145 TARGET_ADX_P: New Macro.
146 TARGET_FXSR_P: New Macro.
147 TARGET_XSAVE_P: New Macro.
148 TARGET_XSAVEOPT_P: New Macro.
149 TARGET_LP64_P: New Macro.
150 TARGET_X32_P: New Macro.
151 TARGET_FPMATH_DEFAULT_P: New Macro.
152 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
153 * config/i386/i386.c (ix86_option_override_internal): New parameters.
155 Change ix86_tune_string to access opts->x_ix86_tune_string.
156 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
157 Change ix86_arch_string to access opts->x_ix86_arch_string.
158 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
159 Change ix86_pmode to access opts->x_ix86_pmode.
160 Change ix86_abi to access opts->x_ix86_abi.
161 Change ix86_cmodel to access opts->x_ix86_cmodel.
162 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
163 Change ix86_isa_flags_explicit to access
164 opts->x_ix86_isa_flags_explicit.
165 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
166 Change ix86_regparm to access opts->x_ix86_regparm.
167 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
168 Change ix86_preferred_stack_boundary_arg to access
169 opts->x_ix86_preferred_stack_boundary_arg.
170 Change ix86_force_align_arg_pointer to access
171 opts->x_ix86_force_align_arg_pointer.
172 Change ix86_incoming_stack_boundar_arg to access
173 opts->x_ix86_incoming_stack_boundar_arg.
174 Change ix86_fpmath to access opts->x_ix86_fpmath.
175 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
176 Change ix86_recip_name to access opts->x_ix86_recip_name.
177 Change ix86_stack_protector_guard to access
178 opts->x_ix86_stack_protector_guard.
179 Change ix86_tune_memcpy_strategy to access
180 opts->x_ix86_tune_memcpy_strategy.
181 Change ix86_tune_memset_strategy to access
182 opts->x_ix86_tune_memset_strategy.
183 Change global_options to access opts.
184 Change global_options_set to access opts_set.
185 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
186 Change TARGET_MMX to TARGET_MMX_P (opts->...)
187 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...)
188 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...)
189 Change TARGET_SSE to TARGET_SSE_P (opts->...)
190 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
191 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...)
192 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...)
193 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...)
194 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...)
195 Change TARGET_AVX to TARGET_AVX_P (opts->...)
196 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...)
197 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...)
198 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...)
199 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...)
200 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...)
201 Change TARGET_FMA to TARGET_FMA_P (opts->...)
202 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...)
203 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...)
204 Change TARGET_XOP to TARGET_XOP_P (opts->...)
205 Change TARGET_LWP to TARGET_LWP_P (opts->...)
206 Change TARGET_ABM to TARGET_ABM_P (opts->...)
207 Change TARGET_BMI to TARGET_BMI_P (opts->...)
208 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...)
209 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...)
210 Change TARGET_TBM to TARGET_TBM_P (opts->...)
211 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...)
212 Change TARGET_SAHF to TARGET_SAHF_P (opts->...)
213 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...)
214 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...)
215 Change TARGET_AES to TARGET_AES_P (opts->...)
216 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...)
217 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...)
218 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...)
219 Change TARGET_RDRND to TARGET_RDRND_P (opts->...)
220 Change TARGET_F16C to TARGET_F16C_P (opts->...)
221 Change TARGET_RTM to TARGET_RTM_P (opts->...)
222 Change TARGET_HLE to TARGET_HLE_P (opts->...)
223 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...)
224 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...)
225 Change TARGET_ADX to TARGET_ADX_P (opts->...)
226 Change TARGET_FXSR to TARGET_FXSR_P (opts->...)
227 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...)
228 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...)
229 Change TARGET_LP64 to TARGET_LP64_P (opts->...)
230 Change TARGET_X32 to TARGET_X32_P (opts->...)
231 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...)
232 Change TARGET_FLOAT_RETURNS_IN_80387 to
233 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...)
234 (ix86_function_specific_save): New parameter. Use opts-> fields
235 to replace global fields.
236 (ix86_function_specific_restore): Ditto.
237 (ix86_valid_target_attribute_inner_p): New parameters.
239 Fix call to ix86_handle_option and set_option.
240 (ix86_valid_target_attribute_tree): New parameters.
241 Change global_options to access opts.
242 Change global_options_set to access opts_set.
243 Fix call to ix86_valid_target_attribute_inner_p.
244 Change ix86_tune_string to access opts->x_ix86_tune_string.
245 Change ix86_arch_string to access opts->x_ix86_arch_string.
246 Change ix86_fpmath to access opts->x_ix86_fpmath
247 Fix call to ix86_option_override_internal.
248 Fix call to ix86_add_new_builtins.
249 Fix calls to build_optimization_node and build_target_option_node.
250 (ix86_valid_target_attribute_p): Remove access to global_options.
251 Use new gcc_options structure func_options.
252 Fix call to ix86_valid_target_attribute_tree.
253 Fix call to build_optimization_node.
254 (get_builtin_code_for_version): Fix call to
255 ix86_valid_target_attribute_tree.
257 2013-10-15 David Malcolm <dmalcolm@redhat.com>
259 * Makefile.in (PICFLAG): New.
260 (enable_host_shared): New.
261 (INTERNAL_CFLAGS): Use PICFLAG.
262 (LIBIBERTY): Use pic build of libiberty.a if configured with
263 --enable-host-shared.
264 * configure.ac: Add --enable-host-shared, setting up new
266 * configure: Regenerate.
267 * doc/install.texi (--enable-shared): Add note contrasting it
269 (--enable-host-shared): New option.
271 2013-10-15 Richard Biener <rguenther@suse.de>
273 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
274 for built-in functions.
276 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
278 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
279 (optimize_range_tests_1): New function,
280 extracted from optimize_range_tests.
281 (optimize_range_tests_xor): Similarly.
282 (optimize_range_tests_diff): New function.
283 (optimize_range_tests): Use optimize_range_tests_1.
285 2013-10-15 Cong Hou <congh@google.com>
287 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
288 requirement of the reduction pattern so that one operand of the
289 reduction operation can come from outside of the loop.
291 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
293 * config/arm/neon-schedgen.ml: Remove.
294 * config/arm/cortex-a9-neon.md: Remove comment regarding
297 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
299 * config/arm/types: Remove old neon types.
301 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
303 * config/arm/cortex-a7.md
304 (cortex_a7_neon_type): New.
305 (cortex_a7_neon_mul): Update for new types.
306 (cortex_a7_neon_mla): Likewise.
307 (cortex_a7_neon): Likewise.
309 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
311 * config/arm/cortex-a15-neon.md
312 (cortex_a15_neon_type): New,
314 (cortex_a15_neon_int_1): Remove.
315 (cortex_a15_neon_int_2): Likewise.
316 (cortex_a15_neon_int_3): Likewise.
317 (cortex_a15_neon_int_4): Likewise.
318 (cortex_a15_neon_int_5): Likewise.
319 (cortex_a15_neon_vqneg_vqabs): Likewise.
320 (cortex_a15_neon_vmov): Likewise.
321 (cortex_a15_neon_vaba): Likewise.
322 (cortex_a15_neon_vaba_qqq): Likewise.
323 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
324 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
325 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
327 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
328 (cortex_a15_neon_mla_qqq_8_16): Likewise.
329 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
330 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
331 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
332 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
333 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
334 (cortex_a15_neon_shift_1): Likewise.
335 (cortex_a15_neon_shift_2): Likewise.
336 (cortex_a15_neon_shift_3): Likewise.
337 (cortex_a15_neon_vshl_ddd): Likewise.
338 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
339 (cortex_a15_neon_vsra_vrsra): Likewise.
340 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
341 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
342 (cortex_a15_neon_bp_3cycle): Likewise.
343 (cortex_a15_neon_ldm_2): Likewise.
344 (cortex_a15_neon_stm_2): Likewise.
345 (cortex_a15_neon_mcr): Likewise.
346 (cortex_a15_neon_mrc): Likewise.
347 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
348 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
349 (cortex_a15_neon_fp_vmul_ddd): Likewise.
350 (cortex_a15_neon_fp_vmul_qqd): Likewise.
351 (cortex_a15_neon_fp_vmla_ddd): Likewise.
352 (cortex_a15_neon_fp_vmla_qqq): Likewise.
353 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
354 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
355 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
356 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
357 (cortex_a15_neon_bp_simple): Likewise.
358 (cortex_a15_neon_bp_2cycle): Likewise.
359 (cortex_a15_neon_bp_3cycle): Likewise.
360 (cortex_a15_neon_vld1_1_2_regs): Likewise.
361 (cortex_a15_neon_vld1_3_4_regs): Likewise.
362 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
363 (cortex_a15_neon_vld2_4_regs): Likewise.
364 (cortex_a15_neon_vld3_vld4): Likewise.
365 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
366 (cortex_a15_neon_vst1_3_4_regs): Likewise.
367 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
368 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
369 (cortex_a15_neon_vst3_vst4): Rename to...
370 (cortex_a15_neon_vst4): This, update for new attributes.
371 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
372 (cortex_a15_neon_vld3_vld4_lane): Likewise.
373 (cortex_a15_neon_vst1_vst2_lane): Likewise.
374 (cortex_a15_neon_vst3_vst4_lane): Likewise.
375 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
376 (cortex_a15_neon_ldm_2): Likewise.
377 (cortex_a15_neon_stm_2): Likewise.
378 (cortex_a15_neon_mcr): Likewise.
379 (cortex_a15_neon_mcr_2_mcrr): Likewise.
380 (cortex_a15_neon_mrc): Likewise.
381 (cortex_a15_neon_mrrc): Likewise.
383 (cortex_a15_neon_abd): New.
384 (cortex_a15_neon_abd_q): Likewise.
385 (cortex_a15_neon_aba): Likewise.
386 (cortex_a15_neon_aba_q): Likewise.
387 (cortex_a15_neon_acc): Likewise.
388 (cortex_a15_neon_acc_q): Likewise.
389 (cortex_a15_neon_arith_basic): Likewise.
390 (cortex_a15_neon_arith_complex): Likewise.
391 (cortex_a15_neon_multiply): Likewise.
392 (cortex_a15_neon_multiply_q): Likewise.
393 (cortex_a15_neon_mla): Likewise.
394 (cortex_a15_neon_mla_q): Likewise.
395 (cortex_a15_neon_sat_mla_long): Likewise.
396 (cortex_a15_neon_shift_acc): Likewise.
397 (cortex_a15_neon_shift_imm_basic): Likewise.
398 (cortex_a15_neon_shift_imm_complex): Likewise.
399 (cortex_a15_neon_shift_reg_basic): Likewise.
400 (cortex_a15_neon_shift_reg_basic_q): Likewise.
401 (cortex_a15_neon_shift_reg_complex): Likewise.
402 (cortex_a15_neon_shift_reg_complex_q): Likewise.
403 (cortex_a15_neon_fp_negabs): Likewise
404 (cortex_a15_neon_fp_arith): Likewise
405 (cortex_a15_neon_fp_arith_q): Likewise
406 (cortex_a15_neon_fp_cvt_int): Likewise
407 (cortex_a15_neon_fp_cvt_int_q): Likewise
408 (cortex_a15_neon_fp_cvt_16): Likewise
409 (cortex_a15_neon_fp_mul): Likewise
410 (cortex_a15_neon_fp_mul_q): Likewise
411 (cortex_a15_neon_fp_mla): Likewise
412 (cortex_a15_neon_fp_mla_q): Likewise
413 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
414 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
415 (cortex_a15_neon_bitops): Likewise.
416 (cortex_a15_neon_bitops_q): Likewise.
417 (cortex_a15_neon_from_gp): Likewise.
418 (cortex_a15_neon_from_gp_q): Likewise.
419 (cortex_a15_neon_tbl3_tbl4): Likewise.
420 (cortex_a15_neon_zip_q): Likewise.
421 (cortex_a15_neon_to_gp): Likewise.
422 (cortex_a15_neon_load_a): Likewise.
423 (cortex_a15_neon_load_b): Likewise.
424 (cortex_a15_neon_load_c): Likewise.
425 (cortex_a15_neon_load_d): Likewise.
426 (cortex_a15_neon_load_e): Likewise.
427 (cortex_a15_neon_load_f): Likewise.
428 (cortex_a15_neon_store_a): Likewise.
429 (cortex_a15_neon_store_b): Likewise.
430 (cortex_a15_neon_store_c): Likewise.
431 (cortex_a15_neon_store_d): Likewise.
432 (cortex_a15_neon_store_e): Likewise.
433 (cortex_a15_neon_store_f): Likewise.
434 (cortex_a15_neon_store_g): Likewise.
435 (cortex_a15_neon_store_h): Likewise.
436 (cortex_a15_vfp_to_from_gp): Likewise.
438 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
440 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
442 (cortex_a9_neon_vshl_ddd): Remove.
443 (cortex_a9_neon_vst3_vst4): Likewise.
444 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
446 (cortex_a9_neon_bit_ops_q): New.
448 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
449 (cortex_a9_neon_int_2): Likewise.
450 (cortex_a9_neon_int_3): Likewise.
451 (cortex_a9_neon_int_4): Likewise.
452 (cortex_a9_neon_int_5): Likewise.
453 (cortex_a9_neon_vqneg_vqabs): Likewise.
454 (cortex_a9_neon_vmov): Likewise.
455 (cortex_a9_neon_vaba): Likewise.
456 (cortex_a9_neon_vaba_qqq): Likewise.
457 (cortex_a9_neon_shift_1): Likewise.
458 (cortex_a9_neon_shift_2): Likewise.
459 (cortex_a9_neon_shift_3): Likewise.
460 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
461 (cortex_a9_neon_vsra_vrsra): Likewise.
462 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
463 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
464 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
466 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
467 (cortex_a9_neon_mla_qqq_8_16): Likewise.
468 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
470 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
471 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
472 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
473 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
474 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
475 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
476 (cortex_a9_neon_fp_vsum): Likewise.
477 (cortex_a9_neon_fp_vmul_ddd): Likewise.
478 (cortex_a9_neon_fp_vmul_qqd): Likewise.
479 (cortex_a9_neon_fp_vmla_ddd): Likewise.
480 (cortex_a9_neon_fp_vmla_qqq): Likewise.
481 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
482 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
483 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
484 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
485 (cortex_a9_neon_bp_simple): Likewise.
486 (cortex_a9_neon_bp_2cycle): Likewise.
487 (cortex_a9_neon_bp_3cycle): Likewise.
488 (cortex_a9_neon_ldr): Likewise.
489 (cortex_a9_neon_str): Likewise.
490 (cortex_a9_neon_vld1_1_2_regs): Likewise.
491 (cortex_a9_neon_vld1_3_4_regs): Likewise.
492 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
493 (cortex_a9_neon_vld2_4_regs): Likewise.
494 (cortex_a9_neon_vld3_vld4): Likewise.
495 (cortex_a9_neon_vld1_vld2_lane): Likewise.
496 (cortex_a9_neon_vld3_vld4_lane): Likewise.
497 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
498 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
499 (cortex_a9_neon_vst1_3_4_regs): Likewise.
500 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
501 (cortex_a9_neon_vst1_vst2_lane): Likewise.
502 (cortex_a9_neon_vst3_vst4_lane): Likewise.
503 (cortex_a9_neon_mcr): Likewise.
504 (cortex_a9_neon_mcr_2_mcrr): Likewise.
505 (cortex_a9_neon_mrc): Likewise.
506 (cortex_a9_neon_mrrc): Likewise.
508 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
510 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
512 (cortex_a8_neon_vshl_ddd): Remove.
513 (cortex_a8_neon_vst3_vst4): Likewise.
514 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
516 (cortex_a8_neon_bit_ops_q): New.
518 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
519 (cortex_a8_neon_int_2): Likewise..
520 (cortex_a8_neon_int_3): Likewise.
521 (cortex_a8_neon_int_5): Likewise.
522 (cortex_a8_neon_vqneg_vqabs): Likewise.
523 (cortex_a8_neon_int_4): Likewise.
524 (cortex_a8_neon_vaba): Likewise.
525 (cortex_a8_neon_vaba_qqq): Likewise.
526 (cortex_a8_neon_shift_1): Likewise.
527 (cortex_a8_neon_shift_2): Likewise.
528 (cortex_a8_neon_shift_3): Likewise.
529 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
530 (cortex_a8_neon_vsra_vrsra): Likewise.
531 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
532 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
533 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
535 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
536 (cortex_a8_neon_mla_qqq_8_16): Likewise.
537 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
539 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
540 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
541 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
542 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
543 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
544 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
545 (cortex_a8_neon_fp_vsum): Likewise.
546 (cortex_a8_neon_fp_vmul_ddd): Likewise.
547 (cortex_a8_neon_fp_vmul_qqd): Likewise.
548 (cortex_a8_neon_fp_vmla_ddd): Likewise.
549 (cortex_a8_neon_fp_vmla_qqq): Likewise.
550 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
551 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
552 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
553 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
554 (cortex_a8_neon_bp_simple): Likewise.
555 (cortex_a8_neon_bp_2cycle): Likewise.
556 (cortex_a8_neon_bp_3cycle): Likewise.
557 (cortex_a8_neon_ldr): Likewise.
558 (cortex_a8_neon_str): Likewise.
559 (cortex_a8_neon_vld1_1_2_regs): Likewise.
560 (cortex_a8_neon_vld1_3_4_regs): Likewise.
561 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
562 (cortex_a8_neon_vld2_4_regs): Likewise.
563 (cortex_a8_neon_vld3_vld4): Likewise.
564 (cortex_a8_neon_vld1_vld2_lane): Likewise.
565 (cortex_a8_neon_vld3_vld4_lane): Likewise.
566 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
567 (cortex_a8_neon_vst1_3_4_regs): Likewise.
568 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
569 (cortex_a8_neon_vst1_vst2_lane): Likewise.
570 (cortex_a8_neon_vst3_vst4_lane): Likewise.
571 (cortex_a8_neon_mcr): Likewise.
572 (cortex_a8_neon_mcr_2_mcrr): Likewise.
573 (cortex_a8_neon_mrc): Likewise.
574 (cortex_a8_neon_mrrc): Likewise.
576 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
578 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
580 * config/aarch64/aarch64-simd.md (neon_type): Remove.
581 (aarch64_simd_dup<mode>): Add "type" attribute.
582 (aarch64_dup_lane<mode>): Likewise.
583 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
584 (*aarch64_simd_mov<mode>): Likewise.
585 (aarch64_simd_mov_from_<mode>low): Likewise.
586 (aarch64_simd_mov_from_<mode>high): Likewise.
587 (orn<mode>3): Likewise.
588 (bic<mode>3): Likewise.
589 (add<mode>3): Likewise.
590 (sub<mode>3): Likewise.
591 (mul<mode>3): Likewise.
592 (*aarch64_mul3_elt<mode>): Likewise.
593 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
594 (*aarch64_mul3_elt_to_128df): Likewise.
595 (*aarch64_mul3_elt_to_64v2df): Likewise.
596 (neg<mode>2): Likewise.
597 (abs<mode>2): Likewise.
598 (abd<mode>_3): Likewise.
599 (aba<mode>_3): Likewise.
600 (fabd<mode>_3): Likewise.
601 (*fabd_scalar<mode>3): Likewise.
602 (and<mode>3): Likewise.
603 (ior<mode>3): Likewise.
604 (xor<mode>3): Likewise.
605 (one_cmpl<mode>2): Likewise.
606 (aarch64_simd_vec_set<mode>): Likewise.
607 (aarch64_simd_lshr<mode>): Likewise.
608 (aarch64_simd_ashr<mode>): Likewise.
609 (aarch64_simd_imm_shl<mode>): Likewise.
610 (aarch64_simd_reg_sshl<mode): Likewise.
611 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
612 (aarch64_simd_reg_shl<mode>_signed): Likewise.
613 (aarch64_simd_vec_setv2di): Likewise.
614 (aarch64_simd_vec_set<mode>): Likewise.
615 (aarch64_mla<mode>): Likewise.
616 (*aarch64_mla_elt<mode>): Likewise.
617 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
618 (aarch64_mls<mode>): Likewise.
619 (*aarch64_mls_elt<mode>): Likewise.
620 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
621 (<su><maxmin><mode>3): Likewise.
622 (move_lo_quad_<mode>): Likewise.
623 (aarch64_simd_move_hi_quad_<mode>): Likewise.
624 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
625 (vec_pack_trunc_<mode>): Likewise.
626 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
627 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
628 (*aarch64_<su>mlal_lo<mode>): Likewise.
629 (*aarch64_<su>mlal_hi<mode>): Likewise.
630 (*aarch64_<su>mlsl_lo<mode>): Likewise.
631 (*aarch64_<su>mlsl_hi<mode>): Likewise.
632 (*aarch64_<su>mlal<mode>): Likewise.
633 (*aarch64_<su>mlsl<mode>): Likewise.
634 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
635 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
636 (add<mode>3): Likewise.
637 (sub<mode>3): Likewise.
638 (mul<mode>3): Likewise.
639 (div<mode>3): Likewise.
640 (neg<mode>2): Likewise.
641 (abs<mode>2): Likewise.
642 (fma<mode>4): Likewise.
643 (*aarch64_fma4_elt<mode>): Likewise.
644 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
645 (*aarch64_fma4_elt_to_128df): Likewise.
646 (*aarch64_fma4_elt_to_64v2df): Likewise.
647 (fnma<mode>4): Likewise.
648 (*aarch64_fnma4_elt<mode>): Likewise.
649 (*aarch64_fnma4_elt_<vswap_width_name><mode>
650 (*aarch64_fnma4_elt_to_128df): Likewise.
651 (*aarch64_fnma4_elt_to_64v2df): Likewise.
652 (<frint_pattern><mode>2): Likewise.
653 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
654 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
655 (vec_unpacks_lo_v4sf): Likewise.
656 (aarch64_float_extend_lo_v2df): Likewise.
657 (vec_unpacks_hi_v4sf): Likewise.
658 (aarch64_float_truncate_lo_v2sf): Likewise.
659 (aarch64_float_truncate_hi_v4sf): Likewise.
660 (aarch64_vmls<mode>): Likewise.
661 (<su><maxmin><mode>3): Likewise.
662 (<maxmin_uns><mode>3): Likewise.
663 (reduc_<sur>plus_<mode>): Likewise.
664 (reduc_<sur>plus_v2di): Likewise.
665 (reduc_<sur>plus_v2si): Likewise.
666 (reduc_<sur>plus_<mode>): Likewise.
667 (aarch64_addpv4sf): Likewise.
668 (clz<mode>2): Likewise.
669 (reduc_<maxmin_uns>_<mode>): Likewise.
670 (reduc_<maxmin_uns>_v2di): Likewise.
671 (reduc_<maxmin_uns>_v2si): Likewise.
672 (reduc_<maxmin_uns>_<mode>): Likewise.
673 (reduc_<maxmin_uns>_v4sf): Likewise.
674 (aarch64_simd_bsl<mode>_internal): Likewise.
675 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
676 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
677 (aarch64_get_lane<mode>): Likewise.
678 (*aarch64_combinez<mode>): Likewise.
679 (aarch64_combine<mode>): Likewise.
680 (aarch64_simd_combine<mode>): Likewise.
681 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
682 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
683 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
684 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
685 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
686 (aarch64_<sur>h<addsub><mode>): Likewise.
687 (aarch64_<sur><addsub>hn<mode>): Likewise.
688 (aarch64_<sur><addsub>hn2<mode>): Likewise.
689 (aarch64_pmul<mode>): Likewise.
690 (aarch64_<su_optab><optab><mode>): Likewise.
691 (aarch64_<sur>qadd<mode>): Likewise.
692 (aarch64_sqmovun<mode>): Likewise.
693 (aarch64_<sur>qmovn<mode>): Likewise.
694 (aarch64_s<optab><mode>): Likewise.
695 (aarch64_sq<r>dmulh<mode>): Likewise.
696 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
697 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
698 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
699 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
700 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
701 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
702 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
703 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
704 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
705 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
706 (aarch64_sqdmull<mode>): Likewise.
707 (aarch64_sqdmull_lane<mode>_internal): Likewise.
708 (aarch64_sqdmull_n<mode>): Likewise.
709 (aarch64_sqdmull2<mode>_internal): Likewise.
710 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
711 (aarch64_sqdmull2_n<mode>_internal): Likewise.
712 (aarch64_<sur>shl<mode>): Likewise.
713 (aarch64_<sur>q<r>shl<mode>
714 (aarch64_<sur>shll_n<mode>): Likewise.
715 (aarch64_<sur>shll2_n<mode>): Likewise.
716 (aarch64_<sur>shr_n<mode>): Likewise.
717 (aarch64_<sur>sra_n<mode>): Likewise.
718 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
719 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
720 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
721 (aarch64_cm<optab><mode>): Likewise.
722 (aarch64_cm<optab>di): Likewise.
723 (aarch64_cm<optab><mode>): Likewise.
724 (aarch64_cm<optab>di): Likewise.
725 (aarch64_cmtst<mode>): Likewise.
726 (aarch64_cmtstdi): Likewise.
727 (aarch64_cm<optab><mode>): Likewise.
728 (*aarch64_fac<optab><mode>): Likewise.
729 (aarch64_addp<mode>): Likewise.
730 (aarch64_addpdi): Likewise.
731 (sqrt<mode>2): Likewise.
732 (vec_load_lanesoi<mode>): Likewise.
733 (vec_store_lanesoi<mode>): Likewise.
734 (vec_load_lanesci<mode>): Likewise.
735 (vec_store_lanesci<mode>): Likewise.
736 (vec_load_lanesxi<mode>): Likewise.
737 (vec_store_lanesxi<mode>): Likewise.
738 (*aarch64_mov<mode>): Likewise.
739 (aarch64_ld2<mode>_dreg): Likewise.
740 (aarch64_ld2<mode>_dreg): Likewise.
741 (aarch64_ld3<mode>_dreg): Likewise.
742 (aarch64_ld3<mode>_dreg): Likewise.
743 (aarch64_ld4<mode>_dreg): Likewise.
744 (aarch64_ld4<mode>_dreg): Likewise.
745 (aarch64_tbl1<mode>): Likewise.
746 (aarch64_tbl2v16qi): Likewise.
747 (aarch64_combinev16qi): Likewise.
748 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
749 (aarch64_st2<mode>_dreg): Likewise.
750 (aarch64_st2<mode>_dreg): Likewise.
751 (aarch64_st3<mode>_dreg): Likewise.
752 (aarch64_st3<mode>_dreg): Likewise.
753 (aarch64_st4<mode>_dreg): Likewise.
754 (aarch64_st4<mode>_dreg): Likewise.
755 (*aarch64_simd_ld1r<mode>): Likewise.
756 (aarch64_frecpe<mode>): Likewise.
757 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
758 (aarch64_frecps<mode>): Likewise.
760 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
762 * config/arm/iterators.md (V_elem_ch): New.
764 (VQH_type): Likewise.
765 * config/arm/arm.md (is_neon_type): New.
766 (conds): Use is_neon_type.
767 (anddi3_insn): Update type attribute.
768 (xordi3_insn): Likewise.
769 (one_cmpldi2): Likewise.
770 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
771 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
772 (*movmisalign<mode>_neon_store): Likewise.
773 (*movmisalign<mode>_neon_load): Likewise.
774 (vec_set<mode>_internal): Likewise.
775 (vec_set<mode>_internal): Likewise.
776 (vec_setv2di_internal): Likewise.
777 (vec_extract<mode>): Likewise.
778 (vec_extract<mode>): Likewise.
779 (vec_extractv2di): Likewise.
780 (*add<mode>3_neon): Likewise.
781 (adddi3_neon): Likewise.
782 (*sub<mode>3_neon): Likewise.
783 (subdi3_neon): Likewise.
784 (fma<VCVTF:mode>4): Likewise.
785 (fma<VCVTF:mode>4_intrinsic): Likewise.
786 (*fmsub<VCVTF:mode>4): Likewise.
787 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
788 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
789 (ior<mode>3): Likewise.
790 (and<mode>3): Likewise.
791 (orn<mode>3_neon): Likewise.
792 (orndi3_neon): Likewise.
793 (bic<mode>3_neon): Likewise.
794 (bicdi3_neon): Likewise.
795 (xor<mode>3): Likewise.
796 (one_cmpl<mode>2): Likewise.
797 (abs<mode>2): Likewise.
798 (neg<mode>2): Likewise.
799 (negdi2_neon): Likewise.
800 (*umin<mode>3_neon): Likewise.
801 (*umax<mode>3_neon): Likewise.
802 (*smin<mode>3_neon): Likewise.
803 (*smax<mode>3_neon): Likewise.
804 (vashl<mode>3): Likewise.
805 (vashr<mode>3_imm): Likewise.
806 (vlshr<mode>3_imm): Likewise.
807 (ashl<mode>3_signed): Likewise.
808 (ashl<mode>3_unsigned): Likewise.
809 (neon_load_count): Likewise.
810 (ashldi3_neon_noclobber): Likewise.
811 (ashldi3_neon): Likewise.
812 (signed_shift_di3_neon): Likewise.
813 (unsigned_shift_di3_neon): Likewise.
814 (ashrdi3_neon_imm_noclobber): Likewise.
815 (lshrdi3_neon_imm_noclobber): Likewise.
816 (<shift>di3_neon): Likewise.
817 (widen_ssum<mode>3): Likewise.
818 (widen_usum<mode>3): Likewise.
819 (quad_halves_<code>v4si): Likewise.
820 (quad_halves_<code>v4sf): Likewise.
821 (quad_halves_<code>v8hi): Likewise.
822 (quad_halves_<code>v16qi): Likewise.
823 (reduc_splus_v2di): Likewise.
824 (neon_vpadd_internal<mode>): Likewise.
825 (neon_vpsmin<mode>): Likewise.
826 (neon_vpsmax<mode>): Likewise.
827 (neon_vpumin<mode>): Likewise.
828 (neon_vpumax<mode>): Likewise.
829 (*ss_add<mode>_neon): Likewise.
830 (*us_add<mode>_neon): Likewise.
831 (*ss_sub<mode>_neon): Likewise.
832 (*us_sub<mode>_neon): Likewise.
833 (neon_vadd<mode>_unspec): Likewise.
834 (neon_vaddl<mode>): Likewise.
835 (neon_vaddw<mode>): Likewise.
836 (neon_vhadd<mode>): Likewise.
837 (neon_vqadd<mode>): Likewise.
838 (neon_vaddhn<mode>): Likewise.
839 (neon_vmul<mode>): Likewise.
840 (neon_vfms<VCVTF:mode>): Likewise.
841 (neon_vmlal<mode>): Likewise.
842 (neon_vmls<mode>): Likewise.
843 (neon_vmlsl<mode>): Likewise.
844 (neon_vqdmulh<mode>): Likewise.
845 (neon_vqdmlal<mode>): Likewise.
846 (neon_vqdmlsl<mode>): Likewise.
847 (neon_vmull<mode>): Likewise.
848 (neon_vqdmull<mode>): Likewise.
849 (neon_vsub<mode>_unspec): Likewise.
850 (neon_vsubl<mode>): Likewise.
851 (neon_vsubw<mode>): Likewise.
852 (neon_vqsub<mode>): Likewise.
853 (neon_vhsub<mode>): Likewise.
854 (neon_vsubhn<mode>): Likewise.
855 (neon_vceq<mode>): Likewise.
856 (neon_vcge<mode>): Likewise.
857 (neon_vcgeu<mode>): Likewise.
858 (neon_vcgt<mode>): Likewise.
859 (neon_vcgtu<mode>): Likewise.
860 (neon_vcle<mode>): Likewise.
861 (neon_vclt<mode>): Likewise.
862 (neon_vcage<mode>): Likewise.
863 (neon_vcagt<mode>): Likewise.
864 (neon_vtst<mode>): Likewise.
865 (neon_vabd<mode>): Likewise.
866 (neon_vabdl<mode>): Likewise.
867 (neon_vaba<mode>): Likewise.
868 (neon_vabal<mode>): Likewise.
869 (neon_vmax<mode>): Likewise.
870 (neon_vmin<mode>): Likewise.
871 (neon_vpaddl<mode>): Likewise.
872 (neon_vpadal<mode>): Likewise.
873 (neon_vpmax<mode>): Likewise.
874 (neon_vpmin<mode>): Likewise.
875 (neon_vrecps<mode>): Likewise.
876 (neon_vrsqrts<mode>): Likewise.
877 (neon_vqabs<mode>): Likewise.
878 (neon_vqneg<mode>): Likewise.
879 (neon_vcls<mode>): Likewise.
880 (clz<mode>2): Likewise.
881 (popcount<mode>2): Likewise.
882 (neon_vrecpe<mode>): Likewise.
883 (neon_vrsqrte<mode>): Likewise.
884 (neon_vget_lane<mode>_sext_internal): Likewise.
885 (neon_vget_lane<mode>_zext_internal): Likewise.
886 (neon_vdup_n<mode>): Likewise.
887 (neon_vdup_n<mode>): Likewise.
888 (neon_vdup_nv2di): Likewise.
889 (neon_vdup_lane<mode>_interal): Likewise.
890 (*neon_vswp<mode>): Likewise.
891 (neon_vcombine<mode>): Likewise.
892 (float<mode><V_cvtto>2): Likewise.
893 (floatuns<mode><V_cvtto>2): Likewise.
894 (fix_trunc<mode><V_cvtto>2): Likewise.
895 (fixuns_trunc<mode><V_cvtto>2
896 (neon_vcvt<mode>): Likewise.
897 (neon_vcvt<mode>): Likewise.
898 (neon_vcvtv4sfv4hf): Likewise.
899 (neon_vcvtv4hfv4sf): Likewise.
900 (neon_vcvt_n<mode>): Likewise.
901 (neon_vcvt_n<mode>): Likewise.
902 (neon_vmovn<mode>): Likewise.
903 (neon_vqmovn<mode>): Likewise.
904 (neon_vqmovun<mode>): Likewise.
905 (neon_vmovl<mode>): Likewise.
906 (neon_vmul_lane<mode>): Likewise.
907 (neon_vmul_lane<mode>): Likewise.
908 (neon_vmull_lane<mode>): Likewise.
909 (neon_vqdmull_lane<mode>): Likewise.
910 (neon_vqdmulh_lane<mode>): Likewise.
911 (neon_vqdmulh_lane<mode>): Likewise.
912 (neon_vmla_lane<mode>): Likewise.
913 (neon_vmla_lane<mode>): Likewise.
914 (neon_vmlal_lane<mode>): Likewise.
915 (neon_vqdmlal_lane<mode>): Likewise.
916 (neon_vmls_lane<mode>): Likewise.
917 (neon_vmls_lane<mode>): Likewise.
918 (neon_vmlsl_lane<mode>): Likewise.
919 (neon_vqdmlsl_lane<mode>): Likewise.
920 (neon_vext<mode>): Likewise.
921 (neon_vrev64<mode>): Likewise.
922 (neon_vrev32<mode>): Likewise.
923 (neon_vrev16<mode>): Likewise.
924 (neon_vbsl<mode>_internal): Likewise.
925 (neon_vshl<mode>): Likewise.
926 (neon_vqshl<mode>): Likewise.
927 (neon_vshr_n<mode>): Likewise.
928 (neon_vshrn_n<mode>): Likewise.
929 (neon_vqshrn_n<mode>): Likewise.
930 (neon_vqshrun_n<mode>): Likewise.
931 (neon_vshl_n<mode>): Likewise.
932 (neon_vqshl_n<mode>): Likewise.
933 (neon_vqshlu_n<mode>): Likewise.
934 (neon_vshll_n<mode>): Likewise.
935 (neon_vsra_n<mode>): Likewise.
936 (neon_vsri_n<mode>): Likewise.
937 (neon_vsli_n<mode>): Likewise.
938 (neon_vtbl1v8qi): Likewise.
939 (neon_vtbl2v8qi): Likewise.
940 (neon_vtbl3v8qi): Likewise.
941 (neon_vtbl4v8qi): Likewise.
942 (neon_vtbl1v16qi): Likewise.
943 (neon_vtbl2v16qi): Likewise.
944 (neon_vcombinev16qi): Likewise.
945 (neon_vtbx1v8qi): Likewise.
946 (neon_vtbx2v8qi): Likewise.
947 (neon_vtbx3v8qi): Likewise.
948 (neon_vtbx4v8qi): Likewise.
949 (*neon_vtrn<mode>_insn): Likewise.
950 (*neon_vzip<mode>_insn): Likewise.
951 (*neon_vuzp<mode>_insn): Likewise.
952 (neon_vld1<mode>): Likewise.
953 (neon_vld1_lane<mode>): Likewise.
954 (neon_vld1_lane<mode>): Likewise.
955 (neon_vld1_dup<mode>): Likewise.
956 (neon_vld1_dup<mode>): Likewise.
957 (neon_vld1_dupv2di): Likewise.
958 (neon_vst1<mode>): Likewise.
959 (neon_vst1_lane<mode>): Likewise.
960 (neon_vst1_lane<mode>): Likewise.
961 (neon_vld2<mode>): Likewise.
962 (neon_vld2<mode>): Likewise.
963 (neon_vld2_lane<mode>): Likewise.
964 (neon_vld2_lane<mode>): Likewise.
965 (neon_vld2_dup<mode>): Likewise.
966 (neon_vst2<mode>): Likewise.
967 (neon_vst2<mode>): Likewise.
968 (neon_vst2_lane<mode>): Likewise.
969 (neon_vst2_lane<mode>): Likewise.
970 (neon_vld3<mode>): Likewise.
971 (neon_vld3qa<mode>): Likewise.
972 (neon_vld3qb<mode>): Likewise.
973 (neon_vld3_lane<mode>): Likewise.
974 (neon_vld3_lane<mode>): Likewise.
975 (neon_vld3_dup<mode>): Likewise.
976 (neon_vst3<mode>): Likewise.
977 (neon_vst3qa<mode>): Likewise.
978 (neon_vst3qb<mode>): Likewise.
979 (neon_vst3_lane<mode>): Likewise.
980 (neon_vst3_lane<mode>): Likewise.
981 (neon_vld4<mode>): Likewise.
982 (neon_vld4qa<mode>): Likewise.
983 (neon_vld4qb<mode>): Likewise.
984 (neon_vld4_lane<mode>): Likewise.
985 (neon_vld4_lane<mode>): Likewise.
986 (neon_vld4_dup<mode>): Likewise.
987 (neon_vst4<mode>): Likewise.
988 (neon_vst4qa<mode>): Likewise.
989 (neon_vst4qb<mode>): Likewise.
990 (neon_vst4_lane<mode>): Likewise.
991 (neon_vst4_lane<mode>): Likewise.
992 (neon_vec_unpack<US>_lo_<mode>): Likewise.
993 (neon_vec_unpack<US>_hi_<mode>): Likewise.
994 (neon_vec_<US>mult_lo_<mode>): Likewise.
995 (neon_vec_<US>mult_hi_<mode>): Likewise.
996 (neon_vec_<US>shiftl_<mode>): Likewise.
997 (neon_unpack<US>_<mode>): Likewise.
998 (neon_vec_<US>mult_<mode>): Likewise.
999 (vec_pack_trunc_<mode>): Likewise.
1000 (neon_vec_pack_trunc_<mode>): Likewise.
1001 (neon_vabd<mode>_2): Likewise.
1002 (neon_vabd<mode>_3): Likewise.
1004 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1006 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1007 (load_pair): Update type attribute.
1008 (store_pair): Update type attribute.
1009 * config/aarch64/iterators.md (q): New.
1011 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1013 * config/arm/types.md: Add new types for Neon insns.
1015 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1016 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1017 Sergey Lega <sergey.s.lega@intel.com>
1018 Anna Tikhonova <anna.tikhonova@intel.com>
1019 Ilya Tocar <ilya.tocar@intel.com>
1020 Andrey Turetskiy <andrey.turetskiy@intel.com>
1021 Ilya Verbin <ilya.verbin@intel.com>
1022 Kirill Yukhin <kirill.yukhin@intel.com>
1023 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1025 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1026 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1027 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1029 (srcp14<mode>): Ditto.
1030 (rsqrt14<mode>): Ditto.
1031 (rsqrt14<mode>): Ditto.
1032 (avx512f_vmscalef<mode>): Ditto.
1033 (avx512f_scalef<mode>): Ditto.
1034 (avx512f_getexp<mode>): Ditto.
1035 (avx512f_sgetexp<mode>): Ditto.
1036 (avx512f_fixupimm<mode>): Ditto.
1037 (avx512f_sfixupimm<mode>): Ditto.
1038 (avx512f_rndscale<mode>): Ditto.
1039 (*avx512er_exp2<mode>): Ditto.
1040 (*avx512er_rcp28<mode>): Ditto.
1041 (avx512er_rsqrt28<mode>): Ditto.
1042 (avx512f_getmant<mode>): Ditto.
1043 (avx512f_getmant<mode>): Ditto.
1044 (avx512f_rndscale<mode>): Fix formatting.
1046 2013-10-15 Martin Jambor <mjambor@suse.cz>
1048 * ipa-utils.h (ipa_edge_within_scc): Declare.
1049 * ipa-cp.c (edge_within_scc): Moved...
1050 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1052 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1053 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1054 Sergey Lega <sergey.s.lega@intel.com>
1055 Anna Tikhonova <anna.tikhonova@intel.com>
1056 Ilya Tocar <ilya.tocar@intel.com>
1057 Andrey Turetskiy <andrey.turetskiy@intel.com>
1058 Ilya Verbin <ilya.verbin@intel.com>
1059 Kirill Yukhin <kirill.yukhin@intel.com>
1060 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1062 * config/i386/predicates.md (const_8_to_15_operand): New.
1063 (const_16_to_31_operand): Ditto.
1064 * config/i386/sse.md (V8FI): New.
1066 (reduc_splus_v8df): Ditto.
1067 (reduc_splus_v16sf): Ditto.
1068 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1069 (vec_extract_hi_<mode>): Ditto.
1070 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1071 (vec_set_lo_<mode>): Ditto.
1072 (vec_set_hi_<mode>): Ditto.
1073 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1074 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1075 (avx512f_pshufd_1): Ditto.
1076 (avx512f_broadcast<mode>): Ditto.
1077 (avx512f_broadcast<mode>): Ditto.
1078 (define_split): Split vec_extract_lo into move.
1079 (ssequartermode): Ditto.
1080 (ssedoublemode): Extened with wider modes.
1081 (vec_extract_lo_<mode>): Ditto.
1083 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1084 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1085 Sergey Lega <sergey.s.lega@intel.com>
1086 Anna Tikhonova <anna.tikhonova@intel.com>
1087 Ilya Tocar <ilya.tocar@intel.com>
1088 Andrey Turetskiy <andrey.turetskiy@intel.com>
1089 Ilya Verbin <ilya.verbin@intel.com>
1090 Kirill Yukhin <kirill.yukhin@intel.com>
1091 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1093 * config/i386/predicates.md (register_or_constm1_operand): New.
1094 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1095 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1096 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1097 UNSPEC_SCATTER_PREFETCH
1099 (avx512f_ucmp<mode>3): Ditto.
1100 (avx512f_vternlog<mode>): Ditto.
1101 (avx512f_align<mode>): Ditto.
1102 (<shift_insn><mode>3): Ditto.
1103 (avx512f_<rotate>v<mode>): Ditto.
1104 (avx512f_<rotate><mode>): Ditto.
1105 (avx512f_eq<mode>3): Ditto.
1106 (avx512f_eq<mode>3_1): Ditto.
1107 (avx512f_gt<mode>3): Ditto.
1108 (avx512f_testm<mode>3): Ditto.
1109 (avx512f_testnm<mode>3): Ditto.
1110 (avx512pf_gatherpf<mode>): Ditto.
1111 (*avx512pf_gatherpf<mode>_mask): Ditto.
1112 (*avx512pf_gatherpf<mode>): Ditto.
1113 (avx512pf_scatterpf<mode>): Ditto.
1114 (*avx512pf_scatterpf<mode>_mask): Ditto.
1115 (*avx512pf_scatterpf<mode>): Ditto.
1116 (avx512f_vec_dup_gpr<mode>): Ditto.
1117 (clz<mode>2): Ditto.
1118 (conflict<mode>): Ditto.
1119 (REDUC_SMINMAX_MODE): Extened with wider modes.
1120 (reduc_<code>_<mode>): Ditto.
1121 (vlshr<mode>3): Ditto.
1122 (vashl<mode>3): Ditto.
1124 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1125 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1126 Sergey Lega <sergey.s.lega@intel.com>
1127 Anna Tikhonova <anna.tikhonova@intel.com>
1128 Ilya Tocar <ilya.tocar@intel.com>
1129 Andrey Turetskiy <andrey.turetskiy@intel.com>
1130 Ilya Verbin <ilya.verbin@intel.com>
1131 Kirill Yukhin <kirill.yukhin@intel.com>
1132 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1134 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1137 (avx512fmaskmode): Ditto.
1138 (bcstscalarsuff): Ditto.
1139 (avx512f_blendm<mode>): Ditto.
1140 (cmp_imm_predicate): Ditto.
1141 (avx512f_cmp<mode>3): Ditto.
1142 (avx512f_vec_dup<mode>): Ditto.
1143 (avx512f_vec_dup_mem<mode>): Ditto.
1144 (avx512f_vpermi2var<mode>3): Ditto.
1145 (avx512f_vpermt2var<mode>3): Ditto.
1146 (vec_init<mode>): Ditto.
1147 (avx512f_gathersi<mode>): Ditto.
1148 (*avx512f_gathersi<mode>): Ditto.
1149 (*avx512f_gathersi<mode>_2): Ditto.
1150 (avx512f_gatherdi<mode>): Ditto.
1151 (*avx512f_gatherdi<mode>): Ditto.
1152 (*avx512f_gatherdi<mode>_2): Ditto.
1153 (avx512f_scattersi<mode>): Ditto.
1154 (*avx512f_scattersi<mode>): Ditto.
1155 (avx512f_scatterdi<mode>): Ditto.
1156 (*avx512f_scatterdi<mode>): Ditto.
1157 (sseintprefix): Extened with wider modes.
1158 (VEC_GATHER_IDXSI): Ditto.
1159 (VEC_GATHER_IDXDI): Ditto.
1160 (VEC_GATHER_SRCDI): Ditto.
1162 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1163 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1165 * config/arm/t-aprofile: New file.
1166 * config.gcc: Handle --with-multilib-list option.
1168 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1170 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
1173 2013-10-15 Richard Biener <rguenther@suse.de>
1175 * tree-loop-distribution.c (build_empty_rdg): Inline into
1177 (rdg_flag_vertex): Inline into single user.
1178 (rdg_flag_vertex_and_dependent): Likewise.
1179 (build_rdg_partition_for_vertex): Remove processed bitmap.
1180 (rdg_build_partitions): Simplify.
1182 2013-10-15 Richard Biener <rguenther@suse.de>
1184 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1185 Restructure forwarding through conversions and copies to
1186 avoid performing copy-propagation the wrong way. Adjust
1187 recursion invocations.
1188 (forward_propagate_addr_expr): Add argument stating if we
1189 are recursing from a single-use.
1190 (ssa_forward_propagate_and_combine): Adjust.
1192 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1194 * dumpfile.h (gcc::dump_manager): New class, to hold state
1195 relating to dumpfile management.
1196 (get_dump_file_name): Remove in favor of method of dump_manager.
1197 (dump_initialized_p): Likewise.
1198 (dump_start): Likewise.
1199 (dump_finish): Likewise.
1200 (dump_switch_p): Likewise.
1201 (dump_register): Likewise.
1202 (get_dump_file_info): Likewise.
1203 * context.c (gcc::context::context): Construct the dump_manager
1205 * context.h (gcc::context::get_dumps): New.
1206 (gcc::context::m_dumps): New.
1207 * coverage.c (coverage_init): Port to dump_manager API.
1208 * dumpfile.c (extra_dump_files): Convert to field of
1210 (extra_dump_files_in_use): Likewise.
1211 (extra_dump_files_alloced): Likewise.
1212 (gcc::dump_manager::dump_manager): New.
1213 (dump_register): Convert to...
1214 (gcc::dump_manager::dump_register): ...method, replacing
1215 function-static next_dump with m_next_dump field.
1216 (get_dump_file_info): Convert to...
1217 (gcc::dump_manager::get_dump_file_info): ...method.
1218 (get_dump_file_name): Convert to...
1219 (gcc::dump_manager::get_dump_file_name): ...method.
1220 (dump_start): Convert to...
1221 (gcc::dump_manager::dump_start): ...method.
1222 (dump_finish): Convert to...
1223 (gcc::dump_manager::dump_finish): ...method.
1224 (dump_begin): Replace body with...
1225 (gcc::dump_manager::dump_begin): ...new method.
1226 (dump_phase_enabled_p): Convert to...
1227 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1228 (dump_phase_enabled_p): Convert to...
1229 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1230 (dump_initialized_p): Convert to...
1231 (gcc::dump_manager::dump_initialized_p): ...method.
1232 (dump_flag_name): Replace body with...
1233 (gcc::dump_manager::dump_flag_name): ...new method.
1234 (dump_enable_all): Convert to...
1235 (gcc::dump_manager::dump_enable_all): ...new method.
1236 (opt_info_enable_passes): Convert to...
1237 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1238 (dump_switch_p_1): Convert to...
1239 (gcc::dump_manager::dump_switch_p_1): ...new method.
1240 (dump_switch_p): Convert to...
1241 (gcc::dump_manager::dump_switch_p): ...new method.
1242 (opt_info_switch_p): Port to dump_manager API.
1243 (enable_rtl_dump_file): Likewise.
1244 * opts-global.c (handle_common_deferred_options): Port to new
1246 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1247 (pass_manager::register_one_dump_file): Likewise.
1248 (pass_manager::register_pass): Likewise.
1249 (pass_init_dump_file): Likewise.
1250 (pass_fini_dump_file): Likewise.
1251 * statistics.c (statistics_early_init): Likewise.
1253 2013-10-14 Richard Biener <rguenther@suse.de>
1255 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1256 iterative_hash_canonical_type, gimple_canonical_type_hash,
1257 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1258 gimple_register_canonical_type, print_gimple_types_stats,
1259 free_gimple_type_tables): Move to lto/lto.c
1260 (gt-gimple.h): Do not include.
1261 * gimple.h (gimple_register_canonical_type,
1262 print_gimple_types_stats, free_gimple_type_tables): Remove.
1263 * Makefile.in (GTFILES): Remove gimple.c.
1265 2013-10-14 Travis Snoozy <quandary@remstate.com>
1268 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1269 scanning for msp430x targets.
1271 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1274 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1275 (registers_ok_for_ldd_peep): Move around.
1276 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1277 * config/sparc/sparc.md (widening peepholes): Use it.
1279 2013-10-14 Richard Biener <rguenther@suse.de>
1283 * gimple.c (iterative_hash_canonical_type): Make sure to
1284 record the hash into the correct hashtable slot.
1286 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1288 PR rtl-optimization/58662
1289 * combine.c (try_combine): Take into account death nodes on I2 when
1290 splitting a PARALLEL of two independent SETs. Fix dump message.
1292 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1295 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1296 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1297 * config/sh/t-sh (sh_treg_combine.o): New entry.
1298 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1299 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1300 (register_sh_passes): New function. Register sh_treg_combine pass.
1301 (sh_option_override): Invoke it.
1302 (sh_canonicalize_comparison): Handle op0_preserve_value.
1303 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1304 opportunities. Canonicalize branch condition.
1305 (nott): Allow only if pseudos can be created for non-SH2A.
1307 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1310 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1311 (ix86_expand_movmem): Replace copy_addr_to_reg with
1312 ix86_copy_addr_to_reg.
1313 (ix86_expand_setmem): Likewise.
1315 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1317 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1318 provided by ix86_fp_compare_mode instead of CCFPUmode.
1320 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1322 * config/aarch64/arm_neon.h
1323 (vtbx<1,3>_<psu>8): Fix register constriants.
1325 2013-10-11 Jeff Law <law@redhat.com>
1327 PR tree-optimization/58640
1328 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1329 threading paths that cross over two loop entry points.
1331 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1333 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1334 handle vector float as well.
1335 (*vsx_le_perm_load_v4si): Likewise.
1336 (*vsx_le_perm_store_v2di): Likewise.
1337 (*vsx_le_perm_store_v4si): Likewise.
1339 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1341 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1342 directly to circumvent subtract from splat{31} workaround.
1343 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1345 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1346 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1347 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1348 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1349 separate big and little endian logic.
1350 (*altivec_vperm_<mode>_internal): New define_insn.
1351 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1352 separate big and little endian logic.
1353 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1354 (vec_permv16qi): Add little endian logic.
1356 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1358 * doc/extend.texi (returns_nonnull): Remove arguments.
1360 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1361 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1362 Sergey Lega <sergey.s.lega@intel.com>
1363 Anna Tikhonova <anna.tikhonova@intel.com>
1364 Ilya Tocar <ilya.tocar@intel.com>
1365 Andrey Turetskiy <andrey.turetskiy@intel.com>
1366 Ilya Verbin <ilya.verbin@intel.com>
1367 Kirill Yukhin <kirill.yukhin@intel.com>
1368 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1370 * config/i386/sse.md (VI48F_256_512): New.
1371 (avx2_permvar<mode>): Change to ...
1372 (<avx2_avx512f>_permvar<mode>): This.
1374 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1375 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1376 Sergey Lega <sergey.s.lega@intel.com>
1377 Anna Tikhonova <anna.tikhonova@intel.com>
1378 Ilya Tocar <ilya.tocar@intel.com>
1379 Andrey Turetskiy <andrey.turetskiy@intel.com>
1380 Ilya Verbin <ilya.verbin@intel.com>
1381 Kirill Yukhin <kirill.yukhin@intel.com>
1382 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1384 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1385 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1386 * config/i386/sse.md (VI4_AVX): New.
1387 (sf2simodelower): Ditto.
1388 (sse2_cvtps2dq): Change to ...
1389 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1391 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1392 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1393 Sergey Lega <sergey.s.lega@intel.com>
1394 Anna Tikhonova <anna.tikhonova@intel.com>
1395 Ilya Tocar <ilya.tocar@intel.com>
1396 Andrey Turetskiy <andrey.turetskiy@intel.com>
1397 Ilya Verbin <ilya.verbin@intel.com>
1398 Kirill Yukhin <kirill.yukhin@intel.com>
1399 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1401 * config/i386/sse.md (V_512): New.
1403 (vcond<V_512:mode><VF_512:mode>): Ditto.
1404 (vcond<V_512:mode><VI_512:mode>): Ditto.
1405 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1407 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1408 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1409 Sergey Lega <sergey.s.lega@intel.com>
1410 Anna Tikhonova <anna.tikhonova@intel.com>
1411 Ilya Tocar <ilya.tocar@intel.com>
1412 Andrey Turetskiy <andrey.turetskiy@intel.com>
1413 Ilya Verbin <ilya.verbin@intel.com>
1414 Kirill Yukhin <kirill.yukhin@intel.com>
1415 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1417 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1418 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1420 (fma<mode>4): Removed condition.
1421 (fms<mode>4): Ditto.
1422 (fnma<mode>4): Ditto.
1423 (fnms<mode>4): Ditto.
1424 (fma4i_fmadd_<mode>): Ditto.
1425 (*fma_fmadd_<mode>): Ditto.
1426 (*fma_fmsub_<mode>): Ditto.
1427 (*fma_fnmadd_<mode>): Ditto.
1428 (*fma_fnmsub_<mode>): Ditto.
1429 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1430 (*fma_fmaddsub_<mode>): Ditto.
1431 (*fma_fmsubadd_<mode>): Ditto.
1432 (*fmai_fmadd_<mode>): Ditto.
1433 (*fmai_fmsub_<mode>): Ditto.
1434 (*fmai_fnmadd_<mode>): Ditto.
1435 (*fmai_fnmsub_<mode>): Ditto.
1437 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1438 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1439 Sergey Lega <sergey.s.lega@intel.com>
1440 Anna Tikhonova <anna.tikhonova@intel.com>
1441 Ilya Tocar <ilya.tocar@intel.com>
1442 Andrey Turetskiy <andrey.turetskiy@intel.com>
1443 Ilya Verbin <ilya.verbin@intel.com>
1444 Kirill Yukhin <kirill.yukhin@intel.com>
1445 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1447 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1448 (VI124_256): Changed to ...
1449 (VI124_256_48_512): This.
1450 (ssepackmode): Extended with wider modes.
1451 (<code><mode>3): Changed iterator.
1452 (*avx2_<code><mode>3): Ditto.
1453 (vec_pack_trunc_<mode>): Ditto.
1455 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1456 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1457 Sergey Lega <sergey.s.lega@intel.com>
1458 Anna Tikhonova <anna.tikhonova@intel.com>
1459 Ilya Tocar <ilya.tocar@intel.com>
1460 Andrey Turetskiy <andrey.turetskiy@intel.com>
1461 Ilya Verbin <ilya.verbin@intel.com>
1462 Kirill Yukhin <kirill.yukhin@intel.com>
1463 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1465 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1466 (VI8F_256_512): Ditto.
1467 (abs<mode>2): Changed iterator.
1468 (avx2_perm<mode>): Changed to ...
1469 (<avx2_avx512f>_perm<mode>): This.
1470 (avx2_perm<mode>_1): Changed to ...
1471 (<avx2_avx512f>_perm<mode>_1): This.
1473 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1474 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1475 Sergey Lega <sergey.s.lega@intel.com>
1476 Anna Tikhonova <anna.tikhonova@intel.com>
1477 Ilya Tocar <ilya.tocar@intel.com>
1478 Andrey Turetskiy <andrey.turetskiy@intel.com>
1479 Ilya Verbin <ilya.verbin@intel.com>
1480 Kirill Yukhin <kirill.yukhin@intel.com>
1481 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1483 * config/i386/sse.md (VI48_AVX512F): New.
1484 (VI48_AVX2): Changed to ...
1485 (VI48_AVX2_48_AVX512F): This.
1486 (avx2_ashrv<mode>): Changed to ...
1487 (<avx2_avx512f>_ashrv<mode>): This.
1488 (avx2_<shift_insn>v<mode>): Changed to ...
1489 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1491 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1492 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1493 Sergey Lega <sergey.s.lega@intel.com>
1494 Anna Tikhonova <anna.tikhonova@intel.com>
1495 Ilya Tocar <ilya.tocar@intel.com>
1496 Andrey Turetskiy <andrey.turetskiy@intel.com>
1497 Ilya Verbin <ilya.verbin@intel.com>
1498 Kirill Yukhin <kirill.yukhin@intel.com>
1499 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1501 * config/i386/sse.md (VI4_AVX512F): New.
1502 (VI8_AVX2_AVX512F): Ditto.
1503 (mul<mode>3): Extended with wider modes.
1504 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1505 (mul<mode>3): Ditto.
1506 (vec_widen_<s>mult_odd_<mode>): Ditto.
1508 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1509 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1510 Sergey Lega <sergey.s.lega@intel.com>
1511 Anna Tikhonova <anna.tikhonova@intel.com>
1512 Ilya Tocar <ilya.tocar@intel.com>
1513 Andrey Turetskiy <andrey.turetskiy@intel.com>
1514 Ilya Verbin <ilya.verbin@intel.com>
1515 Kirill Yukhin <kirill.yukhin@intel.com>
1516 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1518 * config/i386/sse.md (VI2_AVX512F): New.
1519 (VI124_AVX512F): Ditto.
1520 (sseunpackmode): Extended with wider modes.
1521 (sseunpackfltmode): Ditto.
1522 (vec_unpacks_float_hi_<mode>): Ditto.
1523 (vec_unpacks_float_lo_<mode>): Ditto.
1524 (vec_unpacku_float_hi_<mode>): Ditto.
1525 (vec_unpacku_float_lo_<mode>): Ditto.
1526 (vec_unpacks_lo_<mode>): Ditto.
1527 (vec_unpacks_hi_<mode>): Ditto.
1528 (vec_unpacku_lo_<mode>): Ditto.
1529 (vec_unpacku_hi_<mode>): Ditto.
1531 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1532 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1533 Sergey Lega <sergey.s.lega@intel.com>
1534 Anna Tikhonova <anna.tikhonova@intel.com>
1535 Ilya Tocar <ilya.tocar@intel.com>
1536 Andrey Turetskiy <andrey.turetskiy@intel.com>
1537 Ilya Verbin <ilya.verbin@intel.com>
1538 Kirill Yukhin <kirill.yukhin@intel.com>
1539 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1541 * config/i386/i386.md (multdiv): New.
1542 (multdiv_mnemonic): Ditto.
1543 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1544 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1545 (<sse>_vmdiv<mode>3): Removed.
1547 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1548 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1549 Sergey Lega <sergey.s.lega@intel.com>
1550 Anna Tikhonova <anna.tikhonova@intel.com>
1551 Ilya Tocar <ilya.tocar@intel.com>
1552 Andrey Turetskiy <andrey.turetskiy@intel.com>
1553 Ilya Verbin <ilya.verbin@intel.com>
1554 Kirill Yukhin <kirill.yukhin@intel.com>
1555 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1557 * config/i386/sse.md (V): Extended with wider modes.
1559 (ssehalfvecmode): Ditto.
1561 (ssepackfltmode): Ditto.
1562 (avx_vec_concat<mode>): Ditto.
1563 (V_256_512): New iterator.
1564 (VF2_512_256): Ditto.
1565 (si2dfmode): New attribute.
1566 (si2dfmodelower): Ditto.
1568 (concat_tg_mode): Ditto.
1569 (floatv4siv4df2): Changed to ...
1570 (float<si2dfmodelower><mode>2): This.
1571 (avx_cvtps2pd256): Changed to ...
1572 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1573 (vec_pack_trunc_v4df): Changed to ...
1574 (vec_pack_trunc_<mode>): This.
1575 (avx_vpermil<mode>): Changed to ...
1576 (<sse2_avx_avx512f>_vpermil<mode>): This.
1577 (<fixsuffix>fix_truncv8dfv8si2): New.
1578 (vec_pack_sfix_trunc_v8df): Ditto.
1579 (avx512f_rndscale<mode>): Ditto.
1580 (avx512f_roundpd512): Ditto.
1581 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1583 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1584 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1585 Sergey Lega <sergey.s.lega@intel.com>
1586 Anna Tikhonova <anna.tikhonova@intel.com>
1587 Ilya Tocar <ilya.tocar@intel.com>
1588 Andrey Turetskiy <andrey.turetskiy@intel.com>
1589 Ilya Verbin <ilya.verbin@intel.com>
1590 Kirill Yukhin <kirill.yukhin@intel.com>
1591 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1593 * config/i386/i386.md (any_fix): New iterator.
1594 (fixsuffix): New attribute.
1595 * config/i386/sse.md (VF1): Extened with wider modes.
1599 (sseintvecmodelower): Ditto.
1600 (ssescalarmode): Ditto.
1601 (ssescalarnum): Ditto.
1603 (ssexmmmode): Ditto.
1604 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1605 (<sse>_rcp<mode>2): Change iterator.
1606 (rsqrt<mode>2): Ditto.
1607 (<sse>_rsqrt<mode>2): Ditto.
1608 (avx2_vec_dup<mode>): Ditto.
1609 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1610 (round<mode>2_sfix): Ditto.
1611 (avx2_pbroadcast<mode>): Ditto.
1612 (*andnot<mode>3): Handle XI mode.
1613 (*<code><mode>3): Ditto.
1614 (AVXTOSSEMODE): Removed.
1615 (avx_vpermil<mode>): Changed to ...
1616 (<sse2_avx_avx512f>_vpermil<mode>): This.
1618 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1619 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1620 Sergey Lega <sergey.s.lega@intel.com>
1621 Anna Tikhonova <anna.tikhonova@intel.com>
1622 Ilya Tocar <ilya.tocar@intel.com>
1623 Andrey Turetskiy <andrey.turetskiy@intel.com>
1624 Ilya Verbin <ilya.verbin@intel.com>
1625 Kirill Yukhin <kirill.yukhin@intel.com>
1626 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1628 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1629 (<sse>_comi): Ditto.
1630 (<sse>_ucomi): Ditto.
1631 (sse_cvtss2siq_2): Ditto.
1632 (sse2_cvtsd2si): Ditto.
1633 (sse2_cvtsd2siq): Ditto.
1634 (sse2_cvttsd2si): Ditto.
1635 (sse2_cvttsd2siq): Ditto.
1636 (<shift_insn><mode>3): Ditto.
1637 (sse2_cvtsi2sdq): Update constraint and prefix.
1638 (sse_cvtsi2ss): Update prefix.
1639 (sse_cvtsi2ssq): Ditto.
1641 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1643 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1644 ignore internal calls.
1646 2013-10-11 Richard Biener <rguenther@suse.de>
1648 * tree-pretty-print.c (dump_generic_node): Allow to dump
1649 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1650 (OVF) if TREE_OVERFLOW is set.
1652 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1654 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1656 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1657 * gimple.h: Likewise.
1659 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1661 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1664 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1666 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1667 * configure: Regenerate.
1669 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1671 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1672 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1673 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1674 similarly pp_right_paren.
1675 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1676 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1677 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1678 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1680 (omp_declare_simd_clauses_equal,
1681 omp_remove_redundant_declare_simd_attrs): New functions.
1682 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1683 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1684 * tree.h (OMP_LOOP_CHECK): Define.
1685 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1686 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1687 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1688 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1689 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1690 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1691 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1692 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1693 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1694 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1695 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1696 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1697 (omp_remove_redundant_declare_simd_attrs): New prototype.
1698 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1699 GIMPLE_OMP_TEAMS): New codes.
1700 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1701 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1703 (target_nesting_level): New variable.
1704 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1705 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1706 for collapse > 1 static schedule unless ordered.
1707 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1708 (determine_parallel_type): Adjust get_ws_args_for caller.
1709 (install_var_field): Handle mask & 4 for double indirection.
1710 (scan_sharing_clauses): Ignore shared clause on teams construct.
1711 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1712 (create_omp_child_function): If inside target or declare target
1713 constructs, set "omp declare target" attribute on the child function.
1714 (find_combined_for): New function.
1715 (scan_omp_parallel): Handle combined loops.
1716 (scan_omp_target, scan_omp_teams): New functions.
1717 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1718 restrictions and set ctx->cancellable for cancellable constructs.
1719 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1720 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1721 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1722 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1723 (omp_clause_aligned_alignment): New function.
1724 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1726 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1727 on teams constructs. Handle user defined reductions and new
1729 (lower_reduction_clauses): Don't set placeholder to address of ref
1730 if it has already the right type.
1731 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1732 (expand_parallel_call): Use the new non-_start suffixed builtins,
1733 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1734 and GOMP_parallel_end after the call.
1735 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1736 (expand_omp_for_init_counts): Handle combined loops.
1737 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1739 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1740 end of cancellable loops.
1741 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1742 Likewise. Handle collapse > 1 loops.
1743 (expand_omp_simd): Handle combined loops.
1744 (expand_omp_for): Add inner_stmt argument, adjust callers of
1745 expand_omp_for* functions, use expand_omp_for_static*chunk even
1746 for collapse > 1 unless ordered.
1747 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1748 of cancellable sections.
1749 (expand_omp_single): Remove need_barrier variable, just rely on
1750 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1751 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1752 (expand_omp_atomic_load, expand_omp_atomic_store,
1753 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1754 (expand_omp_target): New function.
1755 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1756 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1757 (build_omp_regions_1): Immediately close region for
1758 GF_OMP_TARGET_KIND_UPDATE.
1759 (maybe_add_implicit_barrier_cancel): New function.
1760 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1762 (lower_omp_single): Likewise. Add clobber after the barrier.
1763 (lower_omp_taskgroup): New function.
1764 (lower_omp_for): Handle combined loops. Adjust
1765 lower_rec_input_clauses caller. Handle cancellation.
1766 (lower_depend_clauses): New function.
1767 (lower_omp_taskreg): Lower depend clauses. Adjust
1768 lower_rec_input_clauses caller. Add clobber after the call. Handle
1770 (lower_omp_target, lower_omp_teams): New functions.
1771 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1772 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1773 and GOMP_cancellation_point calls.
1774 (lower_omp): Fold stmts inside of target region.
1775 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1776 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1777 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1778 (BT_FN_VOID_OMPFN_PTR_UINT,
1779 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1780 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1781 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1782 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1783 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1784 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1785 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1786 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1787 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1788 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1789 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1790 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1791 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1792 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1793 BUILT_IN_GOMP_PARALLEL_END.
1794 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1795 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1796 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1797 GF_OMP_FOR_KIND_DISTRIBUTE.
1798 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1799 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1800 (dump_gimple_omp_return): Print lhs if it has any.
1801 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1802 gimple_omp_atomic_seq_cst_p.
1803 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1804 and GIMPLE_OMP_TEAMS.
1805 * langhooks.c (lhd_omp_mappable_type): New function.
1806 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1807 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1809 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1810 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1811 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1813 (struct gimplify_omp_ctx): Add combined_loop field.
1814 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1815 on stmts inside of target region.
1816 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1817 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1818 ORT_TARGET and ORT_TARGET_DATA.
1819 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1821 (omp_notice_threadprivate_variable): Complain about threadprivate
1822 variables in target region.
1823 (omp_notice_variable): Complain about vars with non-mappable type
1824 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1825 (omp_check_private): Ignore ORT_TARGET* regions.
1826 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1827 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1828 (find_combined_omp_for): New function.
1829 (gimplify_omp_for): Handle gimplification of combined loops.
1830 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1832 (gimplify_omp_target_update): New function.
1833 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1834 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1835 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1836 (gimplify_body): If fndecl has "omp declare target" attribute, add
1837 implicit ORT_TARGET context around it.
1838 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1839 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1840 * tree-nested.c (convert_nonlocal_reference_stmt,
1841 convert_local_reference_stmt, convert_gimple_call): Handle
1842 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1843 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1844 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1845 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1846 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1847 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1848 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1849 BUILT_IN_GOMP_LOOP_END_CANCEL,
1850 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1851 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1852 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1853 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1854 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1855 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1856 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1857 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1858 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1859 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1860 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1861 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1862 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1863 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1864 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1865 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1866 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1867 gimple_build_omp_teams): New functions.
1868 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1869 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1870 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1871 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1872 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1873 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1874 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1875 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1876 GF_OMP_ATOMIC_SEQ_CST): New.
1877 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1878 gimple_build_omp_teams): New prototypes.
1879 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1880 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1881 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1882 GIMPLE_OMP_SINGLE as end of range.
1883 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1884 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1885 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1886 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1887 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1888 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1889 gimple_omp_target_kind, gimple_omp_target_set_kind,
1890 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1891 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1892 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1893 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1894 gimple_omp_teams_set_clauses): New inlines.
1895 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1896 and GIMPLE_OMP_TASKGROUP.
1897 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
1898 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1899 enum omp_clause_proc_bind_kind): New.
1900 (union omp_clause_subcode): Add depend_kind, map_kind and
1901 proc_bind_kind fields.
1902 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1903 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1904 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1905 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1906 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1908 2013-10-10 Teresa Johnson <tejohnson@google.com>
1910 * predict.c (tree_estimate_probability): Add new parameter
1911 for estimate_bb_frequencies.
1912 (estimate_bb_frequencies): Add new parameter to force estimation.
1913 (rebuild_frequencies): When max frequency in function is small,
1914 recompute counts from frequencies.
1915 * predict.h (estimate_bb_frequencies): New parameter.
1917 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1919 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1920 optimizations are disabled.
1922 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1924 * coverage.c (coverage_finish): Fix leak of da_file_name.
1926 2013-10-10 Jan Hubicka <jh@suse.cz>
1928 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1929 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1932 2013-10-10 Jakub Jelinek <jakub@redhat.com>
1935 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1936 if any labels are in FALLTHRU_BB, use a special label emitted
1937 immediately after the asm goto insn rather than label_rtx
1939 (expand_asm_stmt): Adjust caller.
1940 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1941 edge if the last insn in predecessor is a jump with single successor,
1942 but it isn't simplejump_p.
1944 2013-10-10 Richard Biener <rguenther@suse.de>
1946 PR tree-optimization/58656
1947 * tree-ssa-pre.c (phi_translate): Do not cache failed
1950 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1952 * gimplify.c: Include expr.h and tm_p.h for targets with special
1953 va-arg padding requirements.
1955 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1957 * tree-flow.h: Move some prototypes to gimple.h.
1958 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1959 * gimple.h: Relocate some prototypes from tree-flow.h
1960 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1962 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1963 (build_va_arg_indirect_ref): Relocate and make static.
1964 (std_gimplify_va_arg_expr): Relocate here.
1965 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1966 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1968 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1970 * doc/md.texi: Document the mnemonic attribute.
1972 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1975 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1976 (set_attr_alternative x ...) when searching for user defined
1979 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1981 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1983 2013-10-09 Easwaran Raman <eraman@google.com>
1985 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1986 * cfgexpand.c (defer_stack_allocation): ...use here
1987 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1989 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1991 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1992 (operand_equal_for_value_replacement): New function, extracted from
1993 value_replacement and enhanced to catch more cases.
1994 (value_replacement): Use operand_equal_for_value_replacement.
1996 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1998 * loop-doloop.c (doloop_modify, doloop_optimize): Use
1999 get_max_loop_iterations.
2001 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2003 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2004 Place comment above function.
2006 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2008 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2009 are not related to tree-cfg.c.
2010 * tree-ssa-address.h: New file. Relocate prototypes.
2011 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2012 (addr_for_mem_ref): New. Combine call to get_address_description and
2013 return addr_for_mem_ref.
2014 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2015 * tree-ssa-live.h: Adjust prototypes.
2016 * passes.c: Include tree-ssa-live.h.
2017 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2018 * graphite.c (graphite_transform_loops): Make static.
2019 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2020 make_pass_graphite, pass_data_graphite_transforms,
2021 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2022 * ipa-pure-const.c (warn_function_noreturn): Make static.
2023 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2024 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2025 Relocate from tree-cfg.c
2026 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2028 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2029 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2030 Move to ipa-pure-const.c.
2031 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2032 Relocate from tree-optimize.c.
2033 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2034 make_pass_fixup_cfg): Move to tree-cfg.c.
2035 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2036 Relocate some prototypes.
2037 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2038 * tree-dump.c (dump_function_to_file): Remove prototype.
2039 Add tree-flow.h to the include file.
2040 * tree-dump.h: Remove prototype.
2041 * tree-parloops.h: New File. Add prototypes.
2042 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2043 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2044 from tree-ssa-loop.c.
2045 * tree-predcom.c (run_tree_predictive_commoning,
2046 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2047 Relocate here from tree-ssa-loop.c.
2048 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2049 ssa_name_values.release ().
2050 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2051 (ssa_name_values): Relocate from tree-flow.h.
2052 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2053 * tree-ssa-loop.c (run_tree_predictive_commoning,
2054 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2055 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2056 make_pass_graphite, pass_data_graphite_transforms,
2057 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2058 tree_parallelize_loops, pass_data_parallelize_loops,
2059 make_pass_parallelize_loops): Move to other files.
2060 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2062 * tree.h: Remove prototypes from tree-address.c.
2064 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2066 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2067 (struct int_tree_map): Move to tree-hasher.h
2068 (SCALE, LABEL, PERCENT): Move to gimple.h
2069 * tree-flow-inline.h: Delete. Move functions to other files.
2070 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2071 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2072 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2073 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2074 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2075 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2076 * tree-hasher.h: Don't include tree-flow.h.
2077 (struct int_tree_map): Relocate from tree-flow.h.
2078 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2079 tree-flow-inline.h and make static.
2080 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2082 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2083 tree-flow-inline.h and make static.
2084 * tree.h (is_global_var, may_be_aliased): Relocate from
2086 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2087 * value-prof.c: No longer include tree-flow-inline.h.
2088 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2090 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2092 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2093 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2094 (enum move_pos): Move to tree-ssa-loop-im.h
2095 * cfgloop.h: Move some prototypes.
2096 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2097 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2098 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2099 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2100 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2101 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2102 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2103 (enum move_pos): Relocate here.
2104 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2106 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2107 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2108 make_pass_lim): Relocate here from tree-ssa-loop.c.
2109 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2111 (loop_edge_to_cancel, unloop_loops): Make static.
2112 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2113 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2114 (tree_complete_unroll, gate_tree_complete_unroll,
2115 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2116 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2117 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2118 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2119 (stmt_invariant_in_loop_p): Remove unused function.
2120 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2121 * tree-ssa-loop-manip.h: New file. Add prototypes.
2122 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2123 (gcov_type_to_double_int): Move to cfgloop.h.
2124 (double_int_cmp, bound_index,
2125 estimate_numbers_of_iterations_loop): Make static.
2126 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2127 (max_loop_iterations): Factor out get_max_loop_iterations.
2128 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2130 * tree-ssa-loop-niter.h: New file. Add prototypes.
2131 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2132 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2133 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2134 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2135 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2136 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2137 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2138 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2139 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2140 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2141 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2142 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2143 pass_data_complete_unroll, make_pass_complete_unroll,
2144 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2145 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2146 tree-ssa-loop-ivcanon.c.
2147 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2148 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2149 tree-ssa-loop-prefetch.c.
2150 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2152 (get_lsm_tmp_name): Relocate and add suffix parameter.
2153 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2154 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2155 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2156 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2157 (get_estimated_loop_iterations): Factor out accessor from
2158 estimated_loop_iterations in tree-ssa-loop-niter.c.
2159 (get_max_loop_iterations): Factor out accessor from
2160 _max_loop_iterations in tree-ssa-niter.c.
2161 * loop-unroll.c (decide_unroll_constant_iterations,
2162 decide_unroll_runtime_iterations, decide_peel_simple,
2163 decide_unroll_stupid): Use new get_* accessors.
2165 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2167 PR tree-optimization/20318
2168 * doc/extend.texi (returns_nonnull): New function attribute.
2169 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2171 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2172 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2174 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2177 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2178 false if both components are bitfields.
2180 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2182 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2183 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2184 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2185 * config/aarch64/aarch64.h
2186 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2187 * config/aarch64/aarch64-simd-builtins.def
2188 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2190 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2192 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2194 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2196 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2198 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2200 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2201 (vneg_f64): New intrinsic.
2202 (vneg_s8): Asm replaced with C.
2203 (vneg_s16): Likewise.
2204 (vneg_s32): Likewise.
2205 (vneg_s64): New intrinsic.
2206 (vnegq_f32): Asm replaced with C.
2207 (vnegq_f64): Likewise.
2208 (vnegq_s8): Likewise.
2209 (vnegq_s16): Likewise.
2210 (vnegq_s32): Likewise.
2211 (vnegq_s64): Likewise.
2213 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2215 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2217 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2219 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2220 packed stack special handling.
2221 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2222 back to fixed stack slots for FPRs saved due to stdarg.
2224 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2226 * config/s390/s390.c (s390_frame_info): Restructure function.
2228 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2230 * config/s390/s390.c (struct s390_frame_layout): New field
2232 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2233 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2234 regs_ever_clobbered to char*.
2235 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2236 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2237 (s390_register_info): Enable FPR save slots. Move/Copy some
2238 functionality into ...
2239 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2240 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2242 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2243 not imply a stack frame.
2244 (s390_init_frame_layout): Remove variable clobbered_regs.
2245 (s390_update_register_info): Remove function.
2246 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2247 cannot be used for register renaming.
2248 (s390_hard_regno_scratch_ok): New function.
2249 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2250 (s390_initial_elimination_offset): Change offset calculation of
2251 the return address pointer.
2252 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2253 (restore_gprs): Set frame related flag.
2254 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2255 (s390_emit_prologue): Call s390_register_info instead of
2256 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2257 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2258 (s390_optimize_prologue): Call s390_optimize_register_info.
2259 Try to remove also FPR slot save/restore INSNs. Remove frame
2260 related flags from restore INSNs.
2262 2013-10-08 DJ Delorie <dj@redhat.com>
2264 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2267 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2268 avoid conflict with the MI use of %c.
2269 * config/rl78/rl78-real.md: change %c to %C throughout.
2270 * config/rl78/rl78-virt.md: Likewise.
2272 2013-10-08 Jan Hubicka <jh@suse.cz>
2274 * config/i386/i386.c (ix86_option_override_internal): Switch
2275 to SSE math for -ffast-math when target ISA supports SSE2.
2277 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2279 * tree-flow.h: Remove some prototypes.
2280 * tree.h: Remove some protypes, add a couple.
2281 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2282 using_eh_for_cleanups_p): Add interface routines for front ends.
2283 * tree-eh.h: New file. Add protoptyes.
2284 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2285 (add_stmt_to_eh_lp_fn): Make static.
2286 (lower_try_finally): Use new using_eh_for_cleanups_p.
2287 * emit-rtl.c: Include tree-eh.h.
2288 * gimple.h: Include tree-eh.h.
2290 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2292 PR tree-optimization/58480
2293 * tree-vrp.c (infer_nonnull_range): New function.
2294 (infer_value_range): Call infer_nonnull_range.
2296 2013-10-08 Dehao Chen <dehao@google.com>
2298 PR tree-optimization/58619
2299 * tree-inline.c (copy_phis_for_bb): Combine location data
2302 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2305 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2306 RTX_FRAME_RELATED_P on INSN.
2308 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2310 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2311 (altivec_expand_vec_perm_const): Call it.
2313 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2315 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2316 sequences for LE VSX loads and stores at expand time.
2317 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2319 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2320 (rs6000_gen_le_vsx_permute): New.
2321 (rs6000_gen_le_vsx_load): New.
2322 (rs6000_gen_le_vsx_store): New.
2323 (rs6000_gen_le_vsx_move): New.
2324 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2325 (*vsx_le_perm_load_v4si): New.
2326 (*vsx_le_perm_load_v8hi): New.
2327 (*vsx_le_perm_load_v16qi): New.
2328 (*vsx_le_perm_store_v2di): New.
2329 (*vsx_le_perm_store_v4si): New.
2330 (*vsx_le_perm_store_v8hi): New.
2331 (*vsx_le_perm_store_v16qi): New.
2332 (*vsx_xxpermdi2_le_<mode>): New.
2333 (*vsx_xxpermdi4_le_<mode>): New.
2334 (*vsx_xxpermdi8_le_V8HI): New.
2335 (*vsx_xxpermdi16_le_V16QI): New.
2336 (*vsx_lxvd2x2_le_<mode>): New.
2337 (*vsx_lxvd2x4_le_<mode>): New.
2338 (*vsx_lxvd2x8_le_V8HI): New.
2339 (*vsx_lxvd2x16_le_V16QI): New.
2340 (*vsx_stxvd2x2_le_<mode>): New.
2341 (*vsx_stxvd2x4_le_<mode>): New.
2342 (*vsx_stxvd2x8_le_V8HI): New.
2343 (*vsx_stxvd2x16_le_V16QI): New.
2345 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2347 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2349 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2351 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2352 loop to work also for 31bit ABI.
2353 Save the stack pointer for frame_size > 0.
2355 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2357 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2358 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2359 constraint letters from expanders.
2360 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2361 retry count to general_operand.
2362 ("tabort"): Give operand 0 a mode.
2363 ("tabort_1"): Add mode and constraint letter for operand 0.
2364 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2366 2013-10-04 Jeff Law <law@redhat.com>
2368 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2370 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2372 (thread_across_edge): Here. Call it.
2374 2013-10-04 Cary Coutant <ccoutant@google.com>
2376 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2377 discarding a location list expression (or a piece of one).
2379 2013-10-03 Jan Hubicka <jh@suse.cz>
2381 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2382 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2383 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2385 2013-10-03 Jan Hubicka <jh@suse.cz>
2387 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2388 accumulate-outgoing-args when producing unwind info.
2390 2013-10-03 Wei Mi <wmi@google.com>
2392 * lra-constraints.c (insert_move_for_subreg): New function
2393 extracted from simplify_operand_subreg.
2394 (simplify_operand_subreg): Add reload for paradoxical subreg.
2396 2013-10-03 Rong Xu <xur@google.com>
2398 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2399 the candidate of builtin_expect such that we should fix the
2400 size/time estimation.
2401 (estimate_function_body_sizes): Do the acutally size/time fix-up
2404 2013-10-03 Rong Xu <xur@google.com>
2406 * predict.c (tree_predict_by_opcode): Get the probability
2407 for builtin_expect from param builtin_expect_probability.
2408 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2409 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2410 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2412 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2415 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2416 * calls.c (alloca_call_p): Use get_callee_fndecl.
2417 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2418 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2420 (vrp_visit_stmt): Remove duplicated code.
2422 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2424 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2425 ceildf2, btruncdf2, instead of vsx_* name.
2427 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2428 iterators to only do V2DF and V4SF here. Move the DF code to
2429 rs6000.md where it is combined with SF mode. Replace <VSv> with
2430 just 'v' since only vector operations are handled with these insns
2431 after moving the DF support to rs6000.md.
2432 (vsx_sub<mode>3): Likewise.
2433 (vsx_mul<mode>3): Likewise.
2434 (vsx_div<mode>3): Likewise.
2435 (vsx_fre<mode>2): Likewise.
2436 (vsx_neg<mode>2): Likewise.
2437 (vsx_abs<mode>2): Likewise.
2438 (vsx_nabs<mode>2): Likewise.
2439 (vsx_smax<mode>3): Likewise.
2440 (vsx_smin<mode>3): Likewise.
2441 (vsx_sqrt<mode>2): Likewise.
2442 (vsx_rsqrte<mode>2): Likewise.
2443 (vsx_fms<mode>4): Likewise.
2444 (vsx_nfma<mode>4): Likewise.
2445 (vsx_copysign<mode>3): Likewise.
2446 (vsx_btrunc<mode>2): Likewise.
2447 (vsx_floor<mode>2): Likewise.
2448 (vsx_ceil<mode>2): Likewise.
2449 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2450 (vsx_sminsf3): Likewise.
2451 (vsx_fmadf4): Likewise.
2452 (vsx_fmsdf4): Likewise.
2453 (vsx_nfmadf4): Likewise.
2454 (vsx_nfmsdf4): Likewise.
2455 (vsx_cmpdf_internal1): Likewise.
2457 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2458 simpler to select whether a target has SPE or traditional floating
2459 point support in iterators.
2460 (TARGET_DF_SPE): Likewise.
2461 (TARGET_SF_FPR): Likewise.
2462 (TARGET_DF_FPR): Likewise.
2463 (TARGET_SF_INSN): Macros to say whether floating point support
2464 exists for a given operation for expanders.
2465 (TARGET_DF_INSN): Likewise.
2467 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2468 combining of SF/DF mode operations, using both traditional and VSX
2476 (abs<mode>2): Combine SF/DF modes using traditional floating point
2477 instructions. Add support for using the upper DF registers with
2478 VSX support, and SF registers with power8-vector support. Update
2479 expanders for operations supported by both the SPE and traditional
2480 floating point units.
2481 (abs<mode>2_fpr): Likewise.
2482 (nabs<mode>2): Likewise.
2483 (nabs<mode>2_fpr): Likewise.
2484 (neg<mode>2): Likewise.
2485 (neg<mode>2_fpr): Likewise.
2486 (add<mode>3): Likewise.
2487 (add<mode>3_fpr): Likewise.
2488 (sub<mode>3): Likewise.
2489 (sub<mode>3_fpr): Likewise.
2490 (mul<mode>3): Likewise.
2491 (mul<mode>3_fpr): Likewise.
2492 (div<mode>3): Likewise.
2493 (div<mode>3_fpr): Likewise.
2494 (sqrt<mode>3): Likewise.
2495 (sqrt<mode>3_fpr): Likewise.
2496 (fre<Fs>): Likewise.
2497 (rsqrt<mode>2): Likewise.
2498 (cmp<mode>_fpr): Likewise.
2499 (smax<mode>3): Likewise.
2500 (smin<mode>3): Likewise.
2501 (smax<mode>3_vsx): Likewise.
2502 (smin<mode>3_vsx): Likewise.
2503 (negsf2): Delete SF operations that are merged with DF.
2510 (fmasf4_fpr): Likewise.
2511 (fmssf4_fpr): Likewise.
2512 (nfmasf4_fpr): Likewise.
2513 (nfmssf4_fpr): Likewise.
2514 (sqrtsf2): Likewise.
2515 (rsqrtsf_internal1): Likewise.
2516 (smaxsf3): Likewise.
2517 (sminsf3): Likewise.
2518 (cmpsf_internal1): Likewise.
2519 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2520 (negdf2): Delete DF operations that are merged with SF.
2522 (nabsdf2): Likewise.
2528 (rsqrtdf_internal1): Likewise.
2529 (fmadf4_fpr): Likewise.
2530 (fmsdf4_fpr): Likewise.
2531 (nfmadf4_fpr): Likewise.
2532 (nfmsdf4_fpr): Likewise.
2533 (sqrtdf2): Likewise.
2534 (smaxdf3): Likewise.
2535 (smindf3): Likewise.
2536 (cmpdf_internal1): Likewise.
2537 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2538 (btrunc<mode>2): Delete separate expander, and combine with the
2539 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2540 (btrunc<mode>2_fpr): Likewise.
2541 (ceil<mode>2): Likewise.
2542 (ceil<mode>2_fpr): Likewise.
2543 (floor<mode>2): Likewise.
2544 (floor<mode>2_fpr): Likewise.
2545 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2546 Add support for using the upper registers with VSX and
2547 power8-vector. Move insns to be closer to the define_expands. On
2548 VSX systems, prefer the traditional form of FMA over the VSX
2549 version, since the traditional form allows the target not to
2550 overlap with the inputs.
2551 (fms<mode>4_fpr): Likewise.
2552 (nfma<mode>4_fpr): Likewise.
2553 (nfms<mode>4_fpr): Likewise.
2555 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2556 Richard Earnshaw <richard.earnshaw@arm.com>
2558 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2559 (struct mult_cost_table): Likewise.
2560 (struct mem_cost_table): Likewise.
2561 (struct fp_cost_table): Likewise.
2562 (struct vector_cost_table): Likewise.
2563 (cpu_cost_table): Likewise.
2564 * config/arm/arm.opt (mold-rts-costs): New option.
2565 (mnew-generic-costs): Likewise.
2566 * config/arm/arm.c (generic_extra_costs): New table.
2567 (cortexa15_extra_costs): Likewise.
2568 (arm_slowmul_tune): Use NULL as new costs.
2569 (arm_fastmul_tune): Likewise.
2570 (arm_strongarm_tune): Likewise.
2571 (arm_xscale_tune): Likewise.
2572 (arm_9e_tune): Likewise.
2573 (arm_v6t2_tune): Likewise.
2574 (arm_cortex_a5_tune): Likewise.
2575 (arm_cortex_a9_tune): Likewise.
2576 (arm_v6m_tune): Likewise.
2577 (arm_fa726te_tune): Likewise.
2578 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2579 (arm_cortex_tune): Use generict_extra_costs.
2580 (shifter_op_p): New function.
2581 (arm_unspec_cost): Likewise.
2582 (LIBCALL_COST): Define.
2583 (arm_new_rtx_costs): New function.
2584 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2585 table is available. Use old costs otherwise unless mnew-generic-costs
2587 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2588 (cpu_cost_table): Declare.
2590 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2593 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2594 (*subs_mul_imm_<mode>)
2595 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2596 (*sub_<shift>_<mode>)
2597 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2598 Remove k constraint.
2600 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2602 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2604 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2606 2013-10-02 Teresa Johnson <tejohnson@google.com>
2608 * predict.c (probably_never_executed): New function.
2609 (probably_never_executed_bb_p): Invoke probably_never_executed.
2610 (probably_never_executed_edge_p): Ditto.
2611 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2612 Treat profile insanities conservatively.
2614 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2616 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2618 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2620 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2621 use it. Use smaller increase for scratch. Don't increase reject
2622 for early clobber scratch.
2623 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2624 setting eliminated regs except setting fp from hfp.
2625 (lra_eliminate): Check lra_insn_recog_data on NULL.
2627 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2630 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2631 setting -mvsx-timode by default until the underlying problem is fixed.
2632 (RS6000_CPU, power7 defaults): Likewise.
2634 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2636 * config/x-linux (host-linux.o): Remove header dependencies.
2637 Use $(COMPILE) and $(POSTCOMPILE).
2638 * config/t-linux-android (linux-android.o): Ditto.
2640 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2642 * Makefile.in (expmed.o-warn): Remove.
2644 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2646 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2647 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2649 2013-10-02 Teresa Johnson <tejohnson@google.com>
2651 * dojump.c (do_jump_1): Divide probability between
2652 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2654 2013-10-02 Tom Tromey <tromey@redhat.com>
2656 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2658 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2660 * tree-flow.h: Remove some prototypes.
2661 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2662 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2663 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2664 mark_virtual_phi_result_for_renaming): Relocate here.
2665 * tree-into-ssa.h: Add prototypes.
2666 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2667 single_pred_before_succ_order.
2668 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2669 (nonfreeing_call_p) Move to gimple.c.
2670 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2672 * basic-block.h (single_pred_before_succ_order): Add prototype.
2673 * gimple.c (nonfreeing_call_p): Relocate here.
2674 * gimple.h: Add prototype.
2675 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2676 * tree-ssa-dom.h: New file. Relocate prototypes here.
2677 * tree-ssa.h: Include tree-ssa-dom.h.
2679 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2681 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2682 Use $(COMPILE) and $(POSTCOMPILE).
2684 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2686 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2688 * tree-flow.h: Remove some prototypes.
2689 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2690 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2691 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2692 propagate_tree_value*): Move from here to...
2693 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2694 propagate_tree_value*): Relocate here.
2695 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2696 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2697 * gimple-fold.c: Remove gimple-fold.h from include list.
2698 * tree-vrp.c: Remove gimple-fold.h from include list.
2699 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2700 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2701 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2702 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2703 * sese.c: Add tree-ssa-propagate.h to include list.
2705 2013-10-02 Richard Biener <rguenther@suse.de>
2707 * tree-loop-distribution.c: Include tree-vectorizer.h for
2709 (enum partition_kind): Remove PKIND_REDUCTION.
2710 (struct partition_s): Remove has_writes member, add reduction_p member.
2711 (partition_alloc): Adjust.
2712 (partition_builtin_p): Likewise.
2713 (partition_has_writes): Remove.
2714 (partition_reduction_p): New function.
2715 (partition_merge_into): Likewise.
2716 (generate_code_for_partition): Commonize builtin partition
2718 (rdg_cannot_recompute_vertex_p): Remove.
2719 (already_processed_vertex_p): Likewise.
2720 (rdg_flag_vertex): Do not set has_writes.
2721 (classify_partition): Adjust.
2722 (rdg_build_partitions): Do not set has_writes, treat all
2723 partitions as useful.
2724 (distribute_loop): Record number of library calls generated. Adjust.
2725 (tree_loop_distribution): Report number of loops and library
2726 calls generated as opt-info.
2728 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2730 * tree-flow.h: Include new .h files. Move prototypes.
2731 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2732 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2733 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2734 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2735 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2736 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2737 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2738 * tree.h (get_ref_base_and_extent): Move prototype out.
2739 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2741 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2742 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2743 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2744 * tree-scalar-evolution.c: Include tree.h.
2745 * sese.c: Include tree.h.
2746 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2747 * dwarf2out.c: Include tree-dfa.h.
2748 * tree-chrec.c: Include tree.h.
2749 * tree-data-ref.c: Include tree.h.
2751 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2753 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2756 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2758 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2759 Use $(COMPILE) and $(POSTCOMPILE).
2760 (sol2-cxx.o): Likewise.
2761 (sol2-stubs.o): Likewise.
2763 * config/x-solaris (host-solaris.o): Likewise.
2765 * config/sparc/t-sparc (sparc.o): Remove.
2766 (sparc-c.o): Remove header dependencies.
2767 Use $(COMPILE) and $(POSTCOMPILE).
2768 * config/sparc/x-sparc: Likewise.
2770 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2772 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2773 * config/arc/arc700.md: Likewise.
2774 * config/arc/arc-modes.def: Likewise.
2775 * config/arc/arc-simd.h: Likewise.
2776 * config/arc/t-arc-uClibc: Likewise.
2777 * config/arc/t-arc-newlib: Likewise.
2779 2013-10-02 Renlin Li <renlin.li@arm.com>
2781 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2783 (aarch64_expand_epilogue): Likewise.
2785 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2786 Yufeng Zhang <yufeng.zhang@arm.com>
2788 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2790 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2791 'base_in' represent a conversion and legal_cast_p_1 holds; set
2792 'base_in' with the returned value from get_unwidened.
2794 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2796 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2797 plus_constant is not used.
2799 2013-10-01 Wei Mi <wmi@google.com>
2801 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2802 * config/i386/i386.md: Add define_peephole2 to
2803 break partial reg stall for cvtss2sd/cvtsd2ss.
2805 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2807 * config/arc/arc.c (pass_arc_ifcvt::clone):
2808 Update for ctxt_ -> m_ctxt change.
2810 2013-10-01 Jeff Law <law@redhat.com>
2812 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2813 outgoing_edge and intermediate_edge fields. Instead store the path.
2814 (redirection_data::hash): Hash on the last edge's destination index.
2815 (redirection_data::equal): Check the entire thread path.
2816 (lookup_redirectio_data): Corresponding changes.
2817 (create_edge_and_update_destination_phis): Likewise.
2818 (thread_single_edge): Likewise.
2820 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2821 Diego Novillo <dnovillo@google.com>
2823 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2824 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2825 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2826 versions of these patterns.
2828 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2829 (__builtin_arc_aligned): Likewise.
2831 * config/arc/arc.md: Expand adc_0 comment stating the intended
2832 purpose and why it isn't ready.
2833 Replace commented out call_value_via_label_mixed with a
2834 plain comment about bl_s.
2836 * config/arc/arc.c (stdio.h): Don't include directly.
2837 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2838 Remove if (1) condition.
2839 (arc_encode_section_info): Fix comment.
2841 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2843 * config/arc/arc.c (arc_conditional_register_usage):
2844 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2845 Also set reg_alloc_order for DMA config regs.
2847 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2848 Jeremy Bennett <jeremy.bennett@embecosm.com>
2850 * doc/install.texi (--with-cpu): Mention ARC.
2851 (arc-*-elf32): New paragraph.
2852 (arc-linux-uclibc): Likewise.
2853 * doc/md.texi (Machine Constraints): Add ARC part.
2854 * doc/invoke.texi: (menu): Add ARC Options.
2855 (Machine Dependent Options) <ARC Options>: Add synopsis.
2856 (node ARC Options): Add.
2857 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2858 (ARC Built-in Functions): New section defining
2859 generic ARC built-in functions.
2860 (ARC SIMD Built-in Functions): New section defining SIMD specific
2862 (Declaring Attributes of Functions): Extended
2863 description of short_call and long_call attributes for ARC and
2864 added index entries.
2866 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2867 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2868 Joern Rennecke <joern.rennecke@embecosm.com>
2869 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2870 Brendan Kehoe <brendan@zen.org>
2871 Michael Eager <eager@eagercon.com>
2872 Simon Cook <simon.cook@embecosm.com>
2873 Jeremy Bennett <jeremy.bennett@embecosm.com>
2875 * config/arc, common/config/arc: New directories.
2877 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2878 Brendan Kehoe <brendan@zen.org>
2879 Simon Cook <simon.cook@embecosm.com>
2881 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2883 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2885 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2886 * tree-ssa-coalesce.h: New. Move prototype to here.
2887 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2888 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2889 (gimple_can_coalesce_p): Move to...
2890 * gimple.c (gimple_can_coalesce_p): Here.
2892 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2894 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2895 (dump_decl_set): Move to gimple.c.
2896 * gimple.h: Don't include tree-ssa-operands.h.
2897 (dump_decl_set): Add prototype.
2898 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2899 Move to gimple-ssa.h.
2900 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2901 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2902 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2903 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2904 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2906 (dump_decl_set): Relocate here.
2907 * gimple-ssa.h: New file.
2908 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2909 Relocate from gimple.h.
2910 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2911 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2912 swap_tree_operands and remove non-ssa path.
2913 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2914 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2916 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2917 vect_is_simple_reduction_1): Use swap_ssa_operands.
2918 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2919 (enum need_phi_state): Move to tree-into-ssa.c.
2920 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2921 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2922 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2923 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2924 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2925 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2926 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2927 num_imm_uses): Move to ssa-iterators.h.
2928 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2929 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2931 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2932 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2933 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2934 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2935 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2936 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2937 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2938 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2939 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2940 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2941 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2942 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2943 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2944 (set_phi_nodes): Move to tree-phinodes.h.
2945 * tree-ssa-operands.h (enum ssa_op_iter_type,
2946 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2947 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2948 (dump_decl_set): Remove prototype.
2949 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2950 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2951 (set_phi_nodes): Relocate from tree-flow-inline.h.
2952 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2954 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2955 include list. Temporarily add gimple.h to include list.
2956 * ssa-iterators.h: New file.
2957 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2958 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2959 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2960 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2961 Relocate from tree-ssa-operands.h.
2962 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2963 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2964 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2965 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2966 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2967 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2968 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2969 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2970 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2971 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2972 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2973 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2974 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2975 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2976 Relocate from tree-flow-inline.h.
2977 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2979 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2982 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2983 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2984 Insert '\t' to output template.
2985 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2986 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2987 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2988 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2990 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
2992 * doc/install.texi (Host/target specific installation notes for GCC):
2993 Put @anchor before @heading.
2994 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2995 Use @email for email addresses.
2997 2013-10-01 Jeff Law <law@redhat.com>
2999 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3000 a vec. Only delete the path if we create one without successfully
3001 registering a jump thread.
3002 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3004 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3005 (paths): New vector of jump threading paths.
3006 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3007 (THREAD_PATH): New accessor macro for the entire thread path.
3008 (lookup_redirection_data): Get intermediate and final outgoing edge
3009 from the thread path.
3010 (create_edge_and_update_destination_phis): Copy the threading path.
3011 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3012 jump threading path.
3013 (ssa_redirect_edges): Get edges and block types from the jump threading
3014 path. Free the path vector.
3015 (thread_block): Get edges from the jump threading path. Look at the
3016 entire path to see if we thread to a loop exit. If we cancel a jump
3017 thread request, then free the path vector.
3018 (thread_single_edge): Get edges and block types from the jump threading
3019 path. Free the path vector.
3020 (thread_through_loop_header): Get edges and block types from the jump
3021 threading path. Free the path vector.
3022 (mark_threaded_blocks): Iterate over the vector of paths and store
3023 the path on the appropriate edge. Get edges and block types from the
3024 jump threading path.
3025 (mark_threaded_blocks): Get edges and block types from the jump
3026 threading path. Free the path vector.
3027 (thread_through_all_blocks): Use the vector of paths rather than
3028 a vector of 3-edge sets.
3029 (register_jump_thread): Accept pointer to a path vector rather
3030 than the path vector itself. Store the path vector for later use.
3033 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3034 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3037 * config/s390/s390.c (s390_split_branches): Modify check for table
3039 (s390_chunkify_start): Rearrange table jump insn check in order to
3040 deal with compare and branch insns correctly.
3042 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3046 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3047 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3048 define_insn_and_split.
3049 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3050 (shiftsi3_compare): New pattern.
3052 * config/arm/unspecs.md (UNSPEC_RRX): New.
3054 2013-10-01 Alan Modra <amodra@gmail.com>
3056 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3057 casing inout operands.
3059 2013-10-01 Richard Biener <rguenther@suse.de>
3061 PR tree-optimization/58553
3062 * tree-loop-distribution.c (struct partition_s): Add niter member.
3063 (classify_partition): Populate niter member for the partition
3064 and properly identify whether the relevant store happens before
3065 or after the loop exit.
3066 (generate_memset_builtin): Use niter member from the partition.
3067 (generate_memcpy_builtin): Likewise.
3069 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3071 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3072 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3074 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3076 * basic-block.h (edge_list): Prefix member names with "m_".
3077 * context.h (context): Likewise.
3078 * domwalk.h (dom_walker): Likewise.
3079 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3080 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3081 * hash-table.h (hash_table): Likewise.
3082 * machmode.h (bit_field_mode_iterator): Likewise.
3083 * pass_manager.h (pass_list): Likewise.
3084 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3085 * tree-pass.h (pass_data): Likewise.
3086 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3087 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3088 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3089 * asan.c (pass_data_asan): Update accordingly.
3090 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3091 (control_dependences::control_dependences): Likewise.
3092 (control_dependences::~control_dependences): Likewise.
3093 (control_dependences::~control_dependences): Likewise.
3094 (control_dependences::get_edges_dependent_on): Likewise.
3095 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3096 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3097 * context.c (gcc::context::context): Likewise.
3098 * cprop.c (pass_rtl_cprop::clone): Likewise.
3099 * domwalk.c (dom_walker::walk): Likewise.
3100 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3101 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3102 * mode-switching.c (pass_mode_switching::clone): Likewise.
3103 * passes.c (opt_pass::opt_pass): Likewise.
3104 (pass_manager::pass_manager): Likewise.
3105 * predict.c (pass_strip_predict_hints::clone): Likewise.
3106 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3107 (pass_split_all_insns::clone): Likewise.
3108 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3110 (bit_field_mode_iterator::next_mode): Likewise.
3111 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3112 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3113 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3114 * tree-complex.c (pass_lower_complex::clone): Likewise.
3115 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3116 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3117 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3118 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3119 (pass_fold_builtins::clone): Likewise.
3120 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3121 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3122 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3123 (pass_cd_dce::clone): Likewise.
3124 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3125 (pass_phi_only_cprop::clone): Likewise.
3126 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3127 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3128 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3129 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3130 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3131 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3132 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3133 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3134 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3135 * tree-vrp.c (pass_vrp::clone): Likewise.
3136 * tsan.c (pass_tsan::clone): Likewise.
3138 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3141 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3142 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3145 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3146 optimization, punt if sign_bit_p looked through any zero extension.
3148 2013-09-30 Teresa Johnson <tejohnson@google.com>
3150 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3151 Update redirected out edge count in joiner case.
3152 (ssa_redirect_edges): Common the joiner and non-joiner cases
3153 so that joiner case gets profile updates.
3155 2013-09-30 Richard Biener <rguenther@suse.de>
3157 PR tree-optimization/58554
3158 * tree-loop-distribution.c (classify_partition): Require
3159 unconditionally executed stores for memcpy and memset recognition.
3160 (tree_loop_distribution): Calculate dominance info.
3162 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3164 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3165 (NO_PROFILE_COUNTERS): Likewise.
3166 (PROFILE_HOOK): Likewise.
3167 (FUNCTION_PROFILER): Likewise.
3168 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3170 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3172 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3173 calls and defines in TARGET_MACHO conditional.
3174 (load_macho_picbase_di): Likewise.
3175 (reload_macho_picbase): Likewise.
3176 (reload_macho_picbase_si): Likewise.
3177 (reload_macho_picbase_di): Likewise.
3178 (nonlocal_goto_receiver): Likewise.
3180 2013-09-30 Nick Clifton <nickc@redhat.com>
3182 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3183 that use the MSP430X ISA.
3184 (msp430_option_override): Scan -mmcu command line option for any
3185 MCU name that supports the MSP430X ISA.
3186 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3187 -mmcu options which enable the MSP430X ISA.
3189 2013-09-30 Richard Biener <rguenther@suse.de>
3192 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3193 before looking for setjmp-like calls.
3195 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3198 * config/darwin-protos.h (machopic_get_function_picbase): New.
3199 * config/darwin.c (machopic_get_function_picbase): New.
3200 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3201 label for a new func. (load_macho_picbase_di): Likewise.
3202 (reload_macho_picbase): New expand.
3203 (reload_macho_picbase_si): New insn.
3204 (reload_macho_picbase_di): New insn.
3205 (nonlocal_goto_receiver): New define and split.
3206 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3207 (unspecv enum): Add UNSPECV_NLGR.
3209 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3211 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3212 that altivec registers are correctly sized on Darwin.
3214 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3216 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3217 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3218 * config/x-darwin (host-darwin.o): Likewise.
3219 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3220 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3221 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3223 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3225 * doc/invoke.texi: Fix usage of @tie{} command.
3227 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3229 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3231 2013-09-28 Jan Hubicka <jh@suse.cz>
3233 * config/i386/x86-tune.def: Add documentation for each of the options;
3236 2013-09-28 Jan Hubicka <jh@suse.cz>
3238 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3240 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3241 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3242 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3243 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3244 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3246 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3248 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3249 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3250 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3251 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3252 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3253 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3254 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3255 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3256 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3257 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3258 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3259 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3260 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3261 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3262 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3263 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3264 gimple.h, godump.c, graphite-clast-to-gimple.c,
3265 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3266 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3267 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3268 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3269 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3270 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3271 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3272 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3273 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3274 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3275 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3276 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3277 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3278 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3279 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3280 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3281 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3282 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3283 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3284 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3285 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3286 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3287 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3288 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3289 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3290 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3291 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3292 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3293 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3294 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3295 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3296 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3297 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3298 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3299 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3300 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3301 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3302 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3303 whitespace before "(".
3305 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3307 * expr.h (extract_bit_field): Remove packedp parameter.
3308 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3309 from forward declaration.
3310 (store_split_bit_field): Remove packedp arg from calls to
3311 extract_fixed_bit_field.
3312 (extract_bit_field_1): Remove packedp parameter and packedp
3313 argument from recursive calls and calls to extract_fixed_bit_field.
3314 (extract_bit_field): Remove packedp parameter and corresponding
3315 arg to extract_bit_field_1.
3316 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3318 (extract_split_bit_field): Remove packedp arg from call to
3319 extract_fixed_bit_field.
3320 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3321 (copy_blkmode_from_reg): Likewise.
3322 (copy_blkmode_to_reg): Likewise.
3323 (read_complex_part): Likewise.
3324 (store_field): Likewise.
3325 (expand_expr_real_1): Likewise.
3326 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3327 to extract_bit_field.
3328 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3329 call to extract_bit_field.
3330 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3331 call to extract_bit_field.
3332 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3333 and special packedp behavior from -fstrict-volatile-bitfields
3336 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3338 * lra-eliminations.c (init_elim_table): Guard value_p.
3340 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3342 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3343 DFmode, DImode, and SFmode in the upper VSX registers based on the
3344 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3345 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3346 of the VSX register handling.
3348 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3350 (zero_extendsidi2_lfiwzx): Add support for loading into the
3351 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3352 only do VSX memory options on Altivec registers.
3353 (extendsidi2_lfiwax): Likewise.
3354 (extendsfdf2_fpr): Likewise.
3355 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3356 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3357 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3358 (movdi_internal64): Likewise.
3360 2013-09-27 Xinliang David Li <davidxl@google.com>
3362 * opts.c (finish_options): Adjust parameters
3363 according to vect cost model.
3364 (common_handle_option): Set dynamic vect cost
3366 targhooks.c (default_add_stmt_cost): Compute stmt cost
3368 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3369 Use helper function.
3370 * tree-vectorizer.h (unlimited_cost_model): New function.
3371 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3372 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3374 (vect_enhance_data_refs_alignment): Ditto.
3375 * flag-types.h: New enum.
3376 * common/config/i386/i386-common.c (ix86_option_init_struct):
3377 No need to initialize vect_cost_model flag.
3378 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3381 2013-09-27 Diego Novillo <dnovillo@google.com>
3383 * gimple.h (enum ssa_mode): Remove.
3385 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3387 * cfgloop.h (number_of_loops): Fix typo in check for null.
3389 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3392 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3393 are children of outermost saved_cfun's loop, and set it up to
3394 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3395 if loop != loop0 and SESE region contains bbs that belong to loop0.
3397 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3399 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3400 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3401 (set_address_segment, set_address_base, set_address_index)
3402 (set_address_disp): Accept the argument unconditionally.
3403 (baseness): Remove must_be_base_p and must_be_index_p checks.
3404 (decompose_normal_address): Classify as much as possible in the
3407 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3409 * cse.c (count_reg_usage): Handle INT_LIST.
3410 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3411 * reginfo.c (reg_scan_mark_refs): Likewise.
3412 * reload1.c (eliminate_regs_1): Likewise.
3414 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3417 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3420 2013-09-27 Richard Biener <rguenther@suse.de>
3422 PR tree-optimization/58459
3423 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3424 restriction not propagating into loops.
3426 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3428 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3429 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3430 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3432 2013-09-26 Richard Biener <rguenther@suse.de>
3434 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3436 2013-09-26 Richard Biener <rguenther@suse.de>
3438 * alias.h (component_uses_parent_alias_set): Rename to ...
3439 (component_uses_parent_alias_set_from): ... this.
3440 * alias.c (component_uses_parent_alias_set): Rename to ...
3441 (component_uses_parent_alias_set_from): ... this and return
3443 (reference_alias_ptr_type_1): Use the result from
3444 component_uses_parent_alias_set_from instead of stripping
3445 components one at a time.
3446 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3448 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3450 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3451 Move prototypes to...
3452 * tree-ssa-ter.h: New File. Move prototypes here.
3453 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3454 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3455 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3456 from is_replaceable_p.
3457 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3458 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3459 (process_replaceable): Use ter_is_replaceable_p.
3460 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3461 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3462 newly refactored ssa_is_replaceable_p.
3463 * cfgexpand.c: Include tree-outof-ssa.h.
3464 * ssaexpand.h: Delete.
3466 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3468 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3469 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3471 (create_gimple_tmp): Delete.
3472 (get_expr_type, build_assign, build_type_cast): Move to...
3473 * gimple-builder.c: New File.
3474 (get_expr_type): Relocate from gimple.c.
3475 (build_assign, build_type_cast): Change to only create ssanames.
3476 * gimple.h: Move prototypes to...
3477 * gimple-builder.h: New File. Here.
3478 * tree-ssa.h: And here.
3479 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3480 count_uses_and_derefs): Relocate from gimple.c.
3481 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3482 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3483 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3484 gimple_replace_ssa_lhs.
3485 * asan.c: Include gimple-builder.h.
3486 * Makefile.in: Add gimple-builder.o.
3488 2013-09-26 Richard Biener <rguenther@suse.de>
3490 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3491 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3492 (loe_visit_block): Use gcc_checking_assert.
3493 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3494 gimple_assign_ssa_name_copy_p.
3495 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3498 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3500 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3501 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3503 2013-09-26 Richard Biener <rguenther@suse.de>
3505 PR tree-optimization/58539
3506 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3507 the fact that debug statements are not taking part in loop-closed
3510 2013-09-26 Nick Clifton <nickc@redhat.com>
3512 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3513 time warning message.
3514 (msp430_print_operand_raw): Delete unused letter parameter.
3515 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3516 (msp430_print_operand_address): New function.
3517 (msp430_print_operand): Move address printing code from here to
3519 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3521 (zero_extendpsisi2): Likewise.
3522 (extendpsisi2): New pattern.
3523 (andneghi3): New pattern.
3525 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3527 * config/aarch64/aarch64.opt (mlra): New option.
3528 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3529 (TARGET_LRA_P): Define.
3531 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3533 * expr.c (expand_assignment): Remove obsolete comment.
3535 2013-09-25 Jeff Law <law@redhat.com>
3537 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3538 tree-ssa-threadupdate.h.
3539 (register_jump_thread): Similarly.
3540 * tree-ssa-threadupdate.h: New header file.
3541 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3542 * tree-vrp.c: Likewise.
3543 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3544 (thread_around_empty_blocks): Change type of path vector argument to
3545 an edge,type pair from just an edge. Initialize both elements when
3546 appending to a jump threading path. Tweak references to elements
3548 (thread_across_edge): Similarly. Release memory for the elements
3550 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3551 (dump_jump_thread_path): New function broken out from
3552 register_jump_thread.
3553 (register_jump_thread): Use dump_jump_thread_path. Change type of
3554 path vector entries. Search the path for NULL edges and dump
3555 the path if one is found. Tweak the conversion of path to 3-edge
3556 form to use the block copy type information embedded in the path.
3558 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3560 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3562 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3563 Vladimir Makarov <vmakarov@redhat.com>
3565 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3566 from the least significant bit.
3567 (strip_address_mutations): Add bitfield operations handling.
3568 (must_be_index_p): Add shifting and rotate operations handling.
3569 (set_address_base): Use must_be_base_p predicate.
3570 (set_address_index): Use must_be_index_p predicate.
3572 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3573 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3574 Sergey Lega <sergey.s.lega@intel.com>
3575 Anna Tikhonova <anna.tikhonova@intel.com>
3576 Ilya Tocar <ilya.tocar@intel.com>
3577 Andrey Turetskiy <andrey.turetskiy@intel.com>
3578 Ilya Verbin <ilya.verbin@intel.com>
3579 Kirill Yukhin <kirill.yukhin@intel.com>
3580 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3582 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3584 (ix86_expand_vector_move_misalign): Support new unaligned load and
3585 stores and use new names.
3586 (CODE_FOR_sse2_storedqu): Rename to ...
3587 (CODE_FOR_sse2_storedquv16qi): ... this.
3588 (CODE_FOR_sse2_loaddqu): Rename to ...
3589 (CODE_FOR_sse2_loaddquv16qi): ... this.
3590 (CODE_FOR_avx_loaddqu256): Rename to ...
3591 (CODE_FOR_avx_loaddquv32qi): ... this.
3592 (CODE_FOR_avx_storedqu256): Rename to ...
3593 (CODE_FOR_avx_storedquv32qi): ... this.
3594 * config/i386/i386.md (fpint_logic): New.
3595 * config/i386/sse.md (VMOVE): Extend for AVX512.
3599 (VI_UNALIGNED_LOADSTORE): Ditto.
3600 (sse2_avx_avx512f): Ditto.
3601 (sse2_avx2): Extend for AVX512.
3602 (sse4_1_avx2): Ditto.
3603 (avx2_avx512f): New.
3604 (sse): Extend for AVX512.
3607 (avxsizesuffix): Ditto.
3608 (sseintvecmode): Ditto.
3610 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3611 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3612 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3613 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3614 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3615 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3616 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3617 (STORENT_MODE): Extend for AVX512.
3618 (*absneg<mode>2): Replace constraint "x" with "v".
3619 (*mul<mode>3): Ditto.
3620 (*ieee_smin<mode>3): Ditto.
3621 (*ieee_smax<mode>3): Ditto.
3622 (avx_cmp<mode>3): Replace VF with VF_128_256.
3623 (*<sse>_maskcmp<mode>3_comm): Ditto.
3624 (<sse>_maskcmp<mode>3): Ditto.
3625 (<sse>_andnot<mode>3): Extend for AVX512.
3626 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3627 (<code><mode>3, fpint_logic): New.
3628 (*<code><mode>3): Extend for AVX512.
3629 (avx512flogicsuff): New.
3630 (avx512f_<logic><mode>): Ditto.
3631 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3633 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3634 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3635 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3636 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3637 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3638 (xop_vpermil2<mode>3): Ditto.
3639 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3640 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3641 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3642 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3644 2013-09-25 Tom Tromey <tromey@redhat.com>
3646 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3647 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3648 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3649 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3650 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3651 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3652 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3653 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3654 (GRAPHITE_HTAB_H): Remove.
3656 2013-09-25 Tom Tromey <tromey@redhat.com>
3658 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3660 2013-09-25 Tom Tromey <tromey@redhat.com>
3662 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3664 2013-09-25 Tom Tromey <tromey@redhat.com>
3666 * config/i386/t-i386 (i386.o): Remove.
3667 (i386-c.o): Use COMPILE and POSTCOMPILE.
3669 2013-09-25 Tom Tromey <tromey@redhat.com>
3671 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3673 2013-09-25 Tom Tromey <tromey@redhat.com>
3675 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3676 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3677 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3678 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3679 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3680 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3681 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3682 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3683 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3684 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3685 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3686 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3687 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3688 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3689 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3690 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3691 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3692 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3693 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3694 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3695 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3696 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3697 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3698 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3699 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3700 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3701 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3702 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3703 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3704 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3705 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3706 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3707 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3708 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3709 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3710 (graphite-clast-to-gimple.o, graphite-dependences.o)
3711 (graphite-interchange.o, graphite-poly.o)
3712 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3713 (graphite-optimize-isl.o, tree-vect-loop.o)
3714 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3715 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3716 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3717 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3718 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3719 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3720 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3721 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3722 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3723 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3724 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3725 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3726 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3727 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3728 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3729 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3730 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3731 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3732 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3733 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3734 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3735 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3736 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3737 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3738 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3739 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3740 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3741 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3742 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3743 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3744 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3745 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3746 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3747 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3748 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3749 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3750 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3751 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3752 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3753 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3754 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3755 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3756 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3757 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3758 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3759 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3760 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3761 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3762 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3763 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3764 (gcov-dump.o): Remove.
3765 (default-c.o): Use COMPILE and POSTCOMPILE.
3766 (CFLAGS-gcc.o): New variable.
3767 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3769 2013-09-25 Tom Tromey <tromey@redhat.com>
3771 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3772 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3773 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3774 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3775 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3776 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3777 (c-family/array-notation-common.o, c-family/stub-objc.o)
3778 (c-family/c-ubsan.o): Remove.
3780 2013-09-25 Tom Tromey <tromey@redhat.com>
3782 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3784 2013-09-25 Tom Tromey <tromey@redhat.com>
3786 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3787 to add -DENABLE_SHARED_LIBGCC.
3788 (gcc.o): Don't use subshell.
3790 2013-09-25 Tom Tromey <tromey@redhat.com>
3792 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3793 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3794 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3795 * configure, config.in: Rebuild.
3797 2013-09-25 Tom Tromey <tromey@redhat.com>
3799 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3800 (COMPILE, POSTCOMPILE): New variables.
3801 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3802 (DEPFILES): New variable.
3803 Include ".Po" files.
3804 * configure.ac: Add checks for dependency checking.
3805 * configure, aclocal.m4: Regenerate.
3807 2013-09-25 Tom Tromey <tromey@redhat.com>
3809 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3810 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3812 2013-09-25 Tom Tromey <tromey@redhat.com>
3814 * Makefile.in (generated_files): Add options.h,
3815 target-hooks-def.h, insn-opinit.h,
3816 common/common-target-hooks-def.h, pass-instances.def,
3817 c-family/c-target-hooks-def.h.
3819 2013-09-25 Jeff Law <law@redhat.com>
3821 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3822 than foo[foo.length () - 1] to access last member in a vec.
3823 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3825 2013-09-25 Richard Biener <rguenther@suse.de>
3828 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3830 2013-09-25 Jan Hubicka <jh@suse.cz>
3832 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3835 2013-09-25 Marek Polacek <polacek@redhat.com>
3838 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3839 when determining the type name.
3841 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3843 * config/sh/sh.md: Fix formatting.
3845 2013-09-24 Xinliang David Li <davidxl@google.com>
3847 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3848 max peel iterations parameter.
3849 * param.def: New parameter.
3850 * doc/invoke.texi: Document New parameter.
3852 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3854 * gimple-pretty-print.c: Various whitespace tweaks.
3855 * tree-core.h: Likewise.
3856 * tree-pretty-print.c: Likewise.
3857 * tree-ssa-alias.c: Likewise.
3858 * tree-ssa-copy.c: Likewise.
3859 * tree-ssanames.c: Likewise.
3860 * tree-ssanames.h: Likewise.
3861 * tree-vrp.c: Likewise.
3863 2013-09-24 Alan Modra <amodra@gmail.com>
3867 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3868 for output operands that disallow regs. Don't use EXPAND_WRITE on
3871 2013-09-24 Richard Biener <rguenther@suse.de>
3874 * tree.c (reference_alias_ptr_type): Move ...
3875 * alias.c (reference_alias_ptr_type): ... here and implement
3876 in terms of the new reference_alias_ptr_type_1.
3877 (ref_all_alias_ptr_type_p): New helper.
3878 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3879 use ref_all_alias_ptr_type_p.
3880 (get_deref_alias_set): Add flag_strict_aliasing check here.
3881 (reference_alias_ptr_type_1): New function, split out from ...
3882 (get_alias_set): ... here.
3883 (alias_ptr_types_compatible_p): New function.
3884 * alias.h (reference_alias_ptr_type): Declare.
3885 (alias_ptr_types_compatible_p): Likewise.
3886 * tree.h (reference_alias_ptr_type): Remove.
3887 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3888 to compare MEM_REF alias types.
3890 2013-09-24 Richard Biener <rguenther@suse.de>
3892 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3894 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3896 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3897 reload helper function arrays into a single array reg_addr.
3898 (reload_fpr_gpr): Likewise.
3899 (reload_gpr_vsx): Likewise.
3900 (reload_vsx_gpr): Likewise.
3901 (struct rs6000_reg_addr): Likewise.
3902 (reg_addr): Likewise.
3903 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3904 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3905 (rs6000_init_hard_regno_mode_ok): Likewise.
3906 (rs6000_secondary_reload_direct_move): Likewise.
3907 (rs6000_secondary_reload): Likewise.
3909 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3910 constraints: wu, ww, and wy. Repurpose wv constraint added during
3911 power8 changes. Put wg constraint in alphabetical order.
3913 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3914 for future work to add ISA 2.07 VSX single precision support.
3915 (-mvsx-scalar-double): Change default from -1 to 1, update
3916 documentation comment.
3917 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3918 (-mupper-regs-df): New debug switch to control whether DF values
3919 can go in the traditional Altivec registers.
3920 (-mupper-regs-sf): New debug switch to control whether SF values
3921 can go in the traditional Altivec registers.
3923 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3925 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3926 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3927 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3928 (rs6000_debug_legitimate_address_p): Print if we are running
3929 before, during, or after reload.
3930 (rs6000_secondary_reload): Add a comment.
3931 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3933 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3934 constraints. Update documentation string.
3935 (wd constraint): Likewise.
3936 (wf constraint): Likewise.
3937 (wg constraint): Likewise.
3938 (wn constraint): Likewise.
3939 (ws constraint): Likewise.
3940 (wt constraint): Likewise.
3941 (wx constraint): Likewise.
3942 (wz constraint): Likewise.
3943 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3945 (ww constraint): Likewise.
3946 (wy constraint): Likewise.
3947 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3948 the previous submissions.
3949 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3951 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3953 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3954 (REG_BR_PROB): Say that the probability is stored in an int_list.
3955 * reg-notes.def: Update commentary to mention INT_LIST.
3956 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3957 (INT_LIST): New rtx.
3958 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3959 * rtlanal.c (int_reg_note_p): New function.
3960 (alloc_reg_note): Assert that the note does not have an int argument.
3961 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3962 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3963 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3964 rather than an INSN_LIST. Handle INT_LIST.
3965 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3966 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3967 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3968 Manipulate them as ints rather than rtxes.
3969 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3970 * regmove.c (copy_src_to_dest): Likewise.
3971 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3973 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3974 into the cases that need it.
3975 * config/arm/arm.c (arm_unwind_emit): Likewise.
3977 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3978 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3979 * loop-doloop.c (add_test, doloop_modify): Likewise.
3980 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3981 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3982 * predict.c (combine_predictions_for_insn): Likewise.
3983 * print-rtl.c (print_rtx): Handle INT_LIST.
3984 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3985 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3986 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3987 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3988 (ix86_split_fp_branch, predict_jump): Likewise.
3989 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3990 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3991 * config/spu/spu.c (ea_load_store_inline): Likewise.
3993 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3994 value of a REG_BR_PROB note.
3995 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3996 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3997 * emit-rtl.c (try_split): Likewise.
3998 * predict.c (br_prob_note_reliable_p): Likewise.
3999 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4000 * reorg.c (mostly_true_jump): Likewise.
4001 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4002 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4003 * config/i386/i386.c (ix86_print_operand): Likewise.
4004 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4005 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4006 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4007 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4008 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4009 * config/sparc/sparc.c (output_cbranch): Likewise.
4010 * config/spu/spu.c (get_branch_target): Likewise.
4011 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4012 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4014 2013-09-23 Jan Hubicka <jh@suse.cz>
4016 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4018 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4019 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4021 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4022 (gimple_fold_call): Dump inconsistent devirtualizations; add
4023 sanity check for type based devirtualizations.
4024 * ipa-prop.c: Include ipa-utils.h
4025 (ipa_intraprocedural_devirtualization): Add sanity check.
4026 (try_make_edge_direct_virtual_call): Likewise.
4028 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4030 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4031 assignment statements.
4033 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4035 * gimple-pretty-print.c (dump_ssaname_info): New function.
4036 (dump_gimple_phi): Call it.
4037 (pp_gimple_stmt_1): Likewise.
4038 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4039 (range_info_def): Declare.
4040 * tree-pretty-print.c (pp_double_int): New function.
4041 (dump_generic_node): Call it.
4042 * tree-pretty-print.h (pp_double_int): Declare.
4043 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4044 * tree-ssanames.h (range_info_def): New structure.
4045 (value_range_type): Move definition here.
4046 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4048 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4050 (set_range_info): New function.
4051 (get_range_info): Likewise.
4052 (duplicate_ssa_name_range_info): Likewise.
4053 (duplicate_ssa_name_fn): Check pointer type and call
4054 duplicate_ssa_name_range_info.
4055 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4056 * tree-vrp.c (value_range_type): Remove definition, now in
4058 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4059 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4060 (SSA_NAME_RANGE_INFO): New macro.
4062 2013-09-23 Richard Biener <rguenther@suse.de>
4064 PR tree-optimization/58464
4065 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4066 (phi_trans_add): Change to add conditionally on being not present.
4067 (phi_translate_1): Remove recursion detection here.
4068 (phi_translate): Pre-seed the cache with NULL to catch
4069 recursion here in a more generic way.
4070 (bitmap_find_leader): Adjust comment.
4071 (get_representative_for): Dump value-numbers.
4072 (create_expression_by_pieces): Likewise.
4073 (insert_into_preds_of_block): Likewise.
4075 2013-09-23 Christian Bruel <christian.bruel@st.com>
4078 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4079 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4081 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4084 2013-09-20 Renlin Li <renlin.li@arm.com>
4086 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4088 (aarch64_expand_epilogue): Likewise.
4089 (aarch64_legitimize_reload_address): Likewise.
4091 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4093 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4094 NULL_TREE before pushing them onto the vector. Likewise for labels.
4096 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4098 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4099 comparison operators when -fno-trapping-math is in effect.
4100 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4101 comparison operators in TFmode and assert that unsupported operators
4103 (ia64_print_operand): Likewise.
4105 2013-09-21 Jan Hubicka <jh@suse.cz>
4107 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4108 (sse_typeless_stores): Enable for core
4109 (sse_load0_by_pxor): Likewise.
4110 (four_jump_limit): Disable for core.
4111 (pad_returns): Likewise.
4112 (avoid_vector_decode): Likewise.
4113 (fuse_cmp_and_branch): Enable for cores.
4114 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4116 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4119 * config/pa/pa.c (pa_option_override): Disable auto increment and
4120 decrement instructions until reload is completed.
4122 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4123 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4124 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4126 2013-09-20 DJ Delorie <dj@redhat.com>
4127 Nick Clifton <nickc@redhat.com>
4129 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4130 (need_to_save): Save bank 0 on interrupts.
4131 (characterize_address): Strip far address wrappers.
4132 (rl78_as_legitimate_address): Likewise.
4133 (transcode_memory_rtx): Likewise.
4134 (rl78_peep_movhi_p): Disable this peephole after devirt.
4135 (rl78_propogate_register_origins): Forget all origins when a
4137 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4138 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4140 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4141 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4142 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4144 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4146 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4148 * config/pa/pa.md: In "scc" insn patterns, change output template to
4149 handle const0_rtx in reg_or_0_operand operands.
4151 2013-09-20 Martin Husemann <martin@NetBSD.org>
4154 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4155 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4157 2013-09-20 Richard Biener <rguenther@suse.de>
4160 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4161 remembering SSA name version and block index.
4162 (new_scev_info_str): Adjust.
4163 (hash_scev_info): Likewise. Also hash the block index.
4164 (eq_scev_info): Adjust.
4165 (find_var_scev_info): Likewise.
4166 (struct instantiate_cache_entry): Remove.
4167 (struct instantiate_cache_type): Use a htab to map name, block
4169 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4170 (get_instantiated_value_entry): Likewise.
4171 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4172 (instantiate_scev_name): Adjust.
4174 2013-09-20 Jeff Law <law@redhat.com>
4176 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4178 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4180 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4181 Call aarch64_simd_expand_args to update op[argc].
4183 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4185 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4188 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4190 * gengtype.c (file_rules): Added rule for *.cc files.
4191 (get_output_file_with_visibility): Give fatal message when no
4194 2013-09-20 Renlin Li <renlin.li@arm.com>
4196 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4197 (aarch64_expand_epilogue): Likewise.
4198 (aarch64_legitimize_reload_address): Likewise.
4200 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4203 * expr.c (expand_assignment): Remove misalignp code path.
4205 2013-09-20 Marek Polacek <polacek@redhat.com>
4208 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4209 TYPE_PRECISION. Add asserts.
4211 2013-09-20 Richard Biener <rguenther@suse.de>
4213 PR tree-optimization/58453
4214 * tree-loop-distribution.c (distribute_loop): Apply the cost
4215 model for -ftree-loop-distribute-patterns, too.
4217 2013-09-20 Richard Biener <rguenther@suse.de>
4220 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4221 make type comparison less strict.
4223 2013-09-20 Alan Modra <amodra@gmail.com>
4225 * configure: Regenerate.
4226 * aclocal.m4: Regenerate.
4228 2013-09-20 Marek Polacek <polacek@redhat.com>
4231 * doc/extend.texi: Document that attribute used is meant to be used
4232 on variables with static storage duration.
4234 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4236 PR tree-optimization/58472
4237 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4238 simd_lane_access set inv_p = false.
4239 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4240 the simduid magic VAR_DECL.
4242 2013-09-19 Jan Hubicka <jh@suse.cz>
4244 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4246 2013-09-17 Jeff Law <law@redhat.com>
4248 * tree-ssa-dom.c (record_temporary_equivalences): New function
4249 split out of dom_opt_dom_walker::after_dom_children.
4250 (dom_opt_dom_walker::thread_across_edge): Move common code
4251 in here from dom_opt_dom_walker::after_dom_children.
4252 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4254 2013-09-19 Jan Hubicka <jh@suse.cz>
4256 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4257 (TARGET_GENERIC): Use PROCESOR_GENERIC
4258 (enum processor_type): Unify generic32 and 64.
4259 * i386.md (cpu): Likewise.
4260 * x86-tune.def (use_leave): Enable for generic32.
4261 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4262 * athlon.md: Change generic64 to generic in all occurences.
4263 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4264 (ix86_target_macros_internal): Likewise.
4265 * driver-i386.c (host_detect_local_cpu): Likewise.
4266 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4268 (generic_memcpy, generic_memset, generic_cost): This one.
4269 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4270 (m_GENERIC32, m_GENERIC64): Remove.
4271 (m_GENERIC): Turn into one flag.
4272 (processor_target): Unify generic tunnings.
4273 (ix86_option_override_internal): Replace generic32/64 by generic.
4274 (ix86_issue_rate): Likewise.
4275 (ix86_adjust_cost): Likewise.
4277 2013-09-19 Jan Hubicka <jh@suse.cz>
4279 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4280 of speculative flag.
4282 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4284 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4285 GOMP_sections_start, even if !exit_reachable.
4287 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4289 * lra-constraints.c (need_for_all_save_p): Use macro
4290 HARD_REGNO_CALL_PART_CLOBBERED.
4291 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4292 set up pseudo conflict hard regs.
4294 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4297 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4300 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4302 PR rtl-optimization/58438
4303 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4304 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4305 from previous subpasses.
4307 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4309 * arm.c (arm_get_frame_offsets): Validate architecture supports
4310 LDRD/STRD before accepting the tuning preference.
4311 (arm_expand_prologue): Likewise.
4312 (arm_expand_epilogue): Likewise.
4314 2013-09-18 Richard Biener <rguenther@suse.de>
4316 PR tree-optimization/58417
4317 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4318 have chrecs with symbols defined in the loop as operands.
4319 (chrec_fold_multiply): Likewise.
4320 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4321 parameters before folding binary operations.
4322 (struct instantiate_cache_entry_hasher): Remove.
4323 (struct instantiate_cache_type): Use a pointer-map.
4324 (instantiate_cache_type::instantiate_cache_type): New function.
4325 (instantiate_cache_type::get): Likewise.
4326 (instantiate_cache_type::set): Likewise.
4327 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4328 (get_instantiated_value_entry): Likewise.
4329 (global_cache): New global.
4330 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4331 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4332 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4333 (instantiate_scev_name): Adjust.
4334 (instantiate_scev): Construct global instead of local cache.
4335 (resolve_mixers): Likewise.
4337 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4338 Paolo Carlini <paolo.carlini@oracle.com>
4341 * doc/implement-cxx.texi: Fix references to the C++ standards.
4343 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4345 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4346 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4347 simd_lane_access drs, update also DR_ALIGNED_TO.
4349 2013-09-18 Marek Polacek <polacek@redhat.com>
4352 * doc/extend.texi: Document no_sanitize_undefined attribute.
4353 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4354 no_sanitize_undefined attribute.
4356 2013-09-18 Nick Clifton <nickc@redhat.com>
4358 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4359 (ASM_DECLARE_FUNCTION_NAME): Define.
4361 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4363 * compare-elim.c (find_comparison_dom_walker): New class
4364 (find_comparisons_in_bb): Rename to
4365 find_comparison_dom_walker::before_dom_children
4366 (find_comparisons): Adjust
4367 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4369 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4370 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4371 (init_walk_dominator_tree): Remove declaration.
4372 (fini_walk_dominator_tree): Remove declaration.
4373 * fwprop.c (single_def_use_dom_walker): New class
4374 (single_def_use_enter_block): Convert to
4375 single_def_use_dom_walker::before_dom_children.
4376 (single_def_use_leave_block): Convert to
4377 single_def_use_dom_walker::after_dom_children.
4378 (build_single_def_use_links): Adjust.
4379 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4381 (find_candidates_in_block): Convert to
4382 find_candidates_dom_walker::before_dom_children.
4383 (execute_strength_reduction): Adjust.
4384 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4385 (sese_dom_walker): New class.
4386 (sese_dom_walker::sese_dom_walker): New constructor.
4387 (sese_dom_walker::~sese_dom_walker): New destructor.
4388 (build_sese_conditions_before): Convert to
4389 sese_dom_walker::before_dom_children.
4390 (build_sese_conditions_after): Convert to
4391 sese_dom_walker::after_dom_children.
4392 (build_poly_scop): Adjust
4393 * tree-into-ssa.c (rewrite_dom_walker): New class
4394 (rewrite_enter_block): Convert to
4395 rewrite_dom_walker::before_dom_children.
4396 (rewrite_leave_block): Convert to
4397 rewrite_dom_walker::after_dom_children.
4398 (rewrite_update_dom_walker): New class.
4399 (rewrite_update_enter_block): Convert to
4400 rewrite_update_dom_walker::before_dom_children.
4401 (rewrite_update_leave_block): Convert to
4402 rewrite_update_dom_walker::after_dom_children.
4403 (rewrite_blocks, rewrite_into_ssa): Adjust.
4404 (mark_def_dom_walker): New class.
4405 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4406 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4407 (mark_def_sites_blocks): Convert to
4408 mark_def_dom_walker::before_dom_children.
4409 (mark_def_site_blocks): Remove.
4410 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4411 (tree_ssa_dominator_optimize): Adjust.
4412 (dom_thread_across_edge): Convert to method
4413 dom_opt_dom_walker::thread_across_edge.
4414 (dom_opt_enter_block): Convert to member function
4415 dom_opt_dom_walker::before_dom_children.
4416 (dom_opt_leave_block): Convert to member function
4417 dom_opt_dom_walker::after_dom_children.
4418 * tree-ssa-dse.c (dse_dom_walker): New class.
4419 (dse_enter_block): Convert to member function
4420 dse_dom_walker::before_dom_children.
4421 (tree_ssa_dse): Adjust.
4422 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4423 (determine_invariantness_stmt): Convert to method
4424 invariantness_dom_walker::before_dom_children.
4425 (determine_invariantness): Remove
4426 (move_computations_dom_walker): New class.
4427 (move_computations_stmt): Convert to method
4428 move_computations_dom_walker::before_dom_children.
4429 (move_computations, tree_ssa_lim): Adjust.
4430 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4431 (nt_init_block): Make method
4432 notrappping_dom_walker::before_dom_children.
4433 (nt_fini_block): Make
4434 method nontrapping_dom_walker::after_dom_children.
4435 (get_non_trapping): Adjust.
4436 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4437 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4438 (eliminate_leave_block): Make method.
4439 eliminate_dom_walker::after_dom_children.
4441 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4442 (strlen_enter_block): Make method
4443 strlen_dom_walker::before_dom_children.
4444 (strlen_leave_block): Make
4445 method strlen_dom_walker::after_dom_children.
4446 (tree_ssa_strlen): Adjust.
4447 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4448 (tree_ssa_uncprop): Adjust.
4449 (uncprop_leave_block): Make method
4450 uncprop_dom_walker::after_dom_children.
4451 (uncprop_leave_block): Make method
4452 uncprop_dom_walker::before_dom_children.
4454 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4456 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4457 branch in same basic block. Check both src and dest of the move insn.
4459 2013-09-17 Nick Clifton <nickc@redhat.com>
4461 * config/rl78/rl78-real.md (bf): New pattern.
4463 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4464 (rl78_print_operand): Do not put a # before a %B.
4465 * config/rl78/rl78.opt: Tweak doc strings.
4467 2013-09-17 DJ Delorie <dj@redhat.com>
4469 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4470 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4471 (rl78_option_override): Likewise, if -mallregs.
4472 (is_virtual_register): Likewise.
4473 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4474 (REGNO_OK_FOR_BASE_P): Likewise.
4475 * config/rl78/rl78.opt (-mallregs): New.
4477 2013-09-17 Nick Clifton <nickc@redhat.com>
4479 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4480 For interrupt functions: save all call clobbered registers if the
4481 interrupt handler is not a leaf function.
4482 (rl78_expand_prologue): Always recompute the frame information.
4483 For interrupt functions: only select bank 0 if one of the bank 0
4484 registers is going to be psuhed.
4486 2013-09-17 DJ Delorie <dj@redhat.com>
4488 * config/rl78/constraints.md: For each W* constraint, rename to C*
4489 and create a W* constraint that checks for an optional ES: prefix
4491 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4492 (es_addr): New. Used to wrap far addresses.
4493 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4494 (rl78_es_base): New.
4495 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4496 wrapped far addresses.
4497 (rl78_print_operand_1): Unwrap far addresses before processing.
4498 (rl78_lo16): Wrap far addresses in unspecs.
4499 (rl78_es_addr): New.
4500 (rl78_es_base): New.
4501 (insn_ok_now): Check for not-yet-wrapped far addresses.
4502 (transcode_memory_rtx): Properly re-wrap far addresses.
4504 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4506 * config/sparc/t-rtems: Add leon3 multilibs.
4508 2013-09-17 Cong Hou <congh@google.com>
4510 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4511 when checking the dot production pattern. The type of rhs operand
4512 of multiply is now checked correctly.
4514 2013-09-17 Jeff Law <law@redhat.com>
4516 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4517 edge implied equivalences into successor phis.
4518 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4519 here from tree-ssa-threadedge.c.
4520 (mark_threaded_blocks): When threading through a joiner, if both
4521 successors of the joiner's clone reach the same block, verify the
4522 PHI arguments are equal. If not, cancel the jump threading request.
4523 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4524 tree-ssa-threadupdate.c
4525 (thread_across_edge): Don't check PHI argument equality when
4526 threading through joiner block here.
4528 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4530 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4531 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4532 (warn_uninit, warn_uninitialized_vars,
4533 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4534 Move to tree-ssa-uninit.c.
4535 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4536 (has_undefined_value_p): New. Pass dependant parts of
4537 ssa_undefined_value_p.
4538 (uninit_undefined_value_p): Use has_undefined_value_p.
4539 (warn_uninit, warn_uninitialized_vars,
4540 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4541 Move from tree-ssa.c.
4542 * tree-ssa.h: Adjust prototypes.
4544 2013-09-17 Jan Hubicka <jh@suse.cz>
4547 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4548 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4549 FUNCTION_NOT_OPTIMIZED.
4550 * ipa-inline-analysis.c (compute_inline_parameters): Function
4551 not optimized is not inlinable unless it is alwaysinline.
4552 (inline_analyze_function): Force calls in not optimized
4553 function not inlinable.
4555 2013-09-17 Jan Hubicka <jh@suse.cz>
4558 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4560 * ipa.c (function_and_variable_visibility): Likewise.
4561 * ipa-profile.c (ipa_profile): Likewise.
4563 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4566 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4567 attribute, and don't inline the resolver in this case.
4569 2013-09-17 Teresa Johnson <tejohnson@google.com>
4571 * coverage.c (get_coverage_counts): Add missing newline.
4573 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4575 PR tree-optimization/58088
4576 * fold-const.c (mask_with_trailing_zeros): New function.
4577 (fold_binary_loc): Make sure we don't recurse infinitely
4578 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4579 Use mask_with_trailing_zeros where appropriate.
4581 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4583 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4584 of 'prev' var to get better distance estimation.
4586 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4588 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4589 * tree-inline.c (is_parameter_of): New predicate.
4590 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4591 a parameter has been remapped.
4592 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4593 (optimize_inline_calls): Initialize transform_parameter.
4594 (copy_gimple_seq_and_replace_locals): Likewise.
4595 (tree_function_versioning): Likewise.
4596 (maybe_inline_call_in_expr): Likewise.
4598 2013-09-17 Nick Clifton <nickc@redhat.com>
4600 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4601 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4603 (is_attr_func): New function.
4604 (msp430_is_interrupt_func): New function.
4605 (is_naked_func): New function.
4606 (is_reentrant_func): New function.
4607 (is_critical_func): New function.
4608 (msp430_start_function): Add annotations for function attributes.
4609 (msp430_attr): New function.
4610 (msp430_attribute_table): New.
4611 (msp430_function_section): New function.
4612 (TARGET_ASM_FUNCTION_SECTION): Define.
4613 (msp430_builtin): New enum.
4614 (msp430_init_builtins): New function.
4615 (msp430_builtin_devl): New function.
4616 (msp430_expand_builtin): New function.
4617 (TARGET_INIT_BUILTINS): Define.
4618 (TARGET_EXPAND_BUILTINS): Define.
4619 (TARGET_BUILTIN_DECL): Define.
4620 (msp430_expand_prologue): Add support for naked, interrupt,
4621 critical and reentrant functions.
4622 (msp430_expand_epilogue): Likewise.
4623 (msp430_print_operand): Handle 'O' character.
4624 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4626 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4627 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4628 (pushm): Use a 'n' rather than an 'i' constraint.
4629 (msp_return): Add generation of the interrupt return instruction.
4630 (disable_interrupts): New pattern.
4631 (enable_interrupts): New pattern.
4632 (push_intr_state): New pattern.
4633 (pop_intr_state): New pattern.
4634 (bic_SR): New pattern.
4635 (bis_SR): New pattern.
4636 * doc/extend.texi: Document MSP430 function attributes and builtin
4639 2013-09-17 Richard Biener <rguenther@suse.de>
4641 PR tree-optimization/58432
4642 * tree-loop-distribution.c (tree_loop_distribution): Also
4643 scan PHIs for outside loop uses and seed a partition from them.
4645 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4647 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4648 (restructure_reference): Call backtrace_base_for_ref.
4650 2013-09-17 Alan Modra <amodra@gmail.com>
4653 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4656 2013-09-16 DJ Delorie <dj@redhat.com>
4658 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4659 vregs location for RL78/G10.
4660 (rl78_expand_prologue): Avoid SEL on G10.
4661 (rl78_expand_epilogue): Likewise.
4662 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4663 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4664 __RL78_G10__ when appropriate.
4665 (ASM_SPEC): Pass -mg10 along to the assembler.
4666 * config/rl78/rl78.md (sel_rb): Disable for G10.
4667 * config/rl78/rl78.opt: Add -mg10 option.
4668 * config/rl78/t-rl78: Add -mg10 multilib.
4670 2013-09-16 Xinliang David Li <davidxl@google.com>
4672 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4673 * omp-low.c (omp_max_vf): Ditto.
4674 (expand_omp_simd): Ditto.
4675 * tree-vectorizer.c (vectorize_loops): Ditto.
4676 (gate_vect_slp): Ditto.
4677 (gate_increase_alignment): Ditto.
4678 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4679 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4680 (gate_tree_vectorize): Name change.
4681 (tree_vectorize): Ditto.
4682 (pass_vectorize::gate): Call new function.
4683 (pass_vectorize::execute): Ditto.
4684 * opts.c: O3 default setting change.
4685 (finish_options): Check new flag.
4686 * doc/invoke.texi: Document new flags.
4687 * common.opt: New flags.
4689 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4691 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4692 (Condition Code): Don't reference it.
4693 * doc/tm.texi: Regenerate.
4695 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4698 * lra-constraints.c (undo_optional_reloads): Consider all optional
4699 reload even if it did not get a hard reg.
4701 2013-09-16 Teresa Johnson <tejohnson@google.com>
4703 * dumpfile.c (dump_loc): Remove newline emission.
4704 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4705 emission to dump_printf_loc calls where missing.
4706 (vect_mark_for_runtime_alias_test): Ditto.
4707 (vect_analyze_data_ref_dependence): Ditto.
4708 (vect_analyze_data_ref_dependences): Ditto.
4709 (vect_slp_analyze_data_ref_dependence): Ditto.
4710 (vect_slp_analyze_data_ref_dependences): Ditto.
4711 (vect_compute_data_ref_alignment): Ditto.
4712 (vect_update_misalignment_for_peel): Ditto.
4713 (vect_verify_datarefs_alignment): Ditto.
4714 (vector_alignment_reachable_p): Ditto.
4715 (vect_get_data_access_cost): Ditto.
4716 (vect_enhance_data_refs_alignment): Ditto.
4717 (vect_find_same_alignment_drs): Ditto.
4718 (vect_analyze_data_refs_alignment): Ditto.
4719 (vect_analyze_group_access): Ditto.
4720 (vect_analyze_data_ref_access): Ditto.
4721 (vect_analyze_data_ref_accesses): Ditto.
4722 (vect_prune_runtime_alias_test_list): Ditto.
4723 (vect_analyze_data_refs): Ditto.
4724 (vect_create_addr_base_for_vector_ref): Ditto.
4725 (vect_create_data_ref_ptr): Ditto.
4726 (vect_grouped_store_supported): Ditto.
4727 (vect_grouped_load_supported): Ditto.
4728 * value-prof.c (check_counter): Ditto.
4729 (check_ic_target): Ditto.
4730 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4731 (vect_recog_widen_mult_pattern): Ditto.
4732 (vect_recog_widen_sum_pattern): Ditto.
4733 (vect_recog_over_widening_pattern): Ditto.
4734 (vect_recog_widen_shift_pattern): Ditto.
4735 (vect_recog_rotate_pattern): Ditto.
4736 (vect_recog_vector_vector_shift_pattern): Ditto.
4737 (vect_recog_divmod_pattern): Ditto.
4738 (vect_recog_mixed_size_cond_pattern): Ditto.
4739 (vect_recog_bool_pattern): Ditto.
4740 (vect_pattern_recog_1): Ditto.
4741 (vect_pattern_recog): Ditto.
4742 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4743 (vect_is_simple_iv_evolution): Ditto.
4744 (vect_analyze_scalar_cycles_1): Ditto.
4745 (vect_get_loop_niters): Ditto.
4746 (vect_analyze_loop_1): Ditto.
4747 (vect_analyze_loop_form): Ditto.
4748 (vect_analyze_loop_operations): Ditto.
4749 (vect_analyze_loop_2): Ditto.
4750 (vect_analyze_loop): Ditto.
4751 (report_vect_op): Ditto.
4752 (vect_is_slp_reduction): Ditto.
4753 (vect_is_simple_reduction_1): Ditto.
4754 (vect_get_known_peeling_cost): Ditto.
4755 (vect_estimate_min_profitable_iters): Ditto.
4756 (vect_model_reduction_cost): Ditto.
4757 (vect_model_induction_cost): Ditto.
4758 (get_initial_def_for_induction): Ditto.
4759 (vect_create_epilog_for_reduction): Ditto.
4760 (vectorizable_reduction): Ditto.
4761 (vectorizable_induction): Ditto.
4762 (vectorizable_live_operation): Ditto.
4763 (vect_loop_kill_debug_uses): Ditto.
4764 (vect_transform_loop): Ditto.
4765 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4766 (vect_stmt_relevant_p): Ditto.
4767 (process_use): Ditto.
4768 (vect_mark_stmts_to_be_vectorized): Ditto.
4769 (vect_model_simple_cost): Ditto.
4770 (vect_model_promotion_demotion_cost): Ditto.
4771 (vect_model_store_cost): Ditto.
4772 (vect_get_store_cost): Ditto.
4773 (vect_model_load_cost): Ditto.
4774 (vect_get_load_cost): Ditto.
4775 (vect_init_vector_1): Ditto.
4776 (vect_get_vec_def_for_operand): Ditto.
4777 (vect_finish_stmt_generation): Ditto.
4778 (vectorizable_call): Ditto.
4779 (vectorizable_conversion): Ditto.
4780 (vectorizable_assignment): Ditto.
4781 (vectorizable_shift): Ditto.
4782 (vectorizable_operation): Ditto.
4783 (vectorizable_store): Ditto.
4784 (vectorizable_load): Ditto.
4785 (vectorizable_condition): Ditto.
4786 (vect_analyze_stmt): Ditto.
4787 (vect_transform_stmt): Ditto.
4788 (vect_is_simple_use): Ditto.
4789 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4790 (vect_can_advance_ivs_p): Ditto.
4791 (vect_update_ivs_after_vectorizer): Ditto.
4792 (vect_do_peeling_for_loop_bound): Ditto.
4793 (vect_gen_niters_for_prolog_loop): Ditto.
4794 (vect_update_inits_of_drs): Ditto.
4795 (vect_create_cond_for_alias_checks): Ditto.
4796 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4797 (vect_build_slp_tree_1): Ditto.
4798 (vect_supported_load_permutation_p): Ditto.
4799 (vect_analyze_slp_instance): Ditto.
4800 (vect_analyze_slp): Ditto.
4801 (vect_make_slp_decision): Ditto.
4802 (vect_detect_hybrid_slp): Ditto.
4803 (vect_bb_vectorization_profitable_p): Ditto.
4804 (vect_slp_analyze_bb_1): Ditto.
4805 (vect_update_slp_costs_according_to_vf): Ditto.
4806 (vect_get_mask_element): Ditto.
4807 (vect_transform_slp_perm_load): Ditto.
4808 (vect_schedule_slp_instance): Ditto.
4809 (vect_schedule_slp): Ditto.
4810 (vect_slp_transform_bb): Ditto.
4811 * profile.c (read_profile_edge_counts): Ditto.
4812 (compute_branch_probabilities): Ditto.
4813 * coverage.c (get_coverage_counts): Ditto.
4815 2013-09-16 Diego Novillo <dnovillo@google.com>
4817 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4819 2013-09-16 Jan Hubicka <jh@suse.cz>
4821 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4822 abstract functions; for static functions check the presence of body.
4824 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4826 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4827 * config/aarch64/aarch64-simd.md
4828 (aarch64_mla_elt<mode>): New.
4829 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4830 (aarch64_mls_elt<mode>): Likewise.
4831 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4832 (aarch64_fma4_elt<mode>): Likewise.
4833 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4834 (aarch64_fma4_elt_to_128v2df): Likewise.
4835 (aarch64_fma4_elt_to_64df): Likewise.
4836 (fnma<mode>4): Likewise.
4837 (aarch64_fnma4_elt<mode>): Likewise.
4838 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4839 (aarch64_fnma4_elt_to_128v2df): Likewise.
4840 (aarch64_fnma4_elt_to_64df): Likewise.
4841 * config/aarch64/iterators.md (VDQSF): New.
4842 * config/aarch64/arm_neon.h
4843 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4844 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4846 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4848 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4849 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4850 (aarch64_mul3_elt_to_128df): Likewise.
4851 (aarch64_mul3_elt_to_64v2df): Likewise.
4852 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4854 (VMUL_CHANGE_NLANES) Likewise.
4857 * config/aarch64/arm_neon.h
4858 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4860 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4862 * config/aarch64/arm_neon.h
4863 (vcvtx_high_f32_f64): Fix parameters.
4865 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4866 Uros Bizjak <ubizjak@gmail.com>
4868 * config/alpha.c: Include tree-ssanames.h.
4870 2013-09-16 Richard Biener <rguenther@suse.de>
4872 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4873 (dot_rdg_1): Handle control_dd.
4874 (create_edge_for_control_dependence): New function.
4875 (create_rdg_edges): Add control dependences if asked for.
4876 (build_rdg): Likewise.
4877 (generate_loops_for_partition): If there are not necessary
4878 control stmts remove all their dependencies.
4879 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4880 (distribute_loop): Pass on control dependences.
4881 (tree_loop_distribution): Compute control dependences and remove
4882 restriction on number of loop nodes.
4884 2013-09-16 Jakub Jelinek <jakub@redhat.com>
4886 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4889 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4891 * cse.c (try_const_anchors): Punt on CC modes.
4893 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4895 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4897 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
4900 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4903 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4906 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
4908 (darwin_objc1_section): Likewise.
4909 (darwin_file_end): Emit Image Info section when required.
4911 2013-09-14 Jan Hubicka <jh@suse.cz>
4913 * tree-into-ssa.c (gate_into_ssa): New.
4914 (pass_data_build_ssa): Use it.
4915 * cgraph.h (expand_thunk): Update prototype.
4916 * cgraphunit.c (analyze_function): Expand thunks early.
4917 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4918 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4919 set lowered flag; do not add new function.
4920 (assemble_thunks_and_aliases): Update.
4921 * tree-ssa.c (gate_init_datastructures): New gate.
4922 (pass_data_init_datastructures): Use it.
4924 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4927 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
4928 xmm register set described in the psABI.
4930 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4932 * dwarf2out.c (should_emit_struct_debug): Add check
4933 for type_decl variable is not NULL.
4935 2013-09-13 Jacek Caban <jacek@codeweavers.com>
4937 * config.gcc: Use new winnt-c.c target hooks
4938 * config/t-winnt: New file
4939 * config/winnt-c.c: New file
4940 * doc/tm.texi.in: Document new hook
4941 * doc/tm.texi: Regenerated
4943 2013-09-13 Jan Hubicka <jh@suse.cz>
4946 * ipa-inline.c (check_callers): New function.
4947 (check_caller_edge): Remove.
4948 (want_inline_function_to_all_callers_p): Also permit alises that are
4950 (inline_to_all_callers): Terminate the walk when devirtualization
4951 introduce new calls.
4953 2013-09-13 Jan Hubicka <jh@suse.cz>
4955 * ipa-inline-analysis.c (struct growth_data): Add node.
4956 (do_estimate_growth_1): Fix detection of recursion.
4958 2013-09-13 Jakub Jelinek <jakub@redhat.com>
4960 PR tree-optimization/58392
4961 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4962 to avoid shadowing of outer loop variable. If
4963 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4964 replace_by_duplicate_decl simduid of loops that have it set and
4965 set dest_cfun->has_simduid_loops and/or
4966 dest_cfun->has_force_vect_loops.
4967 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4968 instead of maybe_lookup_decl.
4969 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4970 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4971 call. Copy over force_vect and copy and remap simduid. Set
4972 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4973 (copy_cfg_body): Remove blocks_to_copy argument. Use
4974 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4975 caller. Don't set cfun->has_simduid_loops and/or
4976 cfun->has_force_vect_loops here.
4977 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4979 (expand_call_inline, tree_function_versioning): Adjust copy_body
4982 2013-09-13 Martin Jambor <mjambor@suse.cz>
4985 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4986 the assert if the edge was a speculative one.
4988 2013-09-13 Richard Biener <rguenther@suse.de>
4990 * tree-data-ref.h (known_dependences_p): Move here ...
4991 * tree-loop-distribution.c (known_dependences_p): ... from here.
4992 (dump_rdg_component, debug_rdg_component): Remove.
4994 (generate_loops_for_partition): Use gimple_uid instead of
4995 relying on matching stmt visit order.
4996 (rdg_build_partitions): Take starting stmt vector.
4997 (ldist_gen): Merge into ...
4998 (distribute_loop): ... this function. Do not compute starting
5000 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5002 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5004 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5005 Set type attribute correctly. Set predicable_short_it attribute.
5006 (cmpsi_shiftsi): Remove %? from output template.
5008 2013-09-13 Richard Biener <rguenther@suse.de>
5010 * tree-loop-distribution.c (struct rdg_component,
5011 rdg_defs_used_in_other_loops_p, free_rdg_components,
5012 rdg_build_components): Remove.
5013 (stmts_from_loop): Do not record virtual PHIs.
5014 (generate_loops_for_partition): Skip virtual PHIs.
5015 (build_rdg_partition_for_component): Rename to ...
5016 (build_rdg_partition_for_vertex): ... this and adjust.
5017 (rdg_build_partitions): Take a vector of starting vertices
5018 instead of components. Remove unnecessary leftover handling.
5019 (ldist_gen): Do not build components or record other stores.
5020 (distribute_loop): Do not distribute loops containing stmts
5023 2013-09-13 Christian Bruel <christian.bruel@st.com>
5026 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5028 2013-09-13 Kai Tietz <ktietz@redhat.com>
5030 * config.gcc: Separate cases for mingw and cygwin targets,
5031 and add 64-bit cygwin target case.
5033 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5034 dll-export inline-functions.
5035 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5038 2013-09-13 Jeff Law <law@redhat.com>
5042 2013-09-06 Jeff Law <law@redhat.com>
5044 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5045 edge implied equivalences into successor phis.
5047 2013-09-12 DJ Delorie <dj@redhat.com>
5049 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5052 2013-09-12 Brooks Moses <bmoses@google.com>
5055 * Makefile.in: Do not install driver binaries in $(target)/bin.
5057 2013-09-12 DJ Delorie <dj@redhat.com>
5059 * config/rl78/rl78.opt (mrelax): New.
5060 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5061 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5063 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5065 (rl78_expand_epilogue): Likewise.
5067 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5070 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5071 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5072 pointer elimination with using remove_reg_equal_offset_note.
5074 2013-09-12 DJ Delorie <dj@redhat.com>
5076 * config/msp430/: New port.
5077 * config.gcc (msp430): Added.
5078 * doc/invoke.texi: Document MSP430 options.
5079 * doc/install.texi: Document msp430-elf
5080 * doc/md.texi: Document msp430-elf
5081 * doc/contrib.texi: Document msp430-elf
5083 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5086 2013-09-12 Martin Jambor <mjambor@suse.cz>
5089 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5090 reference descriptor is NULL.
5091 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5094 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5096 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5097 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5098 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5099 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5100 make_temp_ssa_name): move to tree-ssanames.h
5101 * tree-ssa-alias.h: Move prototype.
5102 * tree-ssa.h: Include tree-ssanames.h.
5103 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5104 * tree-ssanames.h: New. Move items from tree-flow*.h
5105 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5107 2013-09-12 Richard Biener <rguenther@suse.de>
5109 PR tree-optimization/58404
5110 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5111 propagate non-invariant addresses into dereferences wrapped
5112 in component references.
5114 2013-09-12 Richard Biener <rguenther@suse.de>
5116 PR tree-optimization/58402
5117 * passes.def: Move pass_late_warn_uninitialized later.
5119 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5121 * tree-ssa.h: New. Move content from tree-flow.h and
5123 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5124 Move prototypes belonging to tree-ssa.c.
5125 * tree-flow-inline.h (redirect_edge_var_map_def,
5126 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5128 * gimple.h: Adjust prototypes.
5129 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5131 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5132 * tree.h: Move prototype to tree-ssa.h.
5133 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5134 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5135 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5136 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5137 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5138 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5139 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5140 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5141 graphite-blocking.c, graphite-clast-to-gimple.c,
5142 graphite-dependences.c, graphite-interchange.c,
5143 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5144 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5145 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5146 ipa-reference.c, ipa-split.c, ipa-utils.c,
5147 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5148 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5149 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5150 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5151 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5152 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5153 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5154 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5155 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5156 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5157 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5158 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5159 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5160 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5161 value-prof.c, var-tracking.c,
5162 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5164 2013-09-12 Richard Biener <rguenther@suse.de>
5166 PR tree-optimization/58396
5167 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5168 (build_rdg): Take a loop-nest parameter, fix memleaks.
5169 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5171 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5173 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5176 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5178 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5179 The FMA instruction names should have a 'v' prefix.
5181 2013-09-12 Richard Biener <rguenther@suse.de>
5183 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5184 (dot_rdg): Use popen instead of system in optional code.
5185 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5186 (already_processed_vertex_p): Adjust.
5187 (has_anti_or_output_dependence, predecessor_has_mem_write,
5188 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5189 rdg_flag_uses): Remove.
5190 (rdg_flag_vertex): Simplify.
5191 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5193 (build_rdg_partition_for_component): Process the first vertex
5194 of a component only.
5195 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5197 2013-09-12 Alan Modra <amodra@gmail.com>
5199 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5201 2013-09-11 DJ Delorie <dj@redhat.com>
5202 Nick Clifton <nickc@redhat.com>
5204 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5205 (rl78_stack_based_mem): New.
5206 * config/rl78/constraints.md (Iv08): New.
5215 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5217 (movsi): Change from expand to insn-and-split.
5218 (ashrsi3): Clobber AX.
5222 * config/rl78/rl78.md (CC_REG): Fix.
5223 (addsi3): Allow memory and immediate operands.
5224 (addsi3_internal): Split into...
5225 (addsi3_internal_virt): ...new, and ...
5226 (addsi3_internal_real): ...new.
5228 (subsi3_internal_virt): New.
5229 (subsi3_internal_real): New.
5230 (mulsi3): Add memory operand.
5231 (mulsi3_rl78): Likewise.
5232 (mulsi3_g13): Likewise.
5233 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5234 (cbranchqi4_real): Add more constraint options.
5235 (cbranchhi4_real): Expand pattern.
5236 (cbranchhi4_real_signed): New.
5237 (cbranchhi4_real_inverted): New.
5238 (cbranchsi4_real_lt): New.
5239 (cbranchsi4_real_ge): New.
5240 (cbranchsi4_real_signed): New.
5241 (cbranchsi4_real): New.
5243 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5245 (lshrsi3_virt): Likewise.
5246 (ashlsi3_virt): Likewise.
5247 (cbranchqi4_virt_signed): New.
5248 (cbranchhi4_virt_signed): New.
5249 (cbranchsi4_virt): New.
5250 * config/rl78/rl78.c: Whitespace fixes throughout.
5251 (move_elim_pass): New.
5252 (pass_data_rl78_move_elim): New.
5253 (pass_rl78_move_elim): New.
5254 (make_pass_rl78_move_elim): New.
5255 (rl78_devirt_info): Run devirt earlier.
5256 (rl78_move_elim_info): New.
5257 (rl78_asm_file_start): Register it.
5258 (rl78_split_movsi): New.
5259 (rl78_as_legitimate_address): Allow virtual base registers when
5261 (rl78_addr_space_convert): Remove spurious debug stuff.
5262 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5263 (rl78_print_operand): More cases for not printing '#'.
5264 (rl78_expand_compare): Remove most of the logic.
5265 (content_memory): New.
5266 (clear_content_memory): New.
5267 (get_content_index): New.
5268 (get_content_name): New.
5269 (display_content_memory): New.
5270 (update_content): New.
5271 (record_content): New.
5272 (already_contains): New.
5273 (insn_ok_now): Re-recog insns with virtual registers.
5274 (add_postponed_content_update): New.
5275 (process_postponed_content_update): New.
5276 (gen_and_emit_move): New.
5277 (transcode_memory_rtx): Record new location content.
5278 Use gen_and_emit_move.
5279 (force_into_acc): New.
5280 (move_to_acc): Use gen_and_emit_move.
5281 (move_from_acc): Likewise.
5282 (move_acc_to_reg): Likewise.
5283 (move_to_x): Likewise.
5284 (move_to_hl): Likewise.
5285 (move_to_de): Likewise.
5286 (rl78_alloc_physical_registers_op1): Record location content.
5287 (has_constraint): New.
5288 (rl78_alloc_physical_registers_op2): Record location content.
5290 (rl78_alloc_physical_registers_ro1): Likewise.
5291 (rl78_alloc_physical_registers_cmp): Likewise.
5292 (rl78_alloc_physical_registers_umul): Likewise.
5293 (rl78_alloc_address_registers_macax): New.
5294 (rl78_alloc_physical_registers): Initialize and set location
5295 content memory as needed.
5296 (rl78_reorg): Make sure split2 is called.
5297 (rl78_rtx_costs): New.
5299 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5301 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5302 for (not (neg ...)) and (neg (not ...)) cases.
5304 2013-09-11 Richard Biener <rguenther@suse.de>
5307 * passes.def: Split critical edges before late uninit warning passes.
5308 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5310 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5312 PR tree-optimization/58385
5313 * fold-const.c (build_range_check): If both low and high are NULL,
5314 use omit_one_operand_loc to preserve exp side-effects.
5316 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5318 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5320 2013-09-11 Richard Biener <rguenther@suse.de>
5322 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5323 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5324 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5325 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5326 stmts_from_loop, known_dependences_p, build_empty_rdg,
5327 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5328 * tree-loop-distribution.c: ... here.
5329 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5330 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5331 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5332 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5333 * tree-loop-distribution.c: ... here.
5334 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5335 (struct partition_s): Add loops member.
5336 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5337 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5338 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5340 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5341 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5342 Sergey Lega <sergey.s.lega@intel.com>
5343 Anna Tikhonova <anna.tikhonova@intel.com>
5344 Ilya Tocar <ilya.tocar@intel.com>
5345 Andrey Turetskiy <andrey.turetskiy@intel.com>
5346 Ilya Verbin <ilya.verbin@intel.com>
5347 Kirill Yukhin <kirill.yukhin@intel.com>
5348 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5350 * config/i386/constraints.md (k): New.
5352 * config/i386/i386.c (const regclass_map): Add new mask registers.
5353 (dbx_register_map): Ditto.
5354 (dbx64_register_map): Ditto.
5355 (svr4_dbx_register_map): Ditto.
5356 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5358 (ix86_preferred_reload_class): Disable constants for mask registers.
5359 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5360 (ix86_hard_regno_mode_ok): Support new mask registers.
5361 (x86_order_regs_for_local_alloc): Ditto.
5362 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5363 (FIXED_REGISTERS): Add new mask registers.
5364 (CALL_USED_REGISTERS): Ditto.
5365 (REG_ALLOC_ORDER): Ditto.
5366 (VALID_MASK_REG_MODE): New.
5367 (FIRST_MASK_REG): Ditto.
5368 (LAST_MASK_REG): Ditto.
5369 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5370 (MAYBE_MASK_CLASS_P): New.
5371 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5372 (REG_CLASS_CONTENTS): Ditto.
5373 (MASK_REGNO_P): New.
5374 (ANY_MASK_REG_P): Ditto.
5375 (HI_REGISTER_NAMES): Add new mask registers.
5376 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5377 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5379 (attribute "type"): Add mskmov, msklog.
5380 (attribute "length_immediate"): Support them.
5381 (attribute "memory"): Ditto.
5382 (attribute "prefix_0f"): Ditto.
5383 (*movhi_internal): Support new mask registers.
5384 (*movqi_internal): Ditto.
5385 (define_split): Split out clobber pattern is a logic
5386 insn on mask registers.
5387 (*k<logic><mode>): New.
5388 (*andhi_1): Extend to support mask regs.
5389 (*andqi_1): Extend to support mask regs.
5391 (define_split): Split and-not to and and not if operands
5393 (*<code><mode>_1): Separate HI mode to new pattern...
5394 (*<code>hi_1): This.
5395 (*<code>qi_1): Extend to support mask regs.
5397 (kortestzhi): Ditto.
5398 (kortestchi): Ditto.
5400 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5401 (*one_cmplhi2_1): ...Here, now with mask registers support.
5402 (*one_cmplqi2_1): Support new mask registers.
5403 (HI/QImode arithmetics splitter): Don't split if mask registers
5405 (HI/QImode not splitter): Ditto.
5406 * config/i386/predicated.md (mask_reg_operand): New.
5407 (general_reg_operand): Ditto.
5409 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5411 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5412 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5414 2013-09-10 Jeff Law <law@redhat.com>
5416 PR tree-optimization/58380
5417 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5418 of threading through a buried loop header.
5420 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5421 return value for single successor case.
5423 2013-09-10 Jan Hubicka <jh@suse.cz>
5425 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5427 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5430 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5431 support conditional execution.
5432 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5434 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5436 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5438 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5439 Check destination too to check move insn.
5440 (undo_optional_reloads): Add check that the original peudo did not
5441 changed its allocation and the optional reload was inherited on last
5442 inheritance pass. Break loop after deciding to keep optional reload.
5443 (lra_undo_inheritance): Add check that inherited pseudo still in
5446 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5448 * config/aarch64/aarch64.md (generic_sched): New.
5449 * config/aarch64/aarch64-generic.md (load): Make conditional
5450 on generic_sched attribute.
5451 (nonload): Likewise.
5453 2013-09-10 Jan Hubicka <jh@suse.cz>
5455 * lto-cgraph.c: Include ipa-utils.h.
5456 (compute_ltrans_boundary): Also add possible targets into the boundary.
5458 2013-09-10 Jan Hubicka <jh@suse.cz>
5460 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5461 VAR_DECL of vtable rather than full expression.
5463 2013-09-10 Jan Hubicka <jh@suse.cz>
5464 Paolo Carlini <paolo.carlini@oracle.com>
5466 * cgraphunit.c (analyze_functions): Save input_location, set it
5467 to UNKNOWN_LOCATION and restore it at the end.
5469 2013-09-10 Martin Jambor <mjambor@suse.cz>
5471 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5472 represented by a thunk.
5474 2013-09-10 Jeff Law <law@redhat.com>
5476 PR tree-optimization/58343
5477 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5478 jump threading requests through loop headers buried in the middle
5479 of a jump threading path.
5481 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5482 in return value/type.
5484 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5486 PR rtl-optimization/58365
5487 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5488 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5491 2013-09-10 Richard Biener <rguenther@suse.de>
5493 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5494 * tree-data-ref.c (create_rdg_vertices): Collect all data
5495 references, signal failure to the caller, use data-ref API.
5496 (build_rdg): Compute data references only once. Maintain lifetime
5497 of data references and data dependences from within RDG.
5498 (free_rdg): Free dependence relations.
5499 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5500 inventing extra dependences.
5501 (distribute_loop): Update for RDG API changes.
5503 2013-09-10 Kai Tietz <ktietz@redhat.com>
5505 * doc/invoke.texi (fms-extensions): Document changed
5506 behavior for ms-abi targets.
5507 * config/i386/i386.c (ix86_option_override_internal):
5508 Set default value of option -fms-extension for ms-abi targets.
5510 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5512 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5514 2013-09-10 Alan Modra <amodra@gmail.com>
5517 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5519 2013-09-10 Alan Modra <amodra@gmail.com>
5521 * config/rs6000/predicates.md (add_cint_operand): New.
5522 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5523 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5524 using add_cint_operand.
5525 (largetoc_high_plus_aix): Likewise.
5527 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5529 PR tree-optimization/58364
5530 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5531 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5532 the current range can't be an unconditional true or false.
5534 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5536 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5538 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5540 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5542 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5545 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5547 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5549 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5550 NO_REGS for immediate that can't be moved directly into FP_REGS.
5552 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5554 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5555 comparison with negated operand.
5556 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5559 2013-09-09 Richard Biener <rguenther@suse.de>
5562 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5563 of a subloop record all its block as affecting loop-closed SSA form.
5565 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5567 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5568 of an rtx/bitpos pair.
5569 (store_fixed_bit_field): Update accordingly.
5571 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5573 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5575 * builtins.c (expand_errno_check): Likewise.
5576 * dwarf2cfi.c (init_return_column_size): Likewise.
5577 * except.c (sjlj_mark_call_sites): Likewise.
5578 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5579 * lra-constraints.c (emit_inc): Likewise.
5580 * ree.c (combine_set_extension): Likewise.
5581 * regmove.c (fixup_match_2): Likewise.
5582 * reload1.c (inc_for_reload): Likewise.
5584 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5586 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5587 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5588 Use gen_int_mode with the mode of the associated simplify_* call.
5589 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5590 * expmed.c (expand_shift_1): Likewise.
5591 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5592 * loop-iv.c (iv_number_of_iterations): Likewise.
5593 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5594 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5596 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5598 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5599 of the associated expand_* call.
5600 (asan_emit_stack_protection): Likewise.
5601 * builtins.c (round_trampoline_addr): Likewise.
5602 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5603 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5604 (emit_store_flag): Likewise.
5605 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5606 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5608 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5609 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5610 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5611 * modulo-sched.c (generate_prolog_epilog): Likewise.
5612 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5613 (expand_ctz, expand_ffs, expand_unop): Likewise.
5615 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5617 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5618 of the associated gen_rtx_* call.
5619 * caller-save.c (init_caller_save): Likewise.
5620 * combine.c (find_split_point, make_extraction): Likewise.
5621 (make_compound_operation): Likewise.
5622 * dwarf2out.c (mem_loc_descriptor): Likewise.
5623 * explow.c (plus_constant, probe_stack_range): Likewise.
5624 * expmed.c (expand_mult_const): Likewise.
5625 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5626 * reload1.c (init_reload): Likewise.
5627 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5628 * var-tracking.c (adjust_mems): Likewise.
5629 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5630 rather than gen_rtx_fmt_ee.
5632 2013-09-09 Jan Hubicka <jh@suse.cz>
5635 * value-prof.c (gimple_ic): Copy also abnormal edges.
5637 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5639 * asan.c (asan_shadow_cst): Use gen_int_mode.
5641 2013-09-08 Jan Hubicka <jh@suse.cz>
5643 * ipa-profile.c: Add toplevel comment.
5644 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5645 (contains_hot_call_p): New function.
5646 (ipa_propagate_frequency): Set frequencies based on counts when
5648 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5649 profile; do not tamper with profile after inlining if it is read.
5651 2013-09-08 Jan Hubicka <jh@suse.cz>
5653 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5656 2013-09-08 Jan Hubicka <jh@suse.cz>
5658 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5661 2013-09-08 Jeff Law <law@redhat.com>
5664 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5667 2013-09-08 Andi Kleen <ak@linux.intel.com>
5669 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5671 2013-09-08 Jan Hubicka <jh@suse.cz>
5673 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5675 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5677 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5678 for non-debug insns.
5679 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5680 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5681 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5684 2013-09-08 Jan Hubicka <jh@suse.cz>
5686 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5687 targets and devirtualize to BUILT_IN_UNREACHABLE.
5688 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5689 * ipa.c (walk_polymorphic_call_targets): New function.
5690 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5691 functions; use the new timevar.
5692 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5693 was removed from the program.
5694 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5695 not consider it in the walk when its vtable is dead.
5696 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5698 (devirt_variable_node_removal_hook): New function.
5699 (possible_polymorphic_call_targets): Also register
5700 devirt_variable_node_removal_hook.
5701 (ipa_devirt): Do not do non-speculative devirtualization.
5702 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5704 2013-09-08 Jan Hubicka <jh@suse.cz>
5706 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5707 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5708 varpool_remove_variable_insertion_hook): Declare.
5709 * varpool.c (varpool_node_hook_list): New structure.
5710 (first_varpool_node_removal_hook,
5711 first_varpool_variable_insertion_hook): New variables.
5712 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5713 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5714 varpool_remove_variable_insertion_hook,
5715 varpool_call_variable_insertion_hooks): New functions.
5716 (varpool_remove_node): Use it.
5718 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5721 * diagnostic.c (diagnostic_build_prefix): When s.file is
5722 "<built-in>" don't output line and column numbers.
5724 2013-09-06 Jan Hubicka <jh@suse.cz>
5726 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5727 * lto-streamer-out.c: Stream thunks, too.
5728 * lto-streamer-in.c (input_function): Pop cfun here
5729 (lto_read_body): Instead of here.
5731 2013-09-06 Caroline Tice <cmtice@google.com>
5733 * doc/install.texi: Add documentation for the --enable-vtable-verify
5734 and the --disable-libvtv configure options.
5736 2013-09-06 Jeff Law <law@redhat.com>
5738 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5739 edge implied equivalences into successor phis.
5741 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5743 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5745 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5747 * resource.c (mark_target_live_regs): Compute resources taking
5748 into account if a call is predicated or not.
5750 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5752 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5753 the compiler in the function names.
5755 2013-09-06 Jan Hubicka <jh@suse.cz>
5758 * ipa-inline.c (has_caller_p): New function.
5759 (want_inline_function_to_all_callers_p): Use it.
5760 (sum_callers, inline_to_all_callers): Break out from ...
5761 (ipa_inline): ... here.
5763 2013-09-06 Jan Hubicka <jh@suse.cz>
5765 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5766 only when AVX is enabled.
5768 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5770 * config/aarch64/aarch64.md
5771 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5772 is fpsimd_<load/store>2.
5773 (load_pair<mode>): Likewise.
5774 (store_pair<mode>): Likewise.
5776 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5778 * config/arm/types.md (type): Add "mrs" type.
5779 * config/aarch64/aarch64.md
5780 (aarch64_load_tp_hard): Make type "mrs".
5782 (load_tp_hard): Make type "mrs".
5783 * config/arm/cortex-a15.md: Update with new attributes.
5784 * config/arm/cortex-a5.md: Update with new attributes.
5785 * config/arm/cortex-a53.md: Update with new attributes.
5786 * config/arm/cortex-a7.md: Update with new attributes.
5787 * config/arm/cortex-a8.md: Update with new attributes.
5788 * config/arm/cortex-a9.md: Update with new attributes.
5789 * config/arm/cortex-m4.md: Update with new attributes.
5790 * config/arm/cortex-r4.md: Update with new attributes.
5791 * config/arm/fa526.md: Update with new attributes.
5792 * config/arm/fa606te.md: Update with new attributes.
5793 * config/arm/fa626te.md: Update with new attributes.
5794 * config/arm/fa726te.md: Update with new attributes.
5796 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5798 * config/aarch64/aarch64.md
5799 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5800 (*movtf_aarch64): Likewise.
5802 (thumb1_movdi_insn): Use "multiple" for type where more than one
5803 instruction is used for a move.
5804 (*arm32_movhf): Likewise.
5805 (*thumb_movdf_insn): Likewise.
5807 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5809 * config/arm/types.md (type): Rename fcpys to fmov.
5811 (*arm_movsi_vfp): Rename type fcpys as fmov.
5812 (*thumb2_movsi_vfp): Likewise
5813 (*movhf_vfp_neon): Likewise
5814 (*movhf_vfp): Likewise
5815 (*movsf_vfp): Likewise
5816 (*thumb2_movsf_vfp): Likewise
5817 (*movsfcc_vfp): Likewise
5818 (*thumb2_movsfcc_vfp): Likewise
5819 * config/aarch64/aarch64-simd.md
5820 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5821 * config/aarch64/aarch64.md
5822 (*movsi_aarch64): Replace type mov_reg with fmovs.
5823 (*movdi_aarch64): Likewise
5824 (*movsf_aarch64): Likewise
5825 (*movdf_aarch64): Likewise
5827 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5828 * config/arm/iwmmxt.md
5829 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5830 * config/arm/arm1020e.md: Update with new attributes.
5831 * config/arm/cortex-a15-neon.md: Update with new attributes.
5832 * config/arm/cortex-a5.md: Update with new attributes.
5833 * config/arm/cortex-a53.md: Update with new attributes.
5834 * config/arm/cortex-a7.md: Update with new attributes.
5835 * config/arm/cortex-a8-neon.md: Update with new attributes.
5836 * config/arm/cortex-a9.md: Update with new attributes.
5837 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5838 * config/arm/cortex-r4f.md: Update with new attributes.
5839 * config/arm/marvell-pj4.md: Update with new attributes.
5840 * config/arm/vfp11.md: Update with new attributes.
5842 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5844 * config/aarch64/aarch64.md
5845 (*madd<mode>): Fix type attribute.
5846 (*maddsi_uxtw): Likewise.
5847 (*msub<mode>): Likewise.
5848 (*msubsi_uxtw): Likewise.
5849 (<su_optab>maddsidi4): Likewise.
5850 (<su_optab>msubsidi4): Likewise.
5852 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5854 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5855 * config/arm/arm.md (core_cycles): Remove fdiv.
5856 * config/arm/vfp.md:
5857 (*sqrtsf2_vfp): Update for attribute changes.
5858 (*sqrtdf2_vfp): Likewise.
5859 * config/aarch64/aarch64.md:
5860 (sqrt<mode>2): Update for attribute changes.
5861 * config/arm/arm1020e.md: Update with new attributes.
5862 * config/arm/cortex-a15-neon.md: Update with new attributes.
5863 * config/arm/cortex-a5.md: Update with new attributes.
5864 * config/arm/cortex-a53.md: Update with new attributes.
5865 * config/arm/cortex-a7.md: Update with new attributes.
5866 * config/arm/cortex-a8-neon.md: Update with new attributes.
5867 * config/arm/cortex-a9.md: Update with new attributes.
5868 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5869 * config/arm/cortex-r4f.md: Update with new attributes.
5870 * config/arm/marvell-pj4.md: Update with new attributes.
5871 * config/arm/vfp11.md: Update with new attributes.
5873 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5875 * config/arm/types.md
5876 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5877 * config/aarch64/aarch64.md
5878 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5880 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5881 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5882 (float<GPI:mode><GPF:mode>2): Likewise.
5884 (*truncsisf2_vfp): Update with new attributes.
5885 (*truncsidf2_vfp): Likewise.
5886 (fixuns_truncsfsi2): Likewise.
5887 (fixuns_truncdfsi2): Likewise.
5888 (*floatsisf2_vfp): Likewise.
5889 (*floatsidf2_vfp): Likewise.
5890 (floatunssisf2): Likewise.
5891 (floatunssidf2): Likewise.
5892 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5893 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5894 * config/arm/arm1020e.md: Update with new attributes.
5895 * config/arm/cortex-a15-neon.md: Update with new attributes.
5896 * config/arm/cortex-a5.md: Update with new attributes.
5897 * config/arm/cortex-a53.md: Update with new attributes.
5898 * config/arm/cortex-a7.md: Update with new attributes.
5899 * config/arm/cortex-a8-neon.md: Update with new attributes.
5900 * config/arm/cortex-a9.md: Update with new attributes.
5901 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5902 * config/arm/cortex-r4f.md: Update with new attributes.
5903 * config/arm/marvell-pj4.md: Update with new attributes.
5904 * config/arm/vfp11.md: Update with new attributes.
5906 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5908 * config/aarch64/arm_neon.h
5909 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5910 (vqtbx<1,2,3,4><q>_s8): Likewise.
5912 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5914 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5915 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
5916 (add<mode>3): Add type attribute.
5917 (add<mode>3): Likewise.
5918 (usadd<mode>3): Likewise.
5919 (ssadd<mode>3): Likewise.
5920 (sub<mode>3): Likewise.
5921 (sub<mode>3): Likewise.
5922 (ussub<mode>3): Likewise.
5923 (sssub<mode>3): Likewise.
5924 (ssmulsa3): Likewise.
5925 (usmulusa3): Likewise.
5926 (arm_usatsihi): Likewise.
5928 (*movdi_vfp): Add types for all instructions.
5929 (*movdi_vfp_cortexa8): Likewise.
5930 (*movhf_vfp_neon): Likewise.
5931 (*movhf_vfp): Likewise.
5932 (*movdf_vfp): Likewise.
5933 (*thumb2_movdf_vfp): Likewise.
5934 (*thumb2_movdfcc_vfp): Likewise.
5935 * config/arm/arm.md: Add type attribute to all insn patterns.
5936 (*thumb1_adddi3): Add type attribute.
5937 (*arm_adddi3): Likewise.
5938 (*adddi_sesidi_di): Likewise.
5939 (*adddi_zesidi_di): Likewise.
5940 (*thumb1_addsi3): Likewise.
5941 (addsi3_compare0): Likewise.
5942 (*addsi3_compare0_scratch): Likewise.
5943 (*compare_negsi_si): Likewise.
5944 (cmpsi2_addneg): Likewise.
5945 (*addsi3_carryin_<optab>): Likewise.
5946 (*addsi3_carryin_alt2_<optab>): Likewise.
5947 (*addsi3_carryin_clobercc_<optab>): Likewise.
5948 (*subsi3_carryin): Likewise.
5949 (*subsi3_carryin_const): Likewise.
5950 (*subsi3_carryin_compare): Likewise.
5951 (*subsi3_carryin_compare_const): Likewise.
5952 (*arm_subdi3): Likewise.
5953 (*thumb_subdi3): Likewise.
5954 (*subdi_di_zesidi): Likewise.
5955 (*subdi_di_sesidi): Likewise.
5956 (*subdi_zesidi_di): Likewise.
5957 (*subdi_sesidi_di): Likewise.
5958 (*subdi_zesidi_ze): Likewise.
5959 (thumb1_subsi3_insn): Likewise.
5960 (*arm_subsi3_insn): Likewise.
5961 (*anddi3_insn): Likewise.
5962 (*anddi_zesidi_di): Likewise.
5963 (*anddi_sesdi_di): Likewise.
5964 (*ne_zeroextracts): Likewise.
5965 (*ne_zeroextracts): Likewise.
5966 (*ite_ne_zeroextr): Likewise.
5967 (*ite_ne_zeroextr): Likewise.
5968 (*anddi_notdi_di): Likewise.
5969 (*anddi_notzesidi): Likewise.
5970 (*anddi_notsesidi): Likewise.
5971 (andsi_notsi_si): Likewise.
5972 (thumb1_bicsi3): Likewise.
5973 (*iordi3_insn): Likewise.
5974 (*iordi_zesidi_di): Likewise.
5975 (*iordi_sesidi_di): Likewise.
5976 (*thumb1_iorsi3_insn): Likewise.
5977 (*xordi3_insn): Likewise.
5978 (*xordi_zesidi_di): Likewise.
5979 (*xordi_sesidi_di): Likewise.
5980 (*arm_xorsi3): Likewise.
5981 (*andsi_iorsi3_no): Likewise.
5982 (*smax_0): Likewise.
5983 (*smax_m1): Likewise.
5984 (*arm_smax_insn): Likewise.
5985 (*smin_0): Likewise.
5986 (*arm_smin_insn): Likewise.
5987 (*arm_umaxsi3): Likewise.
5988 (*arm_uminsi3): Likewise.
5989 (*minmax_arithsi): Likewise.
5990 (*minmax_arithsi_): Likewise.
5991 (*satsi_<SAT:code>): Likewise.
5992 (arm_ashldi3_1bit): Likewise.
5993 (arm_ashrdi3_1bit): Likewise.
5994 (arm_lshrdi3_1bit): Likewise.
5995 (*arm_negdi2): Likewise.
5996 (*thumb1_negdi2): Likewise.
5997 (*arm_negsi2): Likewise.
5998 (*thumb1_negsi2): Likewise.
5999 (*negdi_extendsid): Likewise.
6000 (*negdi_zero_extend): Likewise.
6001 (*arm_abssi2): Likewise.
6002 (*thumb1_abssi2): Likewise.
6003 (*arm_neg_abssi2): Likewise.
6004 (*thumb1_neg_abss): Likewise.
6005 (one_cmpldi2): Likewise.
6006 (extend<mode>di2): Likewise.
6007 (*compareqi_eq0): Likewise.
6008 (*arm_extendhisi2addsi): Likewise.
6009 (*arm_movdi): Likewise.
6010 (*thumb1_movdi_insn): Likewise.
6011 (*arm_movt): Likewise.
6012 (*thumb1_movsi_insn): Likewise.
6013 (pic_add_dot_plus_four): Likewise.
6014 (pic_add_dot_plus_eight): Likewise.
6015 (tls_load_dot_plus_eight): Likewise.
6016 (*thumb1_movhi_insn): Likewise.
6017 (*thumb1_movsf_insn): Likewise.
6018 (*movdf_soft_insn): Likewise.
6019 (*thumb_movdf_insn): Likewise.
6020 (cbranchsi4_insn): Likewise.
6021 (cbranchsi4_scratch): Likewise.
6022 (*negated_cbranchsi4): Likewise.
6023 (*tbit_cbranch): Likewise.
6024 (*tlobits_cbranch): Likewise.
6025 (*tstsi3_cbranch): Likewise.
6026 (*cbranchne_decr1): Likewise.
6027 (*addsi3_cbranch): Likewise.
6028 (*addsi3_cbranch_scratch): Likewise.
6029 (*arm_cmpdi_insn): Likewise.
6030 (*arm_cmpdi_unsig): Likewise.
6031 (*arm_cmpdi_zero): Likewise.
6032 (*thumb_cmpdi_zero): Likewise.
6033 (*deleted_compare): Likewise.
6034 (*mov_scc): Likewise.
6035 (*mov_negscc): Likewise.
6036 (*mov_notscc): Likewise.
6037 (*cstoresi_eq0_thumb1_insn): Likewise.
6038 (cstoresi_nltu_thumb1): Likewise.
6039 (cstoresi_ltu_thu): Likewise.
6040 (thumb1_addsi3_addgeu): Likewise.
6041 (*arm_jump): Likewise.
6042 (*thumb_jump): Likewise.
6043 (*check_arch2): Likewise.
6044 (arm_casesi_internal): Likewise.
6045 (thumb1_casesi_dispatch): Likewise.
6046 (*arm_indirect_jump): Likewise.
6047 (*thumb1_indirect_jump): Likewise.
6049 (*and_scc): Likewise.
6050 (*ior_scc): Likewise.
6051 (*compare_scc): Likewise.
6052 (*cond_move): Likewise.
6053 (*cond_arith): Likewise.
6054 (*cond_sub): Likewise.
6055 (*cmp_ite0): Likewise.
6056 (*cmp_ite1): Likewise.
6057 (*cmp_and): Likewise.
6058 (*cmp_ior): Likewise.
6059 (*ior_scc_scc): Likewise.
6060 (*ior_scc_scc_cmp): Likewise.
6061 (*and_scc_scc): Likewise.
6062 (*and_scc_scc_cmp): Likewise.
6063 (*and_scc_scc_nod): Likewise.
6064 (*negscc): Likewise.
6065 (movcond_addsi): Likewise.
6066 (movcond): Likewise.
6067 (*ifcompare_plus_move): Likewise.
6068 (*if_plus_move): Likewise.
6069 (*ifcompare_move_plus): Likewise.
6070 (*if_move_plus): Likewise.
6071 (*ifcompare_arith_arith): Likewise.
6072 (*if_arith_arith): Likewise.
6073 (*ifcompare_arith_move): Likewise.
6074 (*if_arith_move): Likewise.
6075 (*ifcompare_move_arith): Likewise.
6076 (*if_move_arith): Likewise.
6077 (*ifcompare_move_not): Likewise.
6078 (*if_move_not): Likewise.
6079 (*ifcompare_not_move): Likewise.
6080 (*if_not_move): Likewise.
6081 (*ifcompare_shift_move): Likewise.
6082 (*if_shift_move): Likewise.
6083 (*ifcompare_move_shift): Likewise.
6084 (*if_move_shift): Likewise.
6085 (*ifcompare_shift_shift): Likewise.
6086 (*ifcompare_not_arith): Likewise.
6087 (*ifcompare_arith_not): Likewise.
6088 (*if_arith_not): Likewise.
6089 (*ifcompare_neg_move): Likewise.
6090 (*if_neg_move): Likewise.
6091 (*ifcompare_move_neg): Likewise.
6092 (*if_move_neg): Likewise.
6093 (prologue_thumb1_interwork): Likewise.
6094 (*cond_move_not): Likewise.
6095 (*sign_extract_onebit): Likewise.
6096 (*not_signextract_onebit): Likewise.
6097 (stack_tie): Likewise.
6098 (align_4): Likewise.
6099 (align_8): Likewise.
6100 (consttable_end): Likewise.
6101 (consttable_1): Likewise.
6102 (consttable_2): Likewise.
6103 (consttable_4): Likewise.
6104 (consttable_8): Likewise.
6105 (consttable_16): Likewise.
6106 (*thumb1_tablejump): Likewise.
6107 (prefetch): Likewise.
6108 (force_register_use): Likewise.
6109 (thumb_eh_return): Likewise.
6110 (load_tp_hard): Likewise.
6111 (load_tp_soft): Likewise.
6112 (tlscall): Likewise.
6113 (*arm_movtas_ze): Likewise.
6114 (*arm_rev): Likewise.
6115 (*arm_revsh): Likewise.
6116 (*arm_rev16): Likewise.
6117 * config/arm/thumb2.md
6118 (*thumb2_smaxsi3): Likewise.
6119 (*thumb2_sminsi3): Likewise.
6120 (*thumb32_umaxsi3): Likewise.
6121 (*thumb2_uminsi3): Likewise.
6122 (*thumb2_negdi2): Likewise.
6123 (*thumb2_abssi2): Likewise.
6124 (*thumb2_neg_abss): Likewise.
6125 (*thumb2_movsi_insn): Likewise.
6126 (tls_load_dot_plus_four): Likewise.
6127 (*thumb2_movhi_insn): Likewise.
6128 (*thumb2_mov_scc): Likewise.
6129 (*thumb2_mov_negs): Likewise.
6130 (*thumb2_mov_negs): Likewise.
6131 (*thumb2_mov_nots): Likewise.
6132 (*thumb2_mov_nots): Likewise.
6133 (*thumb2_movsicc_): Likewise.
6134 (*thumb2_movsfcc_soft_insn): Likewise.
6135 (*thumb2_indirect_jump): Likewise.
6136 (*thumb2_and_scc): Likewise.
6137 (*thumb2_ior_scc): Likewise.
6138 (*thumb2_ior_scc_strict_it): Likewise.
6139 (*thumb2_cond_move): Likewise.
6140 (*thumb2_cond_arith): Likewise.
6141 (*thumb2_cond_ari): Likewise.
6142 (*thumb2_cond_sub): Likewise.
6143 (*thumb2_negscc): Likewise.
6144 (*thumb2_movcond): Likewise.
6145 (thumb2_casesi_internal): Likewise.
6146 (thumb2_casesi_internal_pic): Likewise.
6147 (*thumb2_alusi3_short): Likewise.
6148 (*thumb2_mov<mode>_shortim): Likewise.
6149 (*thumb2_addsi_short): Likewise.
6150 (*thumb2_subsi_short): Likewise.
6151 (thumb2_addsi3_compare0): Likewise.
6152 (*thumb2_cbz): Likewise.
6153 (*thumb2_cbnz): Likewise.
6154 (*thumb2_one_cmplsi2_short): Likewise.
6155 (*thumb2_negsi2_short): Likewise.
6156 (*orsi_notsi_si): Likewise.
6157 * config/arm/arm1020e.md: Update with new attributes.
6158 * config/arm/arm1026ejs.md: Update with new attributes.
6159 * config/arm/arm1136jfs.md: Update with new attributes.
6160 * config/arm/arm926ejs.md: Update with new attributes.
6161 * config/arm/cortex-a15.md: Update with new attributes.
6162 * config/arm/cortex-a5.md: Update with new attributes.
6163 * config/arm/cortex-a53.md: Update with new attributes.
6164 * config/arm/cortex-a7.md: Update with new attributes.
6165 * config/arm/cortex-a8.md: Update with new attributes.
6166 * config/arm/cortex-a9.md: Update with new attributes.
6167 * config/arm/cortex-m4.md: Update with new attributes.
6168 * config/arm/cortex-r4.md: Update with new attributes.
6169 * config/arm/fa526.md: Update with new attributes.
6170 * config/arm/fa606te.md: Update with new attributes.
6171 * config/arm/fa626te.md: Update with new attributes.
6172 * config/arm/fa726te.md: Update with new attributes.
6174 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6176 * config/aarch64/aarch64-simd.md
6177 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6178 <vwx> iterator to ensure correct register choice.
6179 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6180 (aarch64_sqdmull_n<mode>): Likewise.
6181 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6182 * config/aarch64/arm_neon.h
6183 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6184 (vml<as><q>_n_<su>16): Likewise.
6185 (vml<as>l_high_lane<q>_<su>16): Likewise.
6186 (vml<as>l_high_n_<su>16): Likewise.
6187 (vml<as>l_lane<q>_<su>16): Likewise.
6188 (vml<as>l_n_<su>16): Likewise.
6189 (vmul<q>_lane<q>_<su>16): Likewise.
6190 (vmul<q>_n_<su>16): Likewise.
6191 (vmull_lane<q>_<su>16): Likewise.
6192 (vmull_n_<su>16): Likewise.
6193 (vmull_high_lane<q>_<su>16): Likewise.
6194 (vmull_high_n_<su>16): Likewise.
6195 (vqrdmulh<q>_n_s16): Likewise.
6197 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6199 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6200 have the correct lane parameter.
6202 2013-09-06 Richard Biener <rguenther@suse.de>
6204 * cfganal.c (control_dependences::~control_dependences):
6205 Properly free all of the vector.
6207 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6210 * config/i386/i386.c (ix86_conditional_register_usage):
6211 Proper initialize extended SSE registers.
6213 2013-09-06 Jan Hubicka <jh@suse.cz>
6215 PR tree-optimization/58311
6216 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6218 2013-09-06 Jan Hubicka <jh@suse.cz>
6220 * Makefile.in (tree-sra.o): Update dependencies.
6221 * tree-sra.c: Include ipa-utils.h
6222 (scan_function): Use recursive_call_p.
6223 (has_caller_p): New function.
6224 (cgraph_for_node_and_aliases): Count also callers of aliases.
6226 2013-09-06 Jan Hubicka <jh@suse.cz>
6229 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6230 * tree-tailcall.c: Include ipa-utils.h.
6231 (find_tail_calls): Use it.
6232 * ipa-pure-const.c (check_call): Likewise.
6233 * ipa-utils.c (recursive_call_p): New function.
6234 * ipa-utils.h (recursive_call_p): Dclare.
6235 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6236 (symtab_semantically_equivalent_p): New function.
6237 * Makefile.in (tree-tailcall.o): Update dependencies.
6239 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6241 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6244 2013-09-06 Richard Biener <rguenther@suse.de>
6246 * lto-streamer.h (lto_global_var_decls): Remove.
6247 * Makefile.in (OBJS): Remove lto-symtab.o.
6248 (lto-symtab.o): Remove.
6249 (GTFILES): Remove lto-symtab.c
6250 * lto-symtab.c: Move to lto/
6252 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6254 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6255 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6256 (UNSPEC_FPINT_RINT): New constant definitions.
6257 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6258 definition with 2 attributes.
6259 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6260 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6263 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6265 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6266 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6267 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6268 attribute to "z196_alone".
6269 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6272 2013-09-06 Richard Biener <rguenther@suse.de>
6274 * basic-block.h (class control_dependences): New.
6275 * tree-ssa-dce.c (control_dependence_map): Remove.
6277 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6278 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6279 find_pdom, find_control_dependence, find_all_control_dependences):
6281 (mark_control_dependent_edges_necessary,
6282 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6283 tree_dce_done, perform_tree_ssa_dce): Adjust.
6284 * cfganal.c (set_control_dependence_map_bit,
6285 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6286 find_all_control_dependences): Move from tree-ssa-dce.c and
6287 implement as methods of control_dependences class.
6288 (control_dependences::control_dependences): New.
6289 (control_dependences::~control_dependences): Likewise.
6290 (control_dependences::get_edges_dependent_on): Likewise.
6291 (control_dependences::get_edge): Likewise.
6293 2013-09-04 Jan Hubicka <jh@suse.cz>
6295 * tree.c (types_same_for_odr): Drop overactive check.
6296 * ipa-devirt.c (hash_type_name): Likewise.
6298 2013-09-04 Jan Hubicka <jh@suse.cz>
6300 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6301 (analyze_functions): ... here.
6303 2013-09-04 Jan Hubicka <jh@suse.cz>
6306 * cgraphunit.c (analyze_functions): Clear AUX fields
6307 after processing; initialize assembler name has.
6309 2013-09-05 Jeff Law <law@redhat.com>
6311 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6312 from thread_around_empty_block. Record threading path into PATH.
6313 Recurse if threading through the initial block is successful.
6314 (thread_across_edge): Corresponding changes to slightly simplify.
6316 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6318 * config/aarch64/aarch64.md
6319 (type): Remove frecpe, frecps, frecpx.
6320 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6321 fix to be a TARGET_SIMD instruction.
6322 (aarch64_frecps): Remove.
6323 * config/aarch64/aarch64-simd.md
6324 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6325 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6327 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6328 Sofiane Naci <sofiane.naci@arm.com>
6330 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6331 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6332 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6333 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6334 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6335 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6336 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6337 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6338 into "clz, "rbit". Rename "shift" to "shift_imm".
6339 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6340 changes. Update for attribute changes all occurrences of arlo_* and
6342 * config/arm/arm-fixed.md: Update for attribute changes
6343 all occurrences of arlo_* types.
6344 * config/arm/thumb2.md: Update for attribute changes all occurrences
6346 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6347 (cortexa7_older_only): Likewise.
6348 (cortexa7_younger): Likewise.
6349 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6350 (1020alu_shift_op): Likewise.
6351 (1020alu_shift_reg_op): Likewise.
6352 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6353 (alu_shift_op): Likewise.
6354 (alu_shift_reg_op): Likewise.
6355 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6356 (11_alu_shift_op): Likewise.
6357 (11_alu_shift_reg_op): Likewise.
6358 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6359 (9_alu_shift_reg_op): Likewise.
6360 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6362 (cortex_a15_alu_shift): Likewise.
6363 (cortex_a15_alu_shift_reg): Likewise.
6364 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6366 (cortex_a5_alu_shift): Likewise.
6367 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6369 (cortex_a53_alu_shift): Likewise.
6370 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6372 (cortex_a7_alu_reg): Likewise.
6373 (cortex_a7_alu_shift): Likewise.
6374 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6376 (cortex_a8_alu_shift): Likewise.
6377 (cortex_a8_alu_shift_reg): Likewise.
6378 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6379 (cortex_a9_dp_shift): Likewise.
6380 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6382 * config/arm/cortex-r4.md
6383 (cortex_r4_alu): Update for attribute changes.
6384 (cortex_r4_mov): Likewise.
6385 (cortex_r4_alu_shift_reg): Likewise.
6386 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6387 (526_alu_shift_op): Likewise.
6388 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6389 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6390 (626te_alu_shift_op): Likewise.
6391 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6392 (726te_alu_shift_op): Likewise.
6393 (726te_alu_shift_reg_op): Likewise.
6394 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6395 (mp626_alu_shift_op): Likewise.
6396 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6397 (pj4_alu_conds): Likewise.
6398 (pj4_shift): Likewise.
6399 (pj4_shift_conds): Likewise.
6400 (pj4_alu_shift): Likewise.
6401 (pj4_alu_shift_conds): Likewise.
6402 * config/aarch64/aarch64.md: Update for attribute change
6403 all occurrences of arlo_* and shift* types.
6405 2013-09-05 Mike Stump <mikestump@comcast.net>
6407 * tree.h: Move documentation for tree_function_decl to tree-core.h
6408 with the declaration.
6410 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6413 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6414 looking for widest mode.
6416 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6418 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6420 2013-09-05 Richard Biener <rguenther@suse.de>
6422 PR tree-optimization/58137
6423 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6424 Do not create vectors of pointers.
6425 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6426 types for the components of the vector initializer.
6427 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6428 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6430 2013-09-05 Martin Jambor <mjambor@suse.cz>
6432 * ipa-prop.c (remove_described_reference): Accept missing references,
6433 return false if that hppens, otherwise return true.
6434 (cgraph_node_for_jfunc): New function.
6435 (try_decrement_rdesc_refcount): Likewise.
6436 (try_make_edge_direct_simple_call): Use them.
6437 (ipa_edge_removal_hook): Remove references from rdescs.
6438 (ipa_edge_duplication_hook): Clone rdescs and their references
6439 when the new edge has the same caller as the old one.
6440 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6441 reference before removing any edges.
6443 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6445 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6447 * thumb2.md (thumb2_storewb_parisi): New pattern.
6449 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6451 * config/aarch64/aarch64-option-extensions.def: Add
6452 AARCH64_OPT_EXTENSION of 'crc'.
6453 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6454 (AARCH64_ISA_CRC): Ditto.
6455 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6456 description of the CRC extension.
6458 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6460 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6461 * config/rs6000/linux.h: Ditto.
6462 * alpha/linux.h: Ditto.
6463 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6464 no_c99_libc_has_function.
6465 * config/c6x/uclinux-elf.h: Ditto.
6466 * config/lm32/uclinux-elf.h: Ditto.
6467 * config/m68k/uclinux.h: Ditto.
6468 * config/moxie/uclinux.h: Ditto.
6469 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6470 (crisv32-*-linux*, cris-*-linux*): Ditto.
6471 * config/bfin/bfin.c: Include "tm_p.h".
6473 2013-09-05 Richard Biener <rguenther@suse.de>
6475 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6476 check for a definition without a basic-block.
6478 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6479 Sofiane Naci <sofiane.naci@arm.com>
6481 * config/aarch64/aarch64.md
6482 (*movti_aarch64): Rename r_2_f and f_2_r.
6483 (*movsf_aarch64): Likewise.
6484 (*movdf_aarch64): Likewise.
6485 (*movtf_aarch64): Likewise.
6486 (aarch64_movdi_<mode>low): Likewise.
6487 (aarch64_movdi_<mode>high): Likewise.
6488 (aarch64_mov<mode>high_di): Likewise.
6489 (aarch64_mov<mode>low_di): Likewise.
6490 (aarch64_movtilow_tilow): Likewise.
6491 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6492 values to config/arm/types.md
6493 (attribute "conds"): Update for attribute change.
6494 (anddi3_insn): Likewise.
6495 (iordi3_insn): Likewise.
6496 (xordi3_insn): Likewise.
6497 (one_cmpldi2): Likewise.
6498 * config/arm/types.md (type): Add Neon types.
6499 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6500 use "type" attribute.
6501 (movmisalign<mode>_neon_store): Likewise.
6502 (movmisalign<mode>_neon_load): Likewise.
6503 (vec_set<mode>_internal): Likewise.
6504 (vec_setv2di_internal): Likewise.
6505 (vec_extract<mode>): Likewise.
6506 (vec_extractv2di): Likewise.
6507 (add<mode>3_neon): Likewise.
6508 (adddi3_neon): Likewise.
6509 (sub<mode>3_neon): Likewise.
6510 (subdi3_neon): Likewise.
6511 (mul<mode>3_neon): Likewise.
6512 (mul<mode>3add<mode>_neon): Likewise.
6513 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6514 (fma<VCVTF:mode>4)): Likewise.
6515 (fma<VCVTF:mode>4_intrinsic): Likewise.
6516 (fmsub<VCVTF:mode>4)): Likewise.
6517 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6518 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6519 (ior<mode>3): Likewise.
6520 (and<mode>3): Likewise.
6521 (anddi3_neon): Likewise.
6522 (orn<mode>3_neon): Likewise.
6523 (orndi3_neon): Likewise.
6524 (bic<mode>3_neon): Likewise.
6525 (bicdi3_neon): Likewise.
6526 (xor<mode>3): Likewise.
6527 (one_cmpl<mode>2): Likewise.
6528 (abs<mode>2): Likewise.
6529 (neg<mode>2): Likewise.
6530 (umin<mode>3_neon): Likewise.
6531 (umax<mode>3_neon): Likewise.
6532 (smin<mode>3_neon): Likewise.
6533 (smax<mode>3_neon): Likewise.
6534 (vashl<mode>3): Likewise.
6535 (vashr<mode>3_imm): Likewise.
6536 (vlshr<mode>3_imm): Likewise.
6537 (ashl<mode>3_signed): Likewise.
6538 (ashl<mode>3_unsigned): Likewise.
6539 (neon_load_count): Likewise.
6540 (ashldi3_neon_noclobber): Likewise.
6541 (signed_shift_di3_neon): Likewise.
6542 (unsigned_shift_di3_neon): Likewise.
6543 (ashrdi3_neon_imm_noclobber): Likewise.
6544 (lshrdi3_neon_imm_noclobber): Likewise.
6545 (widen_ssum<mode>3): Likewise.
6546 (widen_usum<mode>3): Likewise.
6547 (quad_halves_<code>v4si): Likewise.
6548 (quad_halves_<code>v4sf): Likewise.
6549 (quad_halves_<code>v8hi): Likewise.
6550 (quad_halves_<code>v16qi): Likewise.
6551 (reduc_splus_v2di): Likewise.
6552 (neon_vpadd_internal<mode>): Likewise.
6553 (neon_vpsmin<mode>): Likewise.
6554 (neon_vpsmax<mode>): Likewise.
6555 (neon_vpumin<mode>): Likewise.
6556 (neon_vpumax<mode>): Likewise.
6557 (ss_add<mode>_neon): Likewise.
6558 (us_add<mode>_neon): Likewise.
6559 (ss_sub<mode>_neon): Likewise.
6560 (us_sub<mode>_neon): Likewise.
6561 (neon_vadd<mode>_unspec): Likewise.
6562 (neon_vaddl<mode>): Likewise.
6563 (neon_vaddw<mode>): Likewise.
6564 (neon_vhadd<mode>): Likewise.
6565 (neon_vqadd<mode>): Likewise.
6566 (neon_vaddhn<mode>): Likewise.
6567 (neon_vmul<mode>): Likewise.
6568 (neon_vmla<mode>): Likewise.
6569 (neon_vmlal<mode>): Likewise.
6570 (neon_vmls<mode>): Likewise.
6571 (neon_vmlsl<mode>): Likewise.
6572 (neon_vqdmulh<mode>): Likewise.
6573 (neon_vqdmlal<mode>): Likewise.
6574 (neon_vqdmlsl<mode>): Likewise.
6575 (neon_vmull<mode>): Likewise.
6576 (neon_vqdmull<mode>): Likewise.
6577 (neon_vsub<mode>_unspec): Likewise.
6578 (neon_vsubl<mode>): Likewise.
6579 (neon_vsubw<mode>): Likewise.
6580 (neon_vqsub<mode>): Likewise.
6581 (neon_vhsub<mode>): Likewise.
6582 (neon_vsubhn<mode>): Likewise.
6583 (neon_vceq<mode>): Likewise.
6584 (neon_vcge<mode>): Likewise.
6585 (neon_vcgeu<mode>): Likewise.
6586 (neon_vcgt<mode>): Likewise.
6587 (neon_vcgtu<mode>): Likewise.
6588 (neon_vcle<mode>): Likewise.
6589 (neon_vclt<mode>): Likewise.
6590 (neon_vcage<mode>): Likewise.
6591 (neon_vcagt<mode>): Likewise.
6592 (neon_vtst<mode>): Likewise.
6593 (neon_vabd<mode>): Likewise.
6594 (neon_vabdl<mode>): Likewise.
6595 (neon_vaba<mode>): Likewise.
6596 (neon_vabal<mode>): Likewise.
6597 (neon_vmax<mode>): Likewise.
6598 (neon_vmin<mode>): Likewise.
6599 (neon_vpaddl<mode>): Likewise.
6600 (neon_vpadal<mode>): Likewise.
6601 (neon_vpmax<mode>): Likewise.
6602 (neon_vpmin<mode>): Likewise.
6603 (neon_vrecps<mode>): Likewise.
6604 (neon_vrsqrts<mode>): Likewise.
6605 (neon_vqabs<mode>): Likewise.
6606 (neon_vqneg<mode>): Likewise.
6607 (neon_vcls<mode>): Likewise.
6608 (clz<mode>2): Likewise.
6609 (popcount<mode>2): Likewise.
6610 (neon_vrecpe): Likewise.
6611 (neon_vrsqrte): Likewise.
6612 (neon_vget_lane<mode>_sext_internal): Likewise.
6613 (neon_vget_lane<mode>_zext_internal): Likewise.
6614 (neon_vdup_n<mode>): Likewise.
6615 (neon_vdup_nv2di): Likewise.
6616 (neon_vdpu_lane<mode>_internal): Likewise.
6617 (neon_vswp<mode>): Likewise.
6618 (float<mode><V_cvtto>2): Likewise.
6619 (floatuns<mode><V_cvtto>2): Likewise.
6620 (fix_trunc<mode><V_cvtto>)2): Likewise
6621 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6622 (neon_vcvt<mode>): Likewise.
6623 (neon_vcvtv4sfv4hf): Likewise.
6624 (neon_vcvtv4hfv4sf): Likewise.
6625 (neon_vcvt_n<mode>): Likewise.
6626 (neon_vmovn<mode>): Likewise.
6627 (neon_vqmovn<mode>): Likewise.
6628 (neon_vqmovun<mode>): Likewise.
6629 (neon_vmovl<mode>): Likewise.
6630 (neon_vmul_lane<mode>): Likewise.
6631 (neon_vmull_lane<mode>): Likewise.
6632 (neon_vqdmull_lane<mode>): Likewise.
6633 (neon_vqdmulh_lane<mode>): Likewise.
6634 (neon_vmla_lane<mode>): Likewise.
6635 (neon_vmlal_lane<mode>): Likewise.
6636 (neon_vqdmlal_lane<mode>): Likewise.
6637 (neon_vmls_lane<mode>): Likewise.
6638 (neon_vmlsl_lane<mode>): Likewise.
6639 (neon_vqdmlsl_lane<mode>): Likewise.
6640 (neon_vext<mode>): Likewise.
6641 (neon_vrev64<mode>): Likewise.
6642 (neon_vrev32<mode>): Likewise.
6643 (neon_vrev16<mode>): Likewise.
6644 (neon_vbsl<mode>_internal): Likewise.
6645 (neon_vshl<mode>): Likewise.
6646 (neon_vqshl<mode>): Likewise.
6647 (neon_vshr_n<mode>): Likewise.
6648 (neon_vshrn_n<mode>): Likewise.
6649 (neon_vqshrn_n<mode>): Likewise.
6650 (neon_vqshrun_n<mode>): Likewise.
6651 (neon_vshl_n<mode>): Likewise.
6652 (neon_vqshl_n<mode>): Likewise.
6653 (neon_vqshlu_n<mode>): Likewise.
6654 (neon_vshll_n<mode>): Likewise.
6655 (neon_vsra_n<mode>): Likewise.
6656 (neon_vsri_n<mode>): Likewise.
6657 (neon_vsli_n<mode>): Likewise.
6658 (neon_vtbl1v8qi): Likewise.
6659 (neon_vtbl2v8qi): Likewise.
6660 (neon_vtbl3v8qi): Likewise.
6661 (neon_vtbl4v8qi): Likewise.
6662 (neon_vtbx1v8qi): Likewise.
6663 (neon_vtbx2v8qi): Likewise.
6664 (neon_vtbx3v8qi): Likewise.
6665 (neon_vtbx4v8qi): Likewise.
6666 (neon_vtrn<mode>_internal): Likewise.
6667 (neon_vzip<mode>_internal): Likewise.
6668 (neon_vuzp<mode>_internal): Likewise.
6669 (neon_vld1<mode>): Likewise.
6670 (neon_vld1_lane<mode>): Likewise.
6671 (neon_vld1_dup<mode>): Likewise.
6672 (neon_vld1_dupv2di): Likewise.
6673 (neon_vst1<mode>): Likewise.
6674 (neon_vst1_lane<mode>): Likewise.
6675 (neon_vld2<mode>): Likewise.
6676 (neon_vld2_lane<mode>): Likewise.
6677 (neon_vld2_dup<mode>): Likewise.
6678 (neon_vst2<mode>): Likewise.
6679 (neon_vst2_lane<mode>): Likewise.
6680 (neon_vld3<mode>): Likewise.
6681 (neon_vld3qa<mode>): Likewise.
6682 (neon_vld3qb<mode>): Likewise.
6683 (neon_vld3_lane<mode>): Likewise.
6684 (neon_vld3_dup<mode>): Likewise.
6685 (neon_vst3<mode>): Likewise.
6686 (neon_vst3qa<mode>): Likewise.
6687 (neon_vst3qb<mode>): Likewise.
6688 (neon_vst3_lane<mode>): Likewise.
6689 (neon_vld4<mode>): Likewise.
6690 (neon_vld4qa<mode>): Likewise.
6691 (neon_vld4qb<mode>): Likewise.
6692 (neon_vld4_lane<mode>): Likewise.
6693 (neon_vld4_dup<mode>): Likewise.
6694 (neon_vst4<mode>): Likewise.
6695 (neon_vst4qa<mode>): Likewise.
6696 (neon_vst4qb<mode>): Likewise.
6697 (neon_vst4_lane<mode>): Likewise.
6698 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6699 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6700 (neon_vec_<US>mult_lo_<mode>): Likewise.
6701 (neon_vec_<US>mult_hi_<mode>): Likewise.
6702 (neon_vec_<US>shiftl_<mode>): Likewise.
6703 (neon_unpack<US>_<mode>): Likewise.
6704 (neon_vec_<US>mult_<mode>): Likewise.
6705 (vec_pack_trunc_<mode>): Likewise.
6706 (neon_vec_pack_trunk_<mode>): Likewise.
6707 (neon_vabd<mode>_2): Likewise.
6708 (neon_vabd<mode>_3): Likewise.
6709 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6710 (thumb2_movsi_vfp): Likewise.
6711 (movdi_vfp): Likewise.
6712 (movdi_vfp_cortexa8): Likewise.
6713 (movhf_vfp_neon): Likewise.
6714 (movhf_vfp): Likewiwse.
6715 (movsf_vfp): Likewiwse.
6716 (thumb2_movsf_vfp): Likewiwse.
6717 (movdf_vfp): Likewise.
6718 (thumb2_movdf_vfp): Likewise.
6719 (movsfcc_vfp): Likewise.
6720 (thumb2_movsfcc_vfp): Likewise.
6721 (movdfcc_vfp): Likewise.
6722 (thumb2_movdfcc_vfp): Likewise.
6723 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6724 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6725 (v10_v2c): Likewise.
6726 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6728 (cortex_a15_neon_int_2): Likewise.
6729 (cortex_a15_neon_int_3): Likewise.
6730 (cortex_a15_neon_int_4): Likewise.
6731 (cortex_a15_neon_int_5): Likewise.
6732 (cortex_a15_neon_vqneg_vqabs): Likewise.
6733 (cortex_a15_neon_vmov): Likewise.
6734 (cortex_a15_neon_vaba): Likewise.
6735 (cortex_a15_neon_vaba_qqq): Likewise.
6736 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6737 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6738 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6739 scalar_64_32_long_scalar): Likewise.
6740 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6741 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6742 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6743 lotype_qdd_64_32_long): Likewise.
6744 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6745 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6746 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6747 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6748 (cortex_a15_neon_shift_1): Likewise.
6749 (cortex_a15_neon_shift_2): Likewise.
6750 (cortex_a15_neon_shift_3): Likewise.
6751 (cortex_a15_neon_vshl_ddd): Likewise.
6752 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6753 (cortex_a15_neon_vsra_vrsra): Likewise.
6754 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6755 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6756 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6757 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6758 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6759 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6760 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6761 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6762 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6763 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6764 (cortex_a15_neon_bp_simple): Likewise.
6765 (cortex_a15_neon_bp_2cycle): Likewise.
6766 (cortex_a15_neon_bp_3cycle): Likewise.
6767 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6768 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6769 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6770 (cortex_a15_neon_vld2_4_regs): Likewise.
6771 (cortex_a15_neon_vld3_vld4): Likewise.
6772 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6773 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6774 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6775 (cortex_a15_neon_vst3_vst4): Likewise.
6776 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6777 (cortex_a15_neon_vld3_vld4_lane" 10
6778 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6779 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6780 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6781 (cortex_a15_neon_ldm_2): Likewise.0
6782 (cortex_a15_neon_stm_2): Likewise.
6783 (cortex_a15_neon_mcr): Likewise.
6784 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6785 (cortex_a15_neon_mrc): Likewise.
6786 (cortex_a15_neon_mrrc): Likewise.
6787 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6789 (cortex_a15_alu_shift): Likewise.
6790 (cortex_a15_alu_shift_reg): Likewise.
6791 (cortex_a15_mult32): Likewise.
6792 (cortex_a15_mult64): Likewise.
6793 (cortex_a15_block): Likewise.
6794 (cortex_a15_branch): Likewise.
6795 (cortex_a15_load1): Likewise.
6796 (cortex_a15_load3): Likewise.
6797 (cortex_a15_store1): Likewise.
6798 (cortex_a15_store3): Likewise.
6799 (cortex_a15_call): Likewise.
6800 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6801 (cortex_a5_f2r): Likewise.
6802 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6804 (cortex_a53_f2r): Likewise.
6805 * config/arm/cortex-a7.md
6806 (cortex_a7_branch): Update for attribute change.
6807 (cortex_a7_call): Likewise.
6808 (cortex_a7_alu_imm): Likewise.
6809 (cortex_a7_alu_reg): Likewise.
6810 (cortex_a7_alu_shift): Likewise.
6811 (cortex_a7_mul): Likewise.
6812 (cortex_a7_load1): Likewise.
6813 (cortex_a7_store1): Likewise.
6814 (cortex_a7_load2): Likewise.
6815 (cortex_a7_store2): Likewise.
6816 (cortex_a7_load3): Likewise.
6817 (cortex_a7_store3): Likewise.
6818 (cortex_a7_load4): Likewise.
6819 (cortex_a7_store4): Likewise.
6820 (cortex_a7_fpalu): Likewise.
6821 (cortex_a7_fconst): Likewise.
6822 (cortex_a7_fpmuls): Likewise.
6823 (cortex_a7_neon_mul): Likewise.
6824 (cortex_a7_fpmacs): Likewise.
6825 (cortex_a7_neon_mla: Likewise.
6826 (cortex_a7_fpmuld: Likewise.
6827 (cortex_a7_fpmacd: Likewise.
6828 (cortex_a7_fpfmad: Likewise.
6829 (cortex_a7_fdivs: Likewise.
6830 (cortex_a7_fdivd: Likewise.
6831 (cortex_a7_r2f: Likewise.
6832 (cortex_a7_f2r: Likewise.
6833 (cortex_a7_f_flags: Likewise.
6834 (cortex_a7_f_loads: Likewise.
6835 (cortex_a7_f_loadd: Likewise.
6836 (cortex_a7_f_stores: Likewise.
6837 (cortex_a7_f_stored: Likewise.
6838 (cortex_a7_neon): Likewise.
6839 * config/arm/cortex-a8-neon.md
6840 (cortex_a8_neon_mrc): Update for attribute change.
6841 (cortex_a8_neon_mrrc): Likewise.
6842 (cortex_a8_neon_int_1): Likewise.
6843 (cortex_a8_neon_int_2): Likewise.
6844 (cortex_a8_neon_int_3): Likewise.
6845 (cortex_a8_neon_int_4): Likewise.
6846 (cortex_a8_neon_int_5): Likewise.
6847 (cortex_a8_neon_vqneg_vqabs): Likewise.
6848 (cortex_a8_neon_vmov): Likewise.
6849 (cortex_a8_neon_vaba): Likewise.
6850 (cortex_a8_neon_vaba_qqq): Likewise.
6851 (cortex_a8_neon_vsma): Likewise.
6852 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6853 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6854 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6855 long_scalar): Likewise.
6856 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6857 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6858 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6859 long_scalar_qdd_64_32_long): Likewise.
6860 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6861 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6862 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6863 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6864 (cortex_a8_neon_shift_1): Likewise.
6865 (cortex_a8_neon_shift_2): Likewise.
6866 (cortex_a8_neon_shift_3): Likewise.
6867 (cortex_a8_neon_vshl_ddd): Likewise.
6868 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6869 (cortex_a8_neon_vsra_vrsra): Likewise.
6870 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6871 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6872 (cortex_a8_neon_fp_vsum): Likewise.
6873 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6874 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6875 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6876 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6877 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6878 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6879 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6880 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6881 (cortex_a8_neon_bp_simple): Likewise.
6882 (cortex_a8_neon_bp_2cycle): Likewise.
6883 (cortex_a8_neon_bp_3cycle): Likewise.
6884 (cortex_a8_neon_ldr): Likewise.
6885 (cortex_a8_neon_str): Likewise.
6886 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6887 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6888 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6889 (cortex_a8_neon_vld2_4_regs): Likewise.
6890 (cortex_a8_neon_vld3_vld4): Likewise.
6891 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6892 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6893 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6894 (cortex_a8_neon_vst3_vst4): Likewise.
6895 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6896 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6897 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6898 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6899 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6900 (cortex_a8_neon_mcr): Likewise.
6901 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6902 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6903 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6905 (ca9_neon_mrrc): Likewise.
6906 (cortex_a9_neon_int_1): Likewise.
6907 (cortex_a9_neon_int_2): Likewise.
6908 (cortex_a9_neon_int_3): Likewise.
6909 (cortex_a9_neon_int_4): Likewise.
6910 (cortex_a9_neon_int_5): Likewise.
6911 (cortex_a9_neon_vqneg_vqabs): Likewise.
6912 (cortex_a9_neon_vmov): Likewise.
6913 (cortex_a9_neon_vaba): Likewise.
6914 (cortex_a9_neon_vaba_qqq): Likewise.
6915 (cortex_a9_neon_vsma): Likewise.
6916 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6917 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6918 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6919 long_scalar): Likewise.
6920 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6921 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6922 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6923 long_scalar_qdd_64_32_long): Likewise.
6924 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6925 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6926 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6927 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6928 (cortex_a9_neon_shift_1): Likewise.
6929 (cortex_a9_neon_shift_2): Likewise.
6930 (cortex_a9_neon_shift_3): Likewise.
6931 (cortex_a9_neon_vshl_ddd): Likewise.
6932 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6933 (cortex_a9_neon_vsra_vrsra): Likewise.
6934 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6935 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6936 (cortex_a9_neon_fp_vsum): Likewise.
6937 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6938 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6939 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6940 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6941 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6942 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6943 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6944 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6945 (cortex_a9_neon_bp_simple): Likewise.
6946 (cortex_a9_neon_bp_2cycle): Likewise.
6947 (cortex_a9_neon_bp_3cycle): Likewise.
6948 (cortex_a9_neon_ldr): Likewise.
6949 (cortex_a9_neon_str): Likewise.
6950 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6951 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6952 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6953 (cortex_a9_neon_vld2_4_regs): Likewise.
6954 (cortex_a9_neon_vld3_vld4): Likewise.
6955 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6956 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6957 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6958 (cortex_a9_neon_vst3_vst4): Likewise.
6959 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6960 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6961 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6962 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6963 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6964 (cortex_a9_neon_mcr): Likewise.
6965 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6966 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6967 (cortex_a9_fps): Likewise.
6968 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6970 (cortex_m4_fmuls): Likewise.
6971 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6973 (cortex_r4_mrc): Likewise.
6974 * config/arm/iterators.md: Update comment referring to neon_type.
6975 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
6976 (iwmmxt_movsi_insn): Likewise.
6977 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6979 (pj4_core_to_vfp): Likewise.
6980 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6982 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6983 (vfp_fstore): Likewise.
6984 * doc/md.texi: Change references to neon_type to refer to type.
6986 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6988 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6990 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
6992 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6995 2013-09-04 Easwaran Raman <eraman@google.com>
6998 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6999 (build_and_add_sum): Use it.
7000 (appears_later_in_bb): Simplify code.
7002 2013-09-04 Teresa Johnson <tejohnson@google.com>
7004 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7006 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7008 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7010 2013-09-04 Jan Hubicka <jh@suse.cz>
7012 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7013 * ipa-devirt.c: Include diganostic.h
7014 (odr_type_d): Add types and types_set.
7015 (hash_type_name): Work for types with vtables during LTO.
7016 (odr_hasher::remove): Fix comment; destroy types_set.
7017 (add_type_duplicate): New function,
7018 (get_odr_type): Use it.
7019 (dump_type_inheritance_graph): Dump type duplicates.
7020 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7022 * tree.c (types_same_for_odr): Give exact answers on types with
7025 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7027 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7028 explaining their differences.
7030 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7032 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7034 2013-09-03 Jeff Law <law@redhat.com>
7036 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7037 when not threading through a joiner block. Pass joiner/no joiner
7038 state to register_jump_thread.
7039 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7040 state from argument rather than implying on path length.
7041 Dump the entire jump thread path into debugging dump.
7042 * tree-flow.h (register_jump_thread): Update prototype.
7044 2013-08-29 Xinliang David Li <davidxl@google.com>
7046 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7047 Remove a trivial gcc_assert.
7049 2013-08-29 Xinliang David Li <davidxl@google.com>
7051 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7052 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7053 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7054 Delay base decl alignment adjustment.
7055 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7056 * tree-vectorizer.h: New data structure.
7057 (set_dr_misalignment): New function.
7058 (dr_misalignment): Ditto.
7059 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7060 (vectorizable_load): Ditto.
7061 (ensure_base_align): New function.
7062 (vectorize_loops): Add dbg_cnt support.
7063 (execute_vect_slp): Ditto.
7064 * dbgcnt.def: New debug counter.
7065 * Makefile: New dependency.
7067 2013-09-03 Meador Inge <meadori@codesourcery.com>
7071 2013-08-30 Meador Inge <meadori@codesourcery.com>
7073 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7075 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7077 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7078 function descriptor.
7080 2013-09-03 Richard Biener <rguenther@suse.de>
7082 * tree-affine.c (add_elt_to_tree): Fix association issue,
7083 avoid useless converts and make sure to always return a
7084 properly typed result.
7086 2013-09-03 Richard Biener <rguenther@suse.de>
7089 * fold-const.c (negate_expr_p): Fix division case.
7090 (negate_expr): Likewise.
7092 2013-09-03 Richard Biener <rguenther@suse.de>
7095 * tree-streamer-out.c: Include tm.h.
7096 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7098 2013-09-03 Jan Hubicka <jh@suse.cz>
7100 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7102 2013-09-03 Alan Modra <amodra@gmail.com>
7104 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7105 multilibs to big-endian target and vice versa.
7106 * config/rs6000/t-linux64: Use := assignment on all vars.
7107 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7108 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7109 * config/rs6000/t-linux64le: New file.
7110 * config/rs6000/t-linux64bele: New file.
7111 * config/rs6000/t-linux64lebe: New file.
7113 2013-09-02 Jan Hubicka <jh@suse.cz>
7115 * ipa-inline-transform.c (inline_transform): Do not
7116 optimize_inline_calls when not optimizing.
7118 2013-09-02 Jan Hubicka <jh@suse.cz>
7120 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7121 duplicated nodes for assembler names.
7122 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7123 hard registers from assembler name hash.
7125 2013-09-02 Jan Hubicka <jh@suse.cz>
7127 * ipa-split.c (execute_split_functions): Split externally visible
7128 functions called once.
7130 2013-09-02 Martin Jambor <mjambor@suse.cz>
7133 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7134 linked list. When finding the correct duplicate, also consider also
7135 the caller in additon to its inlined_to node.
7137 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7139 * config/aarch64/aarch64-simd-builtins.def
7140 (dup_lane_scalar): Remove.
7141 * config/aarch64/aarch64-simd.md
7142 (aarch64_simd_dup): Add 'w->w' alternative.
7143 (aarch64_dup_lane<mode>): Allow for VALL.
7144 (aarch64_dup_lane_scalar<mode>): Remove.
7145 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7146 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7147 (aarch64_get_lane_unsigned<mode>): Likewise.
7148 (aarch64_get_lane<mode>): Likewise.
7149 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7150 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7151 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7152 (VCON): Change container of V2SF.
7153 (vswap_width_name): Likewise.
7154 * config/aarch64/arm_neon.h
7155 (__aarch64_vdup_lane_any): New.
7156 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7157 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7158 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7160 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7163 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7164 the source or the destination is a single hard register.
7166 2013-09-02 Richard Biener <rguenther@suse.de>
7169 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7172 2013-09-02 Richard Biener <rguenther@suse.de>
7174 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7175 arithmetic to sizetype.
7177 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7179 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7180 Find auto-increment use both before and after candidate.
7182 2013-09-02 Marek Polacek <polacek@redhat.com>
7184 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7186 2013-09-01 Jan Hubicka <jh@suse.cz>
7188 * Makefile.in: Add ipa-profile.o
7189 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7190 * cgraph.c (struct cgraph_propagate_frequency_data,
7191 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7192 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7193 * cgraph.h (cgraph_propagate_frequency): Remove.
7194 * ipa-inline-analysis.c: Include ipa-utils.h;
7195 drop duplicated cfgloop.h.
7196 (inline_update_callee_summaries): Update.
7197 * ipa-profile.c: New file.
7198 * ipa-utils.h (ipa_propagate_frequency): Declare.
7199 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7200 data-streamer.h, value-prof.h.
7201 (symtab_remove_unreachable_nodes): Update profile.
7202 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7203 account_time_size, cmp_counts, dump_histogram,
7204 ipa_profile_generate_summary, ipa_profile_write_summary,
7205 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7206 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7207 Move to ipa-profile.c.
7209 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7211 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7213 2013-09-01 Jan Hubicka <jh@suse.cz>
7215 * common.opt (fdevirtualize-speculatively): New function.
7216 * invoke.texi (fdevirtualize-speculatively): Document.
7217 * ipa-devirt.c: Include ipa-inline.h
7218 (likely_target_p): New function.
7219 (ipa_devirt): New function.
7220 (gate_ipa_devirt): New function.
7221 (pass_data_ipa_devirt): New static var.
7222 (pass_ipa_devirt): Likewise.
7223 (make_pass_ipa_devirt): New function.
7224 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7225 (common_handle_option): Disable devirtualization when
7226 value range profiling is available.
7227 * passes.def (pass_ipa_devirt): Add.
7228 * timever.def (TV_IPA_DEVIRT): New timevar.
7229 * tree-pass.h (make_pass_ipa_devirt):
7231 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7233 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7234 include sanitize(undefined).
7236 2013-08-31 Diego Novillo <dnovillo@google.com>
7238 * Makefile.in (TREE_CORE_H): Define.
7240 (GTFILES): Add tree-core.h.
7241 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7243 * tree-core.h: New file.
7244 Move all data structures, enum, typedefs, global
7245 declarations and constants from ...
7248 2013-08-31 Jan Hubicka <jh@suse.cz>
7250 * bulitins.c (expand_builtin): Do not early exit for gcov
7251 instrumented functions.
7253 2013-08-31 Marek Polacek <polacek@redhat.com>
7255 * ubsan.c: Include tm_p.h.
7257 2013-08-31 Jan Hubicka <jh@suse.cz>
7259 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7262 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7263 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7264 * tree-cfg.c (verify_gimple_label): ... here.
7265 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7266 (ipa_merge_profiles): New function.
7267 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7268 (lto_input_function_body): Likewise.
7269 * ipa-utils.h (ipa_merge_profiles): Declare.
7270 * lto-streamer.h (lto_input_function_body): Update prototype.
7271 (emit_label_in_global_context_p): Remove.
7272 * lto-symtab.c: Include ipa-utils.h
7273 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7275 2013-08-31 Jan Hubicka <jh@suse.cz>
7277 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7279 2013-08-31 Jan Hubicka <jh@suse.cz>
7281 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7283 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7285 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7286 "cmp" RTX before signed_comparison_operator check to account
7289 2013-08-30 Jan Hubicka <jh@suse.cz>
7291 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7292 (detect_type_change_1): Rename to ...
7293 (detect_type_change): ... this one; early return on non-polymorphic
7295 (detect_type_change_ssa): Add comp_type parameter; update
7296 use of detect_type_change.
7297 (compute_complex_assign_jump_func): Add param_type parameter;
7298 update use of detect_type_change_ssa.
7299 (compute_complex_ancestor_jump_func): Likewise.
7300 (ipa_get_callee_param_type): New function.
7301 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7302 update calls to the jump function computation functions.
7304 2013-08-30 Teresa Johnson <tejohnson@google.com>
7305 Steven Bosscher <steven@gcc.gnu.org>
7307 * cfgrtl.c (fixup_new_cold_bb): New routine.
7308 (commit_edge_insertions): Invoke fixup_partitions.
7309 (find_partition_fixes): New routine.
7310 (fixup_partitions): Ditto.
7311 (verify_hot_cold_block_grouping): Update comments.
7312 (rtl_verify_edges): Invoke find_partition_fixes.
7313 (rtl_verify_bb_pointers): Update comments.
7314 (rtl_verify_bb_layout): Ditto.
7315 * basic-block.h (probably_never_executed_edge_p): Declare.
7316 (fixup_partitions): Ditto.
7317 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7318 * bb-reorder.c (sanitize_hot_paths): New function.
7319 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7321 * predict.c (probably_never_executed_edge_p): New routine.
7322 * cfg.c (check_bb_profile): Add partition insanity warnings.
7324 2013-08-30 Meador Inge <meadori@codesourcery.com>
7326 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7328 2013-08-30 Marek Polacek <polacek@redhat.com>
7330 * Makefile.in (ubsan.o): Add.
7331 (c-family/c-ubsan.o): Add.
7332 (builtins.o): Add ubsan.h dependency.
7333 * ubsan.h: New file.
7334 * ubsan.c: New file.
7335 * common.opt: Add -fsanitize=undefined option.
7336 (flag_sanitize): Add variable.
7337 (fsanitize=): Add option. Add Driver.
7338 (fsanitize=thread): Remove option.
7339 (fsanitize=address): Likewise.
7340 (static-libubsan): New option.
7341 * doc/invoke.texi: Document the new flag and -static-libubsan.
7342 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7343 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7344 * builtin-attrs.def (ATTR_COLD): Define.
7345 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7346 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7347 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7348 * flag-types.h (sanitize_code): New enum.
7349 * opts.c (common_handle_option): Parse command line arguments
7350 of -fsanitize=. Add -fsanitize=unreachable option.
7351 * varasm.c (get_variable_section): Adjust.
7352 (assemble_noswitch_variable): Likewise.
7353 (assemble_variable): Likewise.
7354 (output_constant_def_contents): Likewise.
7355 (categorize_decl_for_section): Likewise.
7356 (place_block_symbol): Likewise.
7357 (output_object_block): Likewise.
7358 * builtins.def: Likewise.
7359 * toplev.c (compile_file): Likewise.
7360 (process_options): Likewise.
7361 * cppbuiltin.c: Likewise.
7362 * tsan.c (tsan_pass): Likewise.
7363 (tsan_gate): Likewise.
7364 (tsan_gate_O0): Likewise.
7365 * cfgexpand.c (partition_stack_vars): Likewise.
7366 (expand_stack_vars): Likewise.
7367 (defer_stack_allocation): Likewise.
7368 (expand_used_vars): Likewise.
7369 * cfgcleanup.c (old_insns_match_p): Likewise.
7370 * asan.c (asan_finish_file): Likewise.
7371 (asan_instrument): Likewise.
7372 (gate_asan): Likewise.
7373 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7374 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7375 (asan_global_struct): Use pointer_sized_int_node instead
7376 calling build_nonstandard_integer_type.
7377 (initialize_sanitizer_builtins): Likewise.
7378 (asan_finish_file): Likewise.
7379 * gcc.c: Document %{%:function(args):X}.
7380 (static_spec_functions): Add sanitize.
7381 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7382 store funcval != NULL there.
7383 (do_spec_1): Adjust handle_spec_function caller.
7384 (handle_braces): Allow %:function(args) as condition.
7385 (sanitize_spec_function): New function.
7386 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7387 (LIBUBSAN_SPEC): Likewise.
7388 (LIBUBSAN_EARLY_SPEC): Likewise.
7389 (SANITIZER_SPEC): Handle libubsan.
7390 (SANITIZER_EARLY_SPEC): Likewise.
7391 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7392 instead of fsanitize=address.
7393 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7394 instead of fsanitize=address*.
7395 * builtins.c: Include ubsan.h.
7396 (fold_builtin_0): Instrument __builtin_unreachable.
7397 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7398 instead of flag_asan.
7399 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7400 (pointer_sized_int_node): Define.
7401 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7403 2013-08-30 Mike Stump <mikestump@comcast.net>
7405 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7408 2013-08-29 Jan Hubicka <jh@suse.cz>
7410 * cgraph.c (cgraph_function_body_availability): Handle weakref
7412 * passes.def: Remove pass_fixup_cfg.
7413 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7414 track when we need to remove functions.
7415 (gate_ipa_inline): Execute inlining always; add comment why.
7416 (pass_data_ipa_inline): Remove TODO_remove_functions.
7417 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7418 do not produce summaries.
7419 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7420 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7421 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7422 constant pool and vtable.
7424 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7426 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7427 New arm_neon.h's internal macros to specify 64-bit constants.
7428 Avoid using stdint.h's macros.
7430 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7432 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7433 before applying REG_P.
7435 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7437 PR tree-optimization/58277
7438 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7439 after seeing too many stmts with vdef in between dombb and current
7440 bb, invalidate everything.
7442 2013-08-30 Richard Biener <rguenther@suse.de>
7444 * fold-const.c (fold_single_bit_test): Fix overflow test.
7446 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7448 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7449 and which can live in a register, always retrieve the value on entry.
7450 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7451 passed by invisible reference specially.
7452 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7453 (vt_add_function_parameter): Correctly deal with a parameter passed by
7454 invisible reference.
7456 2013-08-30 Jan Hubicka <jh@suse.cz>
7458 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7460 2013-08-30 Richard Biener <rguenther@suse.de>
7462 PR tree-optimization/58228
7463 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7464 allow invariant loads in nested loop vectorization.
7466 2013-08-30 Richard Biener <rguenther@suse.de>
7468 PR tree-optimization/58223
7469 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7470 (has_anti_or_output_dependence): ... this and adjust to also
7471 look for output dependences.
7472 (mark_nodes_having_upstream_mem_writes): Adjust.
7473 (rdg_flag_uses): Likewise.
7475 2013-08-30 Richard Biener <rguenther@suse.de>
7477 PR tree-optimization/58010
7478 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7479 assert that we have a loop-closed PHI.
7481 2013-08-29 Jan Hubicka <jh@suse.cz>
7483 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7484 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7485 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7486 (lto_free_function_in_decl_state_for_node): New function.
7488 2013-08-29 Xinliang David Li <davidxl@google.com>
7490 * loop-unroll.c (report_unroll_peel): Minor message change.
7491 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7492 Emit alignment peeling message with default -fopt-info.
7493 (vect_loop_versioning): Emit loop version info message.
7494 * tree-vectorizer.c (vectorize_loops): Minor message change.
7495 (execute_vect_slp): Ditto.
7497 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7499 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7500 of the clone from the DECL_NAME of the original function.
7502 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7504 * passes.c (register_pass): Add overload.
7505 * tree-pass.h (register_pass): Forward declare it. Add comment.
7507 2013-08-29 Jan Hubicka <jh@suse.cz>
7509 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7510 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7511 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7512 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7513 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7514 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7515 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7516 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7518 2013-08-29 Teresa Johnson <tejohnson@google.com>
7520 * dumpfile.c (dump_loc): Output column number.
7521 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7522 * doc/invoke.texi: Document optall -fopt-info flag.
7523 * profile.c (read_profile_edge_counts): Use new dump framework.
7524 (compute_branch_probabilities): Ditto.
7525 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7526 when pass not in any opt group.
7527 * pass_manager.h (pass_manager::get_pass_profile): New method.
7528 * value-prof.c (check_counter): Use new dump framework.
7529 (check_ic_target): Ditto.
7530 * coverage.c (get_coverage_counts): Ditto.
7531 (coverage_init): Setup new dump framework.
7533 2013-08-29 Richard Biener <rguenther@suse.de>
7535 PR tree-optimization/58246
7536 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7537 handle the dominance check inside a basic-block.
7539 2013-08-29 Richard Biener <rguenther@suse.de>
7542 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7543 of default defs that appear in abnormal PHI nodes.
7545 2013-08-29 Richard Biener <rguenther@suse.de>
7547 PR tree-optimization/57685
7548 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7549 single-use operands to avoid exponential complexity.
7551 2013-08-28 Dehao Chen <dehao@google.com>
7553 * ipa-inline.c (edge_badness): Fix integer underflow.
7555 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7557 * gtm-builtins.def (_ITM_free): Declare leaf.
7559 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7562 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7563 (*tls_local_dynamic_base_64_largepic): Likewise.
7564 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7565 Remove predicate from call operand.
7566 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7567 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7569 2013-08-28 Jeff Law <law@redhat.com>
7571 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7572 checks for the number of predecessors and successors allowed.
7573 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7574 which require copying a joiner block if there is a request which
7575 is a subpath that requires no joiner block copying.
7577 2013-08-28 Jan Hubicka <jh@suse.cz>
7579 * lto-streamer-out.c (DFS_write_tree_body): Drop
7580 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7581 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7582 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7583 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7584 Do not read DECL_ERROR_ISSUED.
7585 (unpack_ts_decl_with_vis_value_fields): Do not read
7587 (lto_input_ts_binfo_tree_pointers): Do not read
7588 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7589 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7590 write DECL_ERROR_ISSUED..
7591 (pack_ts_decl_with_vis_value_fields): Do not write
7593 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7594 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7595 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7596 * tree.h (tree_decl_common): Update comment.
7597 (DECL_ERROR_ISSUED): Remove.
7599 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7602 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7604 2013-08-28 Jan Hubicka <jh@suse.cz>
7606 * builtins.def (free): Declare leaf.
7608 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7611 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7612 * configure: Regenerate.
7614 2013-08-27 Martin Jambor <mjambor@suse.cz>
7616 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7617 (ipa_ancestor_jf_data): Likewise.
7618 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7619 (ipa_get_jf_pass_through_type_preserved): New function.
7620 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7621 (ipa_get_jf_ancestor_type_preserved): New function.
7622 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7623 (ipa_get_jf_ancestor_result): Likewise.
7624 (propagate_vals_accross_pass_through): Use
7625 ipa_get_jf_pass_through_result to do all the value mappings.
7626 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7627 type_preserved flag.
7628 (ipa_set_jf_cst_copy): New function.
7629 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7630 (ipa_set_jf_arith_pass_through): Likewise.
7631 (ipa_set_ancestor_jf): Likewise.
7632 (compute_complex_assign_jump_func): Set type_preserved instead of
7634 (ipa_compute_jump_functions_for_edge): Likewise.
7635 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7636 (update_jump_functions_after_inlining): Update type_preserved.
7637 Explicitely create jump functions when combining one with pass_through.
7638 (ipa_write_jump_function): Stream the type_preserved flags.
7639 (ipa_read_jump_function): Likewise.
7641 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7642 Aldy Hernandez <aldyh@redhat.com>
7644 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7645 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7646 * function.h (struct function): Add has_force_vect_loops and
7648 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7649 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7651 * gimple.def: Update CLAUSES comments.
7652 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7653 (gimple_build_omp_for): Add argument to prototype.
7654 (gimple_omp_for_kind): New.
7655 (gimple_omp_for_set_kind): New.
7656 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7657 GOVD_DATA_SHARE_CLASS.
7658 (enum omp_region_type): Add ORT_SIMD.
7659 (gimple_add_tmp_var): Handle ORT_SIMD.
7660 (gimplify_var_or_parm_decl): Same.
7661 (is_gimple_stmt): Same.
7662 (omp_firstprivatize_variable): Same.
7663 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7664 (omp_notice_variable): Handle ORT_SIMD.
7665 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7666 (omp_check_private): Handle ORT_SIMD.
7667 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7669 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7670 Handle OMP_CLAUSE_LASTPRIVATE.
7671 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7673 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7674 (gimplify_expr): Handle OMP_SIMD.
7675 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7676 (expand_GOMP_SIMD_VF): New.
7677 (expand_GOMP_SIMD_LAST_LANE): New.
7678 * internal-fn.def (GOMP_SIMD_LANE): New.
7679 (GOMP_SIMD_VF): New.
7680 (GOMP_SIMD_LAST_LANE): New.
7681 * omp-low.c: Include target.h.
7682 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7684 (check_omp_nesting_restrictions): Same.
7686 (lower_rec_simd_input_clauses): New.
7687 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7689 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7690 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7691 (expand_omp_build_assign): New.
7692 (expand_omp_for_init_counts): New.
7693 (expand_omp_for_init_vars): New.
7694 (extract_omp_for_update_vars): New.
7695 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7696 and rewrite accordingly.
7697 (expand_omp_simd): New.
7698 (expand_omp_for): Use expand_omp_simd.
7699 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7700 (lower_omp_for): Do not lower the body.
7701 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7703 * tree-flow.h (find_omp_clause): Remove prototype.
7704 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7705 forcing vectorization of the loop, or if flag_tree_vectorize.
7706 (gate_tree_if_conversion): Similarly.
7707 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7708 gimple_build_omp_for.
7709 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7710 * tree-parloops (create_parallel_loop): Pass kind argument to
7711 gimple_build_omp_for.
7712 * tree-pretty-print.c (dump_omp_clause): Add cases for
7713 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7714 OMP_CLAUSE__SIMDUID_.
7715 (dump_generic_node): Handle OMP_SIMD.
7716 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7717 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7718 unroll OMP_SIMD loops here.
7719 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7720 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7722 (vect_analyze_data_refs): Handle simd loops.
7723 * tree-vect-loop.c (vectorizable_live_operation): Handle
7725 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7726 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7727 (vectorizable_load): Same.
7728 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7729 (struct simduid_to_vf): New.
7730 (simduid_to_vf::hash): New.
7731 (simduid_to-vf::equal): New.
7732 (struct simd_array_to_simduid): New.
7733 (simd_array_to_simduid::hash): New.
7734 (simd_array_to_simduid::equal): New.
7735 (adjust_simduid_builtins): New.
7736 (struct note_simd_array_uses_struct): New.
7737 (note_simd_array_uses_cb): New.
7738 (note_simd_array_uses): New.
7739 (vectorize_loops): Handle simd hints and adjust simd builtins
7741 * tree-vectorizer.h (struct _stmt_vec_info): Add
7742 simd_lane_access_p field.
7743 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7744 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7745 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7746 (omp_clause_code_name): Same.
7747 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7748 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7749 * tree.def (OMP_SIMD): New entry.
7750 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7751 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7752 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7753 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7754 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7755 (OMP_CLAUSE_LINEAR_STEP): New.
7756 (OMP_CLAUSE_SAFELEN_EXPR): New.
7757 (OMP_CLAUSE__SIMDUID__DECL): New.
7758 (find_omp_clause): New prototype.
7760 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7762 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7763 Haswell processor detection.
7765 2013-08-27 Christian Widmer <shadow@umbrox.de>
7768 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7769 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7770 AVX2 capable processors.
7772 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7774 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7775 of vget_low_* with implementations in terms of other intrinsics.
7777 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7780 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7781 values to -1, 0 and 1.
7783 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7785 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7786 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7787 (<optab><mode>3_insn): Remove.
7788 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7789 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7790 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7791 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7792 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7793 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7794 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7795 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7796 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7797 (ror<mode>3_insn): Likewise.
7798 * config/aarch64/predicates.md (aarch64_simd_register): New.
7800 2013-08-27 Richard Biener <rguenther@suse.de>
7802 PR tree-optimization/57521
7803 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7804 one edge is non-critical.
7805 (find_phi_replacement_condition): Make sure to use a non-critical
7806 edge. Cleanup and remove old bug workarounds.
7807 (bb_postdominates_preds): Remove.
7808 (if_convertible_loop_p_1): Do not compute post-dominators.
7809 (combine_blocks): Do not free post-dominators.
7810 (main_tree_if_conversion): Likewise.
7811 (pass_data_if_conversion): Add TODO_verify_ssa.
7813 2013-08-27 DJ Delorie <dj@redhat.com>
7815 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7817 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7819 * function.c (assign_parm_find_data_types): Set passed_mode and
7820 nominal_mode to the TYPE_MODE of nominal_type for the built
7821 pointer type in case of the struct-pass-by-reference.
7823 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7825 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7826 (UINT16_TYPE): Change default to "unsigned int".
7828 * config/avr/avr.opt (mfract-convert-truncate): New option.
7829 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7830 is set, round negative fractional integers according to n1169
7831 when converting to integer types.
7833 2013-08-26 Jan Hubicka <jh@suse.cz>
7835 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7836 methods can not be called indirectly when their address is not taken.
7838 2013-08-26 Jan Hubicka <jh@suse.cz>
7840 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7843 2013-08-26 Jan Hubicka <jh@suse.cz>
7845 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7848 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7850 * reload.c (find_valid_class): Allow classes that do not include
7851 FIRST_PSEUDO_REGISTER - 1.
7853 2013-08-26 Jan Hubicka <jh@suse.cz>
7855 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7856 fix edge count/frequency when speculation failed; fix type check
7857 for the direct call.
7859 2013-08-26 Jan Hubicka <jh@suse.cz>
7861 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7863 2013-08-26 Jan Hubicka <jh@suse.cz>
7865 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7866 to be changed by edge redirection.
7868 2013-08-26 Jan Hubicka <jh@suse.cz>
7870 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7871 formating; add sanity check.
7872 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7873 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7874 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7875 (resolve_noninline_speculation): Update callee keys, too.
7877 2013-08-26 Jan Hubicka <jh@suse.cz>
7879 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7880 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7882 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7884 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7887 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
7889 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7890 (x86_64_elf_select_section): Ditto.
7891 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7893 (ix86_pass_by_reference): Ditto.
7894 (output_set_got): Ditto.
7895 (ix86_unary_operator_ok): Ditto.
7896 (ix86_expand_builtin): Ditto.
7898 2013-08-23 Jan Hubicka <jh@suse.cz>
7900 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
7902 2013-08-23 Jan Hubicka <jh@suse.cz>
7904 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7905 (tree_decl_with_vis): Add FINAL field.
7907 2013-08-23 Jeff Law <law@redhat.com>
7909 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7910 the debugging dump when the expression is fully redundant.
7912 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7914 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7915 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7916 * pretty-print.c (pp_formatted_text_data): Likewise.
7917 (pp_write_text_to_stream): Likewise.
7918 (pp_write_text_as_dot_label_to_stream): Likewise.
7919 (pp_append_r): Likewise.
7920 (pp_format): Likewise.
7921 (pp_flush): Likewise.
7922 (pp_clear_output_area): Likewise.
7923 (pp_append_text): Likewise.
7924 (pp_formatted_text): Likewise.
7925 (pp_remaining_character_count_for_line): Likewise.
7926 (pp_newline): Likewise.
7927 (pp_character): Likewise.
7928 (output_buffer::~output_buffer): Define.
7929 (pretty_printer::~pretty_printer): Destruct output buffer.
7930 * pretty-print.h (output_buffer::~output_buffer): Declare.
7931 (pretty_printer::~pretty_printer): Declare virtual.
7933 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
7936 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7937 HOST_WIDE_INT_M1U): New macros.
7938 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7939 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7940 unsigned -1 for lshift.
7941 * cse.c (cse_insn): Likewise.
7942 * double-int.c (rshift_double, lshift_double): Likewise.
7943 * builtins.c (fold_builtin_bitop): Likewise.
7944 * combine.c (force_to_mode): Likewise.
7945 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7946 * simplify-rtx.c (simplify_const_unary_operation,
7947 simplify_const_binary_operation): Likewise.
7948 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7949 check_va_list_escapes): Likewise.
7950 * rtlanal.c (nonzero_bits1): Likewise.
7951 * expmed.c (expand_smod_pow2): Likewise.
7952 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7954 2013-08-23 Jan Hubicka <jh@suse.cz>
7956 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7957 as having address taken.
7959 2013-08-23 Jan Hubicka <jh@suse.cz>
7961 * ipa-utils.h (method_class_type): Declare.
7962 * ipa-devirt.c (method_class_type): Export.
7964 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7965 do not walk base classes of anonymous types.
7967 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7969 PR rtl-optimization/58220
7971 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7972 handle SEQUENCE insns properly.
7974 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7976 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7978 (pp_newline_and_indent): Likewise.
7979 (pp_separate_with): Likewise.
7980 * pretty-print.c (pp_newline_and_flush): Define.
7981 (pp_newline_and_indent): Likewise.
7982 (pp_separate_with): Likewise.
7984 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7987 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7988 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7990 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7992 * config/i386/predicates.md (ext_sse_reg_operand): New.
7993 * config/i386/i386.md (*movti_internal): Use
7994 predicate to determine if EVEX is needed.
7995 (*movsi_internal): Ditto.
7996 (*movdf_internal): Ditto.
7997 (*movsf_internal): Ditto.
7998 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8000 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8002 PR tree-optimization/58209
8003 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8004 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8005 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8006 emit POINTER_PLUS_EXPR.
8007 (create_tailcall_accumulator): For pointer result type accumulate in
8010 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8012 * configure.ac: Add backslashes missing from the last change.
8013 * configure: Regenerate.
8015 2013-08-22 Jan Hubicka <jh@susue.cz>
8017 * ipa.c (function_and_variable_visibility): First remember function
8018 was global and then make it local.
8020 2013-08-22 Julian Brown <julian@codesourcery.com>
8022 * configure.ac: Add aarch64 to list of arches which use "nop" in
8024 * configure: Regenerate.
8026 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8028 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8029 gnu_libc_has_function.
8030 * config/s390/tpf.h: Likewise.
8032 2013-08-22 Jan Hubicka <jh@susue.cz>
8034 * timevar.c (validate_phases): Add cast.
8036 2013-08-22 Jan Hubicka <jh@susue.cz>
8038 * timevar.c (validate_phases): Use size_t for memory.
8039 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8041 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8043 * pretty-print.h (output_buffer::output_buffer): Declare.
8044 (pretty_printer::pretty_printer): Likewise.
8045 (pp_construct): Remove.
8046 * pretty-print.c (output_buffer::output_buffer): Define.
8047 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8048 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8049 (print_gimple_expr): Likewise.
8050 (print_gimple_seq): Likewise.
8051 (gimple_dump_bb): Likewise.
8052 * sched-vis.c (dump_value_slim): Likewise.
8053 (dump_insn_slim): Likewise.
8054 (dump_rtl_slim): Likewise.
8055 (str_pattern_slim): Likewise.
8056 * tree-mudflap.c (mf_varname_tree): Likewise.
8057 * graph.c (print_graph_cfg): Likewise.
8058 (start_graph_dump): Likewise.
8059 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8061 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8062 pretty printer initialization.
8063 * coretypes.h (diagnostic_context): Remove superflous type alias
8065 (pretty_printer): Likewise. Declare directly as a class.
8066 (pretty_print_info): Remove declaration as class.
8067 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8068 and pp_clear_output_area.
8069 (asan_add_global): Likewise.
8071 2013-08-22 Jan Hubicka <jh@suse.cz>
8073 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8074 * ipa-utils.h (update_type_inheritance_graph): Declare.
8075 (possible_polymorphic_call_target_p): Declare.
8076 (possible_polymorphic_call_target_p): New.
8077 * ipa-devirt.c: Update toplevel comments.
8078 (cached_polymorphic_call_targets): Move up.
8079 (odr_type_d): Move ID down.
8080 (polymorphic_type_binfo_p): Update comment.
8081 (odr_hasher::remove): Likewise;
8082 (get_odr_type): Set anonymous_namespace.
8083 (dump_odr_type): Dump it.
8084 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8085 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8086 (record_binfo): Add comment.
8087 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8089 (devirt_node_removal_hook): Do not iCE when cache is freed.
8090 (possible_polymorphic_call_target_p): New predicate.
8091 (update_type_inheritance_graph): New function.
8093 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8094 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8095 Sergey Lega <sergey.s.lega@intel.com>
8096 Anna Tikhonova <anna.tikhonova@intel.com>
8097 Ilya Tocar <ilya.tocar@intel.com>
8098 Andrey Turetskiy <andrey.turetskiy@intel.com>
8099 Ilya Verbin <ilya.verbin@intel.com>
8100 Kirill Yukhin <kirill.yukhin@intel.com>
8101 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8103 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8104 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8105 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8106 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8107 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8108 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8109 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8110 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8111 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8112 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8113 OPT_mavx512pf, OPT_mavx512er cases.
8114 * config/i386/constraints.md (v): New constraint.
8115 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8116 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8117 (bit_AVX512CD): New.
8118 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8119 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8120 * config/i386/i386-c.c (ix86_target_macros_internal):
8121 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8123 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8124 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8125 * config/i386/i386.c (regclass_map, dbx_register_map)
8126 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8127 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8128 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8130 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8131 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8132 -mavx512pf options. Fix formatting.
8133 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8134 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8135 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8136 -mavx512cd, -mavx512pf options.
8137 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8138 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8139 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8141 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8142 (ix86_set_reg_reg_cost): Ditto.
8143 (x86_order_regs_for_local_alloc): Ditto.
8144 (MAX_VECT_LEN): Extend to 64-byte.
8145 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8146 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8147 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8148 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8149 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8150 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8151 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8152 (SSE_REG_MODE_P): Support new modes.
8153 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8154 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8155 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8156 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8157 (REG_CLASS_CONTENTS): Add new registers.
8158 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8159 (EXT_REX_SSE_REGNO_P): New.
8160 (HI_REGISTER_NAMES): Add new registers.
8161 * config/i386/i386.md: Define constants for new registers.
8162 (mode): Add new 512-bit modes.
8163 (prefix): Support evex prefix.
8164 (isa): Support avx512f, noavx512f, fma_avx512f.
8165 (ssemodesuffix): Add new 512-bit modes.
8167 (*movxi_internal_avx512f): Ditto.
8168 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8170 (*movsi_internal): Likewise.
8171 (*movdf_internal): Likewise.
8172 (*movsf_internal): Likewise.
8173 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8174 (<code><mode>3): Likewise.
8175 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8177 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8178 with the new constraint "v".
8179 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8181 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8182 with the new constraint "v".
8183 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8184 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8185 (*<plusminus_insn><mode>3): Likewise.
8186 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8187 (*mul<mode>3): Likewise.
8188 (<sse>_vmmul<mode>3): Likewise.
8189 (<sse>_div<mode>3): Likewise.
8190 (<sse>_vmdiv<mode>3): Likewise.
8191 (<sse>_sqrt<mode>2): Likewise.
8192 (<sse>_vmsqrt<mode>2): Likewise.
8193 (*<code><mode>3_finite): Likewise.
8194 (*<code><mode>3) <smaxmin>: Likewise.
8195 (<sse>_vm<code><mode>3): Likewise.
8196 (*<code><mode>3) <any_logic>: Likewise.
8197 (*fma_fmadd_<mode>): Likewise.
8198 (*fma_fmsub_<mode>): Likewise.
8199 (*fma_fnmadd_<mode>): Likewise.
8200 (*fma_fnmsub_<mode>): Likewise.
8201 (*fma_fmaddsub_<mode>): Likewise.
8202 (*fma_fmsubadd_<mode>): Likewise.
8203 (*fmai_fmadd_<mode>): Likewise.
8204 (*fmai_fmsub_<mode>): Likewise.
8205 (*fmai_fnmadd_<mode>): Likewise.
8206 (*fmai_fnmsub_<mode>): Likewise.
8207 (sse_cvtsi2ss): Likewise.
8208 (sse_cvtsi2ssq): Likewise.
8209 (sse_cvtss2si): Likewise.
8210 (sse_cvtss2si_2): Likewise.
8211 (sse_cvtss2siq): Likewise.
8212 (sse_cvtss2siq_2): Likewise.
8213 (sse_cvttss2si): Likewise.
8214 (sse_cvtss2siq_2): Likewise.
8215 (float<sseintvecmodelower><mode>2): Likewise.
8216 (sse2_cvtsd2si_2): Likewise.
8217 (sse2_cvtsd2siq_2): Likewise.
8218 (*<plusminus_insn><mode>3): Likewise.
8219 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8220 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8221 (ashr<mode>3): Likewise.
8222 (<shift_insn><mode>3): Likewise.
8223 (avx2_<code><mode>3): Likewise.
8224 (*avx2_<code><mode>3): Likewise.
8225 (*andnot<mode>3): Likewise.
8226 (*<code><mode>3) <any_logic>: Likewise.
8227 (abs<mode>2): Likewise.
8228 (avx2_permvar<mode>): Likewise.
8229 (avx2_perm<mode>_1): Likewise.
8230 (*avx_vpermilp<mode>): Likewise.
8231 (avx_vpermilvar<mode>3): Likewise.
8232 (avx2_ashrv<mode>): Likewise.
8233 (avx2_<shift_insn>v<mode>): Likewise.
8234 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8236 * doc/rtl.texi: Document XImode.
8238 2013-08-21 Jeff Law <law@redhat.com>
8240 * tree-flow.h (register_jump_thread): Pass vector of edges
8241 instead of each important edge.
8242 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8243 thread path into a vector and pass that to register_jump_thread.
8244 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8245 passed in edge vector to the current 3-edge form.
8248 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8250 * dce.c (fini_dce): Call df_analyze again just in case
8251 delete_unmarked_insns removed anything.
8253 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8255 * reload.h (struct reg_equivs): Rename to ..
8256 (struct reg_equivs_s): .. this.
8258 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8260 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8262 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8264 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8266 2013-08-21 Jeff Law <law@redhat.com>
8268 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8269 simplify assignments too. If the RHS collapses to a singleton
8270 range, then return the value for the range.
8272 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8274 * config/i386/sse.md (V16): Rename to...
8276 (mov<mode>): Update iterator name.
8277 (*mov<mode>_internal): Ditto.
8278 (push<mode>1): Ditto.
8279 (movmisalign<mode>): Ditto.
8281 2013-08-20 Jan Hubicka <jh@suse.cz>
8284 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8285 entry for direct edges.
8286 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8288 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8290 Revert my last two changes, r201865 and r201864:
8293 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8295 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8296 instances can own GC refs.
8298 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8299 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8300 (gcc::context::gt_pch_nx): Likewise.
8301 (gcc::context::gt_pch_nx): Likewise.
8302 * ggc.h (gt_ggc_mx <T>): New.
8303 (gt_pch_nx_with_op <T>): New.
8304 (gt_pch_nx <T>): New.
8305 * passes.c (opt_pass::gt_ggc_mx): New.
8306 (opt_pass::gt_pch_nx): New.
8307 (opt_pass::gt_pch_nx_with_op): New.
8308 (pass_manager::gt_ggc_mx): New.
8309 (pass_manager::gt_pch_nx): New.
8310 (pass_manager::gt_pch_nx_with_op): New.
8311 (pass_manager::operator new): Use
8312 ggc_internal_cleared_alloc_stat rather than xcalloc.
8313 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8314 (pass_manager::gt_ggc_mx): New.
8315 (pass_manager::gt_pch_nx): New.
8316 (pass_manager::gt_pch_nx_with_op): New.
8317 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8318 (opt_pass::operator new): New.
8319 (opt_pass::gt_ggc_mx): New.
8320 (opt_pass::gt_pch_nx): New.
8321 (opt_pass::gt_pch_nx_with_op): New.
8324 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8326 * Makefile.in (GTFILES): Add context.h.
8327 * context.c (gcc::context::operator new): New.
8328 (gcc::context::gt_ggc_mx): New.
8329 (gcc::context::gt_pch_nx): New.
8330 (gcc::context::gt_pch_nx): New.
8331 * context.h (gcc::context): Add GTY((user)) marking.
8332 (gcc::context::operator new): New.
8333 (gcc::context::gt_ggc_mx): New.
8334 (gcc::context::gt_pch_nx): New.
8335 (gcc::context::gt_pch_nx): New.
8336 (g): Add GTY marking.
8337 (gt_ggc_mx (gcc::context *)): New.
8338 (gt_pch_nx (gcc::context *)): New.
8339 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8340 void *cookie)): New.
8341 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8343 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8345 * dce.c (fini_dce): Call df_analyze again just in case
8346 delete_unmarked_insns removed anything.
8348 2013-08-20 Teresa Johnson <tejohnson@google.com>
8350 PR rtl-optimizations/57451
8351 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8352 from crossing split section boundaries.
8354 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8356 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8357 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8358 with MULTLIB_DEFAULTS.
8360 2013-08-20 Nick Clifton <nickc@redhat.com>
8362 * target.def (narrow_volatile_bitfield): Note that the default
8363 value is false, not !TARGET_STRICT_ALIGN.
8364 * doc/tm.texi: Regenerate.
8366 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8368 Fix LIB_SPEC for systems without libpthread.
8370 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8371 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8373 * config/i386/linux-common.h: Likewise.
8374 * config/mips/linux-common.h: Likewise.
8376 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8378 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8381 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8383 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8384 instances can own GC refs.
8386 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8387 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8388 (gcc::context::gt_pch_nx): Likewise.
8389 (gcc::context::gt_pch_nx): Likewise.
8390 * ggc.h (gt_ggc_mx <T>): New.
8391 (gt_pch_nx_with_op <T>): New.
8392 (gt_pch_nx <T>): New.
8393 * passes.c (opt_pass::gt_ggc_mx): New.
8394 (opt_pass::gt_pch_nx): New.
8395 (opt_pass::gt_pch_nx_with_op): New.
8396 (pass_manager::gt_ggc_mx): New.
8397 (pass_manager::gt_pch_nx): New.
8398 (pass_manager::gt_pch_nx_with_op): New.
8399 (pass_manager::operator new): Use
8400 ggc_internal_cleared_alloc_stat rather than xcalloc.
8401 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8402 (pass_manager::gt_ggc_mx): New.
8403 (pass_manager::gt_pch_nx): New.
8404 (pass_manager::gt_pch_nx_with_op): New.
8405 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8406 (opt_pass::operator new): New.
8407 (opt_pass::gt_ggc_mx): New.
8408 (opt_pass::gt_pch_nx): New.
8409 (opt_pass::gt_pch_nx_with_op): New.
8411 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8413 * Makefile.in (GTFILES): Add context.h.
8414 * context.c (gcc::context::operator new): New.
8415 (gcc::context::gt_ggc_mx): New.
8416 (gcc::context::gt_pch_nx): New.
8417 (gcc::context::gt_pch_nx): New.
8418 * context.h (gcc::context): Add GTY((user)) marking.
8419 (gcc::context::operator new): New.
8420 (gcc::context::gt_ggc_mx): New.
8421 (gcc::context::gt_pch_nx): New.
8422 (gcc::context::gt_pch_nx): New.
8423 (g): Add GTY marking.
8424 (gt_ggc_mx (gcc::context *)): New.
8425 (gt_pch_nx (gcc::context *)): New.
8426 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8427 void *cookie)): New.
8428 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8430 2013-08-20 Alan Modra <amodra@gmail.com>
8433 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8434 (rs6000_emit_epilogue): Likewise.
8436 2013-08-19 Dehao Chen <dehao@google.com>
8438 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8440 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8441 Jakub Jelinek <jakub@redhat.com>
8443 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8444 (BUILT_IN_FABSD64): Likewise.
8445 (BUILT_IN_FABSD128): Likewise.
8446 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8447 (fold_builtin_1): Likewise.
8448 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8449 destination and source operands.
8450 (*abstd2_fpr): Likewise.
8451 (*nabstd2_fpr): Likewise.
8453 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8455 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8458 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8460 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8463 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8465 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8466 * builtins.c (default_libc_has_function): New.
8467 (gnu_libc_has_function): Ditto.
8468 (no_c99_libc_has_function): Ditto.
8469 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8470 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8471 (fold_builtin_sincos): Likewise.
8472 (fold_builtin_cexp): Likewise.
8473 * builtins.def (DEF_C94_BUILTIN): Likewise.
8474 (DEF_C99_BUILTIN): Likewise.
8475 (DEF_C99_C90RES_BUILTIN): Likewise.
8476 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8477 definitions to using this define.
8478 * config/darwin-protos.h (darwin_libc_has_function): New.
8479 * config/darwin.c (darwin_libc_has_function): Ditto.
8480 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8481 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8482 * config/darwin.h: Ditto.
8483 * config/elfos.h: Ditto.
8484 * config/freebsd.h: Ditto.
8485 * config/i386/cygming.h: Ditto.
8486 * config/i386/djgpp.h: Ditto.
8487 * config/i386/i386-interix.h: Ditto.
8488 * config/microblaze/microblaze.h: Ditto.
8489 * config/mmix/mmix.h: Ditto.
8490 * config/gnu-user.h: Ditto.
8491 * config/ia64/hpux.h: Ditto.
8492 * config/pa/pa-hpux.h: Ditto.
8493 * config/pdp11/pdp11.h: Ditto.
8494 * config/picochip/picochip.h: Ditto.
8495 * config/linux.h: Ditto.
8496 * config/netbsd.h: Ditto.
8497 * config/openbsd.h: Ditto.
8498 * config/rs6000/aix43.h: Ditto.
8499 * config/rs6000/aix51.h: Ditto.
8500 * config/rs6000/aix52.h: Ditto.
8501 * config/rs6000/aix53.h: Ditto.
8502 * config/rs6000/aix61.h: Ditto.
8503 * config/rs6000/darwin.h: Ditto.
8504 * config/rs6000/linux.h: Ditto.
8505 * config/rs6000/linux64.h: Ditto.
8506 * config/s390/tpf.h: Ditto.
8507 * config/sol2-10.h: Ditto.
8508 * config/sol2.h: Ditto.
8509 * config/vms/vms.h: Ditto.
8510 * config/vxworks.h: Ditto.
8511 * config/linux-android.c (linux_android_libc_has_function):
8512 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8513 * config/linux-protos.h (linux_android_libc_has_function):
8515 * config/i386/i386.c (ix86_libc_has_function): New.
8516 * config/i386/i386-protos.h
8517 (ix86_libc_has_function): New declaration.
8518 * config/i386/i386.md
8519 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8520 ("isinf<mode>2): Likewise.
8521 * convert.c (convert_to_integer): Using new target hook
8522 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8523 TARGET_C99_FUNCTIONS.
8524 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8525 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8526 * coretypes.h (function_class): New enum for different
8527 classes of functions.
8528 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8529 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8530 (TARGET_HAS_SINCOS): Likewise.
8531 (TARGET_LIBC_HAS_FUNCTION): New.
8532 * doc/tm.texi: Regenerated.
8533 * targhooks.h (default_libc_has_function): New declaration.
8534 (no_c99_libc_has_function): Ditto.
8535 (gnu_libc_has_function): Ditto.
8536 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8537 and TARGET_HAS_SINCOS.
8539 2013-08-18 Jan Hubicka <jh@suse.cz>
8541 * Makeifle-in (ipa-devirt.o): New.
8542 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8543 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8544 (analyze_functions): Look into possible targets of polymorphic call.
8545 * dumpfile.c (dump_files): Add type-inheritance dump.
8546 * dumpfile.h (TDI_inheritance): New.
8547 * ipa-devirt.c: New file.
8548 * ipa-utils.h (odr_type_d): Forward declare.
8549 (odr_type): New type.
8550 (build_type_inheritance_graph): Declare.
8551 (possible_polymorphic_call_targets): Declare and introduce inline
8552 variant when only edge is pased.
8553 (dump_possible_polymorphic_call_targets): Likewise.
8554 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8555 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8556 (types_same_for_odr): ... here.
8557 * tree.h (type_in_anonymous_namespace_p): Declare.
8559 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8561 PR tree-optimization/58006
8562 * tree-parloops.c (take_address_of): Don't ICE if get_name
8564 (eliminate_local_variables_stmt): Remove clobber stmts.
8566 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8568 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8569 error message is issued for an alias to undefined symbol.
8571 2013-08-18 Jan Hubicka <jh@suse.cz>
8573 * cgraph.c (cgraph_create_indirect_edge): Discover
8574 polymorphic calls and record basic info into indirect_info.
8575 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8576 devirtualization, ignore objc function calls.
8577 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8578 call with no parm index info.
8579 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8580 * tree.c (virtual_method_call_p): New function.
8581 * tree.h (virtual_method_call_p): Declare.
8583 2013-08-16 Jan Hubicka <jh@suse.cz>
8586 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8588 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8590 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8593 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8595 * gengtype.c (type_for_name): Add special-case support for
8596 locating types within the "gcc::" namespace.
8597 (open_base_files): Emit a "using namespace gcc" directive.
8599 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8602 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8603 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8605 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8606 array instead of each individual operand as a separate argument.
8607 (emit_fusion_gpr_load): Likewise.
8608 (expand_fusion_gpr_load): Add new function declaration.
8610 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8611 signature to have the operands passed as an array, instead of as
8612 separate arguments. Allow ZERO_EXTEND to be in the memory
8613 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8614 depend on the register live/dead flags when peepholes are run.
8615 (expand_fusion_gpr_load): New function to be called from the
8616 peephole2 pass, to change the register that addis sets to be the
8618 (emit_fusion_gpr_load): Change the calling signature to have the
8619 operands passed as an array, instead of as separate arguments.
8620 Allow ZERO_EXTEND to be in the memory address, and also
8621 SIGN_EXTEND if -mpower8-fusion-sign.
8623 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8625 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8626 adjust the register addis loads up in the peephole2 pass. Do not
8627 depend on the register live/dead state when the peephole pass is done.
8629 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8631 * gengtype.c (create_user_defined_type): Ensure that the kind
8632 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8633 declaration is seen before the GTY((user)) marking.
8635 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8638 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8640 2013-08-16 Jan Hubicka <jh@suse.cz>
8642 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8643 arugment expected_type.
8644 (gimple_fold_call): Use it.
8645 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8646 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8647 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8648 (try_make_edge_direct_virtual_call): Likewise.
8649 * tree.c (obj_type_ref_class): New.
8650 * tree.h (obj_type_ref_class): Use it.
8652 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8654 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8655 (rtl_slim_pp): Likewise.
8656 (init_rtl_slim_pretty_print): Likewise.
8657 (dump_value_slim): Don't call it. Use local pretty printer.
8658 (dump_insn_slim): Likewise.
8659 (dump_rtl_slim): Likewise.
8660 (str_pattern_slim): Likewise.
8661 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8664 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8666 PR tree-optimization/58164
8667 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8668 walk gimple_goto_dest of GIMPLE_GOTO.
8670 PR tree-optimization/58165
8671 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8672 bi_call must be the last stmt in a bb, don't split_block, instead
8673 use fallthru edge from it and give up if there is none.
8674 Release conds vector when returning early.
8676 2013-08-14 Xinliang David Li <davidxl@google.com>
8678 * config/i386/i386.c (ix86_option_override_internal):
8679 Remove unused variable and field.
8681 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8684 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8685 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8686 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8687 and Linux, correct BLKmode alignment when 128-bit alignment is
8688 required and compatibility flag is not set.
8689 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8690 for zero-size arguments when compatibility flag is not set.
8692 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8694 PR tree-optimization/58145
8695 * tree-sra.c (build_ref_for_offset): If prev_base has
8696 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8698 2013-08-14 Xinliang David Li <davidxl@google.com>
8700 * config/i386/i386.c (ix86_option_override_internal):
8701 Fix uninitialized variable error.
8703 2013-08-14 Xinliang David Li <davidxl@google.com>
8705 * config/i386/i386.opt: Define two new options.
8706 * config/i386/x86-tune.def: Add arch selector field in macros.
8707 * config/i386/i386.h: Adjust macro definition.
8708 * config/i386/i386.c (ix86_option_override_internal):
8710 (parse_mtune_ctrl_str): New function.
8711 (set_ix86_tune_features): New function.
8712 (ix86_function_specific_restore): Call the new helper function.
8714 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8716 PR rtl-optimization/57662
8717 * sel-sched.c (code_motion_process_successors): When the current insn
8718 is removed after the recursive traversal, break from the loop.
8719 Add comments and debug printouts.
8721 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8722 Alexandre Oliva <aoliva@redhat.com>
8725 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8726 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8727 there also UNSPEC_PLTOFF.
8729 2013-08-14 Marek Polacek <polacek@redhat.com>
8731 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8732 NUM_CONDITIONS bit positions.
8734 2013-08-13 Cary Coutant <ccoutant@google.com>
8736 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8737 (attr_checksum): Hash vector contents instead of pointer.
8738 (attr_checksum_ordered): Likewise.
8740 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8742 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8743 when Pmode != word_mode. Add length_address attribute.
8744 (sse3_monitor_<mode>): Merge from sse3_monitor and
8745 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8746 Pmode != word_mode. Update insn length attribute.
8747 * config/i386/i386.c (ix86_option_override_internal): Update
8748 ix86_gen_monitor selection for merged sse3_monitor insn.
8750 2013-08-13 Julian Brown <julian@codesourcery.com>
8752 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8753 perform invalid legitimization on greater-than-word-size modes for
8756 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8758 * ira.c (setup_class_translate_array): Use aclass instead of cl
8759 for classes not fully covered by allocno classes.
8761 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8763 PR tree-optimization/57661
8764 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8765 * tree-inline.c (tree_function_versioning): Initialize it.
8766 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8767 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8768 that is not being copied.
8771 * asan.c (instrument_strlen_call): Fix typo in comment.
8772 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8774 2013-08-13 Steve Ellcey <sellcey@mips.com>
8776 * config/mips/mips.md (prefetch): Use lw instead of ld on
8777 loongson in 32bit mode.
8779 2013-08-13 Nick Clifton <nickc@redhat.com>
8781 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8783 2013-08-13 Jan Hubicka <jh@suse.cz>
8785 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8786 introduced edge; fix typo in sanity check.
8787 (cgraph_resolve_speculation): Export; improve diagnostic.
8788 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8789 speculation at type mismatch.
8790 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8791 (cgraph_resolve_speculation): Declare.
8792 (symtab_can_be_discarded): New function.
8793 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8794 * ipa-inline-transform.c (speculation_removed): New global var.
8795 (clone_inlined_nodes): See if speculation can be removed.
8796 (inline_call): If speculations was removed, we growths may not match.
8797 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8798 (speculation_useful_p): New function.
8799 (resolve_noninline_speculation): New function.
8800 (inline_small_functions): Resolve useless speculations.
8801 * ipa-inline.h (speculation_useful_p): Declare
8802 * ipa.c (can_replace_by_local_alias): Simplify.
8803 (ipa_profile): Produce speculative calls in non-lto, too;
8804 add simple cost model; produce local aliases.
8806 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8808 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8811 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8813 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8815 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8817 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8818 iterator instead of X87MODEF.
8820 2013-08-12 Perez Read <netfirewall@gmail.com>
8823 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8824 operand 0 for intel asm alternative.
8825 (*movabs<mode>_2): Ditto for operand 1.
8827 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8829 * config/aarch64/arm_none.h
8830 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8832 2013-08-12 Nick Clifton <nickc@redhat.com>
8834 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8836 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8838 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8839 expand for QI/HImode operand to produce more effictive code for
8840 unsigned char(short) --> float(double) conversion.
8842 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8844 * doc/invoke.texi: Mention that -ftls-model does not force the final
8847 2013-08-12 Marek Polacek <polacek@redhat.com>
8848 Marc Glisse <marc.glisse@inria.fr>
8850 PR tree-optimization/57980
8851 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8852 when creating -1 constant.
8854 2013-08-10 Jan Hubicka <jh@suse.cz>
8856 Workaround binutils PR14342.
8857 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8858 (gimple_init_edge_profiler): Likewise.
8859 (gimple_gen_ic_func_profiler): Likewise.
8861 2013-08-09 Jan Hubicka <jh@suse.cz>
8863 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8865 2013-08-09 Xinliang David Li <davidxl@google.com>
8867 * config/i386/stringop.def: New file.
8868 * config/i386/stringop.opt: New file.
8869 * config/i386/i386-opts.h: Include stringopt.def.
8870 * config/i386/i386.opt: Include stringopt.opt.
8871 * config/i386/i386.c (ix86_option_override_internal):
8872 Override default size based stringop inline strategies with options.
8873 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8876 2013-08-09 Jan Hubicka <jh@suse.cz>
8878 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
8880 2013-08-09 Jan Hubicka <jh@suse.cz>
8882 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8884 (dump_cgraph_node): Dump profile-id.
8885 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8886 and common_target_probability.
8887 * lto-cgraph.c (lto_output_edge): Stream common targets.
8888 (lto_output_node): Stream profile ids.
8889 (input_node): Stream profile ids.
8890 (input_edge): Stream common targets.
8891 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8892 * ipa.c: Include value-prof.h
8893 (ipa_profile_generate_summary): Turn indirect call statement histograms
8894 into common targets.
8895 (ipa_profile): Turn common targets into speculative edges.
8897 2013-08-09 Jan Hubicka <jh@suse.cz>
8899 * cgraph.h (cgraph_node): Add profile_id.
8900 * value-prof.c (cgraph_node_map): Turn into pointer_map.
8901 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
8902 (del_node_map): Update.
8903 (find_func_by_funcdef_no): Replace by ...
8904 (find_func_by_profile_id): ... this one.
8905 (gimple_ic_transform): Do not remove useful histograms when
8906 speculation is not done; dump info when indirect call removal
8908 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8909 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8910 (__gcov_indirect_call_profiler_v2): .. this one.
8911 * profile.h (init_node_map): Update.
8912 * coverage.c (coverage_compute_profile_id): New function.
8913 * coverage.h (coverage_compute_profile_id): Declare.
8914 * tree-profile.c (init_ic_make_global_vars): Make
8915 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8916 (gimple_init_edge_profiler): Update prototype of
8917 __gcov_indirect_call_profiler.
8918 (gimple_gen_ic_func_profiler): Simplify.
8919 (tree_profiling): Use init_node_map
8921 2013-08-09 Jan Hubicka <jh@suse.cz>
8923 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8924 non-speculative refs.
8925 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8926 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8927 (cgraph_set_call_stmt): Likewise.
8928 (cgraph_create_edge_1): Fix release checking compilatoin;
8930 (cgraph_free_edge): Free indirect info.
8931 (cgraph_turn_edge_to_speculative): New function.
8932 (cgraph_speculative_call_info): New function.
8933 (cgraph_make_edge_direct): Return direct edge; handle speculation.
8934 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
8935 (dump_cgraph_node): Dump speculation.
8936 (verify_edge_count_and_frequency): Accept speculative edges.
8937 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8938 (verify_cgraph_node): Handle speculation.
8939 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8940 (cgraph_set_call_stmt): Update prototype.
8941 (cgraph_make_edge_direct): Update prototype.
8942 (cgraph_speculative_call_info): Declare.
8943 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8944 to change; update call of ipa_find_references.
8945 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8946 lto_stmt_uid and speculative flags.
8947 (ipa_dump_references): Dump speculation.
8948 (ipa_clone_references): Clone speculative flag.
8949 (ipa_clone_referring): Likewise.
8950 (ipa_clone_ref): New function.
8951 (ipa_find_reference): Look into lto_stmt_uids
8952 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8953 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8954 (ipa_find_reference): Update declaration.
8955 (ipa_clone_ref): Declare.
8956 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8957 stream speculative flag.
8958 (lto_output_ref): Stream statements uids and speculation.
8959 (input_ref): Likewise.
8960 (input_edge): Stream speuclation.
8961 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8962 (cgraph_set_call_stmt_including_clones): Handle speculation.
8963 * ipa-inline.c (heap_edge_removal_hook): New function.
8964 (inline_small_functions): Register it.
8965 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8966 also initialize refs.
8967 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8969 (try_make_edge_direct_simple_call): Likewise.
8970 (try_make_edge_direct_simple_call): Likewise.
8971 (update_indirect_edges_after_inlining): Likewise.
8972 (remove_described_reference): Look proper lto_stmt_uid.
8973 (propagate_controlled_uses): Likewise.
8974 (propagate_controlled_uses): Liekwise.
8975 * tree-inline.c (copy_bb): Copy speculative edges.
8976 (redirect_all_calls): New function.
8977 (copy_cfg_body): Do redirection after loop info is updated.
8978 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
8980 2013-08-09 Jan Hubicka <jh@suse.cz>
8982 * lto-streamer-out.c (output_function): Renumber PHIs.
8983 * lto-streamer-in.c (input_function): Likewise.
8985 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8987 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8988 (get_lane_unsigned): Likewise.
8989 (dup_lane_scalar): Likewise.
8990 (get_lane): enable for VALL.
8991 * config/aarch64/aarch64-simd.md
8992 (aarch64_dup_lane_scalar<mode>): Remove.
8993 (aarch64_get_lane_signed<mode>): Likewise.
8994 (aarch64_get_lane_unsigned<mode>): Likewise.
8995 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8996 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8997 (aarch64_get_lane<mode>): Enable for all vector modes.
8998 (aarch64_get_lanedi): Remove misleading constraints.
8999 * config/aarch64/arm_neon.h
9000 (__aarch64_vget_lane_any): Define.
9001 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9002 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9003 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9004 * config/aarch64/iterators.md (VDQQH): New.
9008 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9010 * configure.ac: Add GAS check for LEON instructions on SPARC.
9011 * configure: Regenerate.
9012 * config.in: Likewise.
9013 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9015 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9016 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9018 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9019 (AS_LEON_FLAG): New macro.
9020 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9021 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9022 Deal with LEON and LEON3 for the memory model.
9023 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9024 (atomic_compare_and_swap<mode>_1): Likewise.
9025 (*atomic_compare_and_swap<mode>_1): Likewise.
9027 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9029 * config/arm/neon.md (vcond): Fix floating-point vector
9030 comparisons against 0.
9032 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9034 * lra-constraints.c (emit_spill_move): Remove assert.
9035 (process_alt_operands): Add more debugging
9036 output. Increase reject for spilling into memory. Decrease
9037 reject for reloading scratch.
9038 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9040 2013-08-08 Steve Ellcey <sellcey@mips.com>
9042 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9043 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9044 micromips incompatible. Add nan2008.
9045 (MULTILIB_DIRNAMES): Add nan2008.
9046 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9047 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9048 and micromips incompatible. Add nan2008.
9049 (MULTILIB_DIRNAMES): Add nan2008.
9050 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9052 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9054 PR rtl-optimization/58079
9055 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9056 simplify_comparison has widened a comparison with an integer.
9058 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9060 * config/arm/neon.md (movmisalign<mode>): Disable when we
9061 don't allow unaligned accesses.
9062 (*movmisalign<mode>_neon_store): Likewise.
9063 (*movmisalign<mode>_neon_load): Likewise.
9064 (*movmisalign<mode>_neon_store): Likewise.
9065 (*movmisalign<mode>_neon_load): Likewise.
9067 2013-08-08 Jan Hubicka <jh@suse.cz>
9069 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9070 (make_pass_rebuild_cgraph_edges): Also clear references.
9071 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9072 * ipa-inline-transform.c (inline_transform): Remove all references
9074 * cgraphunit.c (expand_function): Remove all references after
9076 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9077 (ipa_find_reference): Rewrite to iterator.
9078 (remove_stmt_references): Likewise.
9079 (ipa_clear_stmts_in_references): New function.
9080 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9081 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9083 * ipa-split.c (split_function): Remove references in split function.
9085 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9088 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9089 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9091 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9094 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9095 suggested mode for the assignment isn't compatible with the
9098 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9101 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9103 2013-08-07 Xinliang David Li <davidxl@google.com>
9105 * config/i386/i386.opt: New option -mtune-ctrl=.
9106 * config/i386/x86-tune.def: New file.
9107 * config/i386/i386.h: include x86-tune.def.
9108 * config/i386/i386.c (ix86_option_override_internal):
9109 Parsing -mtune-ctrl= option and set tune features.
9111 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9114 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9115 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9118 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9120 * rtl.h (update_alignments): Declare.
9121 * final.c (grow_label_align): New function extracted from...
9122 (shorten_branches): ...here. Call it.
9123 (update_alignments): New function.
9124 * reorg.c (sibling_labels): New variable.
9125 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9126 the new label along with it onto the sibling_labels vector.
9127 (fill_simple_delay_slots): Adjust call to get_label_before.
9128 (fill_slots_from_thread): Likewise.
9129 (relax_delay_slots): Likewise.
9130 (make_return_insns): Likewise.
9131 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9133 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9135 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9136 document its semantics.
9137 (diagnostic_report_diagnostic): Adjust accordingly.
9139 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9141 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9142 (sparc_option_override): ...and port to new C++ pass API.
9143 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9145 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9147 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9149 2013-08-06 Caroline Tice <cmtice@google.com>
9151 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9152 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9153 * tree-pass.h: Add pass_vtable_verify.
9154 * varasm.c (assemble_variable): Add code to properly set the comdat
9155 section and name for the .vtable_map_vars section.
9156 (assemble_vtyv_preinit_initializer): New function.
9157 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9159 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9160 * vtable-verify.c: New file.
9161 * vtable-verify.h: New file.
9162 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9163 initialiation levels.
9164 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9165 * passes.def: Insert pass_vtable_verify.
9166 * aclocal.m4: Reorder includes.
9167 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9168 -fvtv-counts options.
9169 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9170 as appropriate, if -fvtable-verify=... is used.
9171 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9172 -fvtable-verify=... is used.
9173 * Makefile.in (OBJS): Add vtable-verify.o to list.
9174 (vtable-verify.o): Add new build rule.
9175 (GTFILES): Add vtable-verify.c to list.
9176 * common.opt (fvtable-verify=): New flag.
9177 (vtv_priority): Values for fvtable-verify= flag.
9178 (fvtv-counts): New flag.
9179 (fvtv-debug): New flag.
9180 * tree.h (save_vtable_map_decl): New extern function decl.
9182 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9184 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9185 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9186 (pass_data_rl78_devirt): ...new pass_data instance and...
9187 (make_pass_rl78_devirt): ...new function.
9188 (rl78_asm_file_start): Port pass registration to new C++ API.
9190 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9192 * coretypes.h (rtl_opt_pass): Add.
9193 (gcc::context): Add.
9194 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9195 (epiphany_init): Port to new C++ pass API.
9196 (epiphany_optimize_mode_switching): Likewise.
9197 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9198 (pass_manager::get_pass_mode_switching): New.
9199 (pass_manager::get_pass_peephole2): New.
9200 * mode-switching.c (pass_mode_switching): Add clone method.
9201 * recog.c (pass_peephole2): Add clone method.
9202 (pass_split_all_insns): Add clone method.
9204 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9206 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9207 (mips_option_override): ...here, porting to new C++ API for passes.
9209 2013-08-06 Jan Hubicka <jh@suse.cz>
9211 * cgraph.c (cgraph_get_body): New function based on lto.c
9213 * cgraph.h (cgraph_get_body): Declare.
9214 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9216 * cgraphunit.c (expand_function): Get body prior expanding.
9217 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9218 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9220 * passes.c (do_per_function_toporder): Get body.
9221 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9222 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9224 2013-08-06 Martin Jambor <mjambor@suse.cz>
9227 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9228 is not re-finalized. Rename second parameter to no_collect.
9230 2013-08-06 Martin Jambor <mjambor@suse.cz>
9233 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9234 MEM_REF has proper alignment information.
9236 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9239 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9241 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9242 rtx (*) (rtx, ...) with insn_gen_fn.
9243 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9244 insn_gen_fn::stored_funcptr. Add initializer braces.
9246 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9248 Rewrite how instances of passes are cloned to remove assumptions
9249 about their sizes (thus allowing pass subclasses to have
9250 additional data fields, albeit non-GC-managed ones at this point).
9252 * passes.c (make_pass_instance): Now that passes have clone
9253 methods, rewrite this function to eliminate XNEW and memcpy
9254 calls that used hardcoded sizes. Since this function no longer
9255 creates pass instances, rename it to...
9256 (add_pass_instance): ...this. Document the old way that passes were
9257 numbered and flagged, and rework this function to continue using it.
9258 (next_pass_1): Add an initial_pass argument for use by
9260 (position_pass): When adding multiple instances of a pass, use
9261 the pass's clone method, rather than relying on the XNEW/memcpy
9262 within the former make_pass_instance (now add_pass_instance).
9263 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9264 the initial instance of the current pass within the pass manager.
9266 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9268 This is the automated part of the conversion of passes from C
9269 structs to C++ classes.
9271 Patch autogenerated by refactor_passes.py from
9272 https://github.com/davidmalcolm/gcc-refactoring-scripts
9273 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9275 * asan.c (pass_asan): Convert from a global struct to a subclass of
9276 gimple_opt_pass along with...
9277 (pass_data_asan): ...new pass_data instance and...
9278 (make_pass_asan): ...new function.
9279 (pass_asan_O0): Convert from a global struct to a subclass of
9280 gimple_opt_pass along with...
9281 (pass_data_asan_O0): ...new pass_data instance and...
9282 (make_pass_asan_O0): ...new function.
9283 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9284 subclass of rtl_opt_pass along with...
9285 (pass_data_inc_dec): ...new pass_data instance and...
9286 (make_pass_inc_dec): ...new function.
9287 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9288 a subclass of rtl_opt_pass along with...
9289 (pass_data_reorder_blocks): ...new pass_data instance and...
9290 (make_pass_reorder_blocks): ...new function.
9291 (pass_duplicate_computed_gotos): Convert from a global struct to a
9292 subclass of rtl_opt_pass along with...
9293 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9294 (make_pass_duplicate_computed_gotos): ...new function.
9295 (pass_partition_blocks): Convert from a global struct to a subclass of
9296 rtl_opt_pass along with...
9297 (pass_data_partition_blocks): ...new pass_data instance and...
9298 (make_pass_partition_blocks): ...new function.
9299 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9300 struct to a subclass of rtl_opt_pass along with...
9301 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9303 (make_pass_branch_target_load_optimize1): ...new function.
9304 (pass_branch_target_load_optimize2): Convert from a global struct to a
9305 subclass of rtl_opt_pass along with...
9306 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9308 (make_pass_branch_target_load_optimize2): ...new function.
9309 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9310 of rtl_opt_pass along with...
9311 (pass_data_jump): ...new pass_data instance and...
9312 (make_pass_jump): ...new function.
9313 (pass_jump2): Convert from a global struct to a subclass of
9314 rtl_opt_pass along with...
9315 (pass_data_jump2): ...new pass_data instance and...
9316 (make_pass_jump2): ...new function.
9317 * cfgexpand.c (pass_expand): Convert from a global struct to a
9318 subclass of rtl_opt_pass along with...
9319 (pass_data_expand): ...new pass_data instance and...
9320 (make_pass_expand): ...new function.
9321 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9322 of rtl_opt_pass along with...
9323 (pass_data_free_cfg): ...new pass_data instance and...
9324 (make_pass_free_cfg): ...new function.
9325 (pass_into_cfg_layout_mode): Convert from a global struct to a
9326 subclass of rtl_opt_pass along with...
9327 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9328 (make_pass_into_cfg_layout_mode): ...new function.
9329 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9330 subclass of rtl_opt_pass along with...
9331 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9332 (make_pass_outof_cfg_layout_mode): ...new function.
9333 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9334 struct to a subclass of gimple_opt_pass along with...
9335 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9336 (make_pass_build_cgraph_edges): ...new function.
9337 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9338 subclass of gimple_opt_pass along with...
9339 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9340 (make_pass_rebuild_cgraph_edges): ...new function.
9341 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9342 subclass of gimple_opt_pass along with...
9343 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9345 (make_pass_remove_cgraph_callee_edges): ...new function.
9346 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9347 struct to a subclass of rtl_opt_pass along with...
9348 (pass_data_stack_adjustments): ...new pass_data instance and...
9349 (make_pass_stack_adjustments): ...new function.
9350 * combine.c (pass_combine): Convert from a global struct to a subclass
9351 of rtl_opt_pass along with...
9352 (pass_data_combine): ...new pass_data instance and...
9353 (make_pass_combine): ...new function.
9354 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9355 global struct to a subclass of rtl_opt_pass along with...
9356 (pass_data_compare_elim_after_reload): ...new pass_data instance
9358 (make_pass_compare_elim_after_reload): ...new function.
9359 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9360 of rtl_opt_pass along with...
9361 (pass_data_rtl_cprop): ...new pass_data instance and...
9362 (make_pass_rtl_cprop): ...new function.
9363 * cse.c (pass_cse): Convert from a global struct to a subclass of
9364 rtl_opt_pass along with...
9365 (pass_data_cse): ...new pass_data instance and...
9366 (make_pass_cse): ...new function.
9367 (pass_cse2): Convert from a global struct to a subclass of
9368 rtl_opt_pass along with...
9369 (pass_data_cse2): ...new pass_data instance and...
9370 (make_pass_cse2): ...new function.
9371 (pass_cse_after_global_opts): Convert from a global struct to a
9372 subclass of rtl_opt_pass along with...
9373 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9374 (make_pass_cse_after_global_opts): ...new function.
9375 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9376 of rtl_opt_pass along with...
9377 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9378 (make_pass_ud_rtl_dce): ...new function.
9379 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9380 rtl_opt_pass along with...
9381 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9382 (make_pass_fast_rtl_dce): ...new function.
9383 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9384 a subclass of rtl_opt_pass along with...
9385 (pass_data_df_initialize_opt): ...new pass_data instance and...
9386 (make_pass_df_initialize_opt): ...new function.
9387 (pass_df_initialize_no_opt): Convert from a global struct to a
9388 subclass of rtl_opt_pass along with...
9389 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9390 (make_pass_df_initialize_no_opt): ...new function.
9391 (pass_df_finish): Convert from a global struct to a subclass of
9392 rtl_opt_pass along with...
9393 (pass_data_df_finish): ...new pass_data instance and...
9394 (make_pass_df_finish): ...new function.
9395 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9396 rtl_opt_pass along with...
9397 (pass_data_rtl_dse1): ...new pass_data instance and...
9398 (make_pass_rtl_dse1): ...new function.
9399 (pass_rtl_dse2): Convert from a global struct to a subclass of
9400 rtl_opt_pass along with...
9401 (pass_data_rtl_dse2): ...new pass_data instance and...
9402 (make_pass_rtl_dse2): ...new function.
9403 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9404 subclass of rtl_opt_pass along with...
9405 (pass_data_dwarf2_frame): ...new pass_data instance and...
9406 (make_pass_dwarf2_frame): ...new function.
9407 * except.c (pass_set_nothrow_function_flags): Convert from a global
9408 struct to a subclass of rtl_opt_pass along with...
9409 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9411 (make_pass_set_nothrow_function_flags): ...new function.
9412 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9413 subclass of rtl_opt_pass along with...
9414 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9416 (make_pass_convert_to_eh_region_ranges): ...new function.
9417 * final.c (pass_compute_alignments): Convert from a global struct to a
9418 subclass of rtl_opt_pass along with...
9419 (pass_data_compute_alignments): ...new pass_data instance and...
9420 (make_pass_compute_alignments): ...new function.
9421 (pass_final): Convert from a global struct to a subclass of
9422 rtl_opt_pass along with...
9423 (pass_data_final): ...new pass_data instance and...
9424 (make_pass_final): ...new function.
9425 (pass_shorten_branches): Convert from a global struct to a subclass of
9426 rtl_opt_pass along with...
9427 (pass_data_shorten_branches): ...new pass_data instance and...
9428 (make_pass_shorten_branches): ...new function.
9429 (pass_clean_state): Convert from a global struct to a subclass of
9430 rtl_opt_pass along with...
9431 (pass_data_clean_state): ...new pass_data instance and...
9432 (make_pass_clean_state): ...new function.
9433 * function.c (pass_instantiate_virtual_regs): Convert from a global
9434 struct to a subclass of rtl_opt_pass along with...
9435 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9436 (make_pass_instantiate_virtual_regs): ...new function.
9437 (pass_leaf_regs): Convert from a global struct to a subclass of
9438 rtl_opt_pass along with...
9439 (pass_data_leaf_regs): ...new pass_data instance and...
9440 (make_pass_leaf_regs): ...new function.
9441 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9442 subclass of rtl_opt_pass along with...
9443 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9445 (make_pass_thread_prologue_and_epilogue): ...new function.
9446 (pass_match_asm_constraints): Convert from a global struct to a
9447 subclass of rtl_opt_pass along with...
9448 (pass_data_match_asm_constraints): ...new pass_data instance and...
9449 (make_pass_match_asm_constraints): ...new function.
9450 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9451 subclass of rtl_opt_pass along with...
9452 (pass_data_rtl_fwprop): ...new pass_data instance and...
9453 (make_pass_rtl_fwprop): ...new function.
9454 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9455 rtl_opt_pass along with...
9456 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9457 (make_pass_rtl_fwprop_addr): ...new function.
9458 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9459 rtl_opt_pass along with...
9460 (pass_data_rtl_pre): ...new pass_data instance and...
9461 (make_pass_rtl_pre): ...new function.
9462 (pass_rtl_hoist): Convert from a global struct to a subclass of
9463 rtl_opt_pass along with...
9464 (pass_data_rtl_hoist): ...new pass_data instance and...
9465 (make_pass_rtl_hoist): ...new function.
9466 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9467 subclass of gimple_opt_pass along with...
9468 (pass_data_lower_cf): ...new pass_data instance and...
9469 (make_pass_lower_cf): ...new function.
9470 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9471 from a global struct to a subclass of gimple_opt_pass along with...
9472 (pass_data_strength_reduction): ...new pass_data instance and...
9473 (make_pass_strength_reduction): ...new function.
9474 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9475 of rtl_opt_pass along with...
9476 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9477 (make_pass_rtl_ifcvt): ...new function.
9478 (pass_if_after_combine): Convert from a global struct to a subclass of
9479 rtl_opt_pass along with...
9480 (pass_data_if_after_combine): ...new pass_data instance and...
9481 (make_pass_if_after_combine): ...new function.
9482 (pass_if_after_reload): Convert from a global struct to a subclass of
9483 rtl_opt_pass along with...
9484 (pass_data_if_after_reload): ...new pass_data instance and...
9485 (make_pass_if_after_reload): ...new function.
9486 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9487 a subclass of rtl_opt_pass along with...
9488 (pass_data_initialize_regs): ...new pass_data instance and...
9489 (make_pass_initialize_regs): ...new function.
9490 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9491 of ipa_opt_pass_d along with...
9492 (pass_data_ipa_cp): ...new pass_data instance and...
9493 (make_pass_ipa_cp): ...new function.
9494 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9495 global struct to a subclass of gimple_opt_pass along with...
9496 (pass_data_inline_parameters): ...new pass_data instance and...
9497 (make_pass_inline_parameters): ...new function.
9498 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9499 subclass of gimple_opt_pass along with...
9500 (pass_data_early_inline): ...new pass_data instance and...
9501 (make_pass_early_inline): ...new function.
9502 (pass_ipa_inline): Convert from a global struct to a subclass of
9503 ipa_opt_pass_d along with...
9504 (pass_data_ipa_inline): ...new pass_data instance and...
9505 (make_pass_ipa_inline): ...new function.
9506 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9507 struct to a subclass of gimple_opt_pass along with...
9508 (pass_data_local_pure_const): ...new pass_data instance and...
9509 (make_pass_local_pure_const): ...new function.
9510 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9511 ipa_opt_pass_d along with...
9512 (pass_data_ipa_pure_const): ...new pass_data instance and...
9513 (make_pass_ipa_pure_const): ...new function.
9514 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9515 to a subclass of ipa_opt_pass_d along with...
9516 (pass_data_ipa_reference): ...new pass_data instance and...
9517 (make_pass_ipa_reference): ...new function.
9518 * ipa-split.c (pass_split_functions): Convert from a global struct to
9519 a subclass of gimple_opt_pass along with...
9520 (pass_data_split_functions): ...new pass_data instance and...
9521 (make_pass_split_functions): ...new function.
9522 (pass_feedback_split_functions): Convert from a global struct to a
9523 subclass of gimple_opt_pass along with...
9524 (pass_data_feedback_split_functions): ...new pass_data instance and...
9525 (make_pass_feedback_split_functions): ...new function.
9526 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9527 global struct to a subclass of simple_ipa_opt_pass along with...
9528 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9530 (make_pass_ipa_function_and_variable_visibility): ...new function.
9531 (pass_ipa_free_inline_summary): Convert from a global struct to a
9532 subclass of simple_ipa_opt_pass along with...
9533 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9534 (make_pass_ipa_free_inline_summary): ...new function.
9535 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9536 subclass of ipa_opt_pass_d along with...
9537 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9539 (make_pass_ipa_whole_program_visibility): ...new function.
9540 (pass_ipa_profile): Convert from a global struct to a subclass of
9541 ipa_opt_pass_d along with...
9542 (pass_data_ipa_profile): ...new pass_data instance and...
9543 (make_pass_ipa_profile): ...new function.
9544 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9545 ipa_opt_pass_d along with...
9546 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9547 (make_pass_ipa_cdtor_merge): ...new function.
9548 * ira.c (pass_ira): Convert from a global struct to a subclass of
9549 rtl_opt_pass along with...
9550 (pass_data_ira): ...new pass_data instance and...
9551 (make_pass_ira): ...new function.
9552 (pass_reload): Convert from a global struct to a subclass of
9553 rtl_opt_pass along with...
9554 (pass_data_reload): ...new pass_data instance and...
9555 (make_pass_reload): ...new function.
9556 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9557 subclass of rtl_opt_pass along with...
9558 (pass_data_cleanup_barriers): ...new pass_data instance and...
9559 (make_pass_cleanup_barriers): ...new function.
9560 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9561 of rtl_opt_pass along with...
9562 (pass_data_loop2): ...new pass_data instance and...
9563 (make_pass_loop2): ...new function.
9564 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9565 rtl_opt_pass along with...
9566 (pass_data_rtl_loop_init): ...new pass_data instance and...
9567 (make_pass_rtl_loop_init): ...new function.
9568 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9569 rtl_opt_pass along with...
9570 (pass_data_rtl_loop_done): ...new pass_data instance and...
9571 (make_pass_rtl_loop_done): ...new function.
9572 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9573 subclass of rtl_opt_pass along with...
9574 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9575 (make_pass_rtl_move_loop_invariants): ...new function.
9576 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9577 rtl_opt_pass along with...
9578 (pass_data_rtl_unswitch): ...new pass_data instance and...
9579 (make_pass_rtl_unswitch): ...new function.
9580 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9581 subclass of rtl_opt_pass along with...
9582 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9584 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9585 (pass_rtl_doloop): Convert from a global struct to a subclass of
9586 rtl_opt_pass along with...
9587 (pass_data_rtl_doloop): ...new pass_data instance and...
9588 (make_pass_rtl_doloop): ...new function.
9589 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9590 a subclass of rtl_opt_pass along with...
9591 (pass_data_lower_subreg): ...new pass_data instance and...
9592 (make_pass_lower_subreg): ...new function.
9593 (pass_lower_subreg2): Convert from a global struct to a subclass of
9594 rtl_opt_pass along with...
9595 (pass_data_lower_subreg2): ...new pass_data instance and...
9596 (make_pass_lower_subreg2): ...new function.
9597 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9598 struct to a subclass of ipa_opt_pass_d along with...
9599 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9600 (make_pass_ipa_lto_gimple_out): ...new function.
9601 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9602 of ipa_opt_pass_d along with...
9603 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9604 (make_pass_ipa_lto_finish_out): ...new function.
9605 * mode-switching.c (pass_mode_switching): Convert from a global struct
9606 to a subclass of rtl_opt_pass along with...
9607 (pass_data_mode_switching): ...new pass_data instance and...
9608 (make_pass_mode_switching): ...new function.
9609 * modulo-sched.c (pass_sms): Convert from a global struct to a
9610 subclass of rtl_opt_pass along with...
9611 (pass_data_sms): ...new pass_data instance and...
9612 (make_pass_sms): ...new function.
9613 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9614 subclass of gimple_opt_pass along with...
9615 (pass_data_expand_omp): ...new pass_data instance and...
9616 (make_pass_expand_omp): ...new function.
9617 (pass_lower_omp): Convert from a global struct to a subclass of
9618 gimple_opt_pass along with...
9619 (pass_data_lower_omp): ...new pass_data instance and...
9620 (make_pass_lower_omp): ...new function.
9621 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9622 of gimple_opt_pass along with...
9623 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9624 (make_pass_diagnose_omp_blocks): ...new function.
9625 * passes.c (pass_early_local_passes): Convert from a global struct to
9626 a subclass of simple_ipa_opt_pass along with...
9627 (pass_data_early_local_passes): ...new pass_data instance and...
9628 (make_pass_early_local_passes): ...new function.
9629 (pass_all_early_optimizations): Convert from a global struct to a
9630 subclass of gimple_opt_pass along with...
9631 (pass_data_all_early_optimizations): ...new pass_data instance and...
9632 (make_pass_all_early_optimizations): ...new function.
9633 (pass_all_optimizations): Convert from a global struct to a subclass
9634 of gimple_opt_pass along with...
9635 (pass_data_all_optimizations): ...new pass_data instance and...
9636 (make_pass_all_optimizations): ...new function.
9637 (pass_all_optimizations_g): Convert from a global struct to a subclass
9638 of gimple_opt_pass along with...
9639 (pass_data_all_optimizations_g): ...new pass_data instance and...
9640 (make_pass_all_optimizations_g): ...new function.
9641 (pass_rest_of_compilation): Convert from a global struct to a subclass
9642 of rtl_opt_pass along with...
9643 (pass_data_rest_of_compilation): ...new pass_data instance and...
9644 (make_pass_rest_of_compilation): ...new function.
9645 (pass_postreload): Convert from a global struct to a subclass of
9646 rtl_opt_pass along with...
9647 (pass_data_postreload): ...new pass_data instance and...
9648 (make_pass_postreload): ...new function.
9649 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9650 subclass of rtl_opt_pass along with...
9651 (pass_data_gcse2): ...new pass_data instance and...
9652 (make_pass_gcse2): ...new function.
9653 * postreload.c (pass_postreload_cse): Convert from a global struct to
9654 a subclass of rtl_opt_pass along with...
9655 (pass_data_postreload_cse): ...new pass_data instance and...
9656 (make_pass_postreload_cse): ...new function.
9657 * predict.c (pass_profile): Convert from a global struct to a subclass
9658 of gimple_opt_pass along with...
9659 (pass_data_profile): ...new pass_data instance and...
9660 (make_pass_profile): ...new function.
9661 (pass_strip_predict_hints): Convert from a global struct to a subclass
9662 of gimple_opt_pass along with...
9663 (pass_data_strip_predict_hints): ...new pass_data instance and...
9664 (make_pass_strip_predict_hints): ...new function.
9665 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9666 of rtl_opt_pass along with...
9667 (pass_data_peephole2): ...new pass_data instance and...
9668 (make_pass_peephole2): ...new function.
9669 (pass_split_all_insns): Convert from a global struct to a subclass of
9670 rtl_opt_pass along with...
9671 (pass_data_split_all_insns): ...new pass_data instance and...
9672 (make_pass_split_all_insns): ...new function.
9673 (pass_split_after_reload): Convert from a global struct to a subclass
9674 of rtl_opt_pass along with...
9675 (pass_data_split_after_reload): ...new pass_data instance and...
9676 (make_pass_split_after_reload): ...new function.
9677 (pass_split_before_regstack): Convert from a global struct to a
9678 subclass of rtl_opt_pass along with...
9679 (pass_data_split_before_regstack): ...new pass_data instance and...
9680 (make_pass_split_before_regstack): ...new function.
9681 (pass_split_before_sched2): Convert from a global struct to a subclass
9682 of rtl_opt_pass along with...
9683 (pass_data_split_before_sched2): ...new pass_data instance and...
9684 (make_pass_split_before_sched2): ...new function.
9685 (pass_split_for_shorten_branches): Convert from a global struct to a
9686 subclass of rtl_opt_pass along with...
9687 (pass_data_split_for_shorten_branches): ...new pass_data instance
9689 (make_pass_split_for_shorten_branches): ...new function.
9690 * ree.c (pass_ree): Convert from a global struct to a subclass of
9691 rtl_opt_pass along with...
9692 (pass_data_ree): ...new pass_data instance and...
9693 (make_pass_ree): ...new function.
9694 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9695 subclass of rtl_opt_pass along with...
9696 (pass_data_stack_regs): ...new pass_data instance and...
9697 (make_pass_stack_regs): ...new function.
9698 (pass_stack_regs_run): Convert from a global struct to a subclass of
9699 rtl_opt_pass along with...
9700 (pass_data_stack_regs_run): ...new pass_data instance and...
9701 (make_pass_stack_regs_run): ...new function.
9702 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9703 subclass of rtl_opt_pass along with...
9704 (pass_data_cprop_hardreg): ...new pass_data instance and...
9705 (make_pass_cprop_hardreg): ...new function.
9706 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9707 subclass of rtl_opt_pass along with...
9708 (pass_data_reginfo_init): ...new pass_data instance and...
9709 (make_pass_reginfo_init): ...new function.
9710 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9711 of rtl_opt_pass along with...
9712 (pass_data_regmove): ...new pass_data instance and...
9713 (make_pass_regmove): ...new function.
9714 * regrename.c (pass_regrename): Convert from a global struct to a
9715 subclass of rtl_opt_pass along with...
9716 (pass_data_regrename): ...new pass_data instance and...
9717 (make_pass_regrename): ...new function.
9718 * reorg.c (pass_delay_slots): Convert from a global struct to a
9719 subclass of rtl_opt_pass along with...
9720 (pass_data_delay_slots): ...new pass_data instance and...
9721 (make_pass_delay_slots): ...new function.
9722 (pass_machine_reorg): Convert from a global struct to a subclass of
9723 rtl_opt_pass along with...
9724 (pass_data_machine_reorg): ...new pass_data instance and...
9725 (make_pass_machine_reorg): ...new function.
9726 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9727 of rtl_opt_pass along with...
9728 (pass_data_sched): ...new pass_data instance and...
9729 (make_pass_sched): ...new function.
9730 (pass_sched2): Convert from a global struct to a subclass of
9731 rtl_opt_pass along with...
9732 (pass_data_sched2): ...new pass_data instance and...
9733 (make_pass_sched2): ...new function.
9734 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9735 to a subclass of rtl_opt_pass along with...
9736 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9737 (make_pass_stack_ptr_mod): ...new function.
9738 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9739 to a subclass of rtl_opt_pass along with...
9740 (pass_data_rtl_store_motion): ...new pass_data instance and...
9741 (make_pass_rtl_store_motion): ...new function.
9742 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9743 of gimple_opt_pass along with...
9744 (pass_data_tracer): ...new pass_data instance and...
9745 (make_pass_tracer): ...new function.
9746 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9747 to a subclass of gimple_opt_pass along with...
9748 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9749 (make_pass_diagnose_tm_blocks): ...new function.
9750 (pass_lower_tm): Convert from a global struct to a subclass of
9751 gimple_opt_pass along with...
9752 (pass_data_lower_tm): ...new pass_data instance and...
9753 (make_pass_lower_tm): ...new function.
9754 (pass_tm_init): Convert from a global struct to a subclass of
9755 gimple_opt_pass along with...
9756 (pass_data_tm_init): ...new pass_data instance and...
9757 (make_pass_tm_init): ...new function.
9758 (pass_tm_mark): Convert from a global struct to a subclass of
9759 gimple_opt_pass along with...
9760 (pass_data_tm_mark): ...new pass_data instance and...
9761 (make_pass_tm_mark): ...new function.
9762 (pass_tm_edges): Convert from a global struct to a subclass of
9763 gimple_opt_pass along with...
9764 (pass_data_tm_edges): ...new pass_data instance and...
9765 (make_pass_tm_edges): ...new function.
9766 (pass_tm_memopt): Convert from a global struct to a subclass of
9767 gimple_opt_pass along with...
9768 (pass_data_tm_memopt): ...new pass_data instance and...
9769 (make_pass_tm_memopt): ...new function.
9770 (pass_ipa_tm): Convert from a global struct to a subclass of
9771 simple_ipa_opt_pass along with...
9772 (pass_data_ipa_tm): ...new pass_data instance and...
9773 (make_pass_ipa_tm): ...new function.
9774 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9775 subclass of gimple_opt_pass along with...
9776 (pass_data_call_cdce): ...new pass_data instance and...
9777 (make_pass_call_cdce): ...new function.
9778 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9779 subclass of gimple_opt_pass along with...
9780 (pass_data_build_cfg): ...new pass_data instance and...
9781 (make_pass_build_cfg): ...new function.
9782 (pass_split_crit_edges): Convert from a global struct to a subclass of
9783 gimple_opt_pass along with...
9784 (pass_data_split_crit_edges): ...new pass_data instance and...
9785 (make_pass_split_crit_edges): ...new function.
9786 (pass_warn_function_return): Convert from a global struct to a
9787 subclass of gimple_opt_pass along with...
9788 (pass_data_warn_function_return): ...new pass_data instance and...
9789 (make_pass_warn_function_return): ...new function.
9790 (pass_warn_function_noreturn): Convert from a global struct to a
9791 subclass of gimple_opt_pass along with...
9792 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9793 (make_pass_warn_function_noreturn): ...new function.
9794 (pass_warn_unused_result): Convert from a global struct to a subclass
9795 of gimple_opt_pass along with...
9796 (pass_data_warn_unused_result): ...new pass_data instance and...
9797 (make_pass_warn_unused_result): ...new function.
9798 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9799 a subclass of gimple_opt_pass along with...
9800 (pass_data_merge_phi): ...new pass_data instance and...
9801 (make_pass_merge_phi): ...new function.
9802 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9803 a subclass of gimple_opt_pass along with...
9804 (pass_data_lower_complex): ...new pass_data instance and...
9805 (make_pass_lower_complex): ...new function.
9806 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9807 gimple_opt_pass along with...
9808 (pass_data_lower_complex_O0): ...new pass_data instance and...
9809 (make_pass_lower_complex_O0): ...new function.
9810 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9811 subclass of gimple_opt_pass along with...
9812 (pass_data_lower_eh): ...new pass_data instance and...
9813 (make_pass_lower_eh): ...new function.
9814 (pass_refactor_eh): Convert from a global struct to a subclass of
9815 gimple_opt_pass along with...
9816 (pass_data_refactor_eh): ...new pass_data instance and...
9817 (make_pass_refactor_eh): ...new function.
9818 (pass_lower_resx): Convert from a global struct to a subclass of
9819 gimple_opt_pass along with...
9820 (pass_data_lower_resx): ...new pass_data instance and...
9821 (make_pass_lower_resx): ...new function.
9822 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9823 of gimple_opt_pass along with...
9824 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9825 (make_pass_lower_eh_dispatch): ...new function.
9826 (pass_cleanup_eh): Convert from a global struct to a subclass of
9827 gimple_opt_pass along with...
9828 (pass_data_cleanup_eh): ...new pass_data instance and...
9829 (make_pass_cleanup_eh): ...new function.
9830 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9831 to a subclass of simple_ipa_opt_pass along with...
9832 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9833 (make_pass_ipa_lower_emutls): ...new function.
9834 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9835 a subclass of gimple_opt_pass along with...
9836 (pass_data_if_conversion): ...new pass_data instance and...
9837 (make_pass_if_conversion): ...new function.
9838 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9839 subclass of gimple_opt_pass along with...
9840 (pass_data_build_ssa): ...new pass_data instance and...
9841 (make_pass_build_ssa): ...new function.
9842 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9843 global struct to a subclass of gimple_opt_pass along with...
9844 (pass_data_loop_distribution): ...new pass_data instance and...
9845 (make_pass_loop_distribution): ...new function.
9846 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9847 subclass of gimple_opt_pass along with...
9848 (pass_data_mudflap_1): ...new pass_data instance and...
9849 (make_pass_mudflap_1): ...new function.
9850 (pass_mudflap_2): Convert from a global struct to a subclass of
9851 gimple_opt_pass along with...
9852 (pass_data_mudflap_2): ...new pass_data instance and...
9853 (make_pass_mudflap_2): ...new function.
9854 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9855 subclass of gimple_opt_pass along with...
9856 (pass_data_mudflap_1): ...new pass_data instance and...
9857 (make_pass_mudflap_1): ...new function.
9858 (pass_mudflap_2): Convert from a global struct to a subclass of
9859 gimple_opt_pass along with...
9860 (pass_data_mudflap_2): ...new pass_data instance and...
9861 (make_pass_mudflap_2): ...new function.
9862 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9863 gimple_opt_pass along with...
9864 (pass_data_nrv): ...new pass_data instance and...
9865 (make_pass_nrv): ...new function.
9866 (pass_return_slot): Convert from a global struct to a subclass of
9867 gimple_opt_pass along with...
9868 (pass_data_return_slot): ...new pass_data instance and...
9869 (make_pass_return_slot): ...new function.
9870 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9871 to a subclass of gimple_opt_pass along with...
9872 (pass_data_object_sizes): ...new pass_data instance and...
9873 (make_pass_object_sizes): ...new function.
9874 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9875 global struct to a subclass of gimple_opt_pass along with...
9876 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9878 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9879 (pass_fixup_cfg): Convert from a global struct to a subclass of
9880 gimple_opt_pass along with...
9881 (pass_data_fixup_cfg): ...new pass_data instance and...
9882 (make_pass_fixup_cfg): ...new function.
9883 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9884 (make_pass_mudflap_1): ...new function.
9885 (pass_mudflap_2): Replace declaration with that of...
9886 (make_pass_mudflap_2): ...new function.
9887 (pass_asan): Replace declaration with that of...
9888 (make_pass_asan): ...new function.
9889 (pass_asan_O0): Replace declaration with that of...
9890 (make_pass_asan_O0): ...new function.
9891 (pass_tsan): Replace declaration with that of...
9892 (make_pass_tsan): ...new function.
9893 (pass_tsan_O0): Replace declaration with that of...
9894 (make_pass_tsan_O0): ...new function.
9895 (pass_lower_cf): Replace declaration with that of...
9896 (make_pass_lower_cf): ...new function.
9897 (pass_refactor_eh): Replace declaration with that of...
9898 (make_pass_refactor_eh): ...new function.
9899 (pass_lower_eh): Replace declaration with that of...
9900 (make_pass_lower_eh): ...new function.
9901 (pass_lower_eh_dispatch): Replace declaration with that of...
9902 (make_pass_lower_eh_dispatch): ...new function.
9903 (pass_lower_resx): Replace declaration with that of...
9904 (make_pass_lower_resx): ...new function.
9905 (pass_build_cfg): Replace declaration with that of...
9906 (make_pass_build_cfg): ...new function.
9907 (pass_early_tree_profile): Replace declaration with that of...
9908 (make_pass_early_tree_profile): ...new function.
9909 (pass_cleanup_eh): Replace declaration with that of...
9910 (make_pass_cleanup_eh): ...new function.
9911 (pass_sra): Replace declaration with that of...
9912 (make_pass_sra): ...new function.
9913 (pass_sra_early): Replace declaration with that of...
9914 (make_pass_sra_early): ...new function.
9915 (pass_early_ipa_sra): Replace declaration with that of...
9916 (make_pass_early_ipa_sra): ...new function.
9917 (pass_tail_recursion): Replace declaration with that of...
9918 (make_pass_tail_recursion): ...new function.
9919 (pass_tail_calls): Replace declaration with that of...
9920 (make_pass_tail_calls): ...new function.
9921 (pass_tree_loop): Replace declaration with that of...
9922 (make_pass_tree_loop): ...new function.
9923 (pass_tree_loop_init): Replace declaration with that of...
9924 (make_pass_tree_loop_init): ...new function.
9925 (pass_lim): Replace declaration with that of...
9926 (make_pass_lim): ...new function.
9927 (pass_tree_unswitch): Replace declaration with that of...
9928 (make_pass_tree_unswitch): ...new function.
9929 (pass_predcom): Replace declaration with that of...
9930 (make_pass_predcom): ...new function.
9931 (pass_iv_canon): Replace declaration with that of...
9932 (make_pass_iv_canon): ...new function.
9933 (pass_scev_cprop): Replace declaration with that of...
9934 (make_pass_scev_cprop): ...new function.
9935 (pass_empty_loop): Replace declaration with that of...
9936 (make_pass_empty_loop): ...new function.
9937 (pass_record_bounds): Replace declaration with that of...
9938 (make_pass_record_bounds): ...new function.
9939 (pass_graphite): Replace declaration with that of...
9940 (make_pass_graphite): ...new function.
9941 (pass_graphite_transforms): Replace declaration with that of...
9942 (make_pass_graphite_transforms): ...new function.
9943 (pass_if_conversion): Replace declaration with that of...
9944 (make_pass_if_conversion): ...new function.
9945 (pass_loop_distribution): Replace declaration with that of...
9946 (make_pass_loop_distribution): ...new function.
9947 (pass_vectorize): Replace declaration with that of...
9948 (make_pass_vectorize): ...new function.
9949 (pass_slp_vectorize): Replace declaration with that of...
9950 (make_pass_slp_vectorize): ...new function.
9951 (pass_complete_unroll): Replace declaration with that of...
9952 (make_pass_complete_unroll): ...new function.
9953 (pass_complete_unrolli): Replace declaration with that of...
9954 (make_pass_complete_unrolli): ...new function.
9955 (pass_parallelize_loops): Replace declaration with that of...
9956 (make_pass_parallelize_loops): ...new function.
9957 (pass_loop_prefetch): Replace declaration with that of...
9958 (make_pass_loop_prefetch): ...new function.
9959 (pass_iv_optimize): Replace declaration with that of...
9960 (make_pass_iv_optimize): ...new function.
9961 (pass_tree_loop_done): Replace declaration with that of...
9962 (make_pass_tree_loop_done): ...new function.
9963 (pass_ch): Replace declaration with that of...
9964 (make_pass_ch): ...new function.
9965 (pass_ccp): Replace declaration with that of...
9966 (make_pass_ccp): ...new function.
9967 (pass_phi_only_cprop): Replace declaration with that of...
9968 (make_pass_phi_only_cprop): ...new function.
9969 (pass_build_ssa): Replace declaration with that of...
9970 (make_pass_build_ssa): ...new function.
9971 (pass_build_alias): Replace declaration with that of...
9972 (make_pass_build_alias): ...new function.
9973 (pass_build_ealias): Replace declaration with that of...
9974 (make_pass_build_ealias): ...new function.
9975 (pass_dominator): Replace declaration with that of...
9976 (make_pass_dominator): ...new function.
9977 (pass_dce): Replace declaration with that of...
9978 (make_pass_dce): ...new function.
9979 (pass_dce_loop): Replace declaration with that of...
9980 (make_pass_dce_loop): ...new function.
9981 (pass_cd_dce): Replace declaration with that of...
9982 (make_pass_cd_dce): ...new function.
9983 (pass_call_cdce): Replace declaration with that of...
9984 (make_pass_call_cdce): ...new function.
9985 (pass_merge_phi): Replace declaration with that of...
9986 (make_pass_merge_phi): ...new function.
9987 (pass_split_crit_edges): Replace declaration with that of...
9988 (make_pass_split_crit_edges): ...new function.
9989 (pass_pre): Replace declaration with that of...
9990 (make_pass_pre): ...new function.
9991 (pass_profile): Replace declaration with that of...
9992 (make_pass_profile): ...new function.
9993 (pass_strip_predict_hints): Replace declaration with that of...
9994 (make_pass_strip_predict_hints): ...new function.
9995 (pass_lower_complex_O0): Replace declaration with that of...
9996 (make_pass_lower_complex_O0): ...new function.
9997 (pass_lower_complex): Replace declaration with that of...
9998 (make_pass_lower_complex): ...new function.
9999 (pass_lower_vector): Replace declaration with that of...
10000 (make_pass_lower_vector): ...new function.
10001 (pass_lower_vector_ssa): Replace declaration with that of...
10002 (make_pass_lower_vector_ssa): ...new function.
10003 (pass_lower_omp): Replace declaration with that of...
10004 (make_pass_lower_omp): ...new function.
10005 (pass_diagnose_omp_blocks): Replace declaration with that of...
10006 (make_pass_diagnose_omp_blocks): ...new function.
10007 (pass_expand_omp): Replace declaration with that of...
10008 (make_pass_expand_omp): ...new function.
10009 (pass_expand_omp_ssa): Replace declaration with that of...
10010 (make_pass_expand_omp_ssa): ...new function.
10011 (pass_object_sizes): Replace declaration with that of...
10012 (make_pass_object_sizes): ...new function.
10013 (pass_strlen): Replace declaration with that of...
10014 (make_pass_strlen): ...new function.
10015 (pass_fold_builtins): Replace declaration with that of...
10016 (make_pass_fold_builtins): ...new function.
10017 (pass_stdarg): Replace declaration with that of...
10018 (make_pass_stdarg): ...new function.
10019 (pass_early_warn_uninitialized): Replace declaration with that of...
10020 (make_pass_early_warn_uninitialized): ...new function.
10021 (pass_late_warn_uninitialized): Replace declaration with that of...
10022 (make_pass_late_warn_uninitialized): ...new function.
10023 (pass_cse_reciprocals): Replace declaration with that of...
10024 (make_pass_cse_reciprocals): ...new function.
10025 (pass_cse_sincos): Replace declaration with that of...
10026 (make_pass_cse_sincos): ...new function.
10027 (pass_optimize_bswap): Replace declaration with that of...
10028 (make_pass_optimize_bswap): ...new function.
10029 (pass_optimize_widening_mul): Replace declaration with that of...
10030 (make_pass_optimize_widening_mul): ...new function.
10031 (pass_warn_function_return): Replace declaration with that of...
10032 (make_pass_warn_function_return): ...new function.
10033 (pass_warn_function_noreturn): Replace declaration with that of...
10034 (make_pass_warn_function_noreturn): ...new function.
10035 (pass_cselim): Replace declaration with that of...
10036 (make_pass_cselim): ...new function.
10037 (pass_phiopt): Replace declaration with that of...
10038 (make_pass_phiopt): ...new function.
10039 (pass_forwprop): Replace declaration with that of...
10040 (make_pass_forwprop): ...new function.
10041 (pass_phiprop): Replace declaration with that of...
10042 (make_pass_phiprop): ...new function.
10043 (pass_tree_ifcombine): Replace declaration with that of...
10044 (make_pass_tree_ifcombine): ...new function.
10045 (pass_dse): Replace declaration with that of...
10046 (make_pass_dse): ...new function.
10047 (pass_nrv): Replace declaration with that of...
10048 (make_pass_nrv): ...new function.
10049 (pass_rename_ssa_copies): Replace declaration with that of...
10050 (make_pass_rename_ssa_copies): ...new function.
10051 (pass_sink_code): Replace declaration with that of...
10052 (make_pass_sink_code): ...new function.
10053 (pass_fre): Replace declaration with that of...
10054 (make_pass_fre): ...new function.
10055 (pass_check_data_deps): Replace declaration with that of...
10056 (make_pass_check_data_deps): ...new function.
10057 (pass_copy_prop): Replace declaration with that of...
10058 (make_pass_copy_prop): ...new function.
10059 (pass_vrp): Replace declaration with that of...
10060 (make_pass_vrp): ...new function.
10061 (pass_uncprop): Replace declaration with that of...
10062 (make_pass_uncprop): ...new function.
10063 (pass_return_slot): Replace declaration with that of...
10064 (make_pass_return_slot): ...new function.
10065 (pass_reassoc): Replace declaration with that of...
10066 (make_pass_reassoc): ...new function.
10067 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10068 (make_pass_rebuild_cgraph_edges): ...new function.
10069 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10070 (make_pass_remove_cgraph_callee_edges): ...new function.
10071 (pass_build_cgraph_edges): Replace declaration with that of...
10072 (make_pass_build_cgraph_edges): ...new function.
10073 (pass_local_pure_const): Replace declaration with that of...
10074 (make_pass_local_pure_const): ...new function.
10075 (pass_tracer): Replace declaration with that of...
10076 (make_pass_tracer): ...new function.
10077 (pass_warn_unused_result): Replace declaration with that of...
10078 (make_pass_warn_unused_result): ...new function.
10079 (pass_diagnose_tm_blocks): Replace declaration with that of...
10080 (make_pass_diagnose_tm_blocks): ...new function.
10081 (pass_lower_tm): Replace declaration with that of...
10082 (make_pass_lower_tm): ...new function.
10083 (pass_tm_init): Replace declaration with that of...
10084 (make_pass_tm_init): ...new function.
10085 (pass_tm_mark): Replace declaration with that of...
10086 (make_pass_tm_mark): ...new function.
10087 (pass_tm_memopt): Replace declaration with that of...
10088 (make_pass_tm_memopt): ...new function.
10089 (pass_tm_edges): Replace declaration with that of...
10090 (make_pass_tm_edges): ...new function.
10091 (pass_split_functions): Replace declaration with that of...
10092 (make_pass_split_functions): ...new function.
10093 (pass_feedback_split_functions): Replace declaration with that of...
10094 (make_pass_feedback_split_functions): ...new function.
10095 (pass_strength_reduction): Replace declaration with that of...
10096 (make_pass_strength_reduction): ...new function.
10097 (pass_ipa_lower_emutls): Replace declaration with that of...
10098 (make_pass_ipa_lower_emutls): ...new function.
10099 (pass_ipa_function_and_variable_visibility): Replace declaration with
10101 (make_pass_ipa_function_and_variable_visibility): ...new function.
10102 (pass_ipa_tree_profile): Replace declaration with that of...
10103 (make_pass_ipa_tree_profile): ...new function.
10104 (pass_early_local_passes): Replace declaration with that of...
10105 (make_pass_early_local_passes): ...new function.
10106 (pass_ipa_whole_program_visibility): Replace declaration with that
10108 (make_pass_ipa_whole_program_visibility): ...new function.
10109 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10110 (make_pass_ipa_lto_gimple_out): ...new function.
10111 (pass_ipa_increase_alignment): Replace declaration with that of...
10112 (make_pass_ipa_increase_alignment): ...new function.
10113 (pass_ipa_inline): Replace declaration with that of...
10114 (make_pass_ipa_inline): ...new function.
10115 (pass_ipa_free_lang_data): Replace declaration with that of...
10116 (make_pass_ipa_free_lang_data): ...new function.
10117 (pass_ipa_free_inline_summary): Replace declaration with that of...
10118 (make_pass_ipa_free_inline_summary): ...new function.
10119 (pass_ipa_cp): Replace declaration with that of...
10120 (make_pass_ipa_cp): ...new function.
10121 (pass_ipa_reference): Replace declaration with that of...
10122 (make_pass_ipa_reference): ...new function.
10123 (pass_ipa_pure_const): Replace declaration with that of...
10124 (make_pass_ipa_pure_const): ...new function.
10125 (pass_ipa_pta): Replace declaration with that of...
10126 (make_pass_ipa_pta): ...new function.
10127 (pass_ipa_lto_finish_out): Replace declaration with that of...
10128 (make_pass_ipa_lto_finish_out): ...new function.
10129 (pass_ipa_tm): Replace declaration with that of...
10130 (make_pass_ipa_tm): ...new function.
10131 (pass_ipa_profile): Replace declaration with that of...
10132 (make_pass_ipa_profile): ...new function.
10133 (pass_ipa_cdtor_merge): Replace declaration with that of...
10134 (make_pass_ipa_cdtor_merge): ...new function.
10135 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10137 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10138 (pass_init_datastructures): Replace declaration with that of...
10139 (make_pass_init_datastructures): ...new function.
10140 (pass_fixup_cfg): Replace declaration with that of...
10141 (make_pass_fixup_cfg): ...new function.
10142 (pass_expand): Replace declaration with that of...
10143 (make_pass_expand): ...new function.
10144 (pass_instantiate_virtual_regs): Replace declaration with that of...
10145 (make_pass_instantiate_virtual_regs): ...new function.
10146 (pass_rtl_fwprop): Replace declaration with that of...
10147 (make_pass_rtl_fwprop): ...new function.
10148 (pass_rtl_fwprop_addr): Replace declaration with that of...
10149 (make_pass_rtl_fwprop_addr): ...new function.
10150 (pass_jump): Replace declaration with that of...
10151 (make_pass_jump): ...new function.
10152 (pass_jump2): Replace declaration with that of...
10153 (make_pass_jump2): ...new function.
10154 (pass_lower_subreg): Replace declaration with that of...
10155 (make_pass_lower_subreg): ...new function.
10156 (pass_cse): Replace declaration with that of...
10157 (make_pass_cse): ...new function.
10158 (pass_fast_rtl_dce): Replace declaration with that of...
10159 (make_pass_fast_rtl_dce): ...new function.
10160 (pass_ud_rtl_dce): Replace declaration with that of...
10161 (make_pass_ud_rtl_dce): ...new function.
10162 (pass_rtl_dce): Replace declaration with that of...
10163 (make_pass_rtl_dce): ...new function.
10164 (pass_rtl_dse1): Replace declaration with that of...
10165 (make_pass_rtl_dse1): ...new function.
10166 (pass_rtl_dse2): Replace declaration with that of...
10167 (make_pass_rtl_dse2): ...new function.
10168 (pass_rtl_dse3): Replace declaration with that of...
10169 (make_pass_rtl_dse3): ...new function.
10170 (pass_rtl_cprop): Replace declaration with that of...
10171 (make_pass_rtl_cprop): ...new function.
10172 (pass_rtl_pre): Replace declaration with that of...
10173 (make_pass_rtl_pre): ...new function.
10174 (pass_rtl_hoist): Replace declaration with that of...
10175 (make_pass_rtl_hoist): ...new function.
10176 (pass_rtl_store_motion): Replace declaration with that of...
10177 (make_pass_rtl_store_motion): ...new function.
10178 (pass_cse_after_global_opts): Replace declaration with that of...
10179 (make_pass_cse_after_global_opts): ...new function.
10180 (pass_rtl_ifcvt): Replace declaration with that of...
10181 (make_pass_rtl_ifcvt): ...new function.
10182 (pass_into_cfg_layout_mode): Replace declaration with that of...
10183 (make_pass_into_cfg_layout_mode): ...new function.
10184 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10185 (make_pass_outof_cfg_layout_mode): ...new function.
10186 (pass_loop2): Replace declaration with that of...
10187 (make_pass_loop2): ...new function.
10188 (pass_rtl_loop_init): Replace declaration with that of...
10189 (make_pass_rtl_loop_init): ...new function.
10190 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10191 (make_pass_rtl_move_loop_invariants): ...new function.
10192 (pass_rtl_unswitch): Replace declaration with that of...
10193 (make_pass_rtl_unswitch): ...new function.
10194 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10195 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10196 (pass_rtl_doloop): Replace declaration with that of...
10197 (make_pass_rtl_doloop): ...new function.
10198 (pass_rtl_loop_done): Replace declaration with that of...
10199 (make_pass_rtl_loop_done): ...new function.
10200 (pass_web): Replace declaration with that of...
10201 (make_pass_web): ...new function.
10202 (pass_cse2): Replace declaration with that of...
10203 (make_pass_cse2): ...new function.
10204 (pass_df_initialize_opt): Replace declaration with that of...
10205 (make_pass_df_initialize_opt): ...new function.
10206 (pass_df_initialize_no_opt): Replace declaration with that of...
10207 (make_pass_df_initialize_no_opt): ...new function.
10208 (pass_reginfo_init): Replace declaration with that of...
10209 (make_pass_reginfo_init): ...new function.
10210 (pass_inc_dec): Replace declaration with that of...
10211 (make_pass_inc_dec): ...new function.
10212 (pass_stack_ptr_mod): Replace declaration with that of...
10213 (make_pass_stack_ptr_mod): ...new function.
10214 (pass_initialize_regs): Replace declaration with that of...
10215 (make_pass_initialize_regs): ...new function.
10216 (pass_combine): Replace declaration with that of...
10217 (make_pass_combine): ...new function.
10218 (pass_if_after_combine): Replace declaration with that of...
10219 (make_pass_if_after_combine): ...new function.
10220 (pass_ree): Replace declaration with that of...
10221 (make_pass_ree): ...new function.
10222 (pass_partition_blocks): Replace declaration with that of...
10223 (make_pass_partition_blocks): ...new function.
10224 (pass_match_asm_constraints): Replace declaration with that of...
10225 (make_pass_match_asm_constraints): ...new function.
10226 (pass_regmove): Replace declaration with that of...
10227 (make_pass_regmove): ...new function.
10228 (pass_split_all_insns): Replace declaration with that of...
10229 (make_pass_split_all_insns): ...new function.
10230 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10231 (make_pass_fast_rtl_byte_dce): ...new function.
10232 (pass_lower_subreg2): Replace declaration with that of...
10233 (make_pass_lower_subreg2): ...new function.
10234 (pass_mode_switching): Replace declaration with that of...
10235 (make_pass_mode_switching): ...new function.
10236 (pass_sms): Replace declaration with that of...
10237 (make_pass_sms): ...new function.
10238 (pass_sched): Replace declaration with that of...
10239 (make_pass_sched): ...new function.
10240 (pass_ira): Replace declaration with that of...
10241 (make_pass_ira): ...new function.
10242 (pass_reload): Replace declaration with that of...
10243 (make_pass_reload): ...new function.
10244 (pass_clean_state): Replace declaration with that of...
10245 (make_pass_clean_state): ...new function.
10246 (pass_branch_prob): Replace declaration with that of...
10247 (make_pass_branch_prob): ...new function.
10248 (pass_value_profile_transformations): Replace declaration with that
10250 (make_pass_value_profile_transformations): ...new function.
10251 (pass_postreload_cse): Replace declaration with that of...
10252 (make_pass_postreload_cse): ...new function.
10253 (pass_gcse2): Replace declaration with that of...
10254 (make_pass_gcse2): ...new function.
10255 (pass_split_after_reload): Replace declaration with that of...
10256 (make_pass_split_after_reload): ...new function.
10257 (pass_branch_target_load_optimize1): Replace declaration with that
10259 (make_pass_branch_target_load_optimize1): ...new function.
10260 (pass_thread_prologue_and_epilogue): Replace declaration with that
10262 (make_pass_thread_prologue_and_epilogue): ...new function.
10263 (pass_stack_adjustments): Replace declaration with that of...
10264 (make_pass_stack_adjustments): ...new function.
10265 (pass_peephole2): Replace declaration with that of...
10266 (make_pass_peephole2): ...new function.
10267 (pass_if_after_reload): Replace declaration with that of...
10268 (make_pass_if_after_reload): ...new function.
10269 (pass_regrename): Replace declaration with that of...
10270 (make_pass_regrename): ...new function.
10271 (pass_cprop_hardreg): Replace declaration with that of...
10272 (make_pass_cprop_hardreg): ...new function.
10273 (pass_reorder_blocks): Replace declaration with that of...
10274 (make_pass_reorder_blocks): ...new function.
10275 (pass_branch_target_load_optimize2): Replace declaration with that
10277 (make_pass_branch_target_load_optimize2): ...new function.
10278 (pass_leaf_regs): Replace declaration with that of...
10279 (make_pass_leaf_regs): ...new function.
10280 (pass_split_before_sched2): Replace declaration with that of...
10281 (make_pass_split_before_sched2): ...new function.
10282 (pass_compare_elim_after_reload): Replace declaration with that of...
10283 (make_pass_compare_elim_after_reload): ...new function.
10284 (pass_sched2): Replace declaration with that of...
10285 (make_pass_sched2): ...new function.
10286 (pass_stack_regs): Replace declaration with that of...
10287 (make_pass_stack_regs): ...new function.
10288 (pass_stack_regs_run): Replace declaration with that of...
10289 (make_pass_stack_regs_run): ...new function.
10290 (pass_df_finish): Replace declaration with that of...
10291 (make_pass_df_finish): ...new function.
10292 (pass_compute_alignments): Replace declaration with that of...
10293 (make_pass_compute_alignments): ...new function.
10294 (pass_duplicate_computed_gotos): Replace declaration with that of...
10295 (make_pass_duplicate_computed_gotos): ...new function.
10296 (pass_variable_tracking): Replace declaration with that of...
10297 (make_pass_variable_tracking): ...new function.
10298 (pass_free_cfg): Replace declaration with that of...
10299 (make_pass_free_cfg): ...new function.
10300 (pass_machine_reorg): Replace declaration with that of...
10301 (make_pass_machine_reorg): ...new function.
10302 (pass_cleanup_barriers): Replace declaration with that of...
10303 (make_pass_cleanup_barriers): ...new function.
10304 (pass_delay_slots): Replace declaration with that of...
10305 (make_pass_delay_slots): ...new function.
10306 (pass_split_for_shorten_branches): Replace declaration with that of...
10307 (make_pass_split_for_shorten_branches): ...new function.
10308 (pass_split_before_regstack): Replace declaration with that of...
10309 (make_pass_split_before_regstack): ...new function.
10310 (pass_convert_to_eh_region_ranges): Replace declaration with that
10312 (make_pass_convert_to_eh_region_ranges): ...new function.
10313 (pass_shorten_branches): Replace declaration with that of...
10314 (make_pass_shorten_branches): ...new function.
10315 (pass_set_nothrow_function_flags): Replace declaration with that of...
10316 (make_pass_set_nothrow_function_flags): ...new function.
10317 (pass_dwarf2_frame): Replace declaration with that of...
10318 (make_pass_dwarf2_frame): ...new function.
10319 (pass_final): Replace declaration with that of...
10320 (make_pass_final): ...new function.
10321 (pass_rtl_seqabstr): Replace declaration with that of...
10322 (make_pass_rtl_seqabstr): ...new function.
10323 (pass_release_ssa_names): Replace declaration with that of...
10324 (make_pass_release_ssa_names): ...new function.
10325 (pass_early_inline): Replace declaration with that of...
10326 (make_pass_early_inline): ...new function.
10327 (pass_inline_parameters): Replace declaration with that of...
10328 (make_pass_inline_parameters): ...new function.
10329 (pass_update_address_taken): Replace declaration with that of...
10330 (make_pass_update_address_taken): ...new function.
10331 (pass_convert_switch): Replace declaration with that of...
10332 (make_pass_convert_switch): ...new function.
10333 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10334 to a subclass of simple_ipa_opt_pass along with...
10335 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10336 (make_pass_ipa_tree_profile): ...new function.
10337 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10338 subclass of gimple_opt_pass along with...
10339 (pass_data_sra_early): ...new pass_data instance and...
10340 (make_pass_sra_early): ...new function.
10341 (pass_sra): Convert from a global struct to a subclass of
10342 gimple_opt_pass along with...
10343 (pass_data_sra): ...new pass_data instance and...
10344 (make_pass_sra): ...new function.
10345 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10346 gimple_opt_pass along with...
10347 (pass_data_early_ipa_sra): ...new pass_data instance and...
10348 (make_pass_early_ipa_sra): ...new function.
10349 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10350 subclass of gimple_opt_pass along with...
10351 (pass_data_ccp): ...new pass_data instance and...
10352 (make_pass_ccp): ...new function.
10353 (pass_fold_builtins): Convert from a global struct to a subclass of
10354 gimple_opt_pass along with...
10355 (pass_data_fold_builtins): ...new pass_data instance and...
10356 (make_pass_fold_builtins): ...new function.
10357 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10358 subclass of gimple_opt_pass along with...
10359 (pass_data_copy_prop): ...new pass_data instance and...
10360 (make_pass_copy_prop): ...new function.
10361 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10362 global struct to a subclass of gimple_opt_pass along with...
10363 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10364 (make_pass_rename_ssa_copies): ...new function.
10365 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10366 subclass of gimple_opt_pass along with...
10367 (pass_data_dce): ...new pass_data instance and...
10368 (make_pass_dce): ...new function.
10369 (pass_dce_loop): Convert from a global struct to a subclass of
10370 gimple_opt_pass along with...
10371 (pass_data_dce_loop): ...new pass_data instance and...
10372 (make_pass_dce_loop): ...new function.
10373 (pass_cd_dce): Convert from a global struct to a subclass of
10374 gimple_opt_pass along with...
10375 (pass_data_cd_dce): ...new pass_data instance and...
10376 (make_pass_cd_dce): ...new function.
10377 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10378 subclass of gimple_opt_pass along with...
10379 (pass_data_dominator): ...new pass_data instance and...
10380 (make_pass_dominator): ...new function.
10381 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10382 gimple_opt_pass along with...
10383 (pass_data_phi_only_cprop): ...new pass_data instance and...
10384 (make_pass_phi_only_cprop): ...new function.
10385 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10386 subclass of gimple_opt_pass along with...
10387 (pass_data_dse): ...new pass_data instance and...
10388 (make_pass_dse): ...new function.
10389 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10390 a subclass of gimple_opt_pass along with...
10391 (pass_data_forwprop): ...new pass_data instance and...
10392 (make_pass_forwprop): ...new function.
10393 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10394 struct to a subclass of gimple_opt_pass along with...
10395 (pass_data_tree_ifcombine): ...new pass_data instance and...
10396 (make_pass_tree_ifcombine): ...new function.
10397 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10398 subclass of gimple_opt_pass along with...
10399 (pass_data_ch): ...new pass_data instance and...
10400 (make_pass_ch): ...new function.
10401 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10402 subclass of gimple_opt_pass along with...
10403 (pass_data_tree_loop): ...new pass_data instance and...
10404 (make_pass_tree_loop): ...new function.
10405 (pass_tree_loop_init): Convert from a global struct to a subclass of
10406 gimple_opt_pass along with...
10407 (pass_data_tree_loop_init): ...new pass_data instance and...
10408 (make_pass_tree_loop_init): ...new function.
10409 (pass_lim): Convert from a global struct to a subclass of
10410 gimple_opt_pass along with...
10411 (pass_data_lim): ...new pass_data instance and...
10412 (make_pass_lim): ...new function.
10413 (pass_tree_unswitch): Convert from a global struct to a subclass of
10414 gimple_opt_pass along with...
10415 (pass_data_tree_unswitch): ...new pass_data instance and...
10416 (make_pass_tree_unswitch): ...new function.
10417 (pass_predcom): Convert from a global struct to a subclass of
10418 gimple_opt_pass along with...
10419 (pass_data_predcom): ...new pass_data instance and...
10420 (make_pass_predcom): ...new function.
10421 (pass_vectorize): Convert from a global struct to a subclass of
10422 gimple_opt_pass along with...
10423 (pass_data_vectorize): ...new pass_data instance and...
10424 (make_pass_vectorize): ...new function.
10425 (pass_graphite): Convert from a global struct to a subclass of
10426 gimple_opt_pass along with...
10427 (pass_data_graphite): ...new pass_data instance and...
10428 (make_pass_graphite): ...new function.
10429 (pass_graphite_transforms): Convert from a global struct to a subclass
10430 of gimple_opt_pass along with...
10431 (pass_data_graphite_transforms): ...new pass_data instance and...
10432 (make_pass_graphite_transforms): ...new function.
10433 (pass_check_data_deps): Convert from a global struct to a subclass of
10434 gimple_opt_pass along with...
10435 (pass_data_check_data_deps): ...new pass_data instance and...
10436 (make_pass_check_data_deps): ...new function.
10437 (pass_iv_canon): Convert from a global struct to a subclass of
10438 gimple_opt_pass along with...
10439 (pass_data_iv_canon): ...new pass_data instance and...
10440 (make_pass_iv_canon): ...new function.
10441 (pass_scev_cprop): Convert from a global struct to a subclass of
10442 gimple_opt_pass along with...
10443 (pass_data_scev_cprop): ...new pass_data instance and...
10444 (make_pass_scev_cprop): ...new function.
10445 (pass_record_bounds): Convert from a global struct to a subclass of
10446 gimple_opt_pass along with...
10447 (pass_data_record_bounds): ...new pass_data instance and...
10448 (make_pass_record_bounds): ...new function.
10449 (pass_complete_unroll): Convert from a global struct to a subclass of
10450 gimple_opt_pass along with...
10451 (pass_data_complete_unroll): ...new pass_data instance and...
10452 (make_pass_complete_unroll): ...new function.
10453 (pass_complete_unrolli): Convert from a global struct to a subclass of
10454 gimple_opt_pass along with...
10455 (pass_data_complete_unrolli): ...new pass_data instance and...
10456 (make_pass_complete_unrolli): ...new function.
10457 (pass_parallelize_loops): Convert from a global struct to a subclass
10458 of gimple_opt_pass along with...
10459 (pass_data_parallelize_loops): ...new pass_data instance and...
10460 (make_pass_parallelize_loops): ...new function.
10461 (pass_loop_prefetch): Convert from a global struct to a subclass of
10462 gimple_opt_pass along with...
10463 (pass_data_loop_prefetch): ...new pass_data instance and...
10464 (make_pass_loop_prefetch): ...new function.
10465 (pass_iv_optimize): Convert from a global struct to a subclass of
10466 gimple_opt_pass along with...
10467 (pass_data_iv_optimize): ...new pass_data instance and...
10468 (make_pass_iv_optimize): ...new function.
10469 (pass_tree_loop_done): Convert from a global struct to a subclass of
10470 gimple_opt_pass along with...
10471 (pass_data_tree_loop_done): ...new pass_data instance and...
10472 (make_pass_tree_loop_done): ...new function.
10473 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10474 struct to a subclass of gimple_opt_pass along with...
10475 (pass_data_cse_reciprocals): ...new pass_data instance and...
10476 (make_pass_cse_reciprocals): ...new function.
10477 (pass_cse_sincos): Convert from a global struct to a subclass of
10478 gimple_opt_pass along with...
10479 (pass_data_cse_sincos): ...new pass_data instance and...
10480 (make_pass_cse_sincos): ...new function.
10481 (pass_optimize_bswap): Convert from a global struct to a subclass of
10482 gimple_opt_pass along with...
10483 (pass_data_optimize_bswap): ...new pass_data instance and...
10484 (make_pass_optimize_bswap): ...new function.
10485 (pass_optimize_widening_mul): Convert from a global struct to a
10486 subclass of gimple_opt_pass along with...
10487 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10488 (make_pass_optimize_widening_mul): ...new function.
10489 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10490 subclass of gimple_opt_pass along with...
10491 (pass_data_phiopt): ...new pass_data instance and...
10492 (make_pass_phiopt): ...new function.
10493 (pass_cselim): Convert from a global struct to a subclass of
10494 gimple_opt_pass along with...
10495 (pass_data_cselim): ...new pass_data instance and...
10496 (make_pass_cselim): ...new function.
10497 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10498 subclass of gimple_opt_pass along with...
10499 (pass_data_phiprop): ...new pass_data instance and...
10500 (make_pass_phiprop): ...new function.
10501 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10502 subclass of gimple_opt_pass along with...
10503 (pass_data_pre): ...new pass_data instance and...
10504 (make_pass_pre): ...new function.
10505 (pass_fre): Convert from a global struct to a subclass of
10506 gimple_opt_pass along with...
10507 (pass_data_fre): ...new pass_data instance and...
10508 (make_pass_fre): ...new function.
10509 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10510 subclass of gimple_opt_pass along with...
10511 (pass_data_reassoc): ...new pass_data instance and...
10512 (make_pass_reassoc): ...new function.
10513 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10514 subclass of gimple_opt_pass along with...
10515 (pass_data_sink_code): ...new pass_data instance and...
10516 (make_pass_sink_code): ...new function.
10517 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10518 subclass of gimple_opt_pass along with...
10519 (pass_data_strlen): ...new pass_data instance and...
10520 (make_pass_strlen): ...new function.
10521 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10522 struct to a subclass of gimple_opt_pass along with...
10523 (pass_data_build_alias): ...new pass_data instance and...
10524 (make_pass_build_alias): ...new function.
10525 (pass_build_ealias): Convert from a global struct to a subclass of
10526 gimple_opt_pass along with...
10527 (pass_data_build_ealias): ...new pass_data instance and...
10528 (make_pass_build_ealias): ...new function.
10529 (pass_ipa_pta): Convert from a global struct to a subclass of
10530 simple_ipa_opt_pass along with...
10531 (pass_data_ipa_pta): ...new pass_data instance and...
10532 (make_pass_ipa_pta): ...new function.
10533 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10534 subclass of gimple_opt_pass along with...
10535 (pass_data_uncprop): ...new pass_data instance and...
10536 (make_pass_uncprop): ...new function.
10537 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10538 global struct to a subclass of gimple_opt_pass along with...
10539 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10540 (make_pass_late_warn_uninitialized): ...new function.
10541 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10542 to a subclass of gimple_opt_pass along with...
10543 (pass_data_init_datastructures): ...new pass_data instance and...
10544 (make_pass_init_datastructures): ...new function.
10545 (pass_early_warn_uninitialized): Convert from a global struct to a
10546 subclass of gimple_opt_pass along with...
10547 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10548 (make_pass_early_warn_uninitialized): ...new function.
10549 (pass_update_address_taken): Convert from a global struct to a
10550 subclass of gimple_opt_pass along with...
10551 (pass_data_update_address_taken): ...new pass_data instance and...
10552 (make_pass_update_address_taken): ...new function.
10553 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10554 struct to a subclass of gimple_opt_pass along with...
10555 (pass_data_release_ssa_names): ...new pass_data instance and...
10556 (make_pass_release_ssa_names): ...new function.
10557 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10558 subclass of gimple_opt_pass along with...
10559 (pass_data_stdarg): ...new pass_data instance and...
10560 (make_pass_stdarg): ...new function.
10561 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10562 global struct to a subclass of gimple_opt_pass along with...
10563 (pass_data_convert_switch): ...new pass_data instance and...
10564 (make_pass_convert_switch): ...new function.
10565 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10566 to a subclass of gimple_opt_pass along with...
10567 (pass_data_tail_recursion): ...new pass_data instance and...
10568 (make_pass_tail_recursion): ...new function.
10569 (pass_tail_calls): Convert from a global struct to a subclass of
10570 gimple_opt_pass along with...
10571 (pass_data_tail_calls): ...new pass_data instance and...
10572 (make_pass_tail_calls): ...new function.
10573 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10574 struct to a subclass of gimple_opt_pass along with...
10575 (pass_data_lower_vector): ...new pass_data instance and...
10576 (make_pass_lower_vector): ...new function.
10577 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10578 gimple_opt_pass along with...
10579 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10580 (make_pass_lower_vector_ssa): ...new function.
10581 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10582 to a subclass of gimple_opt_pass along with...
10583 (pass_data_slp_vectorize): ...new pass_data instance and...
10584 (make_pass_slp_vectorize): ...new function.
10585 (pass_ipa_increase_alignment): Convert from a global struct to a
10586 subclass of simple_ipa_opt_pass along with...
10587 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10588 (make_pass_ipa_increase_alignment): ...new function.
10589 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10590 gimple_opt_pass along with...
10591 (pass_data_vrp): ...new pass_data instance and...
10592 (make_pass_vrp): ...new function.
10593 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10594 subclass of simple_ipa_opt_pass along with...
10595 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10596 (make_pass_ipa_free_lang_data): ...new function.
10597 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10598 gimple_opt_pass along with...
10599 (pass_data_tsan): ...new pass_data instance and...
10600 (make_pass_tsan): ...new function.
10601 (pass_tsan_O0): Convert from a global struct to a subclass of
10602 gimple_opt_pass along with...
10603 (pass_data_tsan_O0): ...new pass_data instance and...
10604 (make_pass_tsan_O0): ...new function.
10605 * var-tracking.c (pass_variable_tracking): Convert from a global
10606 struct to a subclass of rtl_opt_pass along with...
10607 (pass_data_variable_tracking): ...new pass_data instance and...
10608 (make_pass_variable_tracking): ...new function.
10609 * web.c (pass_web): Convert from a global struct to a subclass of
10610 rtl_opt_pass along with...
10611 (pass_data_web): ...new pass_data instance and...
10612 (make_pass_web): ...new function.
10613 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10614 declaration with that of...
10615 (make_pass_mode_switch_use): ...new function.
10616 (pass_resolve_sw_modes): Replace declaration with that of...
10617 (make_pass_resolve_sw_modes): ...new function.
10618 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10619 from a global struct to a subclass of rtl_opt_pass along with...
10620 (pass_data_mode_switch_use): ...new pass_data instance and...
10621 (make_pass_mode_switch_use): ...new function.
10622 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10623 from a global struct to a subclass of rtl_opt_pass along with...
10624 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10625 (make_pass_resolve_sw_modes): ...new function.
10626 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10627 struct to a subclass of rtl_opt_pass along with...
10628 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10629 (make_pass_insert_vzeroupper): ...new function.
10630 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10631 global struct to a subclass of rtl_opt_pass along with...
10632 (pass_data_work_around_errata): ...new pass_data instance and...
10633 (make_pass_work_around_errata): ...new function.
10634 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10635 struct to a subclass of rtl_opt_pass along with...
10636 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10637 (make_pass_mips_machine_reorg2): ...new function.
10639 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10641 * passes.c (pass_manager::operator new): New.
10643 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10645 Handwritten part of conversion of passes to C++ classes.
10647 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10648 (toplev.o): Add dep on PASS_MANAGER_H.
10649 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10650 of early local pases to reflect this moving from a global to a
10651 member of gcc::pass_manager.
10652 (cgraph_add_new_function): Likewise.
10653 * lto-cgraph.c (lto_output_node): Update for conversion of
10654 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10655 * passes.c (opt_pass::clone): New.
10656 (opt_pass::gate): New.
10657 (opt_pass::execute): New.
10658 (opt_pass::opt_pass): New.
10659 (pass_manager::execute_early_local_passes): New.
10660 (pass_manager::execute_pass_mode_switching): new.
10661 (finish_optimization_passes): Convert to...
10662 (pass_manager::finish_optimization_passes): ...this.
10663 (finish_optimization_passes): Update for conversion of passes to
10665 (register_dump_files_1): Use has_gate since we cannot portably
10666 check a vtable entry against NULL.
10667 (dump_one_pass): Likewise.
10668 (ipa_write_summaries_2): Likewise.
10669 (ipa_write_optimization_summaries_1): Likewise.
10670 (ipa_read_summaries_1): Likewise.
10671 (ipa_read_optimization_summaries_1): Likewise.
10672 (execute_ipa_stmt_fixups): Likewise.
10673 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10674 pass-creation functions rather than wiring up globals, and
10675 storing the results in fields of pass_manager generated using
10676 pass-instances.def.
10677 (pass_manager::dump_profile_report): Update for conversion of
10678 passes to C++ classes.
10679 (pass_manager::execute_ipa_summary_passes): Likewise.
10680 (execute_one_ipa_transform_pass): Likewise.
10681 (execute_one_pass): Use has_gate and has_execute since we cannot
10682 portably check a vtable entry against NULL.
10683 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10684 (pass_manager): Use pass-instances.def to add fields for the
10685 various pass instances.
10686 * toplev.c (finalize): Update for move of
10687 finish_optimization_passes to a method of gcc::pass_manager.
10688 * toplev.h (finish_optimization_passes): Move to method of class
10690 * tree-pass.h (struct pass_data): New.
10691 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10692 (opt_pass::gate): Convert to virtual function.
10693 (opt_pass::~opt_pass): New.
10694 (opt_pass::clone): New.
10695 (opt_pass::execute): Convert to virtual function.
10696 (opt_pass::opt_pass): New.
10697 (opt_pass::ctxt_): new.
10698 (gimple_opt_pass): Convert to subclass of opt_pass.
10699 (gimple_opt_pass::gimple_opt_pass): New.
10700 (rtl_opt_pass): Convert to subclass of opt_pass.
10701 (rtl_opt_pass::rtl_opt_pass): New.
10702 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10703 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10704 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10705 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10706 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10707 invocation of pass_mode_switching to reflect this moving from a
10708 global to a member of gcc::pass_manager.
10709 (ix86_option_override): Rework how pass_insert_vzeroupper is
10710 added to the pass_manager to reflect autogenerated changes.
10711 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10713 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10715 PR rtl-optimization/57708
10716 * recog.c (peep2_find_free_register): Validate all regs in a
10719 2013-08-05 Jan Hubicka <jh@suse.cz>
10722 * cgraph.c (verify_cgraph_node): Accept local flags from other
10724 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10725 (function_and_variable_visibility): Likewise.
10726 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10728 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10730 * graph.c (init_graph_slim_pretty_print): Remove.
10731 (print_graph_cfg): Do not call it. Use local pretty printer.
10732 (start_graph_dump): Likewise.
10734 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10736 * gimple-pretty-print.c (buffer): Remove.
10737 (initialized): Likewise.
10738 (maybe_init_pretty_print): Likewise.
10739 (print_gimple_stmt): Do not call it. Use non-static local
10740 pretty_printer variable.
10741 (print_gimple_expr): Likewise.
10742 (print_gimple_seq): Likewise.
10743 (gimple_dump_bb): Likewise.
10745 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10747 * asan.c (asan_pp): Remove.
10748 (asan_pp_initialized): Likewise.
10749 (asan_pp_initialize): Likewise.
10750 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10751 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10752 (asan_add_global): Likewise.
10754 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10756 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10757 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10758 * pretty-print.c (pp_formatted_text_data): Likewise.
10759 (pp_indent): Rename from pp_base_indent.
10760 (pp_format): Rename from pp_base_format.
10761 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10762 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10763 (pp_flush): Rename from pp_base_flush.
10764 (pp_set_line_maximum_length): Rename from
10765 pp_base_set_line_maximum_length.
10766 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10767 (pp_set_prefix): Rename from pp_base_set_prefix.
10768 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10769 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10770 (pp_append_text): Rename from pp_base_append_text.
10771 (pp_formatted_text): Rename from pp_base_formatted_text.
10772 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10773 (pp_remaining_character_count_for_line): Rename from
10774 pp_base_remaining_character_count_for_line.
10775 (pp_newline): Rename from pp_base_newline.
10776 (pp_character): Rename from pp_base_character.
10777 (pp_string): Rename from pp_base_string.
10778 (pp_maybe_space): Rename from pp_base_maybe_space.
10779 * asan.c (asan_pp_string): Adjust.
10780 (asan_emit_stack_protection): Likewise.
10781 (asan_add_global): Likewise.
10782 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10783 * tree-mudflap.c (mf_varname_tree): Likewise.
10784 * tree-pretty-print.c (pp_tree_identifier): Rename from
10785 pp_base_tree_identifier.
10786 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10787 Declare as function.
10789 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10791 * pretty-print.h (pp_bar_bar): New.
10792 (pp_ampersand_ampersand): Likewise.
10793 (pp_less_equal): Likewise.
10794 (pp_greater_equal): Likewise.
10795 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10796 printer functions instead of pp_string or operators and punctuators.
10797 (dump_gimple_call): Likewise.
10798 (dump_gimple_omp_for): Likewise.
10799 (dump_gimple_transaction): Likewise.
10800 (dump_gimple_phi): Likewise.
10801 (pp_gimple_stmt_1): Likewise.
10802 * sched-vis.c (print_insn): Likewise.
10803 * tree-mudflap.c (mf_varname_tree): Likewise.
10804 * tree-pretty-print.c (dump_block_node): Likewise.
10805 (dump_generic_node): Likewise.
10807 2013-08-02 Jan Hubicka <jh@suse.cz>
10809 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10811 * lto-streamer-out.c (tree_is_indexable): Results decls and
10812 parm decls are not indexable.
10813 (DFS_write_tree_body): Do not follow args and results.
10814 (hash_tree): Likewise.
10815 (output_functions): Rearrange so struct function is needed
10816 only when real body is output; be able to also ouptut abstract
10817 functions; output DECL_ARGUMENTS and DECL_RESULT.
10818 (lto_output): When not in WPA, ale store abstract functions.
10819 (write_symbol): Do not care about RESULT_DECL.
10820 (output_symbol_p): Handle correctly sbtract decls.
10821 * lto-streamer-in.c (input_function): Rearrange so struct
10822 function can be NULL at entry; allow streaming of
10823 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10824 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10825 sanity check during LTO.
10826 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10827 RESULT_DECl and DECL_ARGUMENTS.
10828 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10831 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10833 * pretty-print.h (pp_underscore): New.
10835 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10836 printer functions instead of pp_character.
10837 (dump_binary_rhs): Likewise.
10838 (dump_ternary_rhs): Likewise.
10839 (dump_gimple_call_args): Likewise.
10840 (pp_points_to_solution): Likewise.
10841 (dump_gimple_call): Likewise.
10842 (dump_gimple_switch): Likewise.
10843 (dump_gimple_cond): Likewise.
10844 (dump_gimple_bind): Likewise.
10845 (dump_gimple_try): Likewise.
10846 (dump_gimple_omp_for): Likewise.
10847 (dump_gimple_omp_continue): Likewise.
10848 (dump_gimple_omp_single): Likewise.
10849 (dump_gimple_omp_sections): Likewise.
10850 (dump_gimple_omp_block): Likewise.
10851 (dump_gimple_omp_critical): Likewise.
10852 (dump_gimple_transaction): Likewise.
10853 (dump_gimple_asm): Likewise.
10854 (dump_gimple_phi): Likewise.
10855 (dump_gimple_omp_parallel): Likewise.
10856 (dump_gimple_omp_task): Likewise.
10857 (dump_gimple_omp_atomic_load): Likewise.
10858 (dump_gimple_omp_atomic_store): Likewise.
10859 (dump_gimple_mem_ops): Likewise.
10860 (pp_gimple_stmt_1): Likewise.
10861 (pp_cfg_jump): Likewise.
10862 (dump_implicit_edges): Likewise.
10863 (gimple_dump_bb_for_graph): Likewise.
10864 * graph.c (draw_cfg_node): Likewise.
10865 * langhooks.c (lhd_print_error_function): Likewise.
10866 * sched-vis.c (print_exp): Likewise.
10867 (print_value): Likewise.
10868 (print_pattern): Likewise.
10869 (print_insn): Likewise.
10870 (rtl_dump_bb_for_graph): Likewise.
10871 * tree-pretty-print.c (dump_function_declaration): Likewise.
10872 (dump_array_domain): Likewise.
10873 (dump_omp_clause): Likewise.
10874 (dump_location): Likewise.
10875 (dump_generic_node): Likewise.
10876 (print_struct_decl): Likewise.
10877 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10879 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10881 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10882 candidate table when replacing a candidate statement.
10883 (replace_rhs_if_not_dup): Likewise.
10884 (replace_one_candidate): Likewise.
10886 2013-08-02 Jan Hubicka <jh@suse.cz>
10887 Martin Liska <marxin.liska@gmail.com>
10889 * cgraphunit.c (add_new_function): Fix logic when adding from
10891 (assemble_thunk): Rename to ...
10892 (expand_thunk); .. this one; export; get it working with
10893 general functions; make produced gimple valid.
10894 * cgraph.h (expand_thunk): Declare.
10896 2013-08-02 Jan Hubicka <jh@suse.cz>
10898 * ipa-cp.c (gather_context_independent_values): Use
10899 ipa_get_param_move_cost.
10900 (get_replacement_map): Remove PARAM; move parameter folding
10902 (create_specialized_node): Update.
10903 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10904 assert that we have gimple body; update move_cost.
10905 (count_formal_params): Assert that we have gimple body.
10906 (ipa_dump_param): New function.
10907 (ipa_alloc_node_params): Break out from ...
10908 (ipa_initialize_node_params): ... here.
10909 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10910 (ipa_write_node_info): Stream move costs.
10911 (ipa_read_node_info): Read move costs.
10912 (ipa_update_after_lto_read): Do not recompute node params.
10913 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10914 (ipa_get_param): Check we are not in WPA.
10915 (ipa_get_param_move_cost): New.
10916 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10917 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10918 parm numbers to be present.
10920 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10922 PR rtl-optimization/58048
10923 * lra-constraints.c (process_alt_operands): Don't check asm
10924 operand on register.
10926 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10928 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10929 the implied StoreLoad barrier for atomic operations if before.
10931 2013-08-02 Jan Hubicka <jh@suse.cz>
10932 Martin Liska <marxin.liska@gmail.com>
10934 * cgraph.c (cgraph_function_body_availability): Do not check
10936 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10937 symtab_node_availability): Declare.
10938 * ipa.c (can_replace_by_local_alias): New.
10939 (function_and_variable_visibility): Use it.
10940 * symtab.c (symtab_for_node_and_aliases,
10941 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
10943 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10945 PR rtl-optimization/57963
10946 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
10947 (lra_constraints): Use them.
10949 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10951 * config/arm/types.md (define_attr "type"): Add "load_acq"
10953 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10955 (cortex_a53_store1): Likewise.
10957 2013-08-01 Jan Hubicka <jh@suse.cz>
10959 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10960 partitions are not needed.
10962 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
10964 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10965 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10966 MAYBE_NON_Q_CLASS_P where appropriate.
10968 2013-08-01 Jan Hubicka <jh@suse.cz>
10970 * cgraph.h (release_function_body): Declare.
10971 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10972 of unused delcarations.
10974 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10976 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10977 RTL form when subtracting a constant.
10979 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10981 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10982 Generate canonical plus rtx with negated immediate instead of minus
10984 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10986 2013-08-01 Jan Hubicka <jh@suse.cz>
10988 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10989 (cgraph_release_function_body): Likewise.
10990 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10991 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10992 to used_as_abstract_origin.
10993 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10994 symbols used as abstract origins.
10995 * cgraphunit.c (analyze_functions): Update.
10996 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10997 used_as_abstract_origin.
10998 * tree-inline.c (tree_function_versioning): Update
10999 used_as_abstract_origin; be ready for DECL_RESULT and
11000 DECL_ARGUMENTS to be NULL.
11002 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11003 for abstract functions.
11004 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11007 2013-08-01 Jan Hubicka <jh@suse.cz>
11009 * profile.c (compute_value_histograms): Fix thinko.
11011 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11013 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11014 aarch-common-protos.h to extra_headers.
11015 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11016 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11017 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11019 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11021 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11022 Include "../arm/types.md". Define "type" attribute for all patterns.
11023 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11026 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11028 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11029 to support power8 load fusion.
11030 (fusion_gpr_mem_load): Likewise.
11032 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11034 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11035 declarations for power8 load fusion.
11036 (emit_fusion_gpr_load): Likewise.
11038 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11039 tuning for power8, turn on fusion mode by default. Turn on sign
11040 extending fusion mode if normal fusion mode is on, and we are at
11042 (fusion_gpr_load_p): New function, return true if we can fuse an
11043 addis instruction with a dependent load to a GPR.
11044 (emit_fusion_gpr_load): Emit the instructions for power8 load
11047 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11048 (VSX load fusion peepholes): New peepholes to fuse together an
11049 addi instruction with a VSX load instruction.
11051 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11052 peepholes to fuse an addis instruction with a load to a GPR base
11053 register. If we are supporting sign extending fusions, convert
11054 sign extending loads to zero extending loads and add an explicit
11057 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11059 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11060 aarch-common-protos.h to extra_headers.
11061 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11062 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11063 (arm_early_store_addr_dep): Likewise.
11064 (arm_no_early_alu_shift_dep): Likewise.
11065 (arm_no_early_alu_shift_value_dep): Likewise.
11066 (arm_no_early_mul_dep): Likewise.
11067 (arm_no_early_store_addr_dep): Likewise.
11068 (arm_mac_accumulator_is_mul_result): Likewise.
11069 (arm_mac_accumulator_is_result): Likewise.
11070 * config/arm/aarch-common.c: ... here. New file.
11071 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11073 (arm_early_store_addr_dep): Likewise.
11074 (arm_no_early_alu_shift_dep): Likewise.
11075 (arm_no_early_alu_shift_value_dep): Likewise.
11076 (arm_no_early_mul_dep): Likewise.
11077 (arm_no_early_store_addr_dep): Likewise.
11078 (arm_mac_accumulator_is_mul_result): Likewise.
11079 (arm_mac_accumulator_is_result): Likewise.
11080 * config/arm/aarch-common-protos.h: ... here. New file.
11081 * config/arm/t-arm (aarch-common.o): Define.
11083 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11085 * config/arm/arm.md: Include new file "types.md".
11086 (define_attr "type"): Move from here to ...
11087 (define_attr "mul32"): Likewise.
11088 (define_attr "mul64"): Likewise.
11089 * config/arm/types.md: ... here. New file.
11091 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11093 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11094 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11096 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11098 * gen-pass-instances.awk: Fix offset of substr().
11100 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11102 * Makefile.in (pass-instances.def): New.
11103 (passes.o): Replace dependency on passes.def with one on
11106 * gen-pass-instances.awk: New.
11108 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11109 rather than passes.def, updating local definition of NEXT_PASS
11110 macro to add an extra NUM parameter (currently unused).
11112 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11114 * Makefile.in (PASS_MANAGER_H): New.
11115 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11116 (passes.o): Likewise.
11117 (statistics.o): Likewise.
11118 (cgraphunit.o): Likewise.
11119 (context.o): Depend on PASS_MANAGER_H.
11121 * pass_manager.h: New.
11123 * cgraphunit.c (cgraph_add_new_function): Update for moves
11124 of globals to fields of pass_manager.
11125 (analyze_function): Likewise.
11126 (expand_function): Likewise.
11127 (ipa_passes): Likewise.
11128 (compile): Likewise.
11130 * context.c (context::context): New.
11131 * context.h (context::context): New.
11132 (context::get_passes): New.
11133 (context::passes_): New.
11135 * lto-cgraph.c (input_node): Update for moves of globals to
11136 fields of pass_manager.
11138 * passes.c (all_passes): Remove, in favor of a field of the
11139 same name within the new class pass_manager.
11140 (all_small_ipa_passes): Likewise.
11141 (all_lowering_passes): Likewise.
11142 (all_regular_ipa_passes): Likewise.
11143 (all_late_ipa_passes): Likewise.
11144 (all_lto_gen_passes): Likewise.
11145 (passes_by_id): Likewise.
11146 (passes_by_id_size): Likewise.
11147 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11148 the new class pass_manager.
11149 (set_pass_for_id): Convert to...
11150 (pass_manager::set_pass_for_id): ...method.
11151 (get_pass_for_id): Convert to...
11152 (pass_manager::get_pass_for_id): ...method.
11153 (register_one_dump_file): Move body of implementation into...
11154 (pass_manager::register_one_dump_file): ...here.
11155 (register_dump_files_1): Convert to...
11156 (pass_manager::register_dump_files_1): ...method.
11157 (register_dump_files): Convert to...
11158 (pass_manager::register_dump_files): ...method.
11159 (create_pass_tab): Update for moves of globals to fields of
11161 (dump_passes): Move body of implementation into...
11162 (pass_manager::dump_passes): ...here.
11163 (register_pass): Move body of implementation into...
11164 (pass_manager::register_pass): ...here.
11165 (init_optimization_passes): Convert into...
11166 (pass_manager::pass_manager): ...constructor for new
11167 pass_manager class, and initialize the pass_lists array.
11168 (check_profile_consistency): Update for moves of globals to
11169 fields of pass_manager.
11170 (dump_profile_report): Move body of implementation into...
11171 (pass_manager::dump_profile_report): ...here.
11172 (ipa_write_summaries_1): Update for moves of pass lists from
11173 being globals to fields of pass_manager.
11174 (ipa_write_optimization_summaries): Likewise.
11175 (ipa_read_summaries): Likewise.
11176 (ipa_read_optimization_summaries): Likewise.
11177 (execute_all_ipa_stmt_fixups): Likewise.
11179 * statistics.c (statistics_fini): Update for moves of globals to
11180 fields of pass_manager.
11182 * toplev.c (general_init): Replace call to
11183 init_optimization_passes with construction of the pass_manager
11186 * tree-pass.h (all_passes): Remove, in favor of a field of the
11187 same name within the new class pass_manager.
11188 (all_small_ipa_passes): Likewise.
11189 (all_lowering_passes): Likewise.
11190 (all_regular_ipa_passes): Likewise.
11191 (all_lto_gen_passes): Likewise.
11192 (all_late_ipa_passes): Likewise.
11193 (passes_by_id): Likewise.
11194 (passes_by_id_size): Likewise.
11195 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11196 the new class pass_manager.
11197 (get_pass_for_id): Remove.
11199 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11201 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11204 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11206 * arm.md (mulhi3): New expand pattern.
11208 2013-07-30 Jan Hubicka <jh@suse.cz>
11209 Martin Liska <marxin.liska@gmail.com>
11211 * profile.c (compute_value_histograms): Do not ICE when
11212 there is mismatch only on some counters.
11214 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11216 PR rtl-optimization/57637
11217 * function.c (move_insn_for_shrink_wrap): Also check the
11218 GEN set of the LIVE problem for the liveness analysis
11219 if it exists, otherwise give up.
11221 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11223 PR tree-optimization/57993
11224 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11225 replaced statement in the candidate table.
11226 (phi_add_costs): Return infinite cost when the hidden basis does
11227 not dominate all phis on which the candidate is dependent.
11228 (replace_one_candidate): Record replaced statement in the
11231 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11233 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11234 (ashlv2si3): New expander.
11235 (*ashlv2si3_i): New define_insn_and_split.
11236 * predicates.md (float_operation): Allow patterns with three
11237 basic sub-patterns.
11239 PR rtl-optimization/58021
11240 * mode-switching.c (create_pre_exit): Always split off preceding
11241 insns if we are not at the basic block head.
11243 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11245 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11246 (UCLIBC_DYNAMIC_LINKER): New macro.
11247 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11249 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11250 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11252 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11253 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11254 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11255 for SF and DF modes. Use ieee_quad_format for TF mode.
11256 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11257 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11258 (mips_option_override): Handle `-mnan=legacy'.
11259 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11260 `-mabs=2008' and `-mnan=2008'.
11261 (OPTION_DEFAULT_SPECS): Add "nan" default.
11262 (ASM_SPEC): Handle `-mnan='.
11263 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11264 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11265 comment accordingly.
11266 (neg<mode>2): Likewise.
11267 * config/mips/mips.opt (mabs, mnan): New options.
11268 * doc/install.texi (Configuration): Document `--with-nan=' option.
11269 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11271 (MIPS Options): Document them.
11272 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11273 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11274 * configure: Regenerate.
11275 * config.in: Regenerate.
11277 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11279 * config/i386/i386.md (float post-reload splitters): Do not check
11280 for subregs of SSE registers.
11282 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11283 H.J. Lu <hongjiu.lu@intel.com>
11287 * config/i386/i386.md (post-reload splitter
11288 to avoid partial SSE reg dependency stalls): New pattern.
11290 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11292 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11294 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11296 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11297 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11299 2013-07-26 Cary Coutant <ccoutant@google.com>
11301 * dwarf2out.c (die_checksum_ordered): Don't include template
11302 instantiations in signature.
11303 (is_template_parameter): New function.
11304 (is_template_instantiation): New function.
11305 (generate_skeleton_bottom_up): Don't include template instantiations
11307 (generate_skeleton): Likewise.
11308 (break_out_comdat_types): Move recursive call to break out nested
11310 (prune_unused_types_mark_generic_parms_dies): Call
11311 is_template_parameter.
11313 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11315 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11316 uses vector registers.
11317 * config/aarch64/iterators.md: Add attributes rtn and vas.
11319 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11320 Richard Earnshaw <richard.earnshaw@arm.com>
11322 * combine.c (simplify_comparison): Re-canonicalize operands
11324 * config/arm/arm.md (movcond_addsi): New splitter.
11326 2013-07-25 Sterling Augustine <saugustine@google.com>
11328 * dwarf2out.c (size_of_pubnames): Move code to...
11329 (include_pubname_in_output): ...here. New.
11330 (want_pubnames): Rearrange.
11331 (output_pubnames): Call include_pubname_in_output. Move assertion.
11333 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11335 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11337 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11340 * doc/extend.texi: Remove obsolete builtins. Fix
11341 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11343 2013-07-25 Jan Hubicka <jh@suse.cz>
11345 * cgraph.c (release_function_body): Break out from ...
11346 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11347 and DECL_ARGUMENTS.
11348 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11349 old_tree in the map.
11350 (create_specialized_node): Update.
11351 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11353 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11354 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11355 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11356 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11359 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11361 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11362 addsi3_carryin_alt2_<optab>): Correct output template.
11364 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11366 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11367 Adjust for arm_restrict_it.
11368 Remove trailing whitespace.
11370 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11372 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11373 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11375 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11377 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11379 PR rtl-optimization/57960
11380 * lra-constraints.c (process_alt_operands): Use the right mode
11381 when checking strict_low.
11383 2013-07-25 Jan Hubicka <jh@suse.cz>
11385 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11386 * cgraph.c (cgraph_remove_node): Do not release function body
11387 when in cgraph streaming.
11388 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11389 in other partitions are not considered reachable; fix handling of
11392 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11394 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11396 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11401 * config/arm/arm.md ("*sibcall_insn): Replace use of
11402 Ss with US. Adjust output for v5 and v4t.
11403 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11405 * config/arm/constraints.md ("Ss"): Rename to US.
11407 2013-07-25 Terry Guo <terry.guo@arm.com>
11409 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11410 shift_add/shift_sub0/shift_sub1 RTXs.
11412 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11413 Anton Blanchard <anton@au1.ibm.com>
11415 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11416 (altivec_vpks<VI_char>ss): Likewise.
11417 (altivec_vpks<VI_char>us): Likewise.
11418 (altivec_vpku<VI_char>us): Likewise.
11419 (altivec_vpku<VI_char>um): Likewise.
11421 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11423 Introduce context class.
11425 * Makefile.in (CONTEXT_H): New.
11426 (OBJS): Add context.o.
11427 (toplev.o): Add CONTEXT_H to dependencies.
11430 * toplev.c (general_init): Create the singleton gcc::context instance.
11436 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11438 PR rtl-optimization/57968
11439 * mode-switching.c (create_pre_exit): Allow instructions that
11440 don't set a return register to need a non-exit mode.
11442 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11443 Anton Blanchard <anton@au1.ibm.com>
11445 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11446 operands to vperm for little endian.
11447 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11448 of lvsl to create the control mask for a vperm for little endian.
11450 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11451 Anton Blanchard <anton@au1.ibm.com>
11453 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11454 two operands for little-endian.
11456 2013-07-23 Steve Ellcey <sellcey@mips.com>
11458 * config/mips/mips.c (mips_case_values_threshold): New.
11459 (TARGET_CASE_VALUES_THRESHOLD): Define.
11461 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11462 Anton Blanchard <anton@au1.ibm.com>
11464 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11465 selection of field for vector splat in little endian mode.
11467 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11469 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11470 expanders to rs6000.md.
11471 (ior<mode>3): Likewise.
11472 (and<mode>3): Likewise.
11473 (one_cmpl<mode>2): Likewise.
11474 (nor<mode>3): Likewise.
11475 (andc<mode>3): Likewise.
11476 (eqv<mode>3): Likewise.
11477 (nand<mode>3): Likewise.
11478 (orc<mode>3): Likewise.
11480 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11483 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11484 to split multi-word logical operations.
11485 (rs6000_split_logical_di): Likewise.
11486 (rs6000_split_logical): Likewise.
11488 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11489 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11490 and allow TImode operations in 32-bit.
11491 (vsx_and<mode>3_64bit): Likewise.
11492 (vsx_ior<mode>3_32bit): Likewise.
11493 (vsx_ior<mode>3_64bit): Likewise.
11494 (vsx_xor<mode>3_32bit): Likewise.
11495 (vsx_xor<mode>3_64bit): Likewise.
11496 (vsx_one_cmpl<mode>2_32bit): Likewise.
11497 (vsx_one_cmpl<mode>2_64bit): Likewise.
11498 (vsx_nor<mode>3_32bit): Likewise.
11499 (vsx_nor<mode>3_64bit): Likewise.
11500 (vsx_andc<mode>3_32bit): Likewise.
11501 (vsx_andc<mode>3_64bit): Likewise.
11502 (vsx_eqv<mode>3_32bit): Likewise.
11503 (vsx_eqv<mode>3_64bit): Likewise.
11504 (vsx_nand<mode>3_32bit): Likewise.
11505 (vsx_nand<mode>3_64bit): Likewise.
11506 (vsx_orc<mode>3_32bit): Likewise.
11507 (vsx_orc<mode>3_64bit): Likewise.
11509 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11510 logical types in GPRs.
11512 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11513 logical insns to rs6000.md, and allow TImode operations in
11515 (altivec_ior<mode>3): Likewise.
11516 (altivec_xor<mode>3): Likewise.
11517 (altivec_one_cmpl<mode>2): Likewise.
11518 (altivec_nor<mode>3): Likewise.
11519 (altivec_andc<mode>3): Likewise.
11521 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11522 attributes for moving the 128-bit logical operations into
11524 (BOOL_REGS_OUTPUT): Likewise.
11525 (BOOL_REGS_OP1): Likewise.
11526 (BOOL_REGS_OP2): Likewise.
11527 (BOOL_REGS_UNARY): Likewise.
11528 (BOOL_REGS_AND_CR0): Likewise.
11529 (one_cmpl<mode>2): Add support for DI logical operations on
11530 32-bit, splitting the operations to 32-bit.
11531 (anddi3): Likewise.
11532 (iordi3): Likewise.
11533 (xordi3): Likewise.
11534 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11535 changes to combine the 32/64-bit code, allow logical operations on
11536 TI mode in 32-bit, and to use similar match_operator patterns like
11537 scalar mode uses. Combine the Altivec and VSX code for logical
11538 operations, and move it here.
11539 (ior<mode>3, 128-bit types): Likewise.
11540 (xor<mode>3, 128-bit types): Likewise.
11541 (one_cmpl<mode>3, 128-bit types): Likewise.
11542 (nor<mode>3, 128-bit types): Likewise.
11543 (andc<mode>3, 128-bit types): Likewise.
11544 (eqv<mode>3, 128-bit types): Likewise.
11545 (nand<mode>3, 128-bit types): Likewise.
11546 (orc<mode>3, 128-bit types): Likewise.
11547 (and<mode>3_internal): Likewise.
11548 (bool<mode>3_internal): Likewise.
11549 (boolc<mode>3_internal1): Likewise.
11550 (boolc<mode>3_internal2): Likewise.
11551 (boolcc<mode>3_internal1): Likewise.
11552 (boolcc<mode>3_internal2): Likewise.
11553 (eqv<mode>3_internal1): Likewise.
11554 (eqv<mode>3_internal2): Likewise.
11555 (one_cmpl1<mode>3_internal): Likewise.
11557 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11559 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11560 Rename flag_stack_usage to flag_stack_usage_info.
11562 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11564 * config/microblaze/sync.md: New file.
11565 * config/microblaze/microblaze.md: Include sync.md
11566 * config/microblaze/microblaze.c: Add print_operand 'y'.
11567 * config/microblaze/constraints.md: Add memory_contraint
11568 'Q' which is a single register.
11570 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11572 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11574 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11576 * reload.c (find_reloads): Exit loop once we find this operand
11577 cannot be reloaded somehow for this alternative.
11579 * reload.c (find_reloads): Exit loop once we find a hard register.
11581 * rtlanal.c (computed_jump_p): Exit loop once we find label
11584 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11586 * cfgloopmanip.c (remove_path): Exit loop after setting
11589 * gensupport.c (subst_dup): Avoid loop if code is not
11590 MATCH_DUP nor MATCH_OP_DUP.
11592 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11594 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11596 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11598 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11599 true for SP_REGNUM if mode == ptr_mode.
11600 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11601 with value R0_REGNUM + 31.
11603 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11605 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11606 pad pointer-typed argument downward.
11608 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11610 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11611 and __ILP32__ when the ILP32 model is in use.
11613 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11615 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11616 (aarch64_load_symref_appropriately): In the case of
11617 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11618 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11619 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11620 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11621 if 'mode' doesn't equal to 'ptr_mode'.
11622 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11623 'vcall_offset'; change to call aarch64_emit_move differently depending
11624 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11625 to calculate the upper bound of 'vcall_offset'.
11626 (aarch64_cannot_force_const_mem): Change to also return true if
11628 (aarch64_legitimize_reload_address): In the case of large
11629 displacements, add new local variable 'xmode' and an assertion
11630 based on it; change to use 'xmode' to generate the new rtx and
11632 (aarch64_asm_trampoline_template): Change to generate the template
11633 differently depending on TARGET_ILP32 or not; change to use
11634 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11635 (aarch64_trampoline_size): Removed.
11636 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11637 and replace immediate literals with it. Change to use 'ptr_mode'
11638 instead of 'DImode' and call convert_memory_address if the mode
11639 of 'fnaddr' doesn't equal to 'ptr_mode'.
11640 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11642 (aarch64_elf_asm_destructor): Likewise.
11643 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11644 on TARGET_ILP32 instead of aarch64_trampoline_size.
11645 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11646 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11647 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11648 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11649 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11650 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11651 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11652 depending on the value of 'mode'.
11653 (add_losym_<mode>): New.
11654 (ldr_got_small_<mode>): New, based on ldr_got_small.
11655 (ldr_got_small): Remove.
11656 (ldr_got_small_sidi): New.
11657 * config/aarch64/iterators.md (P): New.
11658 (PTR): Change to 'ptr_mode' in the condition.
11660 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11662 * config.gcc (aarch64*-*-*): Support --with-abi.
11663 (aarch64*-*-elf): Support --with-multilib-list.
11664 (aarch64*-*-linux*): Likewise.
11665 (supported_defaults): Add abi to aarch64*-*-*.
11666 * configure.ac: Mention AArch64 for --with-multilib-list.
11667 * configure: Re-generated.
11668 * config/aarch64/biarchilp32.h: New file.
11669 * config/aarch64/biarchlp64.h: New file.
11670 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11672 (MULTILIB_DEFAULTS): Ditto.
11673 (DRIVER_SELF_SPECS): Ditto.
11674 (ASM_SPEC): Update to also substitute -mabi.
11675 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11676 file whose name depends on -mabi= and -mbig-endian.
11677 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11679 (POINTER_SIZE): New define.
11680 (POINTERS_EXTEND_UNSIGNED): Ditto.
11681 (enum aarch64_abi_type): New enumeration tag.
11682 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11683 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11684 (TARGET_ILP32): New define.
11685 * config/aarch64/aarch64.opt (mabi): New.
11686 (aarch64_abi): New.
11687 (ilp32, lp64): New values for -mabi.
11688 * config/aarch64/t-aarch64 (comma): New define.
11689 (MULTILIB_OPTIONS): Ditto.
11690 (MULTILIB_DIRNAMES): Ditto.
11691 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11692 * doc/invoke.texi: Document -mabi for AArch64.
11694 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11696 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11698 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11699 Anton Blanchard <anton@au1.ibm.com>
11701 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11702 endianness when selecting field to splat.
11704 2013-07-22 Eric Christopher <echristo@gmail.com>
11706 * dwarf2out.c (die_odr_checksum): New function to use
11707 CHECKSUM_ macros and ULEB128 for DIE tag.
11708 (generate_type_signature): Use.
11710 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11712 * config.gcc (sparc*-*-*): Accept leon3 processor.
11713 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11714 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11715 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11716 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11717 (mfix-ut699): Adjust comment.
11718 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11719 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11720 (CPP_CPU_SPEC): Likewise.
11721 (ASM_CPU_SPEC): Likewise.
11722 * config/sparc/sparc.c (leon3_cost): New constant.
11723 (sparc_option_override): Add leon3 support.
11724 (mem_ref): New function.
11725 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11726 (sparc_do_work_around_errata): Look into the instruction in the delay
11727 slot and adjust accordingly. Add fix for the data cache nullify issues
11728 of the UT699. Change insertion position for the NOP.
11729 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11730 (leon3_load): New reservation.
11731 (leon_store): Bump latency to 2.
11732 (grfpu): New automaton.
11733 (grfpu_alu): New unit.
11734 (grfpu_ds): Likewise.
11735 (leon_fp_alu): Adjust.
11736 (leon_fp_mult): Delete.
11737 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11738 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11739 * config/sparc/sparc.md (cpu): Add leon3.
11740 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11741 (swapsi): Likewise.
11742 (atomic_test_and_set): Likewise.
11743 (ldstub): Likewise.
11745 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11747 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11748 default for R5900 targets.
11749 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11750 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11751 * config/mips/mips.c (mips_option_override): Report an error for
11752 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11753 for -march=r5900 -mhard-float.
11755 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11757 * df-problems.c (can_move_insns_across): Exit loop once we
11758 find a non-fixed, non-global register.
11760 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11763 * omega.c (omega_eliminate_red): Break after setting red_found.
11764 (omega_problem_has_red_equations): Similarly after setting found.
11765 (omega_query_variable): Similarly after setting coupled.
11767 2013-07-22 Marek Polacek <polacek@redhat.com>
11769 * gimplify.c: Don't include gimple.h twice.
11771 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11773 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11774 instead of TARGET_THUMB1.
11775 (Pz): New constraint.
11776 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11778 (compare_negsi_si): Likewise.
11779 (compare_addsi2_op0): Likewise.
11780 (compare_addsi2_op1): Likewise.
11781 (addsi3_carryin_<optab>): Likewise.
11782 (addsi3_carryin_alt2_<optab>): Likewise.
11783 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11784 for arm_restrict_it.
11785 (subsi3_carryin): Likewise.
11786 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11787 (minmax_arithsi): Disable for arm_restrict_it.
11788 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11789 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11790 (satsi_<SAT:code>_shift): Likewise.
11791 (arm_shiftsi3): Add alternative for 16-bit encoding.
11792 (arm32_movhf): Disable for arm_restrict_it.
11793 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11794 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11796 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11798 * config/arm/arm.md (attribute "insn"): Delete.
11799 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11800 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11801 (not_shiftsi): Update for attribute change.
11802 (not_shiftsi_compare0): Likewise.
11803 (not_shiftsi_compare0_scratch): Likewise.
11804 (arm_one_cmplsi2): Likewise.
11805 (thumb1_one_cmplsi2): Likewise.
11806 (notsi_compare0): Likewise.
11807 (notsi_compare0_scratch): Likewise.
11808 (thumb1_movdi_insn): Likewise.
11809 (arm_movsi_insn): Likewise.
11810 (movhi_insn_arch4): Likewise.
11811 (movhi_bytes): Likewise.
11812 (arm_movqi_insn): Likewise.
11813 (thumb1_movqi_insn): Likewise.
11814 (arm32_movhf): Likewise.
11815 (thumb1_movhf): Likewise.
11816 (arm_movsf_soft_insn): Likewise.
11817 (thumb1_movsf_insn): Likewise.
11818 (thumb_movdf_insn): Likewise.
11819 (movsicc_insn): Likewise.
11820 (movsfcc_soft_insn): Likewise.
11821 (and_scc): Likewise.
11822 (cond_move): Likewise.
11823 (if_move_not): Likewise.
11824 (if_not_move): Likewise.
11825 (if_shift_move): Likewise.
11826 (if_move_shift): Likewise.
11827 (if_shift_shift): Likewise.
11828 (if_not_arith): Likewise.
11829 (if_arith_not): Likewise.
11830 (cond_move_not): Likewise.
11831 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11832 (neon_mov<mode>): Likewise.
11833 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11834 (thumb2_movsi_vfp): Likewise.
11835 (movsf_vfp): Likewise.
11836 (thumb2_movsf_vfp): Likewise.
11837 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11839 (cortexa7_older_only): Likewise.
11840 (cortexa7_younger): Likewise.
11841 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11842 (1020alu_shift_op): Likewise.
11843 (1020alu_shift_reg_op): Likewise.
11844 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11845 (alu_shift_op): Likewise.
11846 (alu_shift_reg_op): Likewise.
11847 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11848 (11_alu_shift_op): Likewise.
11849 (11_alu_shift_reg_op): Likewise.
11850 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11851 (9_alu_shift_reg_op): Likewise.
11852 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11854 (cortex_a15_alu_shift): Likewise.
11855 (cortex_a15_alu_shift_reg): Likewise.
11856 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11857 (cortex_a5_alu_shift): Likewise.
11858 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11860 (cortex_a53_alu_shift): Likewise.
11861 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11863 (cortex_a7_alu_reg): Likewise.
11864 (cortex_a7_alu_shift): Likewise.
11865 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11866 (cortex_a8_alu_shift): Likewise.
11867 (cortex_a8_alu_shift_reg): Likewise.
11868 (cortex_a8_mov): Likewise.
11869 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11870 (cortex_a9_dp_shift): Likewise.
11871 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11872 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11873 (cortex_r4_mov): Likewise.
11874 (cortex_r4_alu_shift): Likewise.
11875 (cortex_r4_alu_shift_reg): Likewise.
11876 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11877 (526_alu_shift_op): Likewise.
11878 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11879 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11880 (626te_alu_shift_op): Likewise.
11881 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11882 (726te_alu_op): Likewise.
11883 (726te_alu_shift_op): Likewise.
11884 (726te_alu_shift_reg_op): Likewise.
11885 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11886 (mp626_alu_shift_op): Likewise.
11887 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11888 (pj4_alu_e1_conds): Likewise.
11889 (pj4_alu): Likewise.
11890 (pj4_alu_conds): Likewise.
11891 (pj4_shift): Likewise.
11892 (pj4_shift_conds): Likewise.
11893 (pj4_alu_shift): Likewise.
11894 (pj4_alu_shift_conds): Likewise.
11896 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11898 * config/arm/predicates.md (shiftable_operator_strict_it):
11900 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11901 Disable cond_exec version for arm_restrict_it.
11902 (thumb2_smaxsi3): Convert to generate cond_exec.
11903 (thumb2_sminsi3): Likewise.
11904 (thumb32_umaxsi3): Likewise.
11905 (thumb2_uminsi3): Likewise.
11906 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11907 (thumb2_neg_abssi2): Likewise.
11908 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11909 (thumb2_movsicc_insn): Adjust alternatives.
11910 (thumb2_mov_negscc): Disable for arm_restrict_it.
11911 (thumb2_mov_negscc_strict_it): New pattern.
11912 (thumb2_mov_notscc_strict_it): New pattern.
11913 (thumb2_mov_notscc): Disable for arm_restrict_it.
11914 (thumb2_ior_scc): Likewise.
11915 (thumb2_ior_scc_strict_it): New pattern.
11916 (thumb2_cond_move): Adjust for arm_restrict_it.
11917 (thumb2_cond_arith): Disable for arm_restrict_it.
11918 (thumb2_cond_arith_strict_it): New pattern.
11919 (thumb2_cond_sub): Adjust for arm_restrict_it.
11920 (thumb2_movcond): Likewise.
11921 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11922 (thumb2_zero_extendhisi2_v6): Likewise.
11923 (thumb2_zero_extendqisi2_v6): Likewise.
11924 (orsi_notsi_si): Likewise.
11925 (orsi_not_shiftsi_si): Likewise.
11927 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
11929 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11930 instruction sequence is 1 byte shorter.
11932 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
11934 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11935 it is not needed after split.
11937 2013-07-20 Iain Sandoe <iain@codesourcery.com>
11940 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11941 second label for nonlocal goto receivers. Don't output pic base labels
11942 unless we're producing PIC; mark that action unreachable().
11943 (ix86_save_reg): If the function contains a nonlocal label, save the
11945 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
11946 * config/darwin.c (emitted_pic_label_num): New GTY.
11947 (update_pic_label_number_if_needed): New.
11948 (machopic_output_function_base_name): Adjust for nonlocal receiver
11950 (machopic_should_output_picbase_label): New.
11951 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11952 (nonlocal_goto_receiver): New insn and split.
11954 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11956 * config/aarch64/aarch64-builtins.c
11957 (aarch64_fold_builtin): Fold abs in all modes.
11958 * config/aarch64/aarch64-simd-builtins.def
11959 (abs): Enable for all modes.
11960 * config/aarch64/arm_neon.h
11961 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11962 (vabs_f64): Add missing intrinsic.
11964 2013-07-19 Ian Bolton <ian.bolton@arm.com>
11966 * config/aarch64/arm_neon.h (vabs_s64): New function
11968 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
11971 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11972 * config/avr/avr.md (adjust_len): Add `round'.
11973 * config/avr/avr-protos.h (avr_out_round): New prototype.
11974 (avr_out_plus): Add `out_label' argument.
11975 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11976 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11977 Handle the case where `insn' is just a pattern.
11978 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11979 (avr_out_round): New function.
11980 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11982 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
11984 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11985 Add check for flag_stack_usage to handle -fstack-usage support
11987 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
11989 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11990 interaction for new Power8 flags and VSX.
11992 2013-07-18 Sriraman Tallam <tmsriram@google.com>
11994 PR middle-end/57698
11995 * tree-inline.c (expand_call_inline): Emit errors during
11996 early_inlining only if optimization is not turned on.
11998 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12002 * passes.c (init_optimization_passes): Move the construction of
12003 the pass hierarchy into a new passes.def file.
12005 * Makefile.in (passes.o): Add dependency on passes.def.
12007 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12009 * passes.c (init_optimization_passes): Introduce macros for
12010 constructing the tree of passes (INSERT_PASSES_AFTER,
12011 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12012 TERMINATE_PASS_LIST).
12014 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12015 Wei Mi <wmi@google.com>
12017 PR rtl-optimization/57878
12018 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12020 (reload_pseudo_compare_func): Check nregs first for reload
12023 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12025 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12027 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12029 * read-rtl.c (validate_const_int): Once an invalid character is
12030 seen, quit the loop.
12032 * gengtype.c (write_roots): Similarly once we find the "deletable"
12033 or "if_marked" option.
12035 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12037 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12038 "xtab" and "sat". Move value "clz" from here to ...
12039 (attriubte "type"): ... here.
12040 (satsi_<SAT:code>): Delete "insn" attribute.
12041 (satsi_<SAT:code>_shift): Likewise.
12042 (arm_zero_extendqisi2addsi): Likewise.
12043 (arm_extendqisi2addsi): Likewise.
12044 (clzsi2): Update for attribute changes.
12045 (rbitsi2): Likewise.
12046 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12048 (arm_usatsihi): Likewise.
12049 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12051 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12053 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12054 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12055 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12056 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12057 in alphabetical order.
12058 (attribute "core_cycles"): Update for attribute changes.
12059 (arm_addsi3): Likewise.
12060 (addsi3_compare0): Likewise.
12061 (addsi3_compare0_scratch): Likewise.
12062 (addsi3_compare_op1): Likewise.
12063 (addsi3_compare_op2): Likewise.
12064 (compare_addsi2_op0): Likewise.
12065 (compare_addsi2_op1): Likewise.
12066 (addsi3_carryin_shift_<optab>): Likewise.
12067 (subsi3_carryin_shift): Likewise.
12068 (rsbsi3_carryin_shift): Likewise.
12069 (arm_subsi3_insn): Likewise.
12070 (subsi3_compare0): Likewise.
12071 (subsi3_compare): Likewise.
12072 (arm_andsi3_insn): Likewise.
12073 (thumb1_andsi3_insn): Likewise.
12074 (andsi3_compare0): Likewise.
12075 (andsi3_compare0_scratch): Likewise.
12076 (zeroextractsi_compare0_scratch
12077 (andsi_not_shiftsi_si): Likewise.
12078 (iorsi3_insn): Likewise.
12079 (iorsi3_compare0): Likewise.
12080 (iorsi3_compare0_scratch): Likewise.
12081 (arm_xorsi3): Likewise.
12082 (thumb1_xorsi3_insn): Likewise.
12083 (xorsi3_compare0): Likewise.
12084 (xorsi3_compare0_scratch): Likewise.
12085 (satsi_<SAT:code>_shift): Likewise.
12087 (arm_shiftsi3): Likewise.
12088 (shiftsi3_compare0): Likewise.
12089 (not_shiftsi): Likewise.
12090 (not_shiftsi_compare0): Likewise.
12091 (not_shiftsi_compare0_scratch): Likewise.
12092 (arm_one_cmplsi2): Likewise.
12093 (thumb_one_complsi2): Likewise.
12094 (notsi_compare0): Likewise.
12095 (notsi_compare0_scratch): Likewise.
12096 (thumb1_zero_extendhisi2): Likewise.
12097 (arm_zero_extendhisi2): Likewise.
12098 (arm_zero_extendhisi2_v6): Likewise.
12099 (arm_zero_extendhisi2addsi): Likewise.
12100 (thumb1_zero_extendqisi2): Likewise.
12101 (thumb1_zero_extendqisi2_v6): Likewise.
12102 (arm_zero_extendqisi2): Likewise.
12103 (arm_zero_extendqisi2_v6): Likewise.
12104 (arm_zero_extendqisi2addsi): Likewise.
12105 (thumb1_extendhisi2): Likewise.
12106 (arm_extendhisi2): Likewise.
12107 (arm_extendhisi2_v6): Likewise.
12108 (arm_extendqisi): Likewise.
12109 (arm_extendqisi_v6): Likewise.
12110 (arm_extendqisi2addsi): Likewise.
12111 (thumb1_extendqisi2): Likewise.
12112 (thumb1_movdi_insn): Likewise.
12113 (arm_movsi_insn): Likewise.
12114 (movsi_compare0): Likewise.
12115 (movhi_insn_arch4): Likewise.
12116 (movhi_bytes): Likewise.
12117 (arm_movqi_insn): Likewise.
12118 (thumb1_movqi_insn): Likewise.
12119 (arm32_movhf): Likewise.
12120 (thumb1_movhf): Likewise.
12121 (arm_movsf_soft_insn): Likewise.
12122 (thumb1_movsf_insn): Likewise.
12123 (movdf_soft_insn): Likewise.
12124 (thumb_movdf_insn): Likewise.
12125 (arm_cmpsi_insn): Likewise.
12126 (cmpsi_shiftsi): Likewise.
12127 (cmpsi_shiftsi_swp): Likewise.
12128 (arm_cmpsi_negshiftsi_si): Likewise.
12129 (movsicc_insn): Likewise.
12130 (movsfcc_soft_insn): Likewise.
12131 (arith_shiftsi): Likewise.
12132 (arith_shiftsi_compare0
12133 (arith_shiftsi_compare0_scratch
12134 (sub_shiftsi): Likewise.
12135 (sub_shiftsi_compare0
12136 (sub_shiftsi_compare0_scratch
12137 (and_scc): Likewise.
12138 (cond_move): Likewise.
12139 (if_plus_move): Likewise.
12140 (if_move_plus): Likewise.
12141 (if_move_not): Likewise.
12142 (if_not_move): Likewise.
12143 (if_shift_move): Likewise.
12144 (if_move_shift): Likewise.
12145 (if_shift_shift): Likewise.
12146 (if_not_arith): Likewise.
12147 (if_arith_not): Likewise.
12148 (cond_move_not): Likewise.
12149 (thumb1_ashlsi3): Set type attribute.
12150 (thumb1_ashrsi3): Likewise.
12151 (thumb1_lshrsi3): Likewise.
12152 (thumb1_rotrsi3): Likewise.
12153 (shiftsi3_compare0_scratch): Likewise.
12154 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12155 (neon_mov<mode>): Likewise.
12156 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12158 (thumb2_movsi_insn): Likewise.
12159 (thumb2_cmpsi_neg_shiftsi): Likewise.
12160 (thumb2_extendqisi_v6): Likewise.
12161 (thumb2_zero_extendhisi2_v6): Likewise.
12162 (thumb2_zero_extendqisi2_v6): Likewise.
12163 (thumb2_shiftsi3_short): Likewise.
12164 (thumb2_addsi3_compare0_scratch): Likewise.
12165 (orsi_not_shiftsi_si): Likewise.
12166 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12167 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12169 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12170 (1020alu_shift_op): Likewise.
12171 (1020alu_shift_reg_op): Likewise.
12172 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12173 (alu_shift_op): Likewise.
12174 (alu_shift_reg_op): Likewise.
12175 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12176 (11_alu_shift_op): Likewise.
12177 (11_alu_shift_reg_op): Likewise.
12178 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12179 (9_alu_shift_reg_op): Likewise.
12180 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12182 (cortex_a15_alu_shift): Likewise.
12183 (cortex_a15_alu_shift_reg): Likewise.
12184 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12186 (cortex_a5_alu_shift): Likewise.
12187 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12189 (cortex_a53_alu_shift): Likewise.
12190 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12192 (cortex_a7_alu_reg): Likewise.
12193 (cortex_a7_alu_shift): Likewise.
12194 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12196 (cortex_a8_alu_shift): Likewise.
12197 (cortex_a8_alu_shift_reg): Likewise.
12198 (cortex_a8_mov): Likewise.
12199 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12200 (cortex_a9_dp_shift): Likewise.
12201 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12203 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12205 (cortex_r4_mov): Likewise.
12206 (cortex_r4_alu_shift): Likewise.
12207 (cortex_r4_alu_shift_reg): Likewise.
12208 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12209 (526_alu_shift_op): Likewise.
12210 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12211 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12212 (626te_alu_shift_op): Likewise.
12213 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12214 (726te_alu_op): Likewise.
12215 (726te_alu_shift_op): Likewise.
12216 (726te_alu_shift_reg_op): Likewise.
12217 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12218 (mp626_alu_shift_op): Likewise.
12219 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12220 (pj4_alu_e1_conds): Likewise.
12221 (pj4_alu): Likewise.
12222 (pj4_alu_conds): Likewise.
12223 (pj4_shift): Likewise.
12224 (pj4_shift_conds): Likewise.
12225 (pj4_alu_shift): Likewise.
12226 (pj4_alu_shift_conds): Likewise.
12227 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12229 (cortexa7_older_only): Likewise.
12230 (cortexa7_younger): Likewise.
12232 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12234 * ipa-pure-const.c (generate_summary): Rename to...
12235 (pure_const_generate_summary): ... this.
12237 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12239 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12241 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12244 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12247 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12249 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12250 enabled without -march=zEC12.
12251 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12254 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12256 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12257 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12258 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12259 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12260 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12261 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12262 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12264 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12265 and ISA_HAS_NMADD3_NMSUB3.
12266 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12267 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12268 (nmsub4<mode>, nmsub3<mode>): Likewise.
12269 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12271 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12273 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12274 TARGET_MIPS5400 checking.
12276 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12277 Peter Bergner <bergner@vnet.ibm.com>
12279 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12280 registers in the comment.
12281 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12282 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12283 rather than FIRST_PSEUDO_REGISTERS.
12285 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12287 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12288 enable extra ISA flags with TARGET_HTM.
12290 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12292 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12295 2013-07-15 Cong Hou <congh@google.com>
12297 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12298 in compare function for sorting.
12300 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12302 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12303 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12304 * config/rs6000/rs6000.opt: Add -mhtm option.
12305 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12306 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12307 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12308 __HTM__ if the HTM instructions are available.
12309 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12310 htm_spr_reg_operand): New define_predicates.
12311 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12312 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12314 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12315 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12316 HTM builtin functions.
12317 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12318 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12319 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12320 (rs6000_builtin_mask_calculate): Likewise.
12321 (rs6000_option_override_internal): Likewise.
12322 (bdesc_htm): Add new HTM builtin support.
12323 (htm_spr_num): New function.
12324 (htm_spr_regno): Likewise.
12325 (rs6000_htm_spr_icode): Likewise.
12326 (htm_expand_builtin): Likewise.
12327 (htm_init_builtins): Likewise.
12328 (rs6000_expand_builtin): Add support for HTM builtin functions.
12329 (rs6000_init_builtins): Likewise.
12330 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12332 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12333 (TARGET_HTM, MASK_HTM): Define macros.
12334 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12335 (FIXED_REGISTERS): Likewise.
12336 (CALL_USED_REGISTERS): Likewise.
12337 (CALL_REALLY_USED_REGISTERS): Likewise.
12338 (REG_ALLOC_ORDER): Likewise.
12339 (enum reg_class): Likewise.
12340 (REG_CLASS_NAMES): Likewise.
12341 (REG_CLASS_CONTENTS): Likewise.
12342 (REGISTER_NAMES): Likewise.
12343 (ADDITIONAL_REGISTER_NAMES): Likewise.
12344 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12345 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12346 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12347 * config/rs6000/htm.md: New file.
12348 * config/rs6000/htmintrin.h: New file.
12349 * config/rs6000/htmxlintrin.h: New file.
12351 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12353 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12354 Define SYMBOL_TINY_GOT, update comment.
12355 * config/aarch64/aarch64.c
12356 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12357 (aarch64_expand_mov_immediate): Likewise.
12358 (aarch64_print_operand): Likewise.
12359 (aarch64_classify_symbol): Likewise.
12360 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12361 (ldr_got_tiny): Define.
12363 2013-07-13 Tobias Grosser <tobias@grosser.es>
12365 PR tree-optimization/54094
12366 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12367 scheduling dimension for the parallelism check from the polyhedral
12368 information in the AST.
12369 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12372 2013-07-13 Jason Merrill <jason@redhat.com>
12374 * print-tree.c (debug_vec_tree): Use debug_raw.
12375 (debug_raw (vec<tree, va_gc> &)): New.
12376 (debug_raw (vec<tree, va_gc> *)): New.
12377 * tree.h: Declare them.
12379 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12381 * ifcvt.c (ifcvt_after_combine): New static variable.
12382 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12384 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12385 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12386 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12388 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12390 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12392 2013-07-12 Michael Matz <matz@suse.de>
12394 PR middle-end/55771
12395 * convert.c (convert_to_real): Reject non-float inner types.
12397 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12399 * config/aarch64/aarch64-protos.h
12400 (aarch64_simd_immediate_valid_for_move): Remove.
12401 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12402 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12404 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12406 2013-07-11 Steve Ellcey <sellcey@mips.com>
12408 * config/mips/mips.c (mips_conditional_register_usage): Do not
12409 use t[0-7] registers in MIPS16 mode when optimizing for size.
12411 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12413 * config/i386/i386.c (dispatch_function_versions): Fix array
12414 indexing of function_version_info to match actual_versions.
12416 2013-07-11 Teresa Johnson <tejohnson@google.com>
12418 * vec.h (struct va_gc): Move release out-of-line.
12419 (va_gc::release): Call ggc_free on released vec.
12421 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12423 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12424 Require GOT register as additional operand in UNSPEC.
12425 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12426 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12427 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12428 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12429 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12430 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12431 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12433 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12436 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12437 name seen by assembler/linker rather if available.
12439 2013-07-11 Andreas Schwab <schwab@suse.de>
12441 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12443 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12445 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12447 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12449 * doc/tm.texi.in: Move hook documentation to ....
12450 * target.def: ... here.
12452 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12453 text on @hook line.
12454 * doc/tm.texi: Regenerate.
12456 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12459 * doc/invoke.texi: Document -Wconditionally-supported.
12461 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12464 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12467 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12470 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12471 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12472 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12473 Remove duplicate devices.
12474 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12475 * config/avr/t-multilib: Regenerate.
12476 * config/avr/avr-tables.opt: Regenerate.
12477 * doc/avr-mmcu.texi: Regenerate.
12479 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12482 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12484 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12486 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12487 the cost of MULT when optimizing for size.
12489 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12491 * config/cr16/cr16-protos.h: Don't include target.h.
12493 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12495 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12496 adjust register size for TDmode and TFmode for VSX registers.
12498 2013-07-08 Kai Tietz <ktietz@redhat.com>
12501 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12502 hook_bool_const_tree_true.
12504 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12506 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12507 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12508 * config/s390/s390.md: Define FPR*_REGNUM constants.
12509 Fix FPR2_REGNUM constant (18 -> 17).
12510 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12511 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12512 ("trunc<BFP:mode><DFP_ALL:mode>2")
12513 ("trunc<DFP_ALL:mode><BFP:mode>2")
12514 ("*extend<BFP:mode><DFP_ALL:mode>2")
12515 ("*extend<DFP_ALL:mode><BFP:mode>2")
12516 ("extend<BFP:mode><DFP_ALL:mode>2")
12517 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12520 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12522 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12524 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12526 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12527 and cfun_fpr_bit_p to cfun_fpr_save_p.
12528 (s390_frame_area, s390_register_info, s390_frame_info)
12529 (s390_emit_prologue, s390_emit_epilogue)
12530 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12532 * config/s390/s390.h: Define *_REGNUM macros for floating point
12535 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12537 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12539 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12541 PR rtl-optimization/57786
12542 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12543 and break out of the loop when it is set to false.
12545 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12548 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12549 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12550 (const_int 63)) 0)).
12551 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12552 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12553 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12555 PR rtl-optimization/57829
12556 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12557 mask bits outside of mode are just sign-extension from mode to HWI.
12559 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12561 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12562 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12563 adjust_address instead of change_address to keep info about alignment.
12564 (emit_strmov): Remove.
12565 (emit_memmov): New function.
12566 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12567 (expand_movmem_epilogue): Likewise and return updated rtx for
12569 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12570 destination and source.
12571 (decide_alignment): Refactor, handle vector_loop.
12572 (ix86_expand_movmem): Likewise.
12573 (ix86_expand_setmem): Likewise.
12574 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12576 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12578 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12579 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12581 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12583 * config/i386/sse.md (sse_movlhps): Change alternative 3
12584 of operand 2 to "m".
12586 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12589 * config/i386/sse.md (iptr): New mode attribute.
12590 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12591 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12592 (<sse>_vmmul<mode>3): Ditto.
12593 (<sse>_vmdiv<mode>3): Ditto.
12594 (sse_vmrcpv4sf2): Ditto.
12595 (<sse>_vmsqrt<mode>2): Ditto.
12596 (sse_vmrsqrtv4sf2): Ditto.
12597 (<sse>_vm<code><mode>3): Ditto.
12598 (avx_vmcmp<mode>3): Ditto.
12599 (<sse>_vmmaskcmp<mode>3): Ditto.
12600 (<sse>_comi): Ditto.
12601 (<sse>_ucomi): Ditto.
12602 (*xop_vmfrcz_<mode>): Ditto.
12603 (*fmai_fmadd_<mode>): Ditto.
12604 (*fmai_fmsub_<mode>): Ditto.
12605 (*fmai_fnmadd_<mode>): Ditto.
12606 (*fmai_fnmsub_<mode>): Ditto.
12607 (*fma4i_vmfmadd_<mode>): Ditto.
12608 (*fma4i_vmfmsub_<mode>): Ditto.
12609 (*fma4i_vmfnmadd_<mode>): Ditto.
12610 (*fma4i_vmfnmsub_<mode>): Ditto.
12611 (*xop_vmfrcz_<mode>): Ditto.
12612 (sse_cvtps2pi): Ditto.
12613 (sse_cvttps2pi): Ditto.
12614 (sse_cvtss2si): Ditto.
12615 (sse_cvtss2si_2): Ditto.
12616 (sse_cvtss2siq_2): Ditto.
12617 (sse_cvttss2si): Ditto.
12618 (sse_cvttss2siq): Ditto.
12619 (sse_cvtsd2si): Ditto.
12620 (sse_cvtsd2si_2): Ditto.
12621 (sse_cvtsd2siq_2): Ditto.
12622 (sse_cvttsd2si): Ditto.
12623 (sse_cvttsd2siq): Ditto.
12624 (sse_cvtsd2ss): Ditto.
12625 (sse_cvtss2sd): Ditto.
12626 (avx2_pbroadcast<mode>): Ditto.
12627 (avx2_pbroadcast<mode>_1): Ditto.
12628 (*avx_vperm_broadcast_v4sf): Ditto.
12630 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12631 (sse_movlhps): Ditto.
12632 (sse_storehps): Ditto.
12633 (sse_loadhps): Ditto.
12634 (sse_storelps): Ditto.
12635 (sse_loadlps): Ditto.
12636 (*vec_concatv4sf): Ditto.
12637 (*vec_interleave_highv2df): Ditto.
12638 (*vec_interleave_lowv2df): Ditto.
12639 (*vec_extractv2df_1_sse): Ditto.
12640 (*vec_extractv2df_0_sse): Ditto.
12641 (sse2_storelpd): Ditto.
12642 (sse2_loadlpd): Ditto.
12643 (sse2_movsd): Ditto.
12644 (*vec_concatv4si): Ditto.
12645 (vec_concatv2di): Ditto.
12647 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12648 for Intel asm dialect.
12649 (mmx_punpcklwd): Ditto.
12650 (mmx_punpckldq): Ditto.
12652 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12653 for intel assembler dialect.
12655 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12658 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12659 for BUILT_IN_C{LZ,LRSB}*.
12660 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12661 * tree-vrp.c (extract_range_basic): Handle
12662 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12663 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12664 fall thru to code calling set_value*.
12665 * builtins.c (expand_builtin): Remove *IMAX cases.
12666 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12667 if width is bigger than 2*HWI.
12669 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12671 PR rtl-optimization/55342
12672 * lra-int.h (lra_subreg_reload_pseudos): New.
12673 * lra.c: Add undoing optional reloads to the block diagram.
12674 (lra_subreg_reload_pseudos): New.
12675 (lra_optional_reload_pseudos): Change comments.
12676 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12677 lra_optional_reload_pseudos after undo transformations.
12678 * lra-assigns.c (pseudo_prefix_title): New.
12679 (lra_setup_reg_renumber): Use it.
12680 (spill_for): Ditto. Check subreg reload pseudos too.
12681 (assign_by_spills): Consider subreg reload pseudos too.
12682 * lra-constraints.c (simplify_operand_subreg): Use
12683 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12684 (curr_insn_transform): Recognize and do optional reloads.
12685 (undo_optional_reloads): New.
12686 (lra_undo_inheritance): Call undo_optional_reloads.
12688 2013-07-05 Thomas Quinot <quinot@adacore.com>
12690 * tree-complex.c (expand_complex_operations_1): Fix typo.
12692 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12694 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12695 (tune_params): New member 'const vec_costs'.
12696 * config/aarch64/aarch64.c (generic_vector_cost): New.
12697 (generic_tunings): New member 'generic_vector_cost'.
12698 (aarch64_builtin_vectorization_cost): New.
12699 (aarch64_add_stmt_cost): New.
12700 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12701 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12703 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12706 * config/i386/predicates.md (vsib_address_operand): Disallow
12707 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12709 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12711 PR middle-end/55030
12712 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12713 expand_builtin_setjmp_receiver.
12714 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12715 with NULL for the label parameter.
12716 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12717 the frame-pointer. Adjust comments.
12718 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12719 only if LABEL is non-NULL.
12721 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12723 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12724 (ARM_ABI_AAPCS64): Ditto.
12726 (ARM_DEFAULT_ABI): Ditto.
12728 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12730 * config/aarch64/aarch64-builtins.c
12731 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12732 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12734 * config/aarch64/aarch64-simd.md
12735 (aarch64_ld1<VALL:mode>): New.
12736 (aarch64_st1<VALL:mode>): Likewise.
12737 * config/aarch64/arm_neon.h
12738 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12740 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12742 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12744 (ix86_option_override_internal):Turn on all -mavx target flags by
12745 default as they are dependent on AVX anyway.
12747 2013-07-02 Cary Coutant <ccoutant@google.com>
12749 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12750 deterministic hash.
12751 (loc_checksum_ordered): Likewise.
12752 (hash_loc_operands): Remove inline keyword.
12754 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12756 PR tree-optimization/57741
12757 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12758 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12759 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12760 Allow REAL_CST step_exprs if flag_associative_math.
12761 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12763 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12765 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12767 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12769 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12771 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12773 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12775 (iorsi3_insn): Likewise.
12776 (arm_xorsi3): Likewise.
12778 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12780 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12782 (attribute "type"): ... here, and prefix with "wmmx_".
12783 (attribute "core_cycles"): Update for attribute changes.
12784 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12785 (tbcstv4hi): Likewise.
12786 (tbcstv2si): Likewise.
12787 (iwmmxt_iordi3): Likewise.
12788 (iwmmxt_xordi3): Likewise.
12789 (iwmmxt_anddi3): Likewise.
12790 (iwmmxt_nanddi3): Likewise.
12791 (iwmmxt_arm_movdi): Likewise.
12792 (iwmmxt_movsi_insn): Likewise.
12793 (mov<mode>_internal): Likewise.
12794 (and<mode>3_iwmmxt): Likewise.
12795 (ior<mode>3_iwmmxt): Likewise.
12796 (xor<mode>3_iwmmxt): Likewise.
12797 (add<mode>3_iwmmxt): Likewise.
12798 (ssaddv8qi3): Likewise.
12799 (ssaddv4hi3): Likewise.
12800 (ssaddv2si3): Likewise.
12801 (usaddv8qi3): Likewise.
12802 (usaddv4hi3): Likewise.
12803 (usaddv2si3): Likewise.
12804 (sub<mode>3_iwmmxt): Likewise.
12805 (sssubv8qi3): Likewise.
12806 (sssubv4hi3): Likewise.
12807 (sssubv2si3): Likewise.
12808 (ussubv8qi3): Likewise.
12809 (ussubv4hi3): Likewise.
12810 (ussubv2si3): Likewise.
12811 (mulv4hi3_iwmmxt): Likewise.
12812 (smulv4hi3_highpart): Likewise.
12813 (umulv4hi3_highpart): Likewise.
12814 (iwmmxt_wmacs): Likewise.
12815 (iwmmxt_wmacsz): Likewise.
12816 (iwmmxt_wmacu): Likewise.
12817 (iwmmxt_wmacuz): Likewise.
12818 (iwmmxt_clrdi): Likewise.
12819 (iwmmxt_clrv8qi): Likewise.
12820 (iwmmxt_clr4hi): Likewise.
12821 (iwmmxt_clr2si): Likewise.
12822 (iwmmxt_uavgrndv8qi3): Likewise.
12823 (iwmmxt_uavgrndv4hi3): Likewise.
12824 (iwmmxt_uavgv8qi3): Likewise.
12825 (iwmmxt_uavgv4hi3): Likewise.
12826 (iwmmxt_tinsrb): Likewise.
12827 (iwmmxt_tinsrh): Likewise.
12828 (iwmmxt_tinsrw): Likewise.
12829 (iwmmxt_textrmub): Likewise.
12830 (iwmmxt_textrmsb): Likewise.
12831 (iwmmxt_textrmuh): Likewise.
12832 (iwmmxt_textrmsh): Likewise.
12833 (iwmmxt_textrmw): Likewise.
12834 (iwmxxt_wshufh): Likewise.
12835 (eqv8qi3): Likewise.
12836 (eqv4hi3): Likewise.
12837 (eqv2si3): Likewise.
12838 (gtuv8qi3): Likewise.
12839 (gtuv4hi3): Likewise.
12840 (gtuv2si3): Likewise.
12841 (gtv8qi3): Likewise.
12842 (gtv4hi3): Likewise.
12843 (gtv2si3): Likewise.
12844 (smax<mode>3_iwmmxt): Likewise.
12845 (umax<mode>3_iwmmxt): Likewise.
12846 (smin<mode>3_iwmmxt): Likewise.
12847 (umin<mode>3_iwmmxt): Likewise.
12848 (iwmmxt_wpackhss): Likewise.
12849 (iwmmxt_wpackwss): Likewise.
12850 (iwmmxt_wpackdss): Likewise.
12851 (iwmmxt_wpackhus): Likewise.
12852 (iwmmxt_wpackwus): Likewise.
12853 (iwmmxt_wpackdus): Likewise.
12854 (iwmmxt_wunpckihb): Likewise.
12855 (iwmmxt_wunpckihh): Likewise.
12856 (iwmmxt_wunpckihw): Likewise.
12857 (iwmmxt_wunpckilb): Likewise.
12858 (iwmmxt_wunpckilh): Likewise.
12859 (iwmmxt_wunpckilw): Likewise.
12860 (iwmmxt_wunpckehub): Likewise.
12861 (iwmmxt_wunpckehuh): Likewise.
12862 (iwmmxt_wunpckehuw): Likewise.
12863 (iwmmxt_wunpckehsb): Likewise.
12864 (iwmmxt_wunpckehsh): Likewise.
12865 (iwmmxt_wunpckehsw): Likewise.
12866 (iwmmxt_wunpckelub): Likewise.
12867 (iwmmxt_wunpckeluh): Likewise.
12868 (iwmmxt_wunpckeluw): Likewise.
12869 (iwmmxt_wunpckelsb): Likewise.
12870 (iwmmxt_wunpckelsh): Likewise.
12871 (iwmmxt_wunpckelsw): Likewise.
12872 (ror<mode>3): Likewise.
12873 (ashr<mode>3_iwmmxt): Likewise.
12874 (lshr<mode>3_iwmmxt): Likewise.
12875 (ashl<mode>3_iwmmxt): Likewise.
12876 (ror<mode>3_di): Likewise.
12877 (ashr<mode>3_di): Likewise.
12878 (lshr<mode>3_di): Likewise.
12879 (ashl<mode>3_di): Likewise.
12880 (iwmmxt_wmadds): Likewise.
12881 (iwmmxt_wmaddu): Likewise.
12882 (iwmmxt_tmia): Likewise.
12883 (iwmmxt_tmiaph): Likewise.
12884 (iwmmxt_tmiabb): Likewise.
12885 (iwmmxt_tmiatb): Likewise.
12886 (iwmmxt_tmiabt): Likewise.
12887 (iwmmxt_tmiatt): Likewise.
12888 (iwmmxt_tmovmskb): Likewise.
12889 (iwmmxt_tmovmskh): Likewise.
12890 (iwmmxt_tmovmskw): Likewise.
12891 (iwmmxt_waccb): Likewise.
12892 (iwmmxt_wacch): Likewise.
12893 (iwmmxt_waccw): Likewise.
12894 (iwmmxt_waligni): Likewise.
12895 (iwmmxt_walignr): Likewise.
12896 (iwmmxt_walignr0): Likewise.
12897 (iwmmxt_walignr1): Likewise.
12898 (iwmmxt_walignr2): Likewise.
12899 (iwmmxt_walignr3): Likewise.
12900 (iwmmxt_wsadb): Likewise.
12901 (iwmmxt_wsadh): Likewise.
12902 (iwmmxt_wsadbz): Likewise.
12903 (iwmmxt_wsadhz): Likewise.
12904 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12905 (iwmmxt_wabsdiffb): Likewise.
12906 (iwmmxt_wabsdiffh): Likewise.
12907 (iwmmxt_wabsdiffw): Likewise.
12908 (iwmmxt_waddsubhx): Likewise
12909 (iwmmxt_wsubaddhx): Likewise.
12910 (addc<mode>3): Likewise.
12911 (iwmmxt_avg4): Likewise.
12912 (iwmmxt_avg4r): Likewise.
12913 (iwmmxt_wmaddsx): Likewise.
12914 (iwmmxt_wmaddux): Likewise.
12915 (iwmmxt_wmaddsn): Likewise.
12916 (iwmmxt_wmaddun): Likewise.
12917 (iwmmxt_wmulwsm): Likewise.
12918 (iwmmxt_wmulwum): Likewise.
12919 (iwmmxt_wmulsmr): Likewise.
12920 (iwmmxt_wmulumr): Likewise.
12921 (iwmmxt_wmulwsmr): Likewise.
12922 (iwmmxt_wmulwumr): Likewise.
12923 (iwmmxt_wmulwl): Likewise.
12924 (iwmmxt_wqmulm): Likewise.
12925 (iwmmxt_wqmulwm): Likewise.
12926 (iwmmxt_wqmulmr): Likewise.
12927 (iwmmxt_wqmulwmr): Likewise.
12928 (iwmmxt_waddbhusm): Likewise.
12929 (iwmmxt_waddbhusl): Likewise.
12930 (iwmmxt_wqmiabb): Likewise.
12931 (iwmmxt_wqmiabt): Likewise.
12932 (iwmmxt_wqmiatb): Likewise.
12933 (iwmmxt_wqmiatt): Likewise.
12934 (iwmmxt_wqmiabbn): Likewise.
12935 (iwmmxt_wqmiabtn): Likewise.
12936 (iwmmxt_wqmiatbn): Likewise.
12937 (iwmmxt_wqmiattn): Likewise.
12938 (iwmmxt_wmiabb): Likewise.
12939 (iwmmxt_wmiabt): Likewise.
12940 (iwmmxt_wmiatb): Likewise.
12941 (iwmmxt_wmiatt): Likewise.
12942 (iwmmxt_wmiabbn): Likewise.
12943 (iwmmxt_wmiabtn): Likewise.
12944 (iwmmxt_wmiatbn): Likewise.
12945 (iwmmxt_wmiattn): Likewise.
12946 (iwmmxt_wmiawbb): Likewise.
12947 (iwmmxt_wmiawbt): Likewise.
12948 (iwmmxt_wmiawtb): Likewise.
12949 (iwmmxt_wmiawtt): Likewise.
12950 (iwmmxt_wmiawbbn): Likewise.
12951 (iwmmxt_wmiawbtn): Likewise.
12952 (iwmmxt_wmiawtbn): Likewise.
12953 (iwmmxt_wmiawttn): Likewise.
12954 (iwmmxt_wmerge): Likewise.
12955 (iwmmxt_tandc<mode>3): Likewise.
12956 (iwmmxt_torc<mode>3): Likewise.
12957 (iwmmxt_torvsc<mode>3): Likewise.
12958 (iwmmxt_textrc<mode>3): Likewise.
12959 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12960 (wmmxt_pack): Likewise.
12961 (wmmxt_mult_c1): Likewise.
12962 (wmmxt_mult_c2): Likewise.
12963 (wmmxt_alu_c1): Likewise.
12964 (wmmxt_alu_c2): Likewise.
12965 (wmmxt_alu_c3): Likewise.
12966 (wmmxt_transfer_c1): Likewise.
12967 (wmmxt_transfer_c2): Likewise.
12968 (wmmxt_transfer_c3): Likewise.
12969 (marvell_f_iwmmxt_wstr): Likewise.
12970 (marvell_f_iwmmxt_wldr): Likewise.
12972 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12974 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
12976 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12979 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12980 * lra-constraints.c (need_for_split_p): Check call used hard regs
12981 living through calls.
12983 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12984 call used regs for call insn.
12986 2013-06-28 Jakub Jelinek <jakub@redhat.com>
12989 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12990 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
12992 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12994 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
12996 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12998 * lra-constraints.c (need_for_split_p): Check call used hard regs
12999 living through calls.
13001 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13004 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13005 to tie with any other modes. Eliminate Altivec vector mode tests,
13006 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13007 code, to return 0 if testing MODE2 for a condition, if we've
13008 already tested MODE1 for the same condition.
13010 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13012 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13015 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13017 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13018 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13020 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13022 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13024 (aarch64_symbolic_constant_p): Remove.
13025 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13026 static. Fix line length and white space.
13027 (aarch64_symbolic_constant_p): Remove.
13028 * config/aarch64/predicates.md (aarch64_valid_symref):
13029 Use aarch64_classify_symbol_expression.
13031 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13033 * config/arm/constraints.md (Ts): New constraint.
13034 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13036 (compare_scc): Use "Ts" constraint for operand 0.
13037 (ior_scc_scc): Likewise.
13038 (and_scc_scc): Likewise.
13039 (and_scc_scc_nodom): Likewise.
13040 (ior_scc_scc_cmp): Likewise for operand 7.
13041 (and_scc_scc_cmp): Likewise.
13042 * config/arm/thumb2.md (thumb2_movsi_insn):
13043 Add alternatives for 16-bit encodings.
13044 (thumb2_movhi_insn): Likewise.
13045 (thumb2_movsicc_insn): Likewise.
13046 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13047 (thumb2_negscc): Use "Ts" constraint.
13048 Move mvn instruction outside cond_exec block.
13049 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13050 for 16-bit encodings.
13052 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13054 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13056 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13057 (mulsi3subsi): Likewise.
13058 (mulsidi3adddi): Likewise.
13059 (mulsidi3_v6): Likewise.
13060 (umulsidi3_v6): Likewise.
13061 (umulsidi3adddi_v6): Likewise.
13062 (smulsi3_highpart_v6): Likewise.
13063 (umulsi3_highpart_v6): Likewise.
13064 (mulhisi3tb): Likewise.
13065 (mulhisi3bt): Likewise.
13066 (mulhisi3tt): Likewise.
13067 (maddhisi4): Likewise.
13068 (maddhisi4tb): Likewise.
13069 (maddhisi4tt): Likewise.
13070 (maddhidi4): Likewise.
13071 (maddhidi4tb): Likewise.
13072 (maddhidi4tt): Likewise.
13073 (zeroextractsi_compare0_scratch): Likewise.
13074 (insv_zero): Likewise.
13075 (insv_t2): Likewise.
13076 (anddi_notzesidi_di): Likewise.
13077 (anddi_notsesidi_di): Likewise.
13078 (andsi_notsi_si): Likewise.
13079 (iordi_zesidi_di): Likewise.
13080 (xordi_zesidi_di): Likewise.
13081 (andsi_iorsi3_notsi): Likewise.
13082 (smax_0): Likewise.
13083 (smax_m1): Likewise.
13084 (smin_0): Likewise.
13085 (not_shiftsi): Likewise.
13086 (unaligned_loadsi): Likewise.
13087 (unaligned_loadhis): Likewise.
13088 (unaligned_loadhiu): Likewise.
13089 (unaligned_storesi): Likewise.
13090 (unaligned_storehi): Likewise.
13091 (extv_reg): Likewise.
13092 (extzv_t2): Likewise.
13093 (divsi3): Likewise.
13094 (udivsi3): Likewise.
13095 (arm_zero_extendhisi2addsi): Likewise.
13096 (arm_zero_extendqisi2addsi): Likewise.
13097 (compareqi_eq0): Likewise.
13098 (arm_extendhisi2_v6): Likewise.
13099 (arm_extendqisi2addsi): Likewise.
13100 (arm_movt): Likewise.
13101 (thumb2_ldrd): Likewise.
13102 (thumb2_ldrd_base): Likewise.
13103 (thumb2_ldrd_base_neg): Likewise.
13104 (thumb2_strd): Likewise.
13105 (thumb2_strd_base): Likewise.
13106 (thumb2_strd_base_neg): Likewise.
13107 (arm_negsi2): Add alternative for 16-bit encoding.
13108 (arm_one_cmplsi2): Likewise.
13110 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13112 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13113 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13114 (movdfcc): Likewise.
13115 * config/arm/vfp.md (*thumb2_movsf_vfp):
13116 Disable predication for arm_restrict_it.
13117 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13118 (*thumb2_movdfcc_vfp): Likewise.
13119 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13120 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13121 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13122 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13123 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13124 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13125 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13126 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13127 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13128 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13129 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13130 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13131 Disable predication for arm_restrict_it.
13133 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13135 * config/i386/bmiintrin.h (_bextr_u32): New.
13136 (_bextr_u64): Ditto.
13138 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13140 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13141 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13142 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13143 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13144 * config/mips/n32-elf.h: ...this new file.
13146 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13149 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13150 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13152 2013-06-27 Catherine Moore <clm@codesourcery.com>
13154 * config/mips/mips-tables.opt: Regenerate.
13155 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13156 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13157 * doc/invoke.texi: Add -m14kc.
13159 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13162 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13163 constraints of operand 1 and 2.
13166 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13167 to match RTL canonicalization. Swap predicates and
13168 constraints of operand 1 and 2.
13170 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13172 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13173 Process OP_INOUT regs for splitting too.
13175 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13177 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13178 decl before the loop, initialize to NULL.
13179 (vectorizable_load): Initialize ptr_incr to NULL.
13181 2013-06-27 Martin Jambor <mjambor@suse.cz>
13184 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13185 * ipa-ref.c (ipa_maybe_record_reference): New function.
13186 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13187 * ipa-cp.c (create_specialized_node): Record potential references from
13189 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13191 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13193 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13194 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13195 'mode' to force_reg.
13196 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13197 (aarch64_expand_mov_immediate): Likewise.
13199 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13201 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13202 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13204 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13206 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13207 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13208 (struct machine_function): Add tbegin_p.
13209 (s390_canonicalize_comparison): Fold CC mode compares to
13210 conditional jump if possible.
13211 (s390_emit_jump): Return the emitted jump.
13212 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13213 Handle CCRAWmode compares.
13214 (s390_option_override): Default to -mhtm if available.
13215 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13216 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13217 FPRs instead of df_regs_ever_live_p.
13218 (s390_optimize_nonescaping_tx): New function.
13219 (s390_init_frame_layout): Extend clobbered_regs array to cover
13221 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13222 (s390_expand_tbegin): New function.
13223 (enum s390_builtin): New enum definition.
13224 (code_for_builtin): New array definition.
13225 (s390_init_builtins): New function.
13226 (s390_expand_builtin): New function.
13227 (TARGET_INIT_BUILTINS): Define.
13228 (TARGET_EXPAND_BUILTIN): Define.
13229 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13230 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13231 (s390_alc_comparison): Likewise.
13232 * config/s390/s390-modes.def: Add CCRAWmode.
13233 * config/s390/s390.h (processor_flags): Add PF_TX.
13234 (TARGET_CPU_HTM): Define macro.
13235 (TARGET_HTM): Define macro.
13236 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13237 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13238 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13239 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13240 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13241 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13242 ("*cc_to_int"): Move up.
13243 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13244 constants other than 0.
13245 ("*ccraw_to_int"): New insn and splitter definition.
13246 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13247 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13248 ("tx_assist"): New expander.
13249 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13250 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13251 * config/s390/s390.opt: Add -mhtm option.
13252 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13253 * config/s390/htmxlintrin.h: New file.
13254 * config/s390/htmintrin.h: New file.
13255 * config/s390/s390intrin.h: New file.
13256 * doc/extend.texi: Document htm builtins.
13257 * config.gcc: Add the new header files to extra_headers.
13259 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13261 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13262 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13264 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13265 Pat Haugen <pthaugen@us.ibm.com>
13266 Peter Bergner <bergner@vnet.ibm.com>
13268 * config/rs6000/power8.md: New.
13269 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13270 setting for power8 entry.
13271 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13272 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13273 test for Power4/Power5 only.
13274 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13276 (force_new_group): Adjust comment.
13277 * config/rs6000/rs6000.md: Include power8.md.
13279 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13281 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13282 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13284 (tune_params): Update comment.
13285 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13286 (arm_max_conditional_execute): New function.
13287 (thumb2_final_prescan_insn): Use max_insn_skipped and
13288 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13290 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13292 PR tree-optimization/57705
13293 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13294 SSA_NAME step, provided that it is not defined inside the loop.
13295 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13296 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13298 2013-06-25 Martin Jambor <mjambor@suse.cz>
13300 PR middle-end/57670
13301 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13302 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13304 (ipa_note_param_call): Initialize member_ptr flag.
13305 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13306 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13307 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13308 (ipa_read_indirect_edge_info): Likewise.
13310 2013-06-25 Richard Biener <rguenther@suse.de>
13312 PR middle-end/56977
13313 * passes.c (init_optimization_passes): Move pass_fold_builtins
13314 and pass_dce earlier with -Og.
13316 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13318 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13319 <BIT_FIELD_REF>: Remove trailing TAB.
13320 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13323 2013-06-24 Martin Jambor <mjambor@suse.cz>
13325 PR tree-optimization/57358
13326 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13327 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13328 (ipa_analyze_params_uses): Generate pessimistic info when true.
13330 2013-06-24 Martin Jambor <mjambor@suse.cz>
13332 PR tree-optimization/57539
13333 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13334 global.inlined_to of the new node to it. All callers changed.
13335 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13336 inlining_into, pass it to cgraph_clone_node.
13337 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13338 ipa_free_edge_args_substructures.
13339 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13340 rdesc linked list. Do not assert rdesc edges have inlined caller.
13341 Assert we have found an rdesc in the rdesc list.
13343 2013-06-24 Richard Biener <rguenther@suse.de>
13345 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13346 (pointer_set_lookup): Declare.
13347 (class pointer_map): New template class implementing a
13348 generic pointer to T map.
13349 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13350 pointer_map<T>::contains, pointer_map<T>::insert,
13351 pointer_map<T>::traverse): New functions.
13352 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13353 (pointer_set_lookup): New function.
13354 (pointer_set_contains): Use pointer_set_lookup.
13355 (pointer_set_insert): Likewise.
13356 (insert_aux): Remove.
13357 (struct pointer_map_t): Embed a pointer_set_t.
13358 (pointer_map_create): Adjust.
13359 (pointer_map_destroy): Likewise.
13360 (pointer_map_contains): Likewise.
13361 (pointer_map_insert): Likewise.
13362 (pointer_map_traverse): Likewise.
13363 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13364 pointer_map<unsigned> instead of a pointer_map_t.
13365 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13366 (streamer_tree_cache_lookup): Likewise.
13367 (streamer_tree_cache_create): Likewise.
13368 (streamer_tree_cache_delete): Likewise.
13369 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13370 pointer_map<unsigned> instead of a pointer_map_t.
13371 (lto_init_tree_ref_encoder): Adjust.
13372 (lto_destroy_tree_ref_encoder): Likewise.
13373 * lto-section-out.c (lto_output_decl_index): Likewise.
13374 (lto_record_function_out_decl_state): Likewise.
13375 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13377 2013-06-24 Richard Biener <rguenther@suse.de>
13379 PR tree-optimization/57488
13380 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13382 2013-06-24 Alan Modra <amodra@gmail.com>
13384 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13385 (gen_easy_altivec_constant): Likewise.
13386 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13387 easy_vector_constant_msb): Likewise.
13389 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13392 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13393 add missing return true.
13395 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13398 * config/sh/predicates.md (general_extend_operand): Invoke
13399 general_movsrc_operand for memory operands.
13400 (general_movsrc_operand): Allow reg+reg addressing, do not use
13401 general_operand for memory operands.
13403 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13405 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13406 when current target options does not apply.
13407 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13408 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13409 * config/i386/bmiintrin.h: Pass appropriate target
13410 attributes to header.
13411 * config/i386/mmintrin.h: Ditto.
13412 * config/i386/nmmintrin.h: Ditto.
13413 * config/i386/avx2intrin.h: Ditto.
13414 * config/i386/fxsrintrin.h: Ditto.
13415 * config/i386/tbmintrin.h: Ditto.
13416 * config/i386/xsaveintrin.h: Ditto.
13417 * config/i386/f16cintrin.h: Ditto.
13418 * config/i386/xtestintrin.h: Ditto.
13419 * config/i386/xsaveoptintrin.h: Ditto.
13420 * config/i386/bmi2intrin.h: Ditto.
13421 * config/i386/lzcntintrin.h: Ditto.
13422 * config/i386/smmintrin.h: Ditto.
13423 * config/i386/wmmintrin.h: Ditto.
13424 * config/i386/x86intrin.h: Remove all header include guards.
13425 * config/i386/prfchwintrin.h: Ditto.
13426 * config/i386/pmmintrin.h: Ditto.
13427 * config/i386/tmmintrin.h: Ditto.
13428 * config/i386/xmmintrin.h: Ditto.
13429 * config/i386/popcntintrin.h: Ditto.
13430 * config/i386/rdseedintrin.h: Ditto.
13431 * config/i386/ammintrin.h: Ditto.
13432 * config/i386/emmintrin.h: Ditto.
13433 * config/i386/immintrin.h: Remove all header include guards.
13434 * config/i386/fma4intrin.h: Ditto.
13435 * config/i386/lwpintrin.h: Ditto.
13436 * config/i386/xopintrin.h: Ditto.
13437 * config/i386/ia32intrin.h: Ditto.
13438 * config/i386/avxintrin.h: Ditto.
13439 * config/i386/rtmintrin.h: Ditto.
13440 * config/i386/fmaintrin.h: Ditto.
13441 * config/i386/mm3dnow.h: Ditto.
13443 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13445 * common/config/i386/i386-common.c: Handle LZCNT.
13447 2013-06-22 Andi Kleen <ak@linux.intel.com>
13449 * doc/extend.texi: Use __atomic_store_n instead of
13450 __atomic_store in HLE example.
13452 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13454 * config/sh/sh.c: Remove <cstdlib> workaround.
13456 2013-06-21 Andi Kleen <ak@linux.intel.com>
13458 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13460 2013-06-21 Andi Kleen <ak@linux.intel.com>
13462 * doc/extend.texi: Document that __atomic_clear and
13463 __atomic_test_and_set should only be used with bool.
13465 2013-06-20 Jan Hubicka <jh@suse.cz>
13467 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13468 types_same_for_odr.
13469 * tree.c (decls_same_for_odr): New function.
13470 (same_for_edr): New function.
13471 (types_same_for_odr): New function.
13472 (get_binfo_at_offset): Use it.
13473 * tree.h (types_same_for_odr): Declare.
13475 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13476 Jason Merrill <jason@redhat.com>
13478 * system.h: Include <cstdlib> as well as <stdlib.h>.
13480 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13483 * config/i386/i386.c (construct_container): Report error if
13484 long double is used with disabled x87 float returns.
13486 2013-06-20 Jan Hubicka <jh@suse.cz>
13488 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13490 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13492 PR rtl-optimization/57425
13493 PR rtl-optimization/57569
13494 * alias.c (write_dependence_p): Remove parameters mem_mode and
13495 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13496 Changed all callers.
13497 (canon_anti_dependence): Get comments and semantics in sync.
13498 Add parameter mem_canonicalized. Changed all callers.
13499 * rtl.h (canon_anti_dependence): Update prototype.
13501 2013-06-20 Richard Biener <rguenther@suse.de>
13503 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13504 byte case, inline streamer_read_uchar and defer section
13507 2013-06-20 Richard Biener <rguenther@suse.de>
13509 PR tree-optimization/57584
13510 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13511 SSA names into the expanded expression that take part in
13512 abnormal coalescing.
13514 2013-06-19 Sharad Singhai <singhai@google.com>
13516 * gcov.c (print_usage): Handle new option.
13517 (process_args): Ditto.
13518 (get_gcov_intermediate_filename): New function.
13519 (output_intermediate_file): New function.
13520 (output_gcov_file): New function
13521 (generate_results): Handle new option.
13522 (release_function): Relase demangled name.
13523 (read_graph_file): Handle demangled name.
13524 (output_lines): Ditto.
13525 * doc/gcov.texi: Document gcov intermediate format.
13527 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13530 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13531 (lra_emit_add): Use the functions. Add comment about Y as an
13534 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13537 * collect2.c (collect_atexit): New.
13538 (collect_exit): Delete.
13539 (main): Register collect_atexit with atexit.
13540 (collect_wait): Change collect_exit to exit.
13542 * collect2.h (collect_exit): Delete.
13543 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13545 2013-06-19 Wei Mi <wmi@google.com>
13547 PR rtl-optimization/57518
13548 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13549 if regno is used in paradoxical subreg.
13550 (update_equiv_regs): Check pdx_subregs[regno] before
13551 set a reg to be equivalent with a mem.
13553 2013-06-19 Matthias Klose <doko@ubuntu.com>
13556 * file-find.h (find_a_file): Add a mode parameter.
13557 * file-find.c (find_a_file): Likewise.
13558 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13559 with X_OK for the executables.
13560 * collect2.c (main): Call find_a_file with X_OK.
13562 2013-06-19 Steve Ellcey <sellcey@mips.com>
13565 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13566 Use NEXT_INSN instead of next_real_insn.
13568 2013-06-19 Jan Hubicka <jh@suse.cz>
13570 * cgraph.h (const_value_known_p): Replace by ...
13571 (ctor_for_folding): .. this one.
13572 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13573 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13574 * expr.c (expand_expr_real_1): Likewise.
13575 (string_constant): Likewise.
13576 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13577 * ipa.c (process_references): Likewise.
13578 (symtab_remove_unreachable_nodes): Likewise.
13579 * ipa-inline-analysis.c (param_change_prob): Likewise.
13580 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13581 (get_base_constructor): Likwise.
13582 * varpool.c (varpool_remove_node): Likewise.
13583 (varpool_remove_initializer): LIkewise.
13584 (dump_varpool_node): LIkwise.
13585 (const_value_known_p): Rewrite to ...
13586 (ctor_for_folding): ... this one.
13588 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13591 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13592 PERSONALITY in $PATH derived prefixes.
13594 2013-06-19 Jeff Law <law@redhat.com>
13596 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13599 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13600 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13603 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13605 * config/arm/vfp.md: Move VFP instruction classification documentation
13607 * config/arm/arm.md: ... here. Update instruction classification
13610 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13612 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13613 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13614 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13616 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13619 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13620 with NEXT_INSN. Use tablejump_p to check for jump table data
13623 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13626 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13627 that now in C++ the value is correct per the C++ standards.
13629 2013-06-19 Richard Biener <rguenther@suse.de>
13631 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13632 for global context.
13634 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13637 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13640 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13641 with next_active_insn.
13643 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13645 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13646 functions are inlined during failures to flag an error.
13647 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13648 in early inline pass.
13650 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13652 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13655 2013-06-18 Julian Brown <julian@codesourcery.com>
13657 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13658 Permit virtual register pre-reload if !strict.
13659 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13661 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13663 * config/arm/neon.md (movmisalign<mode>): Use
13664 neon_perm_struct_or_reg_operand instead of
13665 neon_struct_or_register_operand.
13666 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13667 neon_permissive_struct_operand instead of neon_struct_operand.
13668 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13669 neon_vector_mem_operand.
13670 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13671 neon_vector_mem_operand.
13672 (neon_permissive_struct_operand): New.
13673 (neon_struct_or_register_operand): Rename to...
13674 (neon_perm_struct_or_reg_operand): This. Adjust call to
13675 neon_vector_mem_operand.
13677 2013-06-18 Richard Biener <rguenther@suse.de>
13679 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13680 * lto-streamer.h: Include pointer-set.h.
13681 (struct lto_decl_slot): Remove.
13682 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13683 Remove next_index entry.
13684 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13685 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13686 (lto_init_tree_ref_encoder): Adjust.
13687 (lto_destroy_tree_ref_encoder): Likewise.
13688 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13689 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13690 (lto_output_decl_index): Adjust.
13691 (lto_new_out_decl_state): Likewise.
13692 (lto_record_function_out_decl_state): Likewise.
13693 * lto-streamer-out.c (copy_function): Likewise.
13695 2013-06-18 Richard Biener <rguenther@suse.de>
13697 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13698 * cgraphunit.c: Include cfgloop.h.
13699 (init_lowered_empty_function): Initialize the loop tree.
13700 (assemble_thunk): Insert new BBs into loops.
13702 2013-06-18 Richard Biener <rguenther@suse.de>
13704 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13705 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13706 the map from cache entry to cache index optional.
13707 (streamer_tree_cache_replace_tree): Adjust accordingly.
13708 (streamer_tree_cache_append): Likewise.
13709 (streamer_tree_cache_delete): Likewise.
13710 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13711 streamer cache map from cache entry to cache index.
13712 * lto-streamer-out.c (create_output_block): Adjust.
13714 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13716 * config/arm/arm.md (attribute "insn"): Move multiplication and
13717 division attributes to...
13718 (attribute "type"): ... here. Remove mult.
13719 (attribute "mul32"): New attribute.
13720 (attribute "mul64"): Add umaal.
13721 (*arm_mulsi3): Update attributes.
13722 (*arm_mulsi3_v6): Likewise.
13723 (*thumb_mulsi3): Likewise.
13724 (*thumb_mulsi3_v6): Likewise.
13725 (*mulsi3_compare0): Likewise.
13726 (*mulsi3_compare0_v6): Likewise.
13727 (*mulsi_compare0_scratch): Likewise.
13728 (*mulsi_compare0_scratch_v6): Likewise.
13729 (*mulsi3addsi): Likewise.
13730 (*mulsi3addsi_v6): Likewise.
13731 (*mulsi3addsi_compare0): Likewise.
13732 (*mulsi3addsi_compare0_v6): Likewise.
13733 (*mulsi3addsi_compare0_scratch): Likewise.
13734 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13735 (*mulsi3subsi): Likewise.
13736 (*mulsidi3adddi): Likewise.
13737 (*mulsi3addsi_v6): Likewise.
13738 (*mulsidi3adddi_v6): Likewise.
13739 (*mulsidi3_nov6): Likewise.
13740 (*mulsidi3_v6): Likewise.
13741 (*umulsidi3_nov6): Likewise.
13742 (*umulsidi3_v6): Likewise.
13743 (*umulsidi3adddi): Likewise.
13744 (*umulsidi3adddi_v6): Likewise.
13745 (*smulsi3_highpart_nov6): Likewise.
13746 (*smulsi3_highpart_v6): Likewise.
13747 (*umulsi3_highpart_nov6): Likewise.
13748 (*umulsi3_highpart_v6): Likewise.
13749 (mulhisi3): Likewise.
13750 (*mulhisi3tb): Likewise.
13751 (*mulhisi3bt): Likewise.
13752 (*mulhisi3tt): Likewise.
13753 (maddhisi4): Likewise.
13754 (*maddhisi4tb): Likewise.
13755 (*maddhisi4tt): Likewise.
13756 (maddhidi4): Likewise.
13757 (*maddhidi4tb): Likewise.
13758 (*maddhidi4tt): Likewise.
13759 (divsi3): Likewise.
13760 (udivsi3): Likewise.
13761 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13762 (thumb2_mulsi_short_compare0): Likewise.
13763 (thumb2_mulsi_short_compare0_scratch): Likewise.
13764 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13765 (1020mult2): Likewise.
13766 (1020mult3): Likewise.
13767 (1020mult4): Likewise.
13768 (1020mult5): Likewise.
13769 (1020mult6): Likewise.
13770 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13772 (cortex_a15_mult64): Likewise.
13773 (cortex_a15_sdiv): Likewise.
13774 (cortex_a15_udiv): Likewise.
13775 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13781 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13782 (pj4_ir_div): Likewise.
13783 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13784 (11_mult2): Likewise.
13785 (11_mult3): Likewise.
13786 (11_mult4): Likewise.
13787 (11_mult5): Likewise.
13788 (11_mult6): Likewise.
13789 (11_mult7): Likewise.
13790 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13791 (cortex_a8_mla): Likewise.
13792 (cortex_a8_mull): Likewise.
13793 (cortex_a8_smulwy): Likewise.
13794 (cortex_a8_smlald): Likewise.
13795 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13796 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13797 (cortex_r4_mul_3): Likewise.
13798 (cortex_r4_mla_4): Likewise.
13799 (cortex_r4_mla_3): Likewise.
13800 (cortex_r4_smlald): Likewise.
13801 (cortex_r4_mull): Likewise.
13802 (cortex_r4_sdiv): Likewise.
13803 (cortex_r4_udiv): Likewise.
13804 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13805 (cortex_a7_idiv): Likewise.
13806 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13807 (9_mult2): Likewise.
13808 (9_mult3): Likewise.
13809 (9_mult4): Likewise.
13810 (9_mult5): Likewise.
13811 (9_mult6): Likewise.
13812 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13813 (cortex_a53_sdiv): Likewise.
13814 (cortex_a53_udiv): Likewise.
13815 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13816 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13817 (mp626_mult2): Likewise.
13818 (mp626_mult3): Likewise.
13819 (mp626_mult4): Likewise.
13820 * config/arm/fa526.md (526_mult1): Update attribute change.
13821 (526_mult2): Likewise.
13822 * config/arm/arm-generic.md (mult): Update attribute change.
13823 (mult_ldsched_strongarm): Likewise.
13824 (mult_ldsched): Likewise.
13825 (multi_cycle): Likewise.
13826 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13827 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13828 (606te_mult2): Likewise.
13829 (606te_mult3): Likewise.
13830 (606te_mult4): Likewise.
13831 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13832 (cortex_a9_mac16): Likewise.
13833 (cortex_a9_multiply): Likewise.
13834 (cortex_a9_mac): Likewise.
13835 (cortex_a9_multiply_long): Likewise.
13836 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13837 (626te_mult2): Likewise.
13838 (626te_mult3): Likewise.
13839 (626te_mult4): Likewise.
13841 2013-06-18 Richard Biener <rguenther@suse.de>
13844 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13846 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13849 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13850 with next_active_insn.
13852 2013-06-18 Alan Modra <amodra@gmail.com>
13854 * config/rs6000/rs6000.h (enum data_align): New.
13855 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13856 (DATA_ABI_ALIGNMENT): Define.
13857 (CONSTANT_ALIGNMENT): Correct comment.
13858 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13859 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13861 2013-06-17 David Malcolm <dmalcolm@redhat.com>
13863 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13864 ATTRIBUTE_UNUSED marking.
13866 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13868 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13869 alternative and update.
13870 (aarch64_dup_lanedi): Delete.
13871 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13872 * config/aarch64/aarch64-simd-builtins.def: Update.
13874 2013-06-17 Richard Biener <rguenther@suse.de>
13876 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13877 (lto_input_scc): Declare.
13878 (lto_input_tree_1): Likewise.
13879 (struct lto_stats_d): Add num_tree_bodies_output and
13880 num_pickle_refs_output.
13881 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13882 (lto_read_tree_1): Split out from ...
13883 (lto_read_tree): ... this.
13884 (lto_input_scc): New function.
13885 (lto_input_tree_1): Split out from ...
13886 (lto_input_tree): ... this. Handle LTO_tree_scc.
13887 (lto_data_in_create): Create the streamer cache without hashes.
13888 * lto-streamer-out.c (create_output_block): Create the streamer
13889 cache with hashes when not doing WPA.
13890 (lto_write_tree_1): Split out from ...
13891 (lto_write_tree): ... this.
13892 (get_symbol_initial_value): New function.
13893 (lto_output_tree_1): Split out from ...
13894 (lto_output_tree): ... this. Write trees as series of SCCs
13895 using a DFS walk via DFS_write_tree.
13896 (struct sccs, struct scc_entry): New types.
13897 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13898 (DFS_write_tree_body): New function.
13899 (DFS_write_tree): Likewise.
13900 (hash_tree): Likewise.
13901 (scc_entry_compare): Likewise.
13902 (hash_scc): Likewise.
13903 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13904 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13905 TREE_CHAIN as regular reference.
13906 (streamer_read_integer_cst): Remove.
13907 (streamer_get_pickled_tree): Adjust.
13908 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13909 of DECL_EXTERNALs in BLOCK_VARS for now.
13910 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13912 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13913 Add hash value argument and record that if hashes are recorded
13915 (streamer_tree_cache_insert_1): Adjust.
13916 (streamer_tree_cache_insert): Likewise.
13917 (streamer_tree_cache_insert_at): Rename to ...
13918 (streamer_tree_cache_replace_tree): ... this and adjust.
13919 (streamer_tree_cache_append): Adjust.
13920 (record_common_node): Likewise.
13921 (streamer_tree_cache_create): Add argument whether to
13922 record hash values together with trees.
13923 (streamer_tree_cache_delete): Adjust.
13924 * tree-streamer.h (struct streamer_tree_cache_d): Add
13926 (streamer_read_integer_cst): Remove.
13927 (streamer_tree_cache_insert): Adjust.
13928 (streamer_tree_cache_append): Likewise.
13929 (streamer_tree_cache_insert_at): Rename to ...
13930 (streamer_tree_cache_replace_tree): ... this and adjust.
13931 (streamer_tree_cache_create): Add argument whether to record hashes.
13932 (streamer_tree_cache_get): Rename to ...
13933 (streamer_tree_cache_get_tree): ... this.
13934 (streamer_tree_cache_get_hash): New function.
13935 * tree.c (cache_integer_cst): New function.
13936 * tree.h (cache_integer_cst): Declare.
13937 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13938 * lto-symtab.c (lto_varpool_replace_node): Only release
13939 DECL_INITIAL of non-prevailing decls.
13940 * varpool.c (varpool_remove_initializer): Do not release
13941 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13943 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13945 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13946 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13947 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13948 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13949 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13950 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13951 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13952 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13953 instead of TARGET_64BIT.
13954 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13955 Require ISA_HAS_<D>DIV.
13957 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13959 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13960 (mips*-*-linux*): Move default with_llsc setting to where other
13962 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13964 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13965 Likewise. Remove default with_tune setting. Move default float
13966 setting to its own block. Handle with_llsc in the same block as above.
13968 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
13970 PR rtl-optimization/57425
13971 PR rtl-optimization/57569
13972 * alias.c (write_dependence_p): Add new parameters mem_mode,
13973 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13974 Changed all callers.
13975 (canon_anti_dependence): New function.
13976 * cse.c (check_dependence): Use canon_anti_dependence.
13977 * cselib.c (cselib_invalidate_mem): Likewise.
13978 * rtl.h (canon_anti_dependence): Declare.
13980 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13982 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13983 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13984 ".set mips3" for 64-bit targets.
13986 2013-06-15 Dehao Chen <dehao@google.com>
13988 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13989 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13990 (gimple_check_call_args): Likewise.
13991 * value-prof.c (check_ic_target): Likewise.
13992 * ipa-inline.c (early_inliner): Likewise.
13993 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13994 * cgraph.c (cgraph_create_edge_1): Likewise.
13995 (cgraph_make_edge_direct): Likewise.
13997 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14000 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14001 rs6000_output_move_128bit to handle emitting quad memory
14002 operations. Set attribute length to 8 bytes.
14004 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14006 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14008 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14009 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14010 (aarch64_<su>mlsl<mode>): Likewise.
14012 2013-06-14 Mike Stump <mikestump@comcast.net>
14014 * Makefile.in (TARGET_H): Add insn-codes.h.
14016 2013-06-14 Alan Modra <amodra@gmail.com>
14018 PR middle-end/57134
14019 PR middle-end/57586
14020 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14021 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14022 bitfield expansion when EXPAND_MEMORY.
14023 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14025 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14027 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14028 test for clearing quad memory on 32-bit later.
14030 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14032 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14033 (fold_negate_expr): Likewise.
14034 (fold_real_zero_addition_p): Handle vectors.
14035 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14037 2013-06-14 Alan Modra <amodra@gmail.com>
14039 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14041 2013-06-13 Jan Hubicka <jh@suse.cz>
14043 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14044 Local comdats are not externally visible.
14045 * symtab.c (dump_symtab_base): Dump externally visible.
14046 (verify_symtab_base): Verify back links in the symtab hash.
14048 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14050 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14051 CONVERT_EXPR as equal nodes.
14053 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14055 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14057 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14059 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14060 Generalize to complex and vector.
14061 * tree.c (build_all_ones_cst): New function.
14062 * tree.h (build_all_ones_cst): Declare it.
14064 2013-06-13 Alan Modra <amodra@gmail.com>
14066 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14067 * config/rs6000/rs6000.md (signbittf2): New insn.
14068 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14069 (abstf2_internal, cmptf_internal2): Likewise.
14070 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14072 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14073 Pat Haugen <pthaugen@us.ibm.com>
14074 Peter Bergner <bergner@vnet.ibm.com>
14076 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14077 power8 byte, half-word, and quad-word atomic instructions.
14078 (emit_store_conditional): Likewise.
14079 (rs6000_expand_atomic_compare_and_swap): Likewise.
14080 (rs6000_expand_atomic_op): Likewise.
14082 * config/rs6000/sync.md (larx): Add new modes for power8.
14084 (AINT): New mode iterator to include TImode as well as normal
14085 integer modes on power8.
14086 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14087 that VSX registers are not considered. Use AINT mode iterator
14088 instead of INT1 to allow inclusion of quad word atomic operations
14090 (load_locked<mode>): Likewise.
14091 (store_conditional<mode>): Likewise.
14092 (atomic_compare_and_swap<mode>): Likewise.
14093 (atomic_exchange<mode>): Likewise.
14094 (atomic_nand<mode>): Likewise.
14095 (atomic_fetch_<fetchop_name><mode>): Likewise.
14096 (atomic_nand_fetch<mode>): Likewise.
14097 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14099 (ATOMIC): On power8, add QImode, HImode modes.
14100 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14101 modes that promote to SImode.
14102 (load_lockedti): Convert TImode arguments to PTImode, so that we
14103 get a guaranteed even/odd register pair.
14104 (load_lockedpti): Likewise.
14105 (store_conditionalti): Likewise.
14106 (store_conditionalpti): Likewise.
14108 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14109 atomic load/store instructions.
14112 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14114 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14116 (insn_count): New attribute, with most cases extracted from...
14117 (length): ...here. Redefine most cases in terms of insn_count.
14118 (single_insn): Delete.
14119 (can_delay): Use insn_count to check for single instructions.
14120 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14121 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14122 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14123 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14124 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14125 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14126 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14127 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14128 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14129 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14130 rather than "length".
14131 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14132 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14133 Use "insn_count" rather than "length".
14134 * config/mips/mips-dsp.md
14135 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14136 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14139 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14141 PR tree-optimization/57361
14142 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14144 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14146 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14148 (aarch64_simd_combine<mode>): New instruction expansion.
14149 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14150 function prototype.
14151 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14152 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14154 2013-06-12 Jan Hubicka <jh@suse.cz>
14156 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14157 decl has when in streaming stage.
14158 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14159 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14161 2013-06-12 Roland Stigge <stigge@antcom.de>
14164 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14166 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14168 PR tree-optimization/57537
14169 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14170 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14172 2013-06-12 Richard Biener <rguenther@suse.de>
14174 * data-streamer.h (streamer_write_char_stream): CSE
14175 obs->current_pointer.
14176 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14177 streamer_write_char_stream manually and optimize the resulting loop.
14178 (streamer_write_hwi_stream): Likewise.
14180 2013-06-12 Jan Hubicka <jh@suse.cz>
14182 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14183 * cgraph.h (varpool_create_empty_node): Declare.
14184 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14186 * symtab.c (symtab_unregister_node): Be lax about missin entries
14188 (symtab_get_node): Update comment.
14189 * varpool.c (varpool_create_empty_node): Break out from ...
14190 (varpool_node_for_decl): ... here.
14191 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14193 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14195 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14196 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14197 part. Use straight-line flow at the end.
14198 <COMPONENT_REF>: Remove superfluous else.
14199 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14201 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14204 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14205 target hook even for !TREE_PUBLIC decls. If no resolution info
14206 is available, return false for common and external decls.
14208 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14210 * config/rl78/constraints.md (U): New constraint.
14211 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14214 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14217 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14218 to allow returning address to AT_PLATFORM name.
14220 2013-06-11 Jan Hubicka <jh@suse.cz>
14222 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14223 * cgraph.h (symtab_node_base): Add weakref flag.
14224 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14225 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14226 (output_weakrefs): Use weakref flag.
14227 * fold-const.c (simple_operand_p): Handle WEAK.
14228 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14229 * ipa.c (varpool_externally_visible_p): Drop weakref.
14230 (function_and_variable_visibility): Update comment; fix weakref
14231 sanity checks; do not clear DECL_WEAK on them.
14232 * lto-cgraph.c (lto_output_node): update.
14233 (lto_output_varpool_node): Update.
14234 (input_overwrite_node): Update.
14235 (input_node): Update.
14236 (input_varpool_node): Update.
14237 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14238 (lto_symtab_merge_symbols): Add sanity check.
14239 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14240 * passes.c (rest_of_decl_compilation): Set static flag, too.
14241 * symtab.c (dump_symtab_base): Dump weakref.
14242 (verify_symtab_base): Sanity check weakrefs.
14243 (symtab_make_decl_local): Remove duplicated code.
14244 (symtab_alias_ultimate_target): Simplify.
14245 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14247 2013-06-11 Tom de Vries <tom@codesourcery.com>
14249 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14250 sequence_vect == NULL.
14252 2013-06-11 DJ Delorie <dj@redhat.com>
14254 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14255 (rl78_unwind_word_mode): New.
14257 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14259 * final.c (debug_prefix_maps): Make static.
14261 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14263 * function.c (initial_trampoline): Remove stray copy.
14265 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14267 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14269 2013-06-11 Martin Jambor <mjambor@suse.cz>
14271 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14272 within bounds at the beginning of the function.
14274 2013-06-11 Alan Modra <amodra@gmail.com>
14276 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14278 (get_named_section): Don't NULL !DECL_P decl.
14280 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14282 * doc/invoke.texi (core-avx2): Document.
14284 (atom): Updated with MOVBE.
14286 2013-06-11 Richard Biener <rguenther@suse.de>
14288 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14290 2013-06-11 Anton Blanchard <anton@samba.org>
14292 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14293 correct shift value in little-endian mode.
14295 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14298 * varasm.c (get_variable_align): Move #endif to the right place.
14300 2013-06-10 Cary Coutant <ccoutant@google.com>
14302 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14303 for hash so that hash table traversal order is deterministic.
14305 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14306 Pat Haugen <pthaugen@us.ibm.com>
14307 Peter Bergner <bergner@vnet.ibm.com>
14309 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14310 of vectors in GPRS if they are direct moves or quad word load or
14313 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14315 (direct_move_p): Likewise.
14316 (quad_load_store_p): Likewise.
14318 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14319 classes into bins based on the physical register type.
14320 (reg_class_to_reg_type): Likewise.
14321 (IS_STD_REG_TYPE): Likewise.
14322 (IS_FP_VECT_REG_TYPE): Likewise.
14323 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14324 use direct move instructions.
14325 (reload_gpr_vsx): Likewise.
14326 (reload_vsx_gpr): Likewise.
14327 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14328 information that is a simplification of register classes. Also
14329 precalculate direct move reload helpers.
14330 (direct_move_p): New function to return true if the operation can
14331 be done as a direct move instruciton.
14332 (quad_load_store_p): New function to return true if the operation
14333 is a quad memory operation.
14334 (rs6000_legitimize_address): If quad memory, only allow register
14335 indirect for TImode addresses.
14336 (rs6000_legitimate_address_p): Likewise.
14337 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14338 (rs6000_reload_register_type): Likewise.
14339 (register_to_reg_type): Return register type.
14340 (rs6000_secondary_reload_simple_move): New helper function for
14341 secondary reload and secondary memory needed to identify anything
14342 that is a simple move, and does not need reloading.
14343 (rs6000_secondary_reload_direct_move): New helper function for
14344 secondary reload to identify cases that can be done with several
14345 instructions via the direct move instructions.
14346 (rs6000_secondary_reload_move): New helper function for secondary
14347 reload to identify moves between register types that can be done.
14348 (rs6000_secondary_reload): Add support for quad memory operations
14349 and for direct move.
14350 (rs6000_secondary_memory_needed): Likewise.
14351 (rs6000_debug_secondary_memory_needed): Change argument names.
14352 (rs6000_output_move_128bit): New function to return the move to
14353 use for 128-bit moves, including knowing about the various
14354 limitations of quad memory operations.
14356 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14357 memory operations. call rs6000_output_move_128bit for the actual
14358 instruciton(s) to generate.
14359 (vsx_movti_64bit): Likewise.
14361 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14362 (UNSPEC_P8V_MTVSRWZ): Likewise.
14363 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14364 (UNSPEC_P8V_MTVSRD): Likewise.
14365 (UNSPEC_P8V_XXPERMDI): Likewise.
14366 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14367 (UNSPEC_FUSION_GPR): Likewise.
14368 (FMOVE128_GPR): New iterator for direct move.
14369 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14370 (f32_sv): Likewise.
14371 (f32_dm): Likewise.
14372 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14373 loads and direct move instructions.
14374 (zero_extendsidi2_lfiwzx): Likewise.
14375 (extendsidi2_lfiwax): Likewise.
14376 (extendsidi2_nocell): Likewise.
14377 (floatsi<mode>2_lfiwax): Likewise.
14378 (lfiwax): Likewise.
14379 (floatunssi<mode>2_lfiwzx): Likewise.
14380 (lfiwzx): Likewise.
14381 (fix_trunc<mode>_stfiwx): Likewise.
14382 (fixuns_trunc<mode>_stfiwx): Likewise.
14383 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14384 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14385 (parity<mode>2_cmpb): Set length/type attr.
14386 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14387 for 'mr.' to fast_compare.
14388 (bpermd_<mode>): Change type attr to popcnt.
14389 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14390 (p8_mtvsrwz_1): Likewise.
14391 (p8_mtvsrwz_2): Likewise.
14392 (reload_fpr_from_gpr<mode>): Likewise.
14393 (p8_mtvsrd_1): Likewise.
14394 (p8_mtvsrd_2): Likewise.
14395 (p8_xxpermdi_<mode>): Likewise.
14396 (reload_vsx_from_gpr<mode>): Likewise.
14397 (reload_vsx_from_gprsf): Likewise.
14398 (p8_mfvsrd_3_<mode>): LIkewise.
14399 (reload_gpr_from_vsx<mode>): Likewise.
14400 (reload_gpr_from_vsxsf): Likewise.
14401 (p8_mfvsrd_4_disf): Likewise.
14402 (multi-word GPR splits): Do not split direct moves or quad memory
14405 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14407 * tree-into-ssa.c (interesting_blocks): Make static.
14409 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14412 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14413 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14414 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14415 (get_variable_align): New function.
14416 (get_variable_section, emit_bss, emit_common,
14417 assemble_variable_contents, place_block_symbol): Use
14418 get_variable_align instead of DECL_ALIGN.
14419 (assemble_noswitch_variable): Add align argument, use it
14420 instead of DECL_ALIGN.
14421 (assemble_variable): Adjust caller. Use get_variable_align
14422 instead of DECL_ALIGN.
14423 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14425 (DATA_ABI_ALIGNMENT): Define.
14426 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14427 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14428 opt is false, only return the psABI mandated alignment increase.
14429 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14430 (DATA_ABI_ALIGNMENT): ... this.
14431 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14432 (DATA_ABI_ALIGNMENT): ... this.
14433 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14434 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14435 (DATA_ABI_ALIGNMENT): ... this.
14436 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14437 * doc/tm.texi: Regenerated.
14439 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14441 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14442 cmp_code to construct REG_EQUAL note.
14444 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14447 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14448 that operands[2] doesn't overlap with operands[0].
14450 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14451 Jan Hubicka <jh@suse.cz>
14453 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14454 hack to mark symbols as used.
14456 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14458 PR rtl-optimization/57559
14459 * lra-constraints.c (process_alt_operands): Don't discourage
14460 memory with known offset for offsetable memory constraint.
14461 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14463 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14465 * varasm.c (struct oc_local_state): Reorder fields.
14466 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14467 and adjust accordingly.
14468 (output_constructor): Reorder initialization code and adjust call to
14469 output_constructor_bitfield.
14471 2013-06-07 Jan Hubicka <jh@suse.cz>
14473 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14475 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14477 * tree-object-size.c (unknown): Make const.
14479 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14481 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14482 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14483 for last alternative in the cpu_facility attribute.
14485 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14488 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14489 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14490 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14491 * config/arm/constraints.md (Dg): New constraint.
14492 * config/arm/neon.md (xordi3_neon): Remove.
14493 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14494 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14496 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14498 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14499 Clean up alternatives.
14501 2013-06-07 Alan Modra <amodra@gmail.com>
14503 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14506 2013-06-07 Alan Modra <amodra@gmail.com>
14508 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14510 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14512 * config/arm/constraints.md (Df): New constraint.
14513 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14514 Correct length attribute for last two alternatives.
14516 2013-06-07 Alan Modra <amodra@gmail.com>
14518 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14519 override user -mfp-in-toc.
14520 (offsettable_ok_by_alignment): Consider just the current access
14521 rather than the whole object, unless BLKmode. Handle
14522 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14523 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14524 for -mcmodel=medium.
14525 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14526 override user -mfp-in-toc or -msum-in-toc. Default to
14527 -mno-fp-in-toc for -mcmodel=medium.
14529 2013-06-06 DJ Delorie <dj@redhat.com>
14531 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14532 TARGET_VALID_POINTER_MODE.
14534 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14535 Pat Haugen <pthaugen@us.ibm.com>
14536 Peter Bergner <bergner@vnet.ibm.com>
14538 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14539 Document new power8 builtins.
14541 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14542 condition code register, to allow 128-bit logical operations to be
14543 done in the VSX or GPR registers.
14544 (nor<mode>3): Use the canonical form for nor.
14545 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14546 vclz*, and vpopcnt* vector instructions.
14547 (nand<mode>3): Likewise.
14548 (orc<mode>3): Likewise.
14549 (clz<mode>2): LIkewise.
14550 (popcount<mode>2): Likewise.
14552 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14553 that only the GPRs are recognized.
14555 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14556 support for new power8 builtins.
14558 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14560 (xscvdpspn): Likewise.
14566 (vpopcnt): Likewise.
14567 (vpopcntb): Likewise.
14568 (vpopcnth): Likewise.
14569 (vpopcntw): Likewise.
14570 (vpopcntd): Likewise.
14572 (vmrgew): Likewise.
14573 (vmrgow): Likewise.
14575 (eqv_v16qi3): Likewise.
14576 (eqv_v8hi3): Likewise.
14577 (eqv_v4si3): Likewise.
14578 (eqv_v2di3): Likewise.
14579 (eqv_v4sf3): Likewise.
14580 (eqv_v2df3): Likewise.
14582 (nand_v16qi3): Likewise.
14583 (nand_v8hi3): Likewise.
14584 (nand_v4si3): Likewise.
14585 (nand_v2di3): Likewise.
14586 (nand_v4sf3): Likewise.
14587 (nand_v2df3): Likewise.
14589 (orc_v16qi3): Likewise.
14590 (orc_v8hi3): Likewise.
14591 (orc_v4si3): Likewise.
14592 (orc_v2di3): Likewise.
14593 (orc_v4sf3): Likewise.
14594 (orc_v2df3): Likewise.
14596 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14597 allow power8 quad mode in 64-bit.
14598 (rs6000_builtin_vectorized_function): Add support to vectorize
14599 ISA 2.07 count leading zeros, population count builtins.
14600 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14601 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14602 (builtin_function_type): Add vgbbd builtin function which takes an
14604 (altivec_expand_vec_perm_const): Add support for new power8 merge
14607 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14608 that does not include TImdoe for use with 32-bit.
14609 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14611 (UNSPEC_VSX_CVDPSPN): Likewise.
14612 (vsx_xscvdpspn): Likewise.
14613 (vsx_xscvspdpn): Likewise.
14614 (vsx_xscvdpspn_scalar): Likewise.
14615 (vsx_xscvspdpn_directmove): Likewise.
14616 (vsx_and<mode>3): Split logical operations into 32-bit and
14617 64-bit. Add support to do logical operations on TImode as well as
14618 VSX vector types. Allow logical operations to be done in either
14619 VSX registers or in general purpose registers in 64-bit mode. Add
14620 splitters if GPRs were used. For AND, add clobber of CCmode to
14621 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14623 (vsx_and<mode>3_32bit): Likewise.
14624 (vsx_and<mode>3_64bit): Likewise.
14625 (vsx_ior<mode>3): Likewise.
14626 (vsx_ior<mode>3_32bit): Likewise.
14627 (vsx_ior<mode>3_64bit): Likewise.
14628 (vsx_xor<mode>3): Likewise.
14629 (vsx_xor<mode>3_32bit): Likewise.
14630 (vsx_xor<mode>3_64bit): Likewise.
14631 (vsx_one_cmpl<mode>2): Likewise.
14632 (vsx_one_cmpl<mode>2_32bit): Likewise.
14633 (vsx_one_cmpl<mode>2_64bit): Likewise.
14634 (vsx_nor<mode>3): Likewise.
14635 (vsx_nor<mode>3_32bit): Likewise.
14636 (vsx_nor<mode>3_64bit): Likewise.
14637 (vsx_andc<mode>3): Likewise.
14638 (vsx_andc<mode>3_32bit): Likewise.
14639 (vsx_andc<mode>3_64bit): Likewise.
14640 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14641 and xxlorc instructions.
14642 (vsx_eqv<mode>3_64bit): Likewise.
14643 (vsx_nand<mode>3_32bit): Likewise.
14644 (vsx_nand<mode>3_64bit): Likewise.
14645 (vsx_orc<mode>3_32bit): Likewise.
14646 (vsx_orc<mode>3_64bit): Likewise.
14648 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14650 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14652 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14653 (p8_vmrgow): Likewise.
14654 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14655 GPRs to be split under VSX.
14656 (p8v_clz<mode>2): Add power8 count leading zero support.
14657 (p8v_popcount<mode>2): Add power8 population count support.
14658 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14661 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14664 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14666 (vec_nand): Likewise.
14667 (vec_vclz): Likewise.
14668 (vec_vclzb): Likewise.
14669 (vec_vclzd): Likewise.
14670 (vec_vclzh): Likewise.
14671 (vec_vclzw): Likewise.
14672 (vec_vgbbd): Likewise.
14673 (vec_vmrgew): Likewise.
14674 (vec_vmrgow): Likewise.
14675 (vec_vpopcnt): Likewise.
14676 (vec_vpopcntb): Likewise.
14677 (vec_vpopcntd): Likewise.
14678 (vec_vpopcnth): Likewise.
14679 (vec_vpopcntw): Likewise.
14681 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14683 PR rtl-optimization/57468
14684 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14687 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14689 PR rtl-optimization/57459
14690 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14691 type when setting live regs.
14693 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14695 * config/s390/s390.opt (mlra): New option.
14696 * config/s390/s390.c (s390_decompose_address): Check displacement
14697 for all registers for LRA.
14698 (s390_secondary_reload): Don't used secondary reloads for LRA.
14699 (s390_lra_p): New function.
14700 (TARGET_LRA_P): Define.
14701 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14702 of attribute cpu_facility to zarch for the last alternative.
14703 (*cmpmem_short): Ditto.
14705 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14707 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14708 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14709 (arm_expand_prologue): Likewise.
14711 2013-06-06 Teresa Johnson <tejohnson@google.com>
14714 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14715 as this is now done by redirect_edge_and_branch_force.
14716 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14717 barriers, and fix interaction with splitting.
14718 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14719 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14720 reflect changes made in the routine.
14721 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14722 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14723 since this is called in cfglayout mode, and replace partition fixup
14724 with assert as that is now done by force_nonfallthru_and_redirect.
14725 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14726 already be marked with region crossing note.
14727 (insert_section_boundary_note): Make non-static, gate on flag
14728 has_bb_partition, rewrite to also check for multiple partitions.
14729 (rest_of_handle_reorder_blocks): Remove call to
14730 insert_section_boundary_note, now done later during free_cfg.
14731 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14732 * bb-reorder.h (insert_section_boundary_note): Declare.
14733 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14734 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14735 invoke insert_section_boundary_note.
14736 (try_redirect_by_replacing_jump): Remove unnecessary
14737 check for region crossing note.
14738 (fixup_partition_crossing): New function.
14739 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14740 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14741 in non-cfglayout mode.
14742 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14743 remove old code that tried to do this. Emit barrier correctly
14744 when we are in cfglayout mode.
14745 (last_bb_in_partition): New function.
14746 (rtl_split_edge): Correctly fixup partition boundaries.
14747 (commit_one_edge_insertion): Remove old code that tried to
14748 fixup region crossing edge since this is now handled in
14749 split_block, and set up insertion point correctly since
14750 block may now end in a jump.
14751 (verify_hot_cold_block_grouping): Guard against checking when not in
14752 linearized RTL mode.
14753 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14755 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14756 rtl_verify_flow_info, so not called in cfglayout mode.
14757 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14758 (fixup_reorder_chain): Remove old code that attempted to fixup region
14759 crossing note as this is now handled in force_nonfallthru_and_redirect.
14760 (duplicate_insn_chain): Don't duplicate switch section notes.
14761 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14763 * basic-block.h (emit_barrier_after_bb): Declare.
14765 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14767 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14768 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14769 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14771 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14773 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14775 * config/arm/ldmstm.md: Regenerate.
14777 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14779 * config/arm/sync.md (atomic_loaddi_1):
14780 Disable predication for arm_restrict_it.
14781 (arm_load_exclusive<mode>): Likewise.
14782 (arm_load_exclusivesi): Likewise.
14783 (arm_load_exclusivedi): Likewise.
14784 (arm_load_acquire_exclusive<mode>): Likewise.
14785 (arm_load_acquire_exclusivesi): Likewise.
14786 (arm_load_acquire_exclusivedi): Likewise.
14787 (arm_store_exclusive<mode>): Likewise.
14788 (arm_store_exclusive<mode>): Likewise.
14789 (arm_store_release_exclusivedi): Likewise.
14790 (arm_store_release_exclusive<mode>): Likewise.
14792 2013-06-06 Richard Biener <rguenther@suse.de>
14794 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14796 (lto_tag_is_tree_code_p): Adjust.
14797 (lto_tag_is_gimple_code_p): Likewise.
14798 (lto_gimple_code_to_tag): Likewise.
14799 (lto_tag_to_gimple_code): Likewise.
14800 (lto_tree_code_to_tag): Likewise.
14801 (lto_tag_to_tree_code): Likewise.
14802 * data-streamer.h (streamer_write_hwi_in_range): Use
14803 uhwi streaming to stream the normalized range.
14804 (streamer_read_hwi_in_range): Likewise.
14806 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14808 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14809 (predicable_short_it): Likewise.
14810 (predicated): Likewise.
14811 (enabled): Handle above.
14812 (define_cond_exec): Set predicated attribute to yes.
14814 2013-06-05 Mike Stump <mikestump@comcast.net>
14816 * gdbinit.in (__FUNCTION__): Add.
14818 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14820 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14821 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14823 2013-06-05 Jan Hubicka <jh@suse.cz>
14825 * varasm.c (mark_decl_referenced): Revert the removal until targets
14828 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14830 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14831 instead of mark_decl_referenced.
14833 2013-06-05 Jan Hubicka <jh@suse.cz>
14835 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14836 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14837 and symtab_used_from_object_file_p.
14838 (cgraph_make_node_local_1): Clear forced_by_abi.
14839 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14840 * cgraph.h (symtab_node_base): Add forced_by_abi.
14841 (decide_is_variable_needed): Remove.
14842 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14843 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14844 (decide_is_symbol_needed): ... this one; handle symbols in general;
14845 always analyze virtuals; honnor forced_by_abi.
14846 (cgraph_finalize_function): Update.
14847 (varpool_finalize_decl): Update.
14848 (symbol_defined_and_needed): Remove.
14849 (analyze_functions): Update.
14850 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14851 output_refs, input_overwrite_node): Handle forced_by_abi.
14852 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14853 (address_taken_from_non_vtable_p): ... this one.
14854 (comdat_can_be_unshared_p_1): New function.
14855 (cgraph_comdat_can_be_unshared_p): Rename to ...
14856 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14857 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14858 (function_and_variable_visibility): Clear forced_by_abi as needed.
14859 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14860 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
14861 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14862 * varpool.c (decide_is_variable_needed): Remove.
14864 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14866 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14867 (arm_option_override): Override arm_restrict_it where appropriate.
14868 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14869 * config/arm/arm.opt (mrestrict-it): New command-line option.
14870 * doc/invoke.texi: Document -mrestrict-it.
14872 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14874 * tsan.c (tsan_atomic_table): Make const.
14876 2013-06-05 Richard Biener <rguenther@suse.de>
14878 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14879 index associated with the tree we are supposed to replace.
14880 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14881 TREE_ASM_WRITTEN as zero for everything but SSA names.
14883 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14885 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14887 2013-06-04 Jan Hubicka <jh@suse.cz>
14889 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14890 (input_node, input_varpool_node): Handle correctly external same
14892 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14893 nodes at ltrans stage.
14895 2013-06-04 Jan Hubicka <jh@suse.cz>
14897 * ipa-inline.c (update_caller_keys): Fix availability test.
14898 (update_callee_keys): Likewise.
14899 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14900 to follow ELF standard.
14902 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14904 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14905 (mips64r5900el-*-elf*): New configurations.
14906 * config/mips/mips-cpus.def (r5900): New processor.
14907 * config/mips/mips-tables.opt: Regenerate.
14908 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14909 (mips_issue_rate): Handle PROCESSOR_R5900.
14910 (mips_reorg_process_insns): Force reorder mode for the R5900.
14911 * config/mips/mips.h (TARGET_MIPS5900): Define.
14912 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14914 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14916 * config/mips/mips.md (processor): Add r5900.
14917 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14919 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14921 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14922 into function to generate MOVI instruction.
14923 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
14924 (aarch64_preferred_simd_mode): Turn into wrapper.
14925 (aarch64_output_scalar_simd_mov_immediate): New function.
14926 * config/aarch64/aarch64-protos.h: Add prototype for above.
14928 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14930 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14931 element_char member.
14932 (sizetochar): Return signed char.
14933 (aarch64_simd_valid_immediate): Remove elchar and other
14934 unnecessary variables.
14935 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14936 Calculate element_char as required.
14937 * config/aarch64/aarch64-protos.h: Update and move prototype
14938 for aarch64_output_simd_mov_immediate.
14939 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14942 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14944 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14945 information completed by aarch64_simd_valid_immediate.
14946 (aarch64_legitimate_constant_p): Update arguments.
14947 (aarch64_simd_valid_immediate): Work with struct rather than many
14949 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14950 (aarch64_simd_make_constant): Update arguments.
14951 (aarch64_output_simd_mov_immediate): Work with struct rather than
14952 many pointers. Output immediate directly rather than as operand.
14953 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14955 * config/aarch64/constraints.md (Dn): Update arguments.
14957 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14959 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14961 (aarch64_simd_immediate_valid_for_move): Remove.
14962 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14963 (aarch64_simd_make_constant): Update call.
14964 (aarch64_output_simd_mov_immediate): Update call.
14965 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14967 * config/aarch64/constraints.md (Dn): Update call.
14969 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14971 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14972 return type to bool for prototype.
14973 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14974 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14975 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14977 2013-06-04 Catherine Moore <clm@codesourcery.com>
14979 * config/mips/mips.opt (meva): New.
14980 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
14981 (ASM_SPEC): Handle -meva.
14982 * doc/invoke.texi (meva): Document.
14984 2013-06-04 Alan Modra <amodra@gmail.com>
14986 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14989 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14991 * rtl.def: Add extra fourth optional field to define_cond_exec.
14992 * gensupport.c (process_one_cond_exec): Process attributes from
14994 * doc/md.texi: Document fourth field in define_cond_exec.
14996 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14998 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14999 out the processing order as in store_bit_field_1.
15001 2013-06-04 Jan Hubicka <jh@suse.cz>
15003 PR middle-end/57500
15004 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15005 non-VAR_DECL node if it does not exist yet.
15007 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15009 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15010 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15011 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15012 target_cpu_default setting.
15014 2013-06-03 Teresa Johnson <tejohnson@google.com>
15016 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15017 default to -fopt-info=optimized instead of all.
15018 * doc/invoke.texi: Ditto.
15019 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15020 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15021 (execute_vect_slp): Emit BB vectorization success under
15022 MSG_OPTIMIZED_LOCATIONS.
15023 * tree-vect-slp.c (vect_slp_transform_bb): Change
15024 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15025 * tree-vect-loop.c (vect_transform_loop): Ditto.
15027 2013-06-03 Jason Merrill <jason@redhat.com>
15030 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15031 Use TARGET_EXPR for C++.
15033 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15035 PR rtl-optimization/57268
15036 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15037 if DEBUG_INSN_P (insn).
15040 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15042 PR rtl-optimization/57268
15043 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15044 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15046 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15048 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15049 (ix86_avoid_lea_for_addr): Likewise.
15050 (exact_dependency_1): Likewise.
15051 (ix86_adjust_cost): Likewise.
15052 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15054 (do_reorder_for_imul): Fix typo, formatting and
15055 !reload_completed check removed.
15056 (ix86_sched_reorder): Fix typo and formatting.
15057 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15060 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15062 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15064 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15066 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15067 <STRING_CST>: Likewise.
15068 <VECTOR_CST>: Likewise.
15070 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15071 Mikael Morin <mikael@gcc.gnu.org>
15073 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15074 * config.in: Regenerated.
15075 * configure: Regenerated.
15077 2013-06-01 Jan Hubicka <jh@suse.cz>
15079 PR middle-end/57366
15080 * cgraphunit.c (compile): When weakref is not supported,
15081 set up transparent aliases before final output pass.
15082 * varasm.c (assemble_alias): Do not try to do it here.
15084 2013-06-01 Jan Hubicka <jh@suse.cz>
15086 PR middle-end/57467
15087 * passes.c (for_per_function): Skip unanalyzed functions.
15089 2013-06-01 Jan Hubicka <jh@suse.cz>
15091 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15092 (lto_symtab_merge_symbols_1): ... this one.
15093 (lto_symtab_merge_cgraph_nodes): Rename to ...
15094 (lto_symtab_merge_symbols): ... this one; simplify.
15095 * cgraph.c (same_body_aliases_done): Rename to ...
15096 (cpp_implicit_aliases_done): ... this one.
15097 (cgraph_create_function_alias): Update.
15098 (cgraph_same_body_alias): Update.
15099 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15100 (verify_edge_corresponds_to_fndecl): Simplify.
15101 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15102 (cgraph_node): Remove same_body_alias.
15103 (varpool_node): Remove alias_of and extra_name_alias.
15104 (same_body_aliases_done): Rename to ..
15105 (cpp_implicit_aliases_done): ... this one.
15106 (symtab_alias_ultimate_target): Add default parameter.
15107 (symtab_resolve_alias): New function.
15108 (fixup_same_cpp_alias_visibility): Declare.
15109 (cgraph_function_node): Add default parameter.
15110 (cgraph_node_asm_name): Likewise.
15111 (cgraph_function_or_thunk_node): Add default parameter; do
15112 not ICE when it is NULL.
15113 (varpool_variable_node): Likewise.
15114 * tree-emutls.c (create_emultls_var): Update.
15115 (ipa_lower_emutls): Update.
15116 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15117 (cgraph_reset_node): Reset alias info.
15118 (cgraph_finalize_function): Update.
15119 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15120 (analyze_function): Simplify.
15121 (cgraph_process_same_body_aliases): Simplify.
15122 (analyze_functions): Fixup same body aliases.
15123 (handle_alias_pairs): Simplify.
15124 (assemble_thunk): Update.
15125 (assemble_thunks_and_aliases): Update.
15126 (output_weakrefs): Rewrite.
15127 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15128 (lto_output_varpool_node): Likewise.
15129 (compute_ltrans_boundary): Remve assert.
15130 (get_alias_symbol): New functoin.
15131 (input_node): Rewrite alias handling.
15132 (input_varpool_node): Likewise.
15133 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15134 * ipa.c (process_references): Handle weakrefs correctly.
15135 (symtab_remove_unreachable_nodes): Likewise.
15136 * trans-mem.c (get_cg_data): Update.
15137 (ipa_tm_create_version_alias): Update.
15138 (ipa_tm_execute): Update.
15139 * symtab.c (dump_symtab_base): Dump aliases.
15140 (verify_symtab_base): Verify aliases.
15141 (symtab_node_availability): New function.
15142 (symtab_alias_ultimate_target): Simplify.
15143 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15144 handle all the fixup cases.
15145 (symtab_resolve_alias): New function.
15146 * passes.c (ipa_write_summaries): Handle weakrefs.
15147 * varpool.c (varpool_analyze_node): Simplify.
15148 (assemble_aliases): Update.
15149 (varpool_create_variable_alias): Simplify.
15150 (varpool_extra_name_alias): Simplify.
15151 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15152 (lto_symtab_merge_symbols): ... this one.
15154 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15157 PR rtl-optimization/57268
15158 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15159 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15161 2013-06-01 Tobias Burnus <burnus@net-b.de>
15163 Partially reverted:
15164 2013-05-31 Tobias Burnus <burnus@net-b.de>
15166 PR middle-end/57073
15167 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15170 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15172 PR rtl-optimization/57268
15173 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15174 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15176 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15178 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15179 unordered comparison operators when -fno-trapping-math is in effect
15181 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15182 and implement unordered comparison operators properly on the e500.
15184 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15186 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15187 for constant scalar integers.
15188 (simplify_relational_operation_1): Likewise.
15190 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15192 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15193 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15196 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15197 Igor Zamyatin <igor.zamyatin@intel.com>
15199 Silvermont (SLM) architecture performance tuning.
15200 * config/i386/i386.h (enum ix86_tune_indices): Add
15201 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15202 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15204 * config/i386/i386.c (initial_ix86_tune_features)
15205 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15206 (ix86_lea_outperforms): Handle Silvermont tuning.
15207 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15209 (ix86_use_lea_for_mov): Likewise.
15210 (ix86_avoid_lea_for_addr): Likewise.
15211 (ix86_lea_for_add_ok): Likewise.
15212 (exact_dependency_1): New function.
15213 (exact_store_load_dependency): Likewise.
15214 (ix86_adjust_cost): Handle Silvermont tuning.
15215 (do_reoder_for_imul): Likewise.
15216 (swap_top_of_ready_list): New function.
15217 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15219 * config/i386/i386.md (peepholes that split memory operand in fp
15222 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15224 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15225 Remove un-necessary braces.
15227 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15229 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15230 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15232 2013-05-31 Tobias Burnus <burnus@net-b.de>
15234 PR middle-end/57073
15235 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15237 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15240 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15241 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15242 * config/arm/neon.md (iordi3_neon): Remove.
15243 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15244 * config/arm/predicates.md (imm_for_neon_logic_operand):
15245 Move to earlier in the file.
15246 (neon_logic_op2): Likewise.
15247 (arm_iordi_operand_neon): New predicate.
15249 2013-05-31 Richard Biener <rguenther@suse.de>
15251 PR tree-optimization/57478
15252 PR tree-optimization/57453
15253 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15256 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15258 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15259 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15261 2013-05-30 Tobias Burnus <burnus@net-b.de>
15262 Thomas Koenig <tkoenig@gcc.gnu.org>
15264 PR middle-end/57073
15265 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15266 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15268 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15270 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15272 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15274 * target.def (register_usage_leveling_p): New hook.
15275 * targhooks.c (default_register_usage_leveling_p): New.
15276 * targhooks.h (default_register_usage_leveling_p): New prototype.
15277 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15278 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15279 * doc/tm.texi: Update.
15280 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15282 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15284 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15285 (*insv_reg<mode>): New define_insn.
15287 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15289 PR rtl-optimization/57439
15290 * postreload.c (move2add_valid_value_p): Check that we have
15291 a zero subreg_regno_offset when accessing the register in
15292 the requested mode.
15294 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15295 Igor Zamyatin <igor.zamyatin@intel.com>
15297 Silvermont (SLM) architecture pipeline model, tuning and
15299 * config.gcc: Add slm config options and target.
15301 * config/i386/slm.md: New.
15303 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15305 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15307 (ix86_target_macros_internal): Likewise.
15309 * config/i386/i386.c (slm_cost): New cost.
15310 (m_SLM): New macro flag.
15311 (initial_ix86_tune_features): Set m_SLM.
15312 (x86_accumulate_outgoing_args): Likewise.
15313 (x86_arch_always_fancy_math_387): Likewise.
15314 (processor_target_table): Add slm cost.
15315 (cpu_names): Add slm cpu name.
15316 (x86_option_override_internal): Set SLM ISA.
15317 (ix86_issue_rate): New case PROCESSOR_SLM.
15318 (ia32_multipass_dfa_lookahead): Likewise.
15319 (fold_builtin_cpu): Add slm.
15321 * config/i386/i386.h (TARGET_SLM): New target macro.
15322 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15323 (processor_type): Add PROCESSOR_SLM.
15325 * config/i386/i386.md (cpu): Add new value "slm".
15326 (slm.md): Include slm.md.
15328 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15329 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15331 * config/arm/arm-protos.h: Add and update function protos.
15332 * config/arm/arm.c (use_simple_return_p): New added.
15333 (thumb2_expand_return): Check simple_return flag.
15334 * config/arm/arm.md: Add simple_return and conditional simple_return.
15335 * config/arm/iterators.md: Add iterator for return and simple_return.
15337 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15339 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15340 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15341 (arm_emit_vfp_multi_reg_pop): Likewise.
15342 (thumb2_emit_ldrd_pop): Likewise.
15343 (arm_expand_epilogue): Add misc REG_CFA notes.
15344 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15346 2013-05-29 Lawrence Crowl <crowl@google.com>
15348 * config/arm/t-arm: Update for below.
15350 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15351 Change type to hash_table. Update dependent calls and types.
15353 * config/i386/t-cygming: Update for below.
15355 * config/i386/t-interix: Update for below.
15357 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15358 Change type to hash_table. Update dependent calls and types.
15359 (i386_find_on_wrapper_list::wrappers): Likewise.
15361 * config/ia64/t-ia64: Update for below.
15363 * config/ia64/ia64.c (bundle_state_table):
15364 Change type to hash_table. Update dependent calls and types.
15366 * config/mips/mips.c (mips_reorg_process_insns::htab):
15367 Change type to hash_table. Update dependent calls and types.
15369 * config/sol2.c (solaris_comdat_htab):
15370 Change type to hash_table. Update dependent calls and types.
15372 * config/t-sol2: Update for above.
15374 2013-05-29 Teresa Johnson <tejohnson@google.com>
15376 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15377 functions are not yet marked as defined.
15379 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15380 Pat Haugen <pthaugen@us.ibm.com>
15381 Peter Bergner <bergner@vnet.ibm.com>
15383 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15387 (vrotl<mode>3): Likewise.
15388 (vashl<mode>3): Likewise.
15389 (vlshr<mode>3): Likewise.
15390 (vashr<mode>3): Likewise.
15392 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15393 support for power8 V2DI builtins.
15395 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15396 power8 V2DI builtins.
15397 (vupkhsw): Likewise.
15398 (vupklsw): Likewise.
15399 (vaddudm): Likewise.
15400 (vminsd): Likewise.
15401 (vmaxsd): Likewise.
15402 (vminud): Likewise.
15403 (vmaxud): Likewise.
15404 (vpkudum): Likewise.
15405 (vpksdss): Likewise.
15406 (vpkudus): Likewise.
15407 (vpksdus): Likewise.
15412 (vsubudm): Likewise.
15413 (vcmpequd): Likewise.
15414 (vcmpgtsd): Likewise.
15415 (vcmpgtud): Likewise.
15416 (vcmpequd_p): Likewise.
15417 (vcmpgtsd_p): Likewise.
15418 (vcmpgtud_p): Likewise.
15419 (vupkhsw): Likewise.
15420 (vupklsw): Likewise.
15421 (vaddudm): Likewise.
15422 (vmaxsd): Likewise.
15423 (vmaxud): Likewise.
15424 (vminsd): Likewise.
15425 (vminud): Likewise.
15426 (vpksdss): Likewise.
15427 (vpksdus): Likewise.
15428 (vpkudum): Likewise.
15429 (vpkudus): Likewise.
15434 (vsubudm): Likewise.
15436 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15437 support for power8 V2DI instructions.
15439 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15440 power8 V2DI instructions. Combine pack and unpack insns to use an
15441 iterator for each mode. Check whether a particular mode supports
15442 Altivec instructions instead of just checking TARGET_ALTIVEC.
15443 (UNSPEC_VPKUWUM): Likewise.
15444 (UNSPEC_VPKSHSS): Likewise.
15445 (UNSPEC_VPKSWSS): Likewise.
15446 (UNSPEC_VPKUHUS): Likewise.
15447 (UNSPEC_VPKSHUS): Likewise.
15448 (UNSPEC_VPKUWUS): Likewise.
15449 (UNSPEC_VPKSWUS): Likewise.
15450 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15451 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15452 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15453 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15454 (UNSPEC_VUPKHSB): Likewise.
15455 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15456 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15457 (UNSPEC_VUPKHSH): Likewise.
15458 (UNSPEC_VUPKLSB): Likewise.
15459 (UNSPEC_VUPKLSH): Likewise.
15461 (VI_char): Likewise.
15462 (VI_scalar): Likewise.
15463 (VI_unit): Likewise.
15465 (VP_small): Likewise.
15466 (VP_small_lc): Likewise.
15467 (VU_char): Likewise.
15468 (add<mode>3): Likewise.
15469 (altivec_vaddcuw): Likewise.
15470 (altivec_vaddu<VI_char>s): Likewise.
15471 (altivec_vadds<VI_char>s): Likewise.
15472 (sub<mode>3): Likewise.
15473 (altivec_vsubcuw): Likewise.
15474 (altivec_vsubu<VI_char>s): Likewise.
15475 (altivec_vsubs<VI_char>s): Likewise.
15476 (altivec_vavgs<VI_char>): Likewise.
15477 (altivec_vcmpbfp): Likewise.
15478 (altivec_eq<mode>): Likewise.
15479 (altivec_gt<mode>): Likewise.
15480 (altivec_gtu<mode>): Likewise.
15481 (umax<mode>3): Likewise.
15482 (smax<mode>3): Likewise.
15483 (umin<mode>3): Likewise.
15484 (smin<mode>3): Likewise.
15485 (altivec_vpkuhum): Likewise.
15486 (altivec_vpkuwum): Likewise.
15487 (altivec_vpkshss): Likewise.
15488 (altivec_vpkswss): Likewise.
15489 (altivec_vpkuhus): Likewise.
15490 (altivec_vpkshus): Likewise.
15491 (altivec_vpkuwus): Likewise.
15492 (altivec_vpkswus): Likewise.
15493 (altivec_vpks<VI_char>ss): Likewise.
15494 (altivec_vpks<VI_char>us): Likewise.
15495 (altivec_vpku<VI_char>us): Likewise.
15496 (altivec_vpku<VI_char>um): Likewise.
15497 (altivec_vrl<VI_char>): Likewise.
15498 (altivec_vsl<VI_char>): Likewise.
15499 (altivec_vsr<VI_char>): Likewise.
15500 (altivec_vsra<VI_char>): Likewise.
15501 (altivec_vsldoi_<mode>): Likewise.
15502 (altivec_vupkhsb): Likewise.
15503 (altivec_vupkhs<VU_char>): Likewise.
15504 (altivec_vupkls<VU_char>): Likewise.
15505 (altivec_vupkhsh): Likewise.
15506 (altivec_vupklsb): Likewise.
15507 (altivec_vupklsh): Likewise.
15508 (altivec_vcmpequ<VI_char>_p): Likewise.
15509 (altivec_vcmpgts<VI_char>_p): Likewise.
15510 (altivec_vcmpgtu<VI_char>_p): Likewise.
15511 (abs<mode>2): Likewise.
15512 (vec_unpacks_hi_v16qi): Likewise.
15513 (vec_unpacks_hi_v8hi): Likewise.
15514 (vec_unpacks_lo_v16qi): Likewise.
15515 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15516 (vec_unpacks_lo_v8hi): Likewise.
15517 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15518 (vec_pack_trunc_v8h): Likewise.
15519 (vec_pack_trunc_v4si): Likewise.
15520 (vec_pack_trunc_<mode>): Likewise.
15522 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15524 (vec_vmaxsd): Likewise.
15525 (vec_vmaxud): Likewise.
15526 (vec_vminsd): Likewise.
15527 (vec_vminud): Likewise.
15528 (vec_vpksdss): Likewise.
15529 (vec_vpksdus): Likewise.
15530 (vec_vpkudum): Likewise.
15531 (vec_vpkudus): Likewise.
15532 (vec_vrld): Likewise.
15533 (vec_vsld): Likewise.
15534 (vec_vsrad): Likewise.
15535 (vec_vsrd): Likewise.
15536 (vec_vsubudm): Likewise.
15537 (vec_vupkhsw): Likewise.
15538 (vec_vupklsw): Likewise.
15540 2013-05-29 Jan Hubicka <jh@suse.cz>
15542 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15543 flags; reorder rest of fields in more consistent way.
15544 (varpool_node): Remove analyzed, finalized and alias.
15545 (cgraph_ndoe): Likewise.
15546 (symtab_alias_ultimate_target): New function.
15547 (cgraph_function_node): Move offline.
15548 (cgraph_reset_node): Declare.
15549 (cgraph_comdat_can_be_unshared_p): Remove.
15550 (varpool_remove_initializer): Declare.
15551 (varpool_first_defined_variable, varpool_next_defined_variable
15552 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15553 (cgraph_function_with_gimple_body_p): Update.
15554 (varpool_all_refs_explicit_p): Update.
15555 (symtab_alias_target): New function.
15556 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15557 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15558 (cgraph_function_or_thunk_node): Simplify using
15559 symtab_alias_ultimate_target.
15560 (varpool_variable_node): Likewise.
15561 * cgraph.c (cgraph_create_function_alias): Update.
15562 (cgraph_add_thunk): Update.
15563 (cgraph_remove_node): Update.
15564 (dump_cgraph_node): Do not dump removed flags.
15565 (cgraph_function_body_availability): Update.
15566 (cgraph_propagate_frequency): Update.
15567 (verify_cgraph_node): Check sanity of local flag.
15568 (cgraph_function_node): Move here from cgraph.h; revamp for
15569 cgraph_function_or_thunk_node.
15570 * lto-symtab.c (lto_varpool_replace_node): Update.
15571 (lto_symtab_resolve_can_prevail_p): Update.
15572 (lto_symtab_merge_cgraph_nodes): Update.
15573 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15574 propagate_constants_accross_call, devirtualization_time_bonus,
15575 ipcp_propagate_stage): Update.
15576 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15577 * ipa-inline-transform.c (clone_inlined_nodes,
15578 preserve_function_body_p): Update.
15579 * ipa-reference.c (propagate): Update.
15580 (write_node_summary_p): Update.
15581 * toplev.c (wrapup_global_declaration_2): Update.
15582 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15583 (analyze_function) ... this one.
15584 (cgraph_process_new_functions): Update.
15585 (cgraph_reset_node): Export.
15586 (cgraph_finalize_function): Update.
15587 (cgraph_add_new_function): Update.
15588 (process_function_and_variable_attributes): Update.
15589 (varpool_finalize_decl): Update.
15590 (symbol_finalized): Remove.
15591 (symbol_finalized_and_needed): Rename to ...
15592 (symbol_defined_and_needed): ... update.
15593 (cgraph_analyze_functions): Update.
15594 (handle_alias_pairs): Update.
15595 (mark_functions_to_output): Update.
15596 (assemble_thunk): Update.
15597 (output_in_order): Update.
15598 (output_weakrefs): Update.
15599 (finalize_compilation_unit): Update.
15600 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15601 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15602 input_node, input_varpool_node): Update.
15603 * dbxout.c (dbxout_expand_expr): Update.
15604 * cgraphclones.c (cgraph_clone_node): Update.
15605 (cgraph_copy_node_for_versioning): Update.
15606 (cgraph_materialize_clone): Update.
15607 (cgraph_materialize_all_clones): Update.
15608 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15609 propagate_pure_const, propagate_nothrow): Update.
15610 * lto-streamer-out.c (lto_output, write_symbol): Update.
15611 * ipa-utils.c (ipa_reverse_postorder): Update.
15612 * ipa-inline.c (can_inline_edge_p): Update.
15613 (update_caller_keys, ipa_inline): Update.
15614 * dwarf2out.c (reference_to_unused,
15615 premark_types_used_by_global_vars_helper): Update.
15616 * tree-eh.c (tree_could_trap_p): Update.
15617 * ipa-split.c (consider_split, execute_split_functions): Update.
15618 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15619 has_addr_references_p): Update; move ahead in file for better
15621 (process_references): Simplify.
15622 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15623 bodies are removed.
15624 (cgraph_comdat_can_be_unshared_p): Make static.
15625 (cgraph_externally_visible_p): Update.
15626 (varpool_externally_visible_p): Update.
15627 (function_and_variable_visibility): Update.
15628 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15629 ipa_tm_mark_force_output_node): Update.
15630 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15631 estimate_edge_devirt_benefit, inline_generate_summary,
15632 inline_write_summary): Update.
15633 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15634 * ipa-prop.c (ipa_compute_jump_functions): Update.
15635 (ipa_print_node_params, ipa_prop_read_section,
15636 ipa_update_after_lto_read, read_replacements_section): Update.
15637 * varasm.c (mark_decl_referenced): Update.
15638 (assemble_alias, dump_tm_clone_pairs): Update.
15639 * tree-inline.c (copy_bb): Update.
15640 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15642 * symtab.c (dump_symtab_base): Print new flags.
15643 (verify_symtab_base): Verify new flags.
15644 (symtab_alias_ultimate_target): New function.
15645 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15646 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15648 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15650 * i386.c (ix86_get_function_versions_dispatcher,
15651 ix86_generate_version_dispatcher_body): Update.
15652 (fold_builtin_cpu): Use varpool_add_new_variable.
15653 * varpool.c (varpool_remove_initializer): Break out from ...
15654 (varpool_remove_node): ... this one.
15655 (dump_varpool_node, varpool_node_for_asm,
15656 cgraph_variable_initializer_availability, varpool_analyze_node,
15657 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15658 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15661 2013-05-29 Jan Hubicka <jh@suse.cz>
15663 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15665 2013-05-29 Easwaran Raman <eraman@google.com>
15667 PR tree-optimization/57442
15668 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15669 when control exits the main loop.
15671 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15673 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15675 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15676 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15677 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15680 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15682 PR tree-optimization/57441
15683 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15684 Don't limit size of incr_vec to number of candidates.
15686 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15688 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15689 and mips16 directories.
15690 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15691 (MULTILIB_DIRNAMES): Ditto.
15692 (MULTILIB_EXCEPTIONS): Add new exceptions.
15693 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15694 (MULTILIB_DIRNAMES): Ditto.
15695 (MULTILIB_EXCEPTIONS): Add new exceptions.
15697 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15698 Marcus Shawcroft <marcus.shawcroft@arm.com>
15700 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15701 SYMBOL_TINY_ABSOLUTE.
15702 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15703 SYMBOL_TINY_ABSOLUTE.
15704 (aarch64_expand_mov_immediate): Likewise.
15705 (aarch64_classify_symbol): Likewise.
15706 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15707 Permit SYMBOL_TINY_ABSOLUTE.
15708 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15710 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15711 Marcus Shawcroft <marcus.shawcroft@arm.com>
15713 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15714 Refactor if/switch. Replace gcc_assert with if.
15716 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15718 * config/i386/i386.c (initial_ix86_tune_features): Enable
15719 FP Reassociation for AMD bdver1 and bdver2.
15721 2013-05-29 Martin Jambor <mjambor@suse.cz>
15723 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15724 and IMAGPART_EXPR do not occur within other handled_components.
15726 2013-05-29 Richard Biener <rguenther@suse.de>
15728 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15729 access on whether the use is in the BB we currently try to
15731 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15732 vectorize to vect_bb_slp_scalar_cost.
15734 2013-05-29 Richard Biener <rguenther@suse.de>
15736 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15737 computing scalar cost offsetted by stmts that are kept live
15739 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15740 for computation of scalar cost.
15742 2013-05-28 Steve Ellcey <sellcey@mips.com>
15744 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15746 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15748 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15749 array notation built-in reduction functions.
15750 * doc/passes.texi (Passes): Added documentation about changes done
15752 * doc/invoke.texi (C Dialect Options): Added documentation about
15753 the -fcilkplus flag.
15754 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15755 (BUILTINS_DEF): Depend on cilkplus.def.
15756 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15757 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15758 * cilkplus.def: New file.
15760 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15762 PR rtl-optimization/57439
15763 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15765 2013-05-28 Easwaran Raman <eraman@google.com>
15767 PR tree-optimization/57337
15768 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15769 (find_insert_point): Correctly identify the insertion point
15770 when two statements with the same UID is compared.
15772 2013-05-28 Richard Biener <rguenther@suse.de>
15774 PR tree-optimization/56787
15775 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15776 from the list of data references.
15777 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15779 (vect_analyze_loop_operations): Likewise.
15780 (vect_transform_loop): Remove clobbers.
15782 2013-05-28 Martin Jambor <mjambor@suse.cz>
15784 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15785 and REALPART_EXPRs have scalar type.
15787 2013-05-28 Richard Biener <rguenther@suse.de>
15789 PR tree-optimization/57411
15790 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15792 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15793 virtual operand propagation.
15795 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15797 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15798 destination register for bmasksi_vis.
15799 (vector_init_bshuffle): Likewise.
15800 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15802 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15804 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15805 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15806 mode if the instruction isn't available in the original mode.
15807 * config/sparc/sparc.opt (mfix-ut699): New option.
15808 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15809 (divdf3): Turn into expander.
15810 (divdf3_nofix): New insn.
15811 (divdf3_fix): Likewise.
15812 (divsf3): Disable if -mfix-ut699.
15813 (sqrtdf2): Turn into expander.
15814 (sqrtdf2_nofix): New insn.
15815 (sqrtdf2_fix): Likewise.
15816 (sqrtsf2): Disable if -mfix-ut699.
15818 2013-05-27 Richard Biener <rguenther@suse.de>
15820 PR middle-end/57412
15821 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15822 block for the new loop.
15824 2013-05-27 Richard Biener <rguenther@suse.de>
15826 PR tree-optimization/57343
15827 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15828 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15829 (number_of_iterations_cond): Do not build the folded tree.
15831 2013-05-27 Richard Biener <rguenther@suse.de>
15834 PR middle-end/57381
15835 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15836 OEP_CONSTANT_ADDRESS_OF retained.
15838 PR tree-optimization/57417
15839 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15840 for unchanged base.
15841 (set_ssa_val_to): Compare addresses using
15842 get_addr_base_and_unit_offset.
15844 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15846 PR rtl-optimization/56833
15847 * postreload.c (move2add_record_mode): New function.
15848 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15849 (move2add_use_add2_insn): Use move2add_record_sym_value.
15850 (move2add_use_add3_insn): Likewise.
15851 (reload_cse_move2add): Use move2add_valid_value_p and
15852 move2add_record_mode. Invalidate call-clobbered and REG_INC
15853 affected regs by setting reg_mode to VOIDmode.
15854 (move2add_note_store): Don't pretend the inside of a SUBREG is
15855 the actual destination. Invalidate single/leading registers by
15856 setting reg_mode to VOIDmode.
15857 Use move2add_record_sym_value, move2add_valid_value_p and
15858 move2add_record_mode.
15860 2013-05-27 Richard Biener <rguenther@suse.de>
15862 PR tree-optimization/57396
15863 * tree-affine.c (double_int_constant_multiple_p): Properly
15864 return false for val == 0 and div != 0.
15866 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15868 * config/mips/mips.h: Use #elif in preprocessor conditions.
15870 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15873 * config/mips/constraints.md (kl): New constraint.
15874 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15875 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15876 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15877 from LO for MIPS16.
15878 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15880 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15883 * config/mips/mips.c (mips_can_inline_p): New function.
15884 (TARGET_CAN_INLINE_P): Define.
15886 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15888 * sched-int.h (ds_t, dw_t): Make unsigned int.
15889 Fix documentation that describes how all the ds_t bits are used.
15890 Reserve the last bit for delayed-branch scheduling.
15891 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15892 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15893 (gen_dep_weak_1): Remove prototype.
15894 * sched-deps.c (get_dep_weak_1): Make static.
15895 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15896 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15897 * doc/tm.texi: Regenerate.
15898 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15900 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15903 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15905 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15906 Sandra Loosemore <sandra@codesourcery.com>
15908 * config.gcc (powerpc-*): Allow native for with-cpu.
15910 2013-05-24 Jeff Law <law@redhat.com>
15912 PR tree-optimization/57124
15913 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15914 conversion feeding a condition if the range has an overflow
15915 if -fstrict-overflow. Add warnings for when we do make the
15918 2013-05-24 Dehao Chen <dehao@google.com>
15920 * tree-cfg.c (locus_discrim_map): Fix the typo.
15921 (locus_discrim_hasher): Likewise.
15922 (locus_discrim_hasher::hash): Likewise.
15923 (locus_discrim_hasher::equal): Likewise.
15925 2013-05-24 Martin Jambor <mjambor@suse.cz>
15927 PR tree-optimization/57294
15928 * cgraph.h (ipa_record_stmt_references): Declare.
15929 * cgraphbuild.c (ipa_record_stmt_references): New function.
15930 (build_cgraph_edges): Use ipa_record_stmt_references.
15931 (rebuild_cgraph_edges): Likewise.
15932 (cgraph_rebuild_references): Likewise.
15933 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15934 associated with the old statement and build references from the
15935 newly built statements.
15936 * ipa-ref.c (ipa_remove_stmt_references): New function.
15937 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15939 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15941 * lra-constraints.c (emit_spill_move): Use smaller mode for
15943 (check_and_process_move): Consider mem-reg moves for secondary
15945 (curr_insn_transform): Don't lose insns emitted before for
15946 secondary memory moves.
15947 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15948 reg set up in the current insn.
15950 2013-05-24 Dehao Chen <dehao@google.com>
15952 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
15954 (locus_descrim_hasher::equal): Likewise.
15955 (build_gimple_cfg): New discriminator assignment algorithm.
15956 (make_edges): Likewise.
15957 (next_discriminator_for_locus): Likewise.
15958 (same_line_p): Likewise.
15959 (assign_discriminators): Likewise.
15960 (make_cond_expr_edges): Likewise.
15961 (make_gimple_switch_edges): Likewise.
15962 (make_goto_expr_edges): Likewise.
15963 (make_gimple_asm_edges): Likewise.
15965 2013-05-24 Ian Bolton <ian.bolton@arm.com>
15967 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15968 X format specifier to only display bottom 16 bits.
15969 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15970 immediate to match for operand 2, since it will be masked.
15972 2013-05-24 Richard Biener <rguenther@suse.de>
15974 PR tree-optimization/57287
15975 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15976 all SSA names that occur in abnormal PHIs.
15978 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15981 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15982 that index is not negative.
15984 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15986 PR rtl-optimization/55177
15987 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15988 (simplify_byte_swapping_operation): New.
15989 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15990 (simplify_relational_operation_1): Deal with BSWAP.
15992 2013-05-23 Richard Henderson <rth@redhat.com>
15995 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15996 (ix86_reorg): Call it.
15998 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16001 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16002 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16003 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16005 2013-05-23 Christian Bruel <christian.bruel@st.com>
16008 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16010 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16011 Marcus Shawcroft <marcus.shawcroft@arm.com>
16013 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16014 * config/aarch64/constraints.md (Usa): Remove.
16015 * doc/md.texi (AArch64 Usa): Remove.
16017 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16018 Marcus Shawcroft <marcus.shawcroft@arm.com>
16020 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16021 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16022 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16023 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16025 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16027 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16028 instruction (AdvSIMD).
16029 * config/aarch64/aarch64-builtins.c
16030 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16031 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16033 2013-05-23 Martin Jambor <mjambor@suse.cz>
16035 PR middle-end/57347
16036 * tree.h (contains_bitfld_component_ref_p): Declare.
16037 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16038 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16040 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16041 not access a bit-field. Assert all final offsets are byte-aligned.
16043 2013-05-23 Richard Biener <rguenther@suse.de>
16045 PR tree-optimization/57380
16046 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16047 least one invariant or re-used load.
16048 * passes.c (init_optimization_passes): Move pass_phiprop before
16051 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16053 * config/aarch64/aarch64-simd.md
16054 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16056 2013-05-23 Richard Biener <rguenther@suse.de>
16058 PR middle-end/57381
16059 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16060 OEP_CONSTANT_ADDRESS_OF retained.
16062 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16064 PR middle-end/57344
16065 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16066 don't lower unit. Handle unit not being always BITS_PER_WORD.
16068 2013-05-23 Richard Biener <rguenther@suse.de>
16070 PR rtl-optimization/57341
16071 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16072 instead of true_dependence.
16074 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16076 * bb-reorder.c (branch_threshold): Make const.
16077 (exec_threshold): Ditto.
16079 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16080 Pat Haugen <pthaugen@us.ibm.com>
16081 Peter Bergner <bergner@vnet.ibm.com>
16083 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16084 documentation for the power8 crypto builtins.
16086 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16088 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16089 macros for defining power8 builtin functions.
16090 (BU_P8V_AV_2): Likewise.
16091 (BU_P8V_AV_P): Likewise.
16092 (BU_P8V_VSX_1): Likewise.
16093 (BU_P8V_OVERLOAD_1): Likewise.
16094 (BU_P8V_OVERLOAD_2): Likewise.
16095 (BU_CRYPTO_1): Likewise.
16096 (BU_CRYPTO_2): Likewise.
16097 (BU_CRYPTO_3): Likewise.
16098 (BU_CRYPTO_OVERLOAD_1): Likewise.
16099 (BU_CRYPTO_OVERLOAD_2): Likewise.
16100 (XSCVSPDP): Fix typo, point to the correct instruction.
16101 (VCIPHER): Add power8 crypto builtins.
16102 (VCIPHERLAST): Likewise.
16103 (VNCIPHER): Likewise.
16104 (VNCIPHERLAST): Likewise.
16105 (VPMSUMB): Likewise.
16106 (VPMSUMH): Likewise.
16107 (VPMSUMW): Likewise.
16108 (VPERMXOR_V2DI): Likewise.
16109 (VPERMXOR_V4SI: Likewise.
16110 (VPERMXOR_V8HI: Likewise.
16111 (VPERMXOR_V16QI: Likewise.
16112 (VSHASIGMAW): Likewise.
16113 (VSHASIGMAD): Likewise.
16114 (VPMSUM): Likewise.
16115 (VPERMXOR): Likewise.
16116 (VSHASIGMA): Likewise.
16118 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16119 __CRYPTO__ if the crypto instructions are available.
16120 (altivec_overloaded_builtins): Add support for overloaded power8
16123 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16124 support for power8 crypto builtins.
16125 (builtin_function_type): Likewise.
16126 (altivec_init_builtins): Add support for builtins that take vector
16127 long long (V2DI) arguments.
16129 * config/rs6000/crypto.md: New file, define power8 crypto
16132 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16133 Pat Haugen <pthaugen@us.ibm.com>
16134 Peter Bergner <bergner@vnet.ibm.com>
16136 * doc/invoke.texi (Option Summary): Add power8 options.
16137 (RS/6000 and PowerPC Options): Likewise.
16139 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16140 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16141 wm, wn, wr documentation.
16143 * config/rs6000/constraints.md (wm): New constraint for VSX
16144 registers if direct move instructions are enabled.
16145 (wn): New constraint for no registers.
16146 (wq): New constraint for quad word even GPR registers.
16147 (wr): New constraint if 64-bit instructions are enabled.
16148 (wv): New constraint if power8 vector instructions are enabled.
16149 (wQ): New constraint for quad word memory locations.
16151 * config/rs6000/predicates.md (const_0_to_15_operand): New
16152 constraint for 0..15 for crypto instructions.
16153 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16154 as GPR and floating point registers.
16155 (int_reg_operand): New predicate to match only GPR registers.
16156 (base_reg_operand): New predicate to match base registers.
16157 (quad_int_reg_operand): New predicate to match even GPR registers
16158 for quad memory operations.
16159 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16160 operations in both GPR and VSX registers.
16161 (quad_memory_operand): New predicate for quad memory operations.
16162 (reg_or_indexed_operand): New predicate for direct move support.
16164 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16165 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16166 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16167 (POWERPC_MASKS): Add power8 options.
16168 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16171 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16172 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16174 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16175 (-mpower8-fusion): New power8 options.
16176 (-mpower8-fusion-sign): Likewise.
16177 (-mpower8-vector): Likewise.
16178 (-mcrypto): Likewise.
16179 (-mdirect-move): Likewise.
16180 (-mquad-memory): Likewise.
16182 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16183 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16185 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16186 (rs6000_debug_vector_unit): Add p8_vector.
16187 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16188 definitions. Also print fusion state.
16189 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16190 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16191 (rs6000_option_override_internal): Add support for power8.
16192 (rs6000_common_init_builtins): Add debugging for skipped builtins
16193 if -mdebug=builtin.
16194 (rs6000_adjust_cost): Add power8 support.
16195 (rs6000_issue_rate): Likewise.
16196 (insn_must_be_first_in_group): Likewise.
16197 (insn_must_be_last_in_group): Likewise.
16198 (force_new_group): Likewise.
16199 (rs6000_register_move_cost): Likewise.
16200 (rs6000_opt_masks): Likewise.
16202 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16203 power8 capable assembler, default to power7 options.
16204 (TARGET_DIRECT_MOVE): Likewise.
16205 (TARGET_CRYPTO): Likewise.
16206 (TARGET_P8_VECTOR): Likewise.
16207 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16208 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16209 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16210 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16211 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16212 (TARGET_XSCVDPSPN): Likewise.
16213 (TARGET_XSCVSPDPN): Likewsie.
16214 (TARGET_SYNC_HI_QI): Likewise.
16215 (TARGET_SYNC_TI): Likewise.
16216 (MASK_CRYPTO): Likewise.
16217 (MASK_DIRECT_MOVE): Likewise.
16218 (MASK_P8_FUSION): Likewise.
16219 (MASK_P8_VECTOR): Likewise.
16220 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16221 temporary used by some of the direct move instructions to get two FP
16222 temporary registers does not force creation of a stack frame.
16223 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16224 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16225 that any VSX registers are tieable, even if they are also an
16226 Altivec vector mode.
16227 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16228 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16229 (RS6000_BTM_CRYPTO): Likewise.
16230 (RS6000_BTM_COMMON): Likewise.
16232 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16233 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16234 (enum rs6000_vector): Add power8 vector support.
16236 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16240 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16241 (any_sibcall_could_use_r3): this and handle indirect calls.
16242 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16244 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16246 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16248 2013-05-22 Richard Biener <rguenther@suse.de>
16250 PR middle-end/57349
16251 * profile.c (branch_prob): Do not split blocks that are
16252 abnormally receiving from ECF_RETURNS_TWICE functions.
16254 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16256 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16257 address mode. Move pointer mode initialization to the same place.
16259 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16261 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16262 while it has any effect.
16264 2013-05-21 Easwaran Raman <eraman@google.com>
16266 PR tree-optimization/57322
16267 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16268 UID of the statement added to the BB to be 1.
16270 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16272 PR tree-optimization/57331
16273 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16274 of conversion from pointer type to integral type with integer.
16276 2013-05-21 Martin Jambor <mjambor@suse.cz>
16279 * ipa-prop.c (ipa_read_node_info): Process param_used and
16280 controlled_uses in the same order as when writing.
16282 2013-05-21 Magnus Granberg <baldrick@free.fr>
16285 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16287 2013-05-21 Richard Biener <rguenther@suse.de>
16289 PR tree-optimization/57318
16290 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16291 estimate stmts with side-effects as likely eliminated.
16293 2013-05-21 Richard Biener <rguenther@suse.de>
16295 PR tree-optimization/57330
16296 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16297 preserve the call stmts fntype.
16299 2013-05-21 Richard Biener <rguenther@suse.de>
16301 PR tree-optimization/57303
16302 * tree-ssa-sink.c (statement_sink_location): Improve killing
16303 stmt detection and properly handle self-assignments.
16305 2013-05-21 Christian Bruel <christian.bruel@st.com>
16307 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16308 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16309 registers. Set register size out of the PARALLEL loop.
16311 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16314 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16315 (*fmasf4, *fmasf4_media): New insns.
16317 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16319 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16320 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16321 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16322 (mips_idiv_insns): Update the comments to say that the returned
16323 instruction counts are in units of BASE_INSN_LENGTH.
16324 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16325 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16326 using 2 rather than 4 as the length of indirect MIPS16 and
16327 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16328 length of a NOP. Don't divide MIPS16 lengths by 2.
16329 (mips16_split_long_branches): Assume a branch is long if the
16330 length is greater than 4 rather than 8.
16331 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16332 rather than multiplying them by 2. Multiply instruction counts
16333 by BASE_INSN_LENGTH rather than 4.
16334 (*jump_mips16, tls_get_tp_mips16_<mode>)
16335 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16337 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16339 * config/mips/mips.md (extended_mips16): Remove branch case.
16340 (length): Remove duplicated extended_mips16 test.
16342 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16344 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16346 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16348 * recog.h (Recog_data): Rename to...
16349 (recog_data_d): ...this.
16350 (recog_data): Update accordingly.
16351 * recog.c (recog_data): Likewise.
16352 * reload.c (save_recog_data): Likewise.
16353 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16354 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16356 2013-05-17 Julian Brown <julian@codesourcery.com>
16358 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16359 found in a REG_EQUAL note, invalidate it.
16361 2013-05-17 Easwaran Raman <eraman@google.com>
16363 * tree-ssa-reassoc.c (find_insert_point): New function.
16364 (insert_stmt_after): Likewise.
16365 (get_def_stmt): Likewise.
16366 (ensure_ops_are_available): Likewise.
16367 (not_dominated_by): Likewise.
16368 (rewrite_expr_tree): Do not move statements beyond what is
16369 necessary. Remove call to swap_ops_for_binary_stmt...
16370 (reassociate_bb): ... and move it here.
16371 (build_and_add_sum): Assign UIDs for new statements.
16372 (linearize_expr): Likewise.
16373 (do_reassoc): Renumber gimple statement UIDs.
16375 2013-05-17 Jan Hubicka <jh@suse.cz>
16377 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16379 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16380 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16382 (output_weakrefs): Update.
16384 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16385 Martin Jambor <mjambor@suse.cz>
16387 PR middle-end/57276
16388 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16389 value that corresponds to the given aggval is found in values vector.
16391 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16393 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16394 sse, sse2, sse3, ssse3 and sse4a flags to options.
16396 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16398 * gengtype-state.c: (s_expr_writer): New class, to handle
16399 prettifying of output layout of s-expressions.
16400 (state_writer): New class, to write out gtype.state.
16401 (state_written_type_count): Move this variable into member data of
16403 (s_expr_writer::s_expr_writer): New code: constructor for new class
16404 (state_writer::state_writer(): ditto
16405 (s_expr_writer::write_new_line): New function
16406 (s_expr_writer::write_any_indent): ditto
16407 (s_expr_writer::begin_s_expr): ditto
16408 (s_expr_writer::end_s_expr): ditto
16409 (write_state_fileloc): convert to method of state_writer...
16410 (state_writer:: write_state_fileloc): ...and use methods of
16411 s_expr_writer to write indentation into the gtype.state output file
16412 to visually represent the hierarchical structure of the list
16414 (write_state_fields): ditto, renaming to...
16415 (state_writer::write_state_fields)
16416 (write_state_a_string): ditto, renaming to...
16417 (state_writer::write_state_a_string)
16418 (write_state_string_option): ditto, renaming to...
16419 (state_writer::write_state_string_option)
16420 (write_state_type_option): ditto, renaming to...
16421 (state_writer::write_state_type_option)
16422 (write_state_nested_option): ditto, renaming to...
16423 (state_writer::write_state_nested_option)
16424 (write_state_option): ditto, renaming to...
16425 (state_writer::write_state_option)
16426 (write_state_options): ditto, renaming to...
16427 (state_writer::write_state_options)
16428 (write_state_lang_bitmap): ditto, renaming to...
16429 (state_writer::write_state_lang_bitmap)
16430 (write_state_version): ditto, renaming to...
16431 (state_writer::write_state_version)
16432 (write_state_scalar_type): ditto, renaming to...
16433 (state_writer::write_state_scalar_type)
16434 (write_state_string_type): ditto, renaming to...
16435 (state_writer::write_state_string_type)
16436 (write_state_undefined_type): ditto, renaming to...
16437 (state_writer::write_state_undefined_type)
16438 (write_state_struct_union_type): ditto, renaming to...
16439 (state_writer::write_state_struct_union_type)
16440 (write_state_struct_type): ditto, renaming to...
16441 (state_writer::write_state_struct_type)
16442 (write_state_user_struct_type): ditto, renaming to...
16443 (state_writer::write_state_user_struct_type)
16444 (write_state_lang_struct_type): ditto, renaming to...
16445 (state_writer::write_state_lang_struct_type)
16446 (write_state_param_struct_type): ditto, renaming to...
16447 (state_writer::write_state_param_struct_type)
16448 (write_state_pointer_type): ditto, renaming to...
16449 (state_writer::write_state_pointer_type)
16450 (write_state_array_type): ditto, renaming to...
16451 (state_writer::write_state_array_type)
16452 (write_state_gc_used): ditto, renaming to...
16453 (state_writer::write_state_gc_used)
16454 (write_state_common_type_content): ditto, renaming to...
16455 (state_writer::write_state_common_type_content)
16456 (write_state_type): ditto, renaming to...
16457 (state_writer::write_state_type)
16458 (write_state_pair_list): ditto, renaming to...
16459 (state_writer::write_state_pair_list)
16460 (write_state_pair): ditto, renaming to...
16461 (state_writer::write_state_pair)
16462 (write_state_typedefs): ditto, renaming to...
16463 (state_writer::write_state_typedefs)
16464 (write_state_structures): ditto, renaming to...
16465 (state_writer::write_state_structures)
16466 (write_state_param_structs): ditto, renaming to...
16467 (state_writer::write_state_param_structs)
16468 (write_state_variables): ditto, renaming to...
16469 (state_writer::write_state_variables)
16470 (write_state_srcdir): ditto, renaming to...
16471 (state_writer::write_state_srcdir)
16472 (write_state_files_list): ditto, renaming to...
16473 (state_writer::write_state_files_list)
16474 (write_state_languages): ditto, renaming to...
16475 (state_writer::write_state_languages)
16476 (write_state): create a state_writer instance and use it when
16477 writing out the state file
16479 2013-05-17 Mike Stump <mikestump@comcast.net>
16481 PR rtl-optimization/57304
16482 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16483 accessing DF_REF_REAL_LOC.
16485 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16487 PR rtl-optimization/57281
16488 PR rtl-optimization/57300
16489 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16490 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16491 what the other splitter did if the registers are dead.
16493 2013-05-17 Richard Biener <rguenther@suse.de>
16495 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16498 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16500 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16503 2013-05-17 Marek Polacek <polacek@redhat.com>
16505 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16506 length when doing non-zero store of storing '\0' to '\0'.
16508 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16510 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16511 vect_external_def oprnd1 with loop_vinfo, try to emit
16512 optional cast, negation and and stmts on the loop preheader
16513 edge instead of into the pattern def seq.
16515 PR tree-optimization/57051
16516 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16517 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16519 2013-05-16 Nick Clifton <nickc@redhat.com>
16521 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16522 (rl78_is_naked_func): New function.
16523 (rl78_expand_prologue): Skip prologue generation for naked functions.
16524 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16525 * doc/extend.texi (naked): Add RL78 to the list of processors
16526 that supports this attribute.
16528 2013-05-16 Jeff Law <law@redhat.com>
16530 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16532 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16534 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16535 cache parameters using detect_caches_amd also for CYRIX,
16536 NSC and TM2 signatures.
16538 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16539 Dzianis Kahanovich <mahatma@eu.by>
16543 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16544 VIA/Centaur processors and determine their cache parameters
16545 using detect_caches_amd.
16547 2013-05-16 Teresa Johnson <tejohnson@google.com>
16549 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16550 (rtl_verify_edges): New function.
16551 (rtl_verify_bb_insns): Ditto.
16552 (rtl_verify_bb_pointers): Ditto.
16553 (rtl_verify_bb_insn_chain): Ditto.
16554 (rtl_verify_fallthru): Ditto.
16555 (rtl_verify_bb_layout): Ditto.
16556 (rtl_verify_flow_info_1): Outline checks into new functions.
16557 (rtl_verify_flow_info): Ditto.
16559 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16561 * cfghooks.c (copy_bbs): Add update_dominance argument.
16562 * cfghooks.h (copy_bbs): Update prototype.
16563 * tree-cfg.c (gimple_duplicate_sese_region):
16564 Add update_dominance argument.
16565 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16566 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16567 gimple_duplicate_sese_region call.
16568 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16569 Update copy_bbs call.
16570 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16571 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16573 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16575 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16576 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16577 vect_recog_rotate_pattern.
16578 (vect_recog_rotate_pattern): New function.
16580 2013-05-16 Jason Merrill <jason@redhat.com>
16582 * Makefile.in (LLINKER): New variable.
16583 (mostlyclean): Remove link mutex.
16584 * configure.ac: Handle --enable-link-mutex.
16585 * lock-and-run.sh: New script.
16587 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16590 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16593 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16595 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16597 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16599 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16600 * config/arm/arm.c (next_consecutive_mem): New function.
16601 (gen_movmem_ldrd_strd): Likewise.
16602 * config/arm/arm.md (movmemqi): Update condition and code.
16603 (unaligned_loaddi, unaligned_storedi): New patterns.
16605 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16607 * config.gcc: Obsolete *-*-solaris2.9*.
16608 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16610 2013-05-16 Richard Biener <rguenther@suse.de>
16612 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16613 earlier, after GRAPHITE transforms and IV canonicalization.
16615 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16617 * omp-low.c (extract_omp_for_data): For collapsed loops,
16618 if at least one of the loops is known at compile time to
16619 iterate zero times, set count to 0.
16620 (expand_omp_regimplify_p): New function.
16621 (expand_omp_for_generic): For collapsed loops, if at least
16622 one of the loops isn't known to iterate at least once,
16623 add runtime check with setting count to 0.
16624 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16625 For unsigned types if it isn't known at compile time that
16626 the loop will iterate at least once, add runtime check to bypass
16627 the whole loop if initial condition isn't true.
16629 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16631 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16633 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16635 PR middle-end/57286
16636 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16637 transformations to avoid an infinite loop.
16639 2013-05-16 Marek Polacek <polacek@redhat.com>
16641 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16643 2013-05-15 Leif Ekblad <leif@rdos.net>
16645 * config/i386/i386.c (ix86_decompose_address): Use
16646 DEFAULT_TLS_SEG_REG to access TLS segment register.
16647 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16648 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16649 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16651 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16654 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16655 sibling calls to functions that would normally be lazily bound,
16656 unless $gp is call-clobbered.
16658 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16660 * config/i386/i386.c (ix86_option_override_internal): Update
16661 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16662 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16663 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16664 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16666 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16668 2013-05-15 Andreas Schwab <schwab@suse.de>
16670 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16671 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16674 2013-05-15 Teresa Johnson <tejohnson@google.com>
16676 * loop-unroll.c (report_unroll_peel): Check decision before
16677 emitting unroll/peel message.
16679 2013-05-15 Teresa Johnson <tejohnson@google.com>
16681 * function.h (has_bb_partition): New rtl_data flag.
16682 (bb_reorder_complete): Ditto.
16683 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16684 instead of flag_reorder_blocks_and_partition.
16685 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16686 with some enhancements.
16687 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16688 * bb-reorder.c (connect_traces): Check for has_bb_partition
16689 instead of flag_reorder_blocks_and_partition.
16690 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16691 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16692 verify_hot_cold_block_grouping.
16693 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16695 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16698 * config/arm/predicates.md (call_insn_operand): New predicate.
16699 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16700 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16701 if insn is not a tail call.
16702 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16704 * config/arm/arm.h (enum reg_class): New caller save register class.
16705 (REG_CLASS_NAMES): Likewise.
16706 (REG_CLASS_CONTENTS): Likewise.
16707 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16710 2013-05-15 Richard Biener <rguenther@suse.de>
16712 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16713 of MSG_OPTIMIZED_LOCATIONS.
16714 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16715 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16717 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16718 of MSG_OPTIMIZED_LOCATIONS.
16719 (execute_vect_slp): Likewise.
16720 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16721 (vect_create_cond_for_alias_checks): Likewise.
16722 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16723 (vect_recog_widen_mult_pattern): Likewise.
16724 (vect_recog_widen_sum_pattern): Likewise.
16725 (vect_recog_over_widening_pattern): Likewise.
16726 (vect_recog_widen_shift_pattern): Likewise.
16727 (vect_recog_vector_vector_shift_pattern): Likewise.
16728 (vect_recog_divmod_pattern): Likewise.
16729 (vect_recog_mixed_size_cond_pattern): Likewise.
16730 (vect_recog_bool_pattern): Likewise.
16731 (vect_pattern_recog_1): Likewise.
16733 2013-05-15 Martin Jambor <mjambor@suse.cz>
16735 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16736 non-functions to builtin_unreachable.
16737 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16738 correct when new direct edges were discovered.
16740 2013-05-15 Martin Jambor <mjambor@suse.cz>
16742 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16743 header, print symbol order instead of node uid, print more information
16744 about indirect edge targets.
16745 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16747 (ipa_make_edge_direct_to_target): Likewise.
16748 (remove_described_reference): Likewise.
16749 (propagate_controlled_uses): Likewise.
16750 (ipa_print_node_params): Also print symbol order.
16751 (ipcp_transform_function): Print symbol order instead of node uids.
16752 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16753 (cgraph_get_create_real_symbol_node): Likewise.
16754 * ipa-cp.c (print_lattice): Likewise.
16755 (print_all_lattices): Likewise.
16756 (determine_versionability): Likewise.
16757 (initialize_node_lattices): Likewise.
16758 (estimate_local_effects): Likewise.
16759 (update_profiling_info): Likewise.
16760 (create_specialized_node): Likewise.
16761 (perhaps_add_new_callers): Likewise.
16762 (decide_about_value): Likewise.
16763 (decide_whether_version_node): Likewise.
16764 (identify_dead_nodes): Likewise.
16765 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16766 (dump_inline_summary): Likewise.
16767 (estimate_node_size_and_time): Likewise.
16768 (inline_analyze_function): Likewise.
16769 * ipa-inline.c (report_inline_failed_reason): Likewise.
16770 (want_early_inline_function_p): Likewise.
16771 (edge_badness): Likewise.
16772 (update_edge_key): Likewise.
16773 (inline_small_functions): Likewise. Add dumping of order to two other
16775 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16776 instead of node uids.
16777 (propagate_pure_const): Likewise.
16778 (propagate_pure_const): Likewise.
16779 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16780 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16781 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16783 * tree-pretty-print.c (dump_function_header): Likewise.
16784 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16785 Print symbol order instead of node uids.
16787 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16789 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16790 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16792 2013-05-15 Richard Biener <rguenther@suse.de>
16794 PR tree-optimization/57275
16795 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16796 return value for fail to do runtime alias checks for gather loads.
16798 2013-05-15 Jan Hubicka <jh@suse.cz>
16802 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16803 weakrefs are not external.
16804 (lto_symtab_merge_decls): Fix thinko when dealing with
16805 non-lto_symtab decls.
16806 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16807 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16808 * varpool.c (dump_varpool_node): Dump more flags.
16810 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16812 * config/i386/i386.c (processor_alias_table): Add instruction
16813 FSGSBASE for AMD bdver3 architecture.
16815 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16817 * tree.c (warn_deprecated_use): Print file:line using locus color.
16818 * diagnostic.c (diagnostic_report_current_module): Print file:line
16819 and file:line:column using locus color.
16821 2013-05-14 Mike Stump <mikestump@comcast.net>
16823 * gdbinit.in: Add __null.
16825 2013-05-14 Mike Stump <mikestump@comcast.net>
16827 * recog.h: Rename struct recog_data to Recog_data.
16828 * recog.c: Likewise.
16829 * reload.c (can_reload_into): Likewise.
16830 * config/picochip/picochip.c: Likewise.
16832 2013-05-14 Mike Stump <mikestump@comcast.net>
16834 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16836 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16838 * resource.h (struct resources): Remove unch_memory member.
16839 (CLEAR_RESOURCE): Don't clear unch_memory.
16840 * resource.c (mark_referenced_resources): Don't set it.
16841 (mark_set_resources): Likewise.
16842 (mark_target_live_regs): Don't clear it.
16843 (init_resource_info): Likewise.
16844 * reorg.c (resource_conflicts_p): Don't compare it.
16845 (redundant_insn): Don't set it.
16847 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16849 * emit-rtl.c (next_label): Remove unused function.
16850 (skip_consecutive_labels, link_cc0_insns): Move to ...
16851 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16852 only place where these functions are used, and make them static.
16854 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16856 * fold-const.c (fold_negate_expr): Handle vectors.
16857 (fold_truth_not_expr): Make it static.
16858 (fold_invert_truthvalue): New static function.
16859 (invert_truthvalue_loc): Handle vectors. Do not call
16860 fold_truth_not_expr directly.
16861 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16862 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16863 (fold_comparison): Handle vector constants.
16864 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16865 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16866 * tree.h (fold_truth_not_expr): Remove declaration.
16868 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16870 * config/aarch64/aarch64-simd.md
16871 (aarch64_vcond_internal<mode>): Rename to...
16872 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16873 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16874 float modes. Clarify all iterator modes.
16875 (vcond<mode><mode>): Use new name for vcond expanders.
16876 (vcond<v_cmp_result><mode>): Likewise.
16877 (vcondu<mode><mode>: Likewise.
16878 * config/aarch64/iterators.md (VDQF_COND): New.
16880 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16883 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16884 variable for the shift amount. Check that we shift by non-negative
16887 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16890 * config/arm/arm.h (EPILOGUE_USES): Only return true
16891 for LR_REGNUM after epilogue_completed.
16893 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
16895 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
16896 is error_mark_node.
16898 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16901 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16902 and Solaris 11+/x86 with gld.
16903 * configure: Regenerate.
16905 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16907 * expmed.c (expand_shift_1): Canonicalize rotates by
16908 constant bitsize / 2 to bitsize - 1.
16909 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
16910 case ROTATERT>: Likewise.
16913 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16915 * config/i386/i386.md (rotateinv): New code attr.
16916 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16917 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16918 roll $31, %eax, etc.
16920 2013-05-14 Richard Biener <rguenther@suse.de>
16922 PR middle-end/57235
16923 * tree-eh.c (sink_clobbers): Give up for successors with
16924 multiple predecessors and no virtual uses.
16926 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16928 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16929 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16931 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16933 PR middle-end/57251
16934 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16935 the case when both op0 and op1 have VOIDmode.
16937 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16939 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16940 in multiply-accumulate mode.
16942 2013-05-13 Guozhi Wei <carrot@google.com>
16944 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16946 2013-05-13 Kai Tietz <ktietz@redhat.com>
16949 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
16950 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
16951 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
16952 * config/i386/i386.c (ix86_option_override_internal): Likewise.
16953 (ix86_expand_prologue): Likewise.
16954 (ix86_expand_split_stack_prologue): Likewise.
16955 (legitimate_pic_address_disp_p): Likewise.
16956 (legitimize_pic_address): Likewise.
16957 (legitimize_tls_address): Likewise.
16958 (legitimize_pe_coff_symbol): Likewise.
16959 (output_pic_addr_const): Likewise.
16960 (construct_plt_address): Likewise.
16961 (ix86_expand_call): Likewise.
16962 (x86_output_mi_thunk): Likewise.
16963 (x86_function_profiler): Likewise.
16965 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16967 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16968 similar switch cases.
16969 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16970 (aarch64_simd_mov_to_<mode>low): Delete.
16971 (aarch64_simd_mov_to_<mode>high): Delete.
16972 (move_lo_quad_<mode>): Add w<-r alternative.
16973 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16974 (aarch64_simd_mov_from_*): Update type attribute.
16975 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16978 2013-05-13 Jan Hubicka <jh@suse.cz>
16980 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16981 * config/i386/i386.c (ix86_compute_frame_layout,
16982 ix86_expand_epilogue, emit_i387_cw_initialization,
16983 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16984 ix86_local_alignment): Fix use of size/speed predicates.
16986 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16988 PR tree-optimization/45216
16989 PR tree-optimization/57157
16990 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16991 the (-Y) & (B - 1) variant if OP is |.
16992 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16993 return shifted. Use (-op1) & (prec - 1) as other_amount
16994 instead of prec - op1.
16996 2013-05-13 Martin Jambor <mjambor@suse.cz>
16998 PR middle-end/42371
16999 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17000 (ipa_constant_data): New type.
17001 (ipa_jump_func): Use ipa_constant_data to hold information about
17002 constant jump functions.
17003 (ipa_get_jf_constant): Adjust to jump function type changes.
17004 (ipa_get_jf_constant_rdesc): New function.
17005 (ipa_param_descriptor): New field controlled_uses.
17006 (ipa_get_controlled_uses): New function.
17007 (ipa_set_controlled_uses): Likewise.
17008 * ipa-ref.h (ipa_find_reference): Declare.
17009 * ipa-prop.c (ipa_cst_ref_desc): New type.
17010 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17012 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17013 New parameter cs. Adjust all callers.
17014 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17015 (remove_described_reference): New function.
17016 (jfunc_rdesc_usable): Likewise.
17017 (try_make_edge_direct_simple_call): Decrement controlled use count,
17018 attempt to remove reference if it hits zero.
17019 (combine_controlled_uses_counters): New function.
17020 (propagate_controlled_uses): Likewise.
17021 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17022 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17023 (ipa_print_node_params): Print described use counter.
17024 (ipa_write_jump_function): Adjust to jump function type changes.
17025 (ipa_read_jump_function): New parameter CS, pass it to
17026 ipa_set_jf_constant. Adjust caller.
17027 (ipa_write_node_info): Stream controlled use count
17028 (ipa_read_node_info): Likewise.
17029 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17031 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17032 count. Remove cloning-added reference if it reaches zero.
17033 * ipa-ref.c (ipa_find_reference): New function.
17035 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17037 * config/i386/i386.c (processor_target_table): Modified default
17038 alignment values for AMD BD and BT architectures.
17040 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17042 * tree-vect-generic.c (uniform_vector_p): Move ...
17043 * tree.c (uniform_vector_p): ... here.
17044 * tree.h (uniform_vector_p): Declare it.
17045 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17048 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17050 PR tree-optimization/57230
17051 * tree-ssa-strlen.c (handle_char_store): Record length for
17052 array store from STRING_CST.
17054 PR tree-optimization/57230
17055 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17058 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17060 * config/epiphany/epiphany.c (epiphany_init): Check size of
17061 NUM_MODES_FOR_MODE_SWITCHING.
17062 (epiphany_expand_prologue):
17063 Remove CONFIG_REGNUM initial value handling code.
17064 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17065 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17066 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17067 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17068 Don't return 1 for FP_MODE_NONE.
17069 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17070 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17071 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17072 * config/epiphany/epiphany.md (save_config): New pattern.
17074 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17076 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17078 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17080 * config/i386/i386.md (memory): Handle sseishft1.
17081 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17082 (*vec_extractv2di_1): Ditto.
17084 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17086 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17089 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17091 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17092 Add mthumb/march=armv7-a multilib.
17093 Add mthumb/march=armv7-r multilib.
17094 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17096 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17098 * config/v850/t-rtems: Add more multilibs.
17100 2013-05-10 Richard Biener <rguenther@suse.de>
17102 PR tree-optimization/57214
17103 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17104 not propagate from SSA names that occur in abnormal PHI nodes.
17106 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17108 * stor-layout.c (element_precision): New function.
17109 * machmode.h (element_precision): Declare it.
17110 * tree.c (build_minus_one_cst): New function.
17111 (element_precision): Likewise.
17112 * tree.h (build_minus_one_cst): Declare new function.
17113 (element_precision): Likewise.
17114 * fold-const.c (operand_equal_p): Use element_precision.
17115 (fold_binary_loc): Handle vector types.
17116 * convert.c (convert_to_integer): Use element_precision.
17117 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17120 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17122 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17123 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17124 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17125 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17126 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17127 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17128 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17129 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17130 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17131 (Uuw8): New constraints.
17132 (Usb4): Move into alphabetical order.
17133 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17134 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17135 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17136 previously unnamed patterns.
17137 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17138 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17139 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17140 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17141 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17142 of specific lengths.
17144 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17146 * config/i386/i386.md (rotateinv): New code attr.
17147 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17148 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17149 roll $31, %eax, etc.
17151 PR tree-optimization/45216
17152 PR tree-optimization/57157
17153 * tree-ssa-forwprop.c (simplify_rotate): New function.
17154 (ssa_forward_propagate_and_combine): Call it.
17156 2013-05-10 Richard Biener <rguenther@suse.de>
17158 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17159 disable peeling when we version for aliasing.
17160 (vector_alignment_reachable_p): Honor explicit user alignment.
17161 (vect_supportable_dr_alignment): Likewise.
17162 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17163 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17164 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17165 then peeling to arrange for the cost-model check to come first.
17167 2013-05-10 Alan Modra <amodra@gmail.com>
17169 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17170 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17171 * configure: Regenerate.
17173 2013-05-10 Alan Modra <amodra@gmail.com>
17176 * varasm.c (default_elf_select_section): Move !DECL_P check..
17177 (get_named_section): ..to here before calling get_section_name.
17179 (default_section_type_flags): Add DECL_P check.
17180 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17181 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17183 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17185 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17186 When using gen_stack_adjust_str with a register offset, add a
17187 REG_FRAME_RELATED_EXPR note.
17189 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17191 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17192 (*vec_extractv4si_zext_mem): Ditto.
17193 (*vec_extractv2di): Add 0->x and x->x alternatives.
17194 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17195 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17197 2013-05-09 Jason Merrill <jason@redhat.com>
17199 N3639 C++1y VLA support
17200 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17203 * tree.c (build_constructor_va): New.
17204 * tree.h: Declare it.
17206 2013-05-09 Martin Jambor <mjambor@suse.cz>
17209 * gimple-fold.c (canonicalize_constructor_val): Call
17210 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17212 2013-05-09 Jan Hubicka <jh@suse.cz>
17213 Richard Biener <rguenther@suse.de>
17216 * symtab.c (symtab_make_decl_local): Do not add private names.
17218 2013-05-09 Jan Hubicka <jh@suse.cz>
17221 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17222 (unlink_from_assembler_name_hash): Likewise.
17223 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17224 symtab_unregister_node, symtab_initialize_asm_name_hash,
17225 change_decl_assembler_name): Update.
17227 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17229 * config/aarch64/aarch64.md: New movtf split.
17230 (*movtf_aarch64): Update.
17231 (aarch64_movdi_tilow): Handle TF modes and rename to
17232 aarch64_movdi_<mode>low.
17233 (aarch64_movdi_tihigh): Handle TF modes and rename to
17234 aarch64_movdi_<mode>high
17235 (aarch64_movtihigh_di): Handle TF modes and rename to
17236 aarch64_mov<mode>high_di
17237 (aarch64_movtilow_di): Handle TF modes and rename to
17238 aarch64_mov<mode>low_di
17239 (aarch64_movtilow_tilow): Remove spurious whitespace.
17240 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17242 (aarch64_print_operand): Update.
17244 2013-05-09 Alan Modra <amodra@gmail.com>
17246 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17248 * configure: Regenerate.
17250 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17252 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17253 splitter preparation statements.
17254 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17255 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17258 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17260 * gimple-ssa-strength-reduction.c (count_candidates): Change
17261 return value to int.
17262 (analyze_candidates_and_replace): Change type of length to int.
17264 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17266 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17267 (*vec_extract<mode>): Use VI12_128 mode iterator.
17268 (*vec_extract<mode>_mem): Ditto.
17269 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17272 2013-05-08 Diego Novillo <dnovillo@google.com>
17277 2012-08-17 Diego Novillo <dnovillo@google.com>
17280 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17281 * config.in: Regenerate.
17282 * configure: Regenerate.
17283 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17285 2013-05-08 Jan Hubicka <jh@suse.cz>
17288 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17289 * cgraph.h (symtab_node_base): Add unique_name.
17290 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17291 input_overwrite_node, input_varpool_node): Stream unique_name.
17292 * cgraphclones.c (cgraph_create_virtual_clone,
17293 cgraph_function_versioning): Set unique_name.
17294 * ipa.c (function_and_variable_visibility): Set unique_name.
17296 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17298 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17299 (alloc_cand_and_find_basis): Restrict conditional candidate
17300 processing to CAND_MULTs.
17302 2013-05-08 Jan Hubicka <jh@suse.cz>
17305 lto-symtab.c (lto_symtab_symbol_p): New function.
17306 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17307 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17308 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17309 Skip static symbols.
17311 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17313 PR tree-optimization/57200
17314 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17315 Only call inform if the preceding warning_at returns true.
17317 2013-05-07 Han Shen <shenhan@google.com>
17319 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17320 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17321 * common.opt (fstack-protector-strong): New option.
17322 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17323 * doc/invoke.texi (Optimization Options): Document
17324 "-fstack-protector-strong".
17325 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17327 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17329 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17331 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17333 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17334 it is not lhs of the insn.
17335 (setup_reg_equiv): Remove insn having equiv memory which it is not
17337 * lra-constraints.c (process_address): Try to improve generation
17338 code for address base + disp.
17339 (lra_constraints): Make correct the code for checking insn setting
17340 up backward equivalence. Remove insn only if it is in the init
17342 * lra-eliminations.c (update_reg_eliminate): Change return value.
17343 (lra_eliminate): Use the result.
17345 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17347 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17348 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17349 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17350 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17351 register target operands.
17352 (*vec_extractv8hi_sse2): New pattern.
17353 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17354 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17355 (*vec_extract<mode>_mem): New insn and split pattern.
17357 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17359 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17360 (TARGET_ASAN_SHADOW_OFFSET): Define.
17361 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17362 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17364 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17366 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17367 (incr_vec_index): Return -1 if increment not found.
17368 (create_add_on_incoming_edge): Assert if increment not found.
17369 (record_increment): Limit number of increments recorded.
17370 (all_phi_incrs_profitable): Return false if an increment not found.
17371 (replace_profitable_candidates): Don't process increments that were
17373 (analyze_candidates_and_replace): Limit size of incr_vec.
17375 2013-05-07 Richard Biener <rguenther@suse.de>
17377 * calls.c (special_function_p): setjmp-like functions are leaf.
17378 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17379 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17381 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17383 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17384 (aarch64_simd_mov<mode>): New expander.
17385 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17386 (aarch64_simd_mov_to_<mode>high): Likewise.
17387 (aarch64_simd_mov_from_<mode>low): Likewise.
17388 (aarch64_simd_mov_from_<mode>high): Likewise.
17389 (aarch64_dup_lane<mode>): Update.
17390 (aarch64_dup_lanedi): New instruction pattern.
17391 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17392 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17394 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17396 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17397 (replace_mult_candidate): Remove unnecessary argument; remove
17398 unnecessary parameter from call to introduce_cast_before_cand.
17399 (replace_unconditional_candidate): Remove unnecessary parameter
17400 from call to replace_mult_candidate.
17401 (replace_conditional_candidate): Likewise.
17402 (insert_initializers): Use make_temp_ssa_name.
17403 (introduce_cast_before_cand): Remove unnecessary argument; use
17404 make_temp_ssa_name.
17405 (replace_one_candidate): Remove unnecessary argument; remove
17406 unnecessary parameter from calls to introduce_cast_before_cand.
17407 (replace_profitable_candidates): Remove unnecessary parameters
17408 from calls to replace_one_candidate.
17410 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17412 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17413 phi def as possibly hiding a basis for a CAND_ADD whose operands
17414 have been commuted in the analysis.
17415 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17417 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17419 * config/aarch64/aarch64.md
17420 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17421 shift value between 0-4.
17423 2013-05-07 Richard Biener <rguenther@suse.de>
17425 * double-int.h (rshift): New overload.
17426 * double-int.c (rshift): New function.
17427 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17428 (create_reference_ops_from_ref): Remove.
17429 (vn_reference_insert): Use shared ops for constructing the
17430 reference and copy it.
17432 2013-05-07 Richard Biener <rguenther@suse.de>
17434 PR middle-end/57190
17435 * tree-eh.c (sink_clobbers): Properly propagate
17436 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17438 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17440 PR tree-optimization/57149
17441 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17442 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17443 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17444 uninit_undefined_value_p instead of ssa_undefined_value_p.
17447 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17448 for modifier == EXPAND_INITIALIZER.
17450 2013-05-07 Anton Blanchard <anton@samba.org>
17452 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17453 for powerpc64 little endian.
17454 * configure: Regenerate.
17456 2013-05-06 Graham Stott <grahams@btinternet.com>
17458 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17459 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17460 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17461 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17463 2013-05-06 Graham Stott <grahams@btinternet.com>
17465 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17466 codes which allow non-lvalues.
17468 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17470 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17471 components are all 1s.
17472 (integer_minus_onep): New function.
17473 * tree.h (integer_minus_onep): Declare it.
17474 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17475 integer_minus_onep instead of integer_all_onesp.
17477 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17480 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17481 variations of these patterns.
17483 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17485 * config/i386/i386.md (isa): Add x64_sse4 member.
17486 (enabled): Handle x64_sse4.
17487 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17488 instruction for 64bit SSE4_1 targets. Update insn attributes.
17489 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17490 instruction for SSE4_1 targets. Update insn attributes.
17491 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17492 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17493 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17495 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17496 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17497 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17499 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17502 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17504 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17506 * final.c (do_assembler_dialects): Don't handle curly braces and
17507 vertical bar escaped by % as dialect delimiters.
17508 (output_asm_insn): Print curly braces and vertical bar if escaped
17509 by % and ASSEMBLER_DIALECT defined.
17510 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17511 * doc/tm.texi: Regenerated.
17513 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17515 * config/mips/mips.c: Include tree-pass.h.
17516 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17517 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17518 (pass_mips_machine_reorg2): New machine specific pass.
17519 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17520 (mips_option_override): Register the new pass.
17521 * rtl.h (cleanup_barriers): Remove prototype.
17522 (dbr_schedule): Likewise.
17523 * jump.c (cleanup_barriers): Make static.
17524 * reorg.c (dbr_schedule): Likewise.
17526 2013-05-06 Richard Biener <rguenther@suse.de>
17528 PR tree-optimization/57185
17529 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17530 SSA names properly.
17532 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17535 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17536 between "first_arg" and "insn", not "last" and "insn".
17538 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17540 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17541 (find_candidates_in_block): Re-enable slsr_process_phi.
17542 (create_phi_basis): Fix double counting of candidate adjustment.
17544 2013-05-06 Richard Biener <rguenther@suse.de>
17546 PR middle-end/57147
17547 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17548 the edge is also fallthru, preserve it and just clear the
17550 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17551 also complex, preserve that and just clear the fallthru flag.
17552 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17553 update virtual operands.
17555 2013-05-06 Alan Modra <amodra@gmail.com>
17557 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17558 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17559 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17560 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17561 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17562 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17564 2013-05-06 Alan Modra <amodra@gmail.com>
17566 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17567 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17568 (DEFAULT_ASM_ENDIAN): Define.
17569 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17570 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17571 Update -K PIC clause from sysv4.h.
17572 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17573 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17575 2013-05-06 Alan Modra <amodra@gmail.com>
17577 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17578 twice for little-endian.
17579 (ashrdi3_no_power, ashrdi3): Support little-endian.
17581 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17584 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17585 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17587 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17588 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17589 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17592 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17593 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17595 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17596 * config/pa/pa.opt: Make mbig-switch a no-op.
17597 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17598 (CASE_VECTOR_MODE): Always return SImode.
17599 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17600 for the !TARGET_BIG_SWITCH case.
17601 * config/pa/pa-linux.h: Likewise.
17602 * config/pa/pa-openbsd.h: Likewise.
17603 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17604 * config/pa/pa.md (short_jump): Remove define_insn.
17605 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17606 (casesi0): Remove define_insn.
17607 (type): Remove btable_branch.
17608 (pa_combine_type): Likewise.
17609 (in_nullified_branch_delay): Likewise.
17610 (in_call_delay): Likewise.
17611 (define_delay): Likewise.
17612 (define_insn_reservation "Z3"): Likewise.
17613 (define_insn_reservation "Z4"): Likewise.
17614 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17615 (pa_adjust_insn_length): Remove adjustment for btable branches.
17616 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17619 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17621 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17622 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17623 Add m->r,x alternatives.
17624 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17625 splitters using SWI48x mode iterator.
17626 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17627 TARGET_64BIT. Add m->x alternative.
17628 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17629 Add o->x alternative. Enable for TARGET_SSE.
17630 (sse_storeq): Remove expander.
17631 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17632 with memory input operand.
17633 (*vec_extractv2di_1 splitter): New.
17634 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17635 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17637 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17639 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17640 (INT_LOWPART): Delete.
17641 (extract_MB): Adjust.
17642 (extract_ME): Adjust.
17643 (print_operand): Adjust.
17645 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17647 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17648 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17649 (reg_or_logical_cint_operand, easy_fp_constant,
17650 logical_const_operand): Delete "CONST_DOUBLE" case.
17651 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17652 "HOST_BITS_PER_WIDE_INT == 64" test.
17653 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17654 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17655 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17656 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17658 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17659 CONST_DOUBLE DImode/VOIDmode case.
17660 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17661 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17662 CONST_DOUBLE VOIDmode case.
17663 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17664 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17665 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17666 Delete CONST_DOUBLE case.
17667 (splitters for mov FMOVE64 const_double): Delete
17668 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17669 "HOST_BITS_PER_WIDE_INT >= 64" test.
17670 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17672 (mov DI const_double): Delete.
17674 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17676 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17677 on op shows all bits zero in mode of a lowpart subreg, return zero.
17679 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17682 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17683 to save TFmode registers and DImode to save TImode registers for
17684 caller save operations.
17685 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17686 mark being partially clobbered since they only use the first
17689 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17690 and TDmode only use the upper 64-bits of each VSX register.
17692 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17694 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17695 (find_candidates_in_block): Disable slsr_process_phi.
17697 2013-05-03 Guozhi Wei <carrot@google.com>
17699 * coverage.c (coverage_obj_init): Move the construction of gcov
17701 (build_init_ctor): ... here.
17703 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17705 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17706 (slsr_cand_d): Redefine def_phi.
17707 (stride_status, phi_adjust_status, count_phis_status): New enums.
17708 (find_phi_def): New.
17709 (find_basis_for_base_expr): New.
17710 (find_basis_for_candidate): Handle hidden bases.
17711 (alloc_cand_and_find_basis): Handle phi candidates.
17712 (slsr_process_phi): New.
17713 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17714 (create_mul_imm_cand): Likewise.
17715 (create_add_ssa_cand): Exclude phi base candidates.
17716 (create_add_imm_cand): Likewise.
17717 (slsr_process_cast): Likewise.
17718 (slsr_process_copy): Likewise.
17719 (find_candidates_in_block): Handle phi candidates.
17720 (dump_candidate): Likewise.
17721 (unconditional_cands): Delete.
17722 (unconditional_cands_with_known_stride_p): Delete.
17723 (phi_dependent_cand_p): New.
17724 (cand_increment): Handle phi-dependent candidates.
17725 (replace_dependent): Delete.
17726 (replace_mult_candidate): New.
17727 (replace_unconditional_candidate): New.
17728 (incr_vec_index): Move to avoid forward reference.
17729 (create_add_on_incoming_edge): New.
17730 (create_phi_basis): New.
17731 (replace_dependents): Delete.
17732 (replace_conditional_candidate): New.
17733 (phi_add_costs): New.
17734 (replace_uncond_cands_and_profitable_phis): New.
17735 (record_increment): Handle phi adjustments.
17736 (record_phi_increments): New.
17737 (record_increments): Handle phi adjustments.
17738 (phi_incr_cost): New.
17739 (lowest_cost_path): Handle phis.
17740 (total_savings): Likewise.
17741 (analyze_increments): Likewise.
17742 (ncd_with_phi): New.
17743 (ncd_of_cand_and_phis): New.
17744 (nearest_common_dominator_for_cands): Handle phi increments.
17745 (all_phi_incrs_profitable): New.
17746 (replace_profitable_candidates): Handle phi-dependent candidates.
17747 (analyze_candidates_and_replace): Likewise.
17749 2013-05-03 Teresa Johnson <tejohnson@google.com>
17752 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17753 do not exceed REG_BR_PROB_BASE.
17755 2013-05-03 Jeff Law <law@redhat.com>
17757 PR tree-optimization/57144
17758 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17759 operand of the condition will bit into the new type when eliminating
17760 a cast feeding a condition.
17762 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17764 PR rtl-optimization/57130
17765 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17766 of COMPARE as in_code to the recursive call if needed.
17768 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17770 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17771 (enabled): Handle new members.
17772 * config/i386/sse.md (*vec_concatv2si): Merge from
17773 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17774 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17776 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17778 PR tree-optimization/57027
17779 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17780 for fnms opportunity, check we got the prerequisite kind
17781 of tree / gimple before using accessor functions.
17783 2013-05-03 Richard Biener <rguenther@suse.de>
17785 * double-int.h (lshift): New overload without precision
17786 and arith argument.
17787 (operator *=, operator +=, operator -=): Move ...
17788 * double-int.c (operator *=, operator +=, operator -=): ... here
17789 and implement more efficiently.
17790 (mul_double_with_sign): Remove.
17791 (lshift_double): Adjust to take unsinged shift argument, push
17792 dispatching code to callers.
17793 (mul_double_wide_with_sign): Add early out for callers that
17794 are not interested in high parts or overflow.
17795 (lshift): New function.
17796 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17797 dispatch code here.
17798 (lrotate, rrotate): Use logical shifts.
17799 * expr.c (get_inner_reference): Use lshift.
17800 * fixed-value.c (do_fixed_divide): Likewise.
17801 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17802 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17803 (indirect_refs_may_alias_p): Likewise.
17804 (stmt_kills_ref_p_1): Likewise.
17806 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17808 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17810 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17812 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17813 scalar form of FABD instruction.
17815 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17817 * lra-constraints.c (process_alt_operands): Add checking alt
17818 number to choose the best alternative.
17820 2013-05-02 Richard Biener <rguenther@suse.de>
17822 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17823 bitmap and its handling.
17824 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17826 2013-05-02 Richard Biener <rguenther@suse.de>
17828 PR middle-end/57140
17829 * tree-inline.c (copy_loops): Properly handle removed loops.
17830 (copy_cfg_body): Mark destination loops for fixup if source
17831 loops needed fixup.
17833 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17836 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17837 generating simple_return for naked functions.
17839 2013-05-02 Martin Jambor <mjambor@suse.cz>
17841 PR middle-end/56988
17842 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17843 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17845 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17846 ipa_agg_replacement_value structures.
17847 (known_aggs_to_agg_replacement_list): Likewise.
17848 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17849 (read_agg_replacement_chain): Likewise.
17850 (ipcp_transform_function): Also check that by_ref flags match.
17852 2013-05-02 Richard Biener <rguenther@suse.de>
17854 * graphds.h (struct graph): Add obstack member.
17855 * graphds.c (new_graph): Initialize obstack and allocate
17857 (add_edge): Allocate edge from the obstack.
17858 (free_graph): Free the obstack instead of all edges and vertices.
17860 2013-05-02 Teresa Johnson <tejohnson@google.com>
17862 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17864 * cfg.c (update_bb_profile_for_threading): Ditto.
17865 * tree-inline.c (copy_bb): Ditto.
17866 (copy_edges_for_bb): Ditto.
17867 (initialize_cfun): Ditto.
17868 (copy_cfg_body): Ditto.
17869 (expand_call_inline): Ditto.
17870 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17871 (estimate_node_size_and_time): Ditto.
17872 (inline_merge_summary): Ditto.
17873 * cgraphclones.c (cgraph_clone_edge): Ditto.
17874 (cgraph_clone_node): Ditto.
17875 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17876 (compute_trg_info): Ditto.
17878 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17880 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17881 S reg when fp attribute set.
17882 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17884 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17886 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17888 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17889 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17890 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17892 2013-05-02 Richard Biener <rguenther@suse.de>
17894 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17895 (struct instantiate_cache_entry): New type.
17896 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17897 (struct instantiate_cache_type): New type.
17898 (set_instantiated_value, get_instantiated_value): Remove.
17899 (get_instantiated_value_entry): New function.
17900 (instantiate_scev_name): Use the new cache and adjust.
17901 (instantiate_scev_poly): Adjust.
17902 (instantiate_scev_binary): Likewise.
17903 (instantiate_array_ref): Likewise.
17904 (instantiate_scev_convert): Likewise.
17905 (instantiate_scev_not): Likewise.
17906 (instantiate_scev_3): Likewise.
17907 (instantiate_scev_2): Likewise.
17908 (instantiate_scev_r): Likewise.
17909 (instantiate_scev): Likewise.
17910 (resolve_mixers): Likewise.
17912 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17915 * lra-constraints.c (best_small_class_operands_num): Remove.
17916 (process_alt_operands): Remove small_class_operands_num. Take
17917 small classes operands into losers and only if the operand is not
17918 matched. Modify debugging output.
17919 (curr_insn_transform): Remove best_small_class_operands_num.
17922 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17924 * config/aarch64/aarch64-builtins.c
17925 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17926 * config/aarch64/aarch64-simd-builtins.def
17927 (reduc_splus_): Add new modes.
17928 (reduc_uplus_): New.
17929 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17930 (reduc_uplus_v4sf): Likewise.
17931 (reduc_splus_v4sf): Likewise.
17932 (aarch64_addv<mode>): Likewise.
17933 (reduc_uplus_<mode>): Likewise.
17934 (reduc_splus_<mode>): Likewise.
17935 (aarch64_addvv2di): Likewise.
17936 (reduc_uplus_v2di): Likewise.
17937 (reduc_splus_v2di): Likewise.
17938 (aarch64_addvv2si): Likewise.
17939 (reduc_uplus_v2si): Likewise.
17940 (reduc_splus_v2si): Likewise.
17941 (reduc_<sur>plus_<mode>): New.
17942 (reduc_<sur>plus_v2di): Likewise.
17943 (reduc_<sur>plus_v2si): Likewise.
17944 (reduc_<sur>plus_v4sf): Likewise.
17945 (aarch64_addpv4sf): Likewise.
17946 * config/aarch64/arm_neon.h
17947 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17948 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17949 add UNSPEC_SADDV, UNSPEC_UADDV.
17951 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17953 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17955 * config/aarch64/arm_neon.h
17956 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17958 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17960 * config/aarch64/aarch64-builtins
17961 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17963 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17965 * config/aarch64/aarch64-simd-builtins.def
17966 (reduc_smax_): New.
17967 (reduc_smin_): Likewise.
17968 (reduc_umax_): Likewise.
17969 (reduc_umin_): Likewise.
17970 (reduc_smax_nan_): Likewise.
17971 (reduc_smin_nan_): Likewise.
17974 (smax): Update for V2SF, V4SF and V2DF modes.
17977 (smin_nan): Likewise.
17978 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17979 (<su><maxmin><mode>3): ...This, refactor.
17980 (s<maxmin><mode>3): New.
17981 (<maxmin_uns><mode>3): Likewise.
17982 (reduc_<maxmin_uns>_<mode>): Refactor.
17983 (reduc_<maxmin_uns>_v4sf): Likewise.
17984 (reduc_<maxmin_uns>_v2si): Likewise.
17985 (aarch64_<fmaxmin><mode>: Remove.
17986 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17988 (vmin<q>_f<32,64>): Likewise.
17989 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17991 (su): Add mappings for smax, smin, umax, umin.
17993 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17994 (FMAXMIN): Rename as...
17995 (FMAXMIN_UNS): ...This.
17997 (fmaxminv): Likewise.
17998 (fmaxmin): Likewise.
18000 (maxmin_uns_op): Likewise.
18002 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18004 * config/aarch64/arm_neon.h
18005 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18006 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18007 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18009 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18011 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18012 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18014 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18016 * config/aarch64/aarch64-simd.md
18017 (vcond<mode>_internal): Handle special cases for constant masks.
18018 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18019 (vcondu<mode><mode>): Likewise.
18020 (vcond<v_cmp_result><mode>): New.
18022 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18024 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18025 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18026 * config/aarch64/aarch64-simd-builtins.def
18027 (cmeq): Update to BUILTIN_VALLDI.
18032 * config/aarch64/arm_neon.h
18033 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18034 to builtins or C as appropriate.
18036 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18038 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18040 (cmhi): Rename to...
18042 * config/aarch64/aarch64-simd.md
18043 (simd_mode): Add SF.
18044 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18045 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18046 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18047 (cstore<mode>_neg): ...This.
18048 * config/aarch64/iterators.md
18050 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18051 (COMPARISONS): New.
18052 (UCOMPARISONS): Likewise.
18053 (optab): Add missing comparisons.
18059 (VCMP_S): Likewise.
18060 (VCMP_U): Likewise.
18061 (V_cmp_result): Add DF, SF modes.
18062 (v_cmp_result): Likewise.
18064 (vmtype): Likewise.
18065 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18067 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18069 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18070 define_insn to define_insn_and_split.
18071 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18072 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18073 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18074 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18075 (thumb2_negscc): Likewise.
18077 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18079 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18081 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18083 * config/arm/thumb2.md: Remove trailing whitespaces.
18085 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18087 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18088 Use gen_int_mode rather than GEN_INT.
18090 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18092 * value-prof.c (stream_in_histogram_value): Remove the strayed
18095 2013-04-30 Richard Biener <rguenther@suse.de>
18097 PR middle-end/57122
18098 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18100 2013-04-30 Richard Biener <rguenther@suse.de>
18102 PR middle-end/57107
18103 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18105 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18107 PR rtl-optimization/56957
18108 PR rtl-optimization/57105
18109 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18110 variable. Use just INSN_UID for determining whether an insn
18111 should be only disconnected from the insn stream.
18112 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18114 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18116 PR tree-optimization/57104
18117 * tsan.c (instrument_expr): Don't instrument accesses to
18118 DECL_HARD_REGISTER VAR_DECLs.
18120 2013-04-30 Richard Biener <rguenther@suse.de>
18122 * function.h (loops_for_fn): New inline function.
18123 (set_loops_for_fn): Likewise.
18124 * cfgloop.h (place_new_loop): Add struct function parameter.
18125 (get_loop): Likewise.
18126 (get_loops): Likewise.
18127 (number_of_loops): Likewise.
18128 (fel_next): Adjust.
18129 (fel_init): Likewise.
18130 * cfg.c (get_loop_copy): Adjust.
18131 * cfgloop.c (flow_loops_dump): Likewise.
18132 (record_loop_exits): Likewise.
18133 (verify_loop_structure): Likewise.
18134 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18135 (estimate_reg_pressure_cost): Likewise.
18136 (mark_loop_exit_edges): Likewise.
18137 * cfgloopmanip.c (place_new_loop): Likewise.
18138 (add_loop): Likewise.
18139 (duplicate_loop): Likewise.
18140 * graph.c (draw_cfg_nodes): Likewise.
18141 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18142 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18143 (extract_affine_chrec): Likewise.
18144 (build_scop_iteration_domain): Likewise.
18145 * graphite.c (graphite_initialize): Likewise.
18146 * ira-build.c (create_loop_tree_nodes): Likewise.
18147 (more_one_region_p): Likewise.
18148 (rebuild_regno_allocno_maps): Likewise.
18149 (mark_loops_for_removal): Likewise.
18150 (mark_all_loops_for_removal): Likewise.
18151 (remove_unnecessary_regions): Likewise.
18152 (ira_build): Likewise.
18153 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18154 * loop-init.c (fix_loop_structure): Likewise.
18155 (gate_rtl_move_loop_invariants): Likewise.
18156 (gate_rtl_unswitch): Likewise.
18157 (gate_rtl_unroll_and_peel_loops): Likewise.
18158 (rtl_doloop): Likewise.
18159 * lto-streamer-in.c (input_cfg): Likewise.
18160 * lto-streamer-out.c (output_cfg): Likewise.
18161 * modulo-sched.c (sms_schedule): Likewise.
18162 * predict.c (tree_estimate_probability): Likewise.
18163 (tree_estimate_probability_driver): Likewise.
18164 (estimate_loops): Likewise.
18165 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18166 (move_sese_region_to_fn): Likewise.
18167 (debug_loop_num): Likewise.
18168 * tree-chrec.c (chrec_evaluate): Likewise.
18169 (hide_evolution_in_other_loops_than_loop): Likewise.
18170 (chrec_component_in_loop_num): Likewise.
18171 (reset_evolution_in_loop): Likewise.
18172 (evolution_function_is_invariant_rec_p): Likewise.
18173 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18174 * tree-inline.c (copy_loops): Likewise.
18175 (copy_cfg_body): Likewise.
18176 (tree_function_versioning): Likewise.
18177 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18178 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18180 (add_to_evolution_1): Likewise.
18181 (scev_const_prop): Likewise.
18182 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18183 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18184 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18185 (tree_ssa_lim_initialize): Likewise.
18186 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18187 (verify_loop_closed_ssa): Likewise.
18188 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18189 (tree_ssa_loop_im): Likewise.
18190 (tree_ssa_loop_unswitch): Likewise.
18191 (tree_vectorize): Likewise.
18192 (check_data_deps): Likewise.
18193 (tree_ssa_loop_ivcanon): Likewise.
18194 (tree_ssa_loop_bounds): Likewise.
18195 (tree_complete_unroll): Likewise.
18196 (tree_complete_unroll_inner): Likewise.
18197 (tree_parallelize_loops): Likewise.
18198 (tree_ssa_loop_prefetch): Likewise.
18199 (tree_ssa_loop_ivopts): Likewise.
18200 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18201 * tree-vectorizer.c (vectorize_loops): Likewise.
18203 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18205 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18206 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18207 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18208 with EABI_LINK_SPEC.
18210 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18213 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18216 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18219 * lra-constraints.c (process_alt_operands): Discourage a bit more
18220 using memory for pseudos. Print cost dump for alternatives.
18221 Modify cost values for conflicts with early clobbers.
18222 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18224 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18227 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18229 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18231 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18232 from/to S register.
18233 (movdi_aarch64): Support LDR/STR from/to D register.
18235 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18237 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18238 or higher by default.
18240 2013-04-29 Richard Biener <rguenther@suse.de>
18242 PR middle-end/57075
18243 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18244 even if not adding abnormal edges for calls that can make
18247 2013-04-29 Richard Biener <rguenther@suse.de>
18249 PR middle-end/57103
18250 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18252 (move_stmt_r): Remove redundant checking.
18254 2013-04-29 Teresa Johnson <tejohnson@google.com>
18257 * basic-block.h (apply_scale): New function.
18258 (apply_probability): Use apply_scale.
18259 * gimple-streamer-in.c (input_bb): Ditto.
18260 * lto-streamer-in.c (input_cfg): Ditto.
18261 * lto-cgraph.c (merge_profile_summaries): Ditto.
18262 * tree-optimize.c (execute_fixup_cfg): Ditto.
18263 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18264 (copy_edges_for_bb): Ditto.
18265 (copy_cfg_body): Ditto.
18267 2013-04-29 Tom de Vries <tom@codesourcery.com>
18269 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18270 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18271 (tail_merge_optimize): Handle current_loops == NULL.
18273 2013-04-26 Jeff Law <law@redhat.com>
18275 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18276 (simplify_cond_using_ranges): Generalize code to simplify
18277 COND_EXPRs where one argument is a constant and the other
18278 is an SSA_NAME created by an integral type conversion.
18280 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18282 * config/arm/arm.md (store_minmaxsi): Use only when
18283 optimize_insn_for_size_p.
18285 2013-04-29 Christian Bruel <christian.bruel@st.com>
18288 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18290 2013-04-29 Richard Biener <rguenther@suse.de>
18292 PR middle-end/57089
18293 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18294 loop tree make sure to schedule a fixup for the child as well.
18295 (expand_omp_for_generic): Properly add loops.
18296 (expand_omp_for_static_nochunk): Likewise.
18297 (expand_omp_for_static_chunk): Likewise.
18298 (expand_omp_for): For the degenerate case fixup loops.
18299 (expand_omp_sections): Fix default bb placement in loops.
18300 (expand_omp_atomic_pipeline): Properly add loops.
18302 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18304 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18306 2013-04-29 Tom de Vries <tom@codesourcery.com>
18308 * tree-ssa-tail-merge.c: Update header comment.
18310 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18312 * config/aarch64/arm_neon.h
18313 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18314 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18315 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18316 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18317 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18318 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18319 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18320 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18322 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18324 * config/aarch64/aarch64-simd.md
18325 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18326 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18327 fix_trunc, fixuns_trunc.
18328 (ftrunc<VDQF:mode>2): New.
18329 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18330 (fix_trunc_optab): New.
18332 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18334 * config/aarch64/aarch64-builtins.c
18335 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18336 iceilf, lround, iroundf.
18338 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18341 * config/i386/i386.h (enum ix86_tune_indices)
18342 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18343 New, split from X86_TUNE_INTER_UNIT_MOVES.
18344 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18345 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18346 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18347 (TARGET_INTER_UNIT_MOVES): Remove.
18348 * config/i386/i386.c (initial_ix86_tune_features): Update.
18349 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18350 (ix86_expand_convert_uns_didf_sse): Use
18351 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18352 (ix86_expand_vector_init_one_nonzero): Ditto.
18353 (ix86_expand_vector_init_interleave): Ditto.
18354 (inline_secondary_memory_needed): Return true for moves from SSE class
18355 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18356 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18357 * config/i386/constraints.md (Yi, Ym): Depend on
18358 TARGET_INTER_UNIT_MOVES_TO_VEC.
18359 (Yj, Yn): New constraints.
18360 * config/i386/i386.md (*movdi_internal): Change constraints of
18361 operand 1 from Yi to Yj and from Ym to Yn.
18362 (*movsi_internal): Ditto.
18363 (*movdf_internal): Ditto.
18364 (*movsf_internal): Ditto.
18365 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18366 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18367 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18368 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18369 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18370 * config/i386/sse.md (movdi_to_sse): Ditto.
18371 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18372 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18373 TARGET_INTER_UNIT_MOVES.
18374 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18375 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18376 instead of TARGET_INTER_UNIT_MOVES.
18377 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18378 operand 1 from Yi to Yj and from Ym to Yn.
18380 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18382 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18383 (float_truncate_hi_): Likewise.
18384 (float_extend_lo_): Likewise.
18385 (float_truncate_lo_): Likewise.
18386 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18387 (aarch64_float_extend_lo_v2df): Likewise.
18388 (vec_unpacks_hi_v4sf): Likewise.
18389 (aarch64_float_truncate_lo_v2sf): Likewise.
18390 (aarch64_float_truncate_hi_v4sf): Likewise.
18391 (vec_pack_trunc_v2df): Likewise.
18392 (vec_pack_trunc_df): Likewise.
18394 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18396 * config/aarch64/aarch64-builtins.c
18397 (aarch64_fold_builtin): Fold float conversions.
18398 * config/aarch64/aarch64-simd-builtins.def
18399 (floatv2si, floatv4si, floatv2di): New.
18400 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18401 * config/aarch64/aarch64-simd.md
18402 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18403 * config/aarch64/iterators.md (FLOATUORS): New.
18404 (optab): Add float, floatuns.
18405 (su_optab): Likewise.
18407 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18409 * config/aarch64/aarch64-builtins.c
18410 (aarch64_builtin_vectorized_function): Use new names for
18412 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18413 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18414 (fcvtzu): Split as...
18415 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18416 (fcvtas): Split as...
18417 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18418 (fcvtau): Split as...
18419 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18420 (fcvtps): Split as...
18421 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18422 (fcvtpu): Split as...
18423 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18424 (fcvtms): Split as...
18425 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18426 (fcvtmu): Split as...
18427 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18428 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18429 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18430 (lfrintnusf, lfrintnudf): Likewise.
18431 * config/aarch64/aarch64-simd.md
18432 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18434 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18435 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18436 (fcvt_pattern): Likewise.
18438 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18440 * config/aarch64/aarch64-simd.md
18441 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18442 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18444 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18446 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18447 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18448 (vrnd<a,m,n,p>_f32): Implement using builtins.
18449 (vrnd<i,x><q>_f<32, 64>): New.
18451 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18453 * config/aarch64/aarch64-builtins.c
18454 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18455 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18456 (frintz): Rename to...
18458 (frintp): Rename to...
18460 (frintm): Rename to...
18462 (frinti): Rename to...
18463 (nearbyint): ...this.
18464 (frintx): Rename to...
18466 (frinta): Rename to...
18468 * config/aarch64/aarch64-simd.md
18469 (aarch64_frint<frint_suffix><mode>): Delete.
18470 (<frint_pattern><mode>2): Convert to insn.
18471 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18472 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18473 (frint_pattern): Likewise.
18474 (frint_suffix): Likewise.
18476 2013-04-29 Richard Biener <rguenther@suse.de>
18478 PR tree-optimization/57081
18479 * loop-init.c: Include tree-flow.h.
18480 (loop_optimizer_finalize): Free number of iteration estimates.
18481 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18483 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18485 PR tree-optimization/57083
18486 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18487 non-singleton shift count range, zero extend low_bound for uns case.
18489 * config/i386/predicates.md (general_vector_operand): New predicate.
18490 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18491 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18492 if they aren't nonimmediate operands. If their original values
18493 satisfy const_vector_equal_evenodd_p, don't shift them.
18494 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18495 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18496 if not nonimmediate_operand.
18497 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18498 instead of register_operand.
18499 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18501 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18503 * stor-layout.c (finalize_size_functions): Allocate a structure and
18504 reset cfun before dumping the functions.
18506 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18508 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18511 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18512 use xop_pmacsdqh if uns_p.
18513 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18514 the immediate rotate count.
18516 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18518 * rtl.h (struct rtx_def): Add comment for field jump.
18519 (LRA_SUBREG_P): New macro.
18520 * recog.c (register_operand): Check LRA_SUBREG_P.
18521 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18523 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18524 (lra_final_code_change): Skip subreg change for operators.
18525 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18526 if there are no operand changes.
18527 * lra-constraints.c (curr_insn_set): New.
18528 (match_reload): Set LRA_SUBREG_P.
18529 (emit_spill_move): Ditto.
18530 (check_and_process_move): Use curr_insn_set. Process only single
18531 set insns. Don't initialize sec_mem_p and change_p.
18532 (simplify_operand_subreg): Use LRA_SUBREG_P.
18533 (reg_in_class_p): New function.
18534 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18535 of #ifdef. Add code to remove cycling.
18536 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18537 non-null disp. Reload inner instead of disp when base and index
18538 are null. Try to put lo_sum into register.
18539 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18540 (check_and_process_move): Move code for move cost check to
18541 simple_move_p. Remove equiv_substitution.
18542 (simple_move_p): New function.
18543 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18544 curr_insn_set. Call check_and_process_move only for single set
18545 insns. Use the new function. Move call of check_and_process_move
18546 after operand equiv substitution and address process.
18548 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18551 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18552 with nonlocal goto receivers or returns twice calls, ignore
18553 unininitialized values from abnormal edges to nl goto receiver
18554 or returns twice call.
18556 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18558 PR tree-optimization/57051
18559 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18560 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18563 2013-04-26 Richard Biener <rguenther@suse.de>
18565 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18566 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18567 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18568 (fixup_loop_arrays_after_move): New function.
18569 (move_sese_region_to_fn): Properly outline the loop tree parts
18570 of the SESE region.
18572 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18574 * config/i386/i386.md (type, unit): Fix long lines.
18576 2013-04-26 Richard Biener <rguenther@suse.de>
18578 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18579 (lto-streamer-out.o): Likewise.
18580 * cfgloop.c (init_loops_structure): Export, add struct function
18581 argument and adjust.
18582 (flow_loops_find): Adjust.
18583 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18584 (init_loops_structure): Declare.
18585 * lto-streamer-in.c: Include cfgloop.h.
18586 (input_cfg): Input the loop tree.
18587 * lto-streamer-out.c: Include cfgloop.h.
18588 (output_cfg): Output the loop tree.
18589 (output_struct_function_base): Do not drop PROP_loops.
18591 2013-03-26 Richard Biener <rguenther@suse.de>
18593 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18594 (pass_build_cfg): Provide PROP_loops.
18595 (move_sese_region_to_fn): Remove loops that are outlined into fn
18597 * tree-inline.c: Include cfgloop.h.
18598 (initialize_cfun): Do not drop PROP_loops.
18599 (copy_loops): New function.
18600 (copy_cfg_body): Copy loop structure.
18601 (tree_function_versioning): Initialize destination loop tree.
18602 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18603 (pass_parallelize_loops): Do IL verification.
18604 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18605 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18606 the CFG make sure we fixup loops as well.
18607 * tree-ssa-tail-merge.c: Include cfgloop.h.
18608 (replace_block_by): When merging loop latches mark loops for fixup.
18609 * lto-streamer-out.c (output_struct_function_base): Drop
18610 PROP_loops for now.
18611 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18612 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18613 * ipa-split.c: Include cfgloop.h.
18614 (split_function): Add the new return block to the loop tree root.
18615 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18616 whether we have removed the forwarder block.
18617 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18618 * cfgloop.h (place_new_loop): Declare.
18619 * cfgloopmanip.c (place_new_loop): Export.
18620 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18621 (tree-switch-conversion.o): Likewise.
18622 (tree-complex.o): Likewise.
18623 (tree-inline.o): Likewise.
18624 (tree-ssa-tailmerge.o): Likewise.
18625 (ipa-split.o): Likewise.
18626 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18627 (tree-ssa-copy.o): Likewise.
18628 * tree-switch-conversion.c: Include cfgloop.h
18629 (process_switch): If we emit a bit-test cascade, schedule loops
18631 * tree-complex.c: Include cfgloop.h.
18632 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18633 * asan.c: Include cfgloop.h.
18634 (create_cond_insert_point): Properly add new basic-blocks to
18635 loops, schedule loop fixup.
18636 * cfgloop.c (verify_loop_structure): Check that looks are not
18638 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18640 (expand_omp_for_generic): Likewise.
18641 (expand_omp_sections): Likewise.
18642 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18643 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18644 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18645 is initialized, not when loops are present.
18646 * tree-parloops.c (parallelize_loops): Remove checking here.
18647 * passes.c (init_optimization_passes): Schedule a copy-propagation
18648 pass before complete unrolling of inner loops.
18650 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18652 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18653 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18654 (colorize_init): Add argument to _WIN32 version.
18655 * toplev.c: Include diagnostic-color.h.
18656 (process_options): Default to -fdiagnostics-color=auto if
18657 GCC_COLORS env var is in the environment.
18658 * common.opt (fdiagnostics-color=): Add Var and Init.
18659 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18660 env var is in the environment, the default is auto rather than never.
18662 * diagnostic.h (file_name_as_prefix): Add context argument.
18663 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18664 the string as locus.
18665 * langhooks.c (lhd_print_error_function): Adjust caller.
18667 2013-04-25 Lawrence Crowl <crowl@google.com>
18669 * var-tracking.c (shared_hash_def::htab):
18670 Change type to hash_table. Update dependent calls and types.
18672 2013-04-25 Lawrence Crowl <crowl@google.com>
18674 * Makefile.in: Update as needed below.
18676 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18677 Move declaration to after the type's method definitons.
18679 * attribs.c (htab_t scoped_attributes::attribute_hash):
18680 Change type to hash_table. Update dependent calls and types.
18682 * bitmap.c (htab_t bitmap_desc_hash):
18683 Change type to hash_table. Update dependent calls and types.
18685 * cselib.c (htab_t cselib_hash_table):
18686 Change type to hash_table. Update dependent calls and types.
18688 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18689 (hash_string_slot_node): Move implementation into lto-streamer.h
18690 struct string_slot_hasher.
18691 (eq_string_slot_node): Likewise.
18693 * data-streamer-out.c: Update output_block::string_hash_table
18694 dependent calls and types.
18696 * dwarf2cfi.c (htab_t trace_index):
18697 Change type to hash_table. Update dependent calls and types.
18699 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18700 Change type to hash_table. Update dependent calls and types.
18701 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18702 (htab_t optimize_external_refs::map): Likewise.
18703 (htab_t output_comp_unit::extern_map): Likewise.
18704 (htab_t output_comdat_type_unit::extern_map): Likewise.
18705 (htab_t output_macinfo::macinfo_htab): Likewise.
18706 (htab_t optimize_location_lists::htab): Likewise.
18707 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18709 * except.c (htab_t ehspec_hash_type):
18710 Change type to hash_table. Update dependent calls and types.
18711 (assign_filter_values::ttypes): Likewise.
18712 (assign_filter_values::ehspec): Likewise.
18713 (sjlj_assign_call_site_values::ar_hash): Likewise.
18714 (convert_to_eh_region_ranges::ar_hash): Likewise.
18716 * gcse.c (htab_t pre_ldst_table):
18717 Change type to hash_table. Update dependent calls and types.
18719 * ggc-common.c (htab_t saving_htab):
18720 Change type to hash_table. Update dependent calls and types.
18721 (htab_t loc_hash): Likewise.
18722 (htab_t ptr_hash): Likewise.
18723 (call_count): Rename ggc_call_count.
18724 (call_alloc): Rename ggc_call_alloc.
18725 (loc_descriptor): Rename make_loc_descriptor.
18726 (add_statistics): Rename ggc_add_statistics.
18728 * ggc-common.c (saving_htab):
18729 Change type to hash_table. Update dependent calls and types.
18731 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18732 (push_gimplify_context): Likewise.
18733 (pop_gimplify_context): Likewise.
18734 (struct gimple_temp_hash_elt): Added.
18735 (struct gimplify_hasher): Likewise.
18736 (struct gimplify_ctx.temp_htab):
18737 Change type to hash_table. Update dependent calls and types.
18739 * gimple-fold.c: Include gimplify-ctx.h.
18741 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18742 Change type to hash_table. Update dependent calls and types.
18743 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18744 avoid potential global name collision.
18746 * gimplify.c: Include gimplify-ctx.h.
18747 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18748 (htab_t gimplify_ctx::temp_htab):
18749 Update dependent calls and types for new type hash_table.
18750 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18751 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18753 * gimplify-ctx.h: New.
18754 (struct gimple_temp_hash_elt): Move from gimplify.c.
18755 (class gimplify_hasher): New.
18756 (struct gimplify_ctx): Move from gimple.h.
18757 (htab_t gimplify_ctx::temp_htab):
18758 Change type to hash_table. Update dependent calls and types.
18760 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18761 (htab_t ivs_params::newivs_index):
18762 Change type to hash_table. Update dependent calls and types.
18763 (htab_t ivs_params::params_index): Likewise.
18764 (htab_t print_generated_program::params_index): Likewise.
18765 (htab_t gloog::newivs_index): Likewise.
18766 (htab_t gloog::params_index): Likewise.
18768 * graphite.c: Include graphite-htab.h.
18769 4htab_t graphite_transform_loops::bb_pbb_mapping):
18770 Change type to hash_table. Update dependent calls and types.
18772 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18773 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18774 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18776 * graphite-dependences.c: Include graphite-htab.h.
18777 (loop_is_parallel_p): Change hash table type of parameter.
18779 * graphite-htab.h: New.
18780 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18781 (extern find_pbb_via_hash): Move from graphite-poly.h.
18782 (extern loop_is_parallel_p): Move from graphite-poly.h.
18783 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18785 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18786 (extern loop_is_parallel_p): Move to graphite-htab.h.
18787 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18789 * haifa-sched.c (htab_t delay_htab):
18790 Change type to hash_table. Update dependent calls and types.
18791 (htab_t delay_htab_i2): Likewise.
18793 * ira-color.c (htab_t allocno_hard_regs_htab):
18794 Change type to hash_table. Update dependent calls and types.
18796 * ira-costs.c (htab_t cost_classes_htab):
18797 Change type to hash_table. Update dependent calls and types.
18799 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18800 Change type to hash_table. Update dependent calls and types.
18802 * loop-iv.c (htab_t bivs):
18803 Change type to hash_table. Update dependent calls and types.
18805 * loop-unroll.c (htab_t opt_info::insns_to_split):
18806 Change type to hash_table. Update dependent calls and types.
18807 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18809 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18810 (struct string_slot_hasher): New.
18811 (htab_t output_block::string_hash_table):
18812 Change type to hash_table. Update dependent calls and types.
18814 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18815 (htab_t file_name_hash_table):
18816 Change type to hash_table. Update dependent calls and types.
18818 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18821 * lto-streamer.c (htab_t tree_htab):
18822 Change type to hash_table. Update dependent calls and types.
18824 * omp-low.c: Include gimplify-ctx.h.
18826 * passes.c (htab_t name_to_pass_map):
18827 Change type to hash_table. Update dependent calls and types.
18828 (pass_traverse): Rename to passes_pass_traverse.
18830 * plugin.c (htab_t event_tab):
18831 Change type to hash_table. Update dependent calls and types.
18833 * postreload-gcse.c (htab_t expr_table):
18834 Change type to hash_table. Update dependent calls and types.
18835 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18837 * sese.c (debug_rename_map_1): Make extern.
18838 (htab_t copy_bb_and_scalar_dependences::rename_map):
18839 Change type to hash_table. Update dependent calls and types.
18841 * sese.h (extern debug_rename_map): Move to .c file.
18843 * store-motion.c (htab_t store_motion_mems_table):
18844 Change type to hash_table. Update dependent calls and types.
18846 * trans-mem.c (htab_t tm_new_mem_hash):
18847 Change type to hash_table. Update dependent calls and types.
18849 * tree-browser.c (htab_t TB_up_ht):
18850 Change type to hash_table. Update dependent calls and types.
18852 * tree-cfg.c (htab_t discriminator_per_locus):
18853 Change type to hash_table. Update dependent calls and types.
18855 * tree-complex.c: Include tree-hasher.h
18856 (htab_t complex_variable_components):
18857 Change type to hash_table. Update dependent calls and types.
18859 * tree-eh.c (htab_t finally_tree):
18860 Change type to hash_table. Update dependent calls and types.
18862 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18863 struct int_tree_hasher.
18864 (extern int_tree_map_eq): Likewise.
18865 (uid_decl_map_hash): Removed.
18866 (extern decl_tree_map_eq): Likewise.
18868 * tree-hasher.h: New.
18869 (struct int_tree_hasher): New.
18870 (typedef int_tree_htab_type): New.
18872 * tree-inline.c: Include gimplify-ctx.h.
18874 * tree-mudflap.c: Include gimplify-ctx.h.
18876 * tree-parloops.c: Include tree-hasher.h.
18877 (htab_t eliminate_local_variables_stmt::decl_address):
18878 Change type to hash_table. Update dependent calls and types.
18879 (htab_t separate_decls_in_region::decl_copies): Likewise.
18881 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18882 Change type to hash_table. Update dependent calls and types.
18884 * tree-sra.c (candidates):
18885 Change type to hash_table. Update dependent calls and types.
18887 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18889 (int_tree_map_hash): Likewise.
18891 * tree-ssa-dom.c (htab_t avail_exprs):
18892 Change type to hash_table. Update dependent calls and types.
18894 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18895 Change type to hash_table. Update dependent calls and types.
18897 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18898 Change type to hash_table. Update dependent calls and types.
18900 * tree-ssa-phiopt.c (seen_ssa_names):
18901 Change type to hash_table. Update dependent calls and types.
18903 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18904 Change type to hash_table. Update dependent calls and types.
18906 * tree-ssa-uncprop.c (equiv):
18907 Change type to hash_table. Update dependent calls and types.
18909 2013-04-25 Jakub Jelinek <jakub@redhat.com>
18911 PR rtl-optimization/57003
18912 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18913 call note_stores with kill_clobbered_value callback again after
18914 killing regs_invalidated_by_call.
18916 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18918 * config/aarch64/aarch64-simd.md
18919 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18920 (aarch64_simd_bsl<mode>): Likewise.
18921 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18923 2013-04-25 Marek Polacek <polacek@redhat.com>
18925 PR tree-optimization/57066
18926 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
18928 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18930 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18932 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18934 * config/aarch64/aarch64-builtins.c
18935 (aarch64_fold_builtin): New.
18936 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18937 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18938 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18939 * config/aarch64/arm_neon.h
18940 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18942 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18943 Tejas Belagod <tejas.belagod@arm.com>
18945 * config/aarch64/aarch64-builtins.c
18946 (aarch64_gimple_fold_builtin): New.
18947 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18948 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18949 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18950 (addvv4sf): Update.
18951 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18953 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18955 * config/aarch64/aarch64.md
18956 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18958 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18960 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18961 (*ngcsi_uxtw): New pattern.
18963 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18964 Julian Brown <julian@codesourcery.com>
18966 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18967 (TB_DREG): Add T_V4HF.
18968 (v4hf_UP): New macro.
18969 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18970 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18971 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18973 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18974 (arm_vector_mode_supported_p): Handle V4HF.
18975 (arm_mangle_map): Handle V4HFmode.
18976 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18977 * config/arm/arm_neon_builtins.def: Add entries for
18978 vcvtv4hfv4sf, vcvtv4sfv4hf.
18979 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18980 (neon_vcvtv4hfv4sf): Likewise.
18981 * config/arm/neon-gen.ml: Handle half-precision floating point
18983 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18984 * config/arm/arm_neon.h: Regenerate.
18985 * config/arm/neon.ml (type elts): Add F16.
18986 (type vectype): Add T_float16x4, T_floatHF.
18987 (type vecmode): Add V4HF.
18988 (type features): Add Requires_FP_bit feature.
18989 (elt_width): Handle F16.
18990 (elt_class): Likewise.
18991 (elt_of_class_width): Likewise.
18992 (mode_of_elt): Refactor.
18993 (type_for_elt): Handle F16, fix error messages.
18994 (vectype_size): Handle T_float16x4.
18995 (vcvt_sh): New function.
18996 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18997 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18998 (string_of_mode): Handle V4HF.
18999 * doc/arm-neon-intrinsics.texi: Regenerate.
19001 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19003 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19004 format specifier in 'X' case.
19006 2013-04-25 Alan Modra <amodra@gmail.com>
19009 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19010 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19011 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19012 Repeat for many other rotate/shift and mask patterns using subregs.
19013 Name lshiftrt insns.
19014 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19015 on WORDS_BIG_ENDIAN.
19017 2013-04-25 Alan Modra <amodra@gmail.com>
19019 * config.gcc: Support little-endian powerpc-linux targets.
19020 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19021 (LINK_OS_LINUX_SPEC): Define.
19022 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19023 Preserve MASK_LITTLE_ENDIAN.
19024 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19025 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19026 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19027 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19028 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19029 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19030 Correct fp word order for little-endian. Don't shift toc entries
19031 smaller than a word for little-endian.
19032 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19033 (bswapdi2 splits): Correct low-part subreg for little-endian.
19034 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19035 low/high where such is correct only for be.
19036 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19037 little-endian for -mcall-aixdesc.
19039 2013-04-25 Alan Modra <amodra@gmail.com>
19041 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19042 replace_equiv_address_nv.
19044 2013-04-25 Alan Modra <amodra@gmail.com>
19046 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19048 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19051 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19052 * rtl.h (struct rtx_def): ...
19054 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19056 PR rtl-optimizations/57046
19057 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19058 for multi-reg splits.
19060 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19062 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19064 2013-04-24 Sterling Augustine <saugustine@google.com>
19066 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19067 (comp_dir_string, debug_str_dwo_section): New.
19068 (DEBUG_STR_DWO_SECTION): Rename to ...
19069 (DEBUG_DWO_STR_SECTION): ... this.
19070 (DEBUG_NORM_STR_SECTION): Delete.
19071 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19072 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19073 (find_AT_string): Move most logic to ...
19074 (find_AT_string_in_table): ... here. New.
19075 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19076 add_skeleton_AT_string. Delete logic.
19077 (output_skeleton_debug_sections): Remove call to
19078 add_top_level_skeleton_die_attrs.
19079 (add_comp_dir_attribute): Move logic to comp_dir_string.
19080 (dwarf2out_init): Initialize debug_str_dwo_section.
19081 (output_indirect_string): Call find_string_form.
19082 (output_indirect_strings): Rewrite.
19083 (prune_unused_types): Empty skeleton_debug_str_hash.
19084 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19085 (dwarf2out_finish): Call output_indirect_strings.
19087 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19089 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19091 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19093 * rtl.h (struct rtx_def): Add comment for field jump.
19094 (LRA_SUBREG_P): New macro.
19095 * recog.c (register_operand): Check LRA_SUBREG_P.
19096 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19098 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19099 (lra_final_code_change): Skip subreg change for operators.
19100 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19101 if there are no operand changes.
19102 * lra-constraints.c (curr_insn_set): New.
19103 (match_reload): Set LRA_SUBREG_P.
19104 (emit_spill_move): Ditto.
19105 (check_and_process_move): Use curr_insn_set. Process only single
19106 set insns. Don't initialize sec_mem_p and change_p.
19107 (simplify_operand_subreg): Use LRA_SUBREG_P.
19108 (reg_in_class_p): New function.
19109 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19110 of #ifdef. Add code to remove cycling.
19111 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19112 non-null disp. Reload inner instead of disp when base and index
19113 are null. Try to put lo_sum into register.
19114 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19115 (check_and_process_move): Move code for move cost check to
19116 simple_move_p. Remove equiv_substitution.
19117 (simple_move_p): New function.
19118 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19119 curr_insn_set. Call check_and_process_move only for single set
19120 insns. Use the new function. Move call of check_and_process_move
19121 after operand equiv substitution and address process.
19123 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19125 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19126 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19127 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19129 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19131 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19133 2013-04-24 Marek Polacek <polacek@redhat.com>
19135 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19136 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19137 (select_loops_exit_conditions): Likewise.
19138 (number_of_iterations_for_all_loops): Likewise.
19139 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19140 (scev_analysis): Likewise.
19142 2013-04-02 Catherine Moore <clm@codesourcery.com>
19143 Chao-ying Fu <fu@mips.com>
19145 * config/mips/micromips.md (jraddiusp): New pattern.
19146 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19147 instruction if possible.
19149 2013-04-24 Alan Modra <amodra@gmail.com>
19151 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19153 2013-04-24 Julian Brown <julian@codesourcery.com>
19154 Chung-Lin Tang <cltang@codesourcery.com>
19156 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19157 dependency behavior in enumeration type DIE generation. Add TODO note
19158 to comments about future DW_FORM_sdata/udata re-work of related code.
19160 2013-04-23 Lawrence Crowl <crowl@google.com>
19162 * Makefile.in: Update as needed below.
19164 * hash-table.h (class hash_table):
19165 Correct many methods with parameter types compare_type to the correct
19166 value_type. (Correct code was unlikely to notice the change.)
19167 (hash_table::elements_with_deleted) New.
19168 (class hashtable::iterator): New.
19169 (hashtable::begin()): New.
19170 (hashtable::end()): New.
19171 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19173 * statistics.c (statistics_hashes):
19174 Change type to hash_table. Update dependent calls and types.
19176 * tree-into-ssa.c (var_infos):
19177 Change type to hash_table. Update dependent calls and types.
19179 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19180 Change type to hash_table. Update dependent calls and types.
19182 * tree-ssa-loop-im.c (struct mem_ref.refs):
19183 Change type to hash_table. Update dependent calls and types.
19185 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19186 Change type to hash_table. Update dependent calls and types.
19188 * tree-ssa-sccvn.c (vn_tables_s::nary):
19189 Change type to hash_table. Update dependent calls and types.
19190 (vn_tables_s::phis): Likewise.
19191 (vn_tables_s::references): Likewise.
19193 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19194 (vn_reference_eq): Update parameter and return types.
19196 * tree-ssa-structalias.c (pointer_equiv_class_table):
19197 Change type to hash_table. Update dependent calls and types.
19198 (location_equiv_class_table): Likewise.
19200 * tree-vect-data-refs.c: Consequential changes for making
19201 peeling a hash_table.
19203 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19204 (destroy_loop_vec_info): Dependent hash_table update.
19206 * tree-vectorizer.h (peeling_htab):
19207 Change type to hash_table. Update dependent calls and types.
19209 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19211 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19212 to check the register content is equal or not.
19213 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19214 to assign register content record.
19215 * lra-eliminations.c (update_reg_eliminate): Use
19216 lra_update_reg_val_offset to update register content offset.
19217 * lra-int.h (struct lra_reg): Add offset member.
19218 (lra_reg_val_equal_p): New static inline function.
19219 (lra_update_reg_val_offset): New static inline function.
19220 (lra_assign_reg_val): New static inline function.
19221 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19222 to assign register content record.
19223 (initialize_lra_reg_info_element): Initial offset to zero.
19225 2013-04-23 Catherine Moore <clm@codesourcery.com>
19227 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19228 operands. Record compression.
19230 2013-04-23 Xinliang David Li <davidxl@google.com>
19232 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19234 2013-04-23 Richard Biener <rguenther@suse.de>
19236 PR middle-end/57036
19237 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19238 parameter, only add abnormal goto edges from the copied body
19239 if the call could perform abnormal gotos.
19240 (copy_cfg_body): Adjust.
19242 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19244 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19246 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19248 * coretypes.h (gimple_stmt_iterator): Add struct to make
19251 2013-04-23 Richard Biener <rguenther@suse.de>
19253 PR tree-optimization/57026
19254 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19255 from SSA names occuring in abnormal PHI nodes.
19257 2013-04-22 Andi Kleen <ak@linux.intel.com>
19259 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19261 2013-04-22 Andi Kleen <ak@linux.intel.com>
19263 * lto/lto.c (print_lto_report_1): Declare early.
19264 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19266 2013-04-22 Andi Kleen <ak@linux.intel.com>
19268 * common.opt (-flto-report-wpa): Add.
19269 * doc/invoke.texi (-flto-report-wpa): Add.
19270 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19273 2013-04-22 Xinliang David Li <davidxl@google.com>
19275 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19276 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19277 * Makefile.in: New dependency
19279 David Daney <ddaney.cavm@gmail.com>
19281 * configure.ac (gcc_cv_as_micromips_support): Use the
19282 --fatal-warnings option.
19283 * configure: Regenerate.
19285 2013-04-22 Marek Polacek <polacek@redhat.com>
19288 * tsan.c (instrument_expr): Don't instrument expression
19289 in case its size is zero.
19291 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19295 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19297 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19299 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19301 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19302 (gimple_stmt_iterator): New typedef.
19303 * gimple.h (gimple_stmt_iterator): Rename to...
19304 (gimple_stmt_iterator_d): ... This.
19305 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19306 trees be valid for GIMPLE and GENERIC.
19307 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19308 * gimple-fold.c (gimple_fold_call): Call target hook
19309 gimple_fold_builtin.
19310 * hooks.c (hook_bool_gsiptr_false): New.
19311 * hooks.h (hook_bool_gsiptr_false): New.
19312 * target.def (fold_stmt): New.
19313 * doc/tm.texi: Regenerate.
19315 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19318 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19319 a set sp if no stack realignment.
19321 2013-04-22 Nick Clifton <nickc@redhat.com>
19323 * config.gcc (tilegx-linux): Extend extra_objs rather than
19325 (tilepro-linux): Likewise.
19327 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19329 * config/aarch64/aarch64-builtins.c
19331 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19332 (VAR<1-12>): Add MAP parameter.
19333 (BUILTIN_*): Likewise.
19334 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19335 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19336 (aarch64_ushl_n<mode>): Likewise.
19337 (aarch64_sshr_n<mode>): Likewise.
19338 (aarch64_ushr_n<mode>): Likewise.
19339 (aarch64_<maxmin><mode>): Likewise.
19340 (aarch64_sqrt<mode>): Likewise.
19341 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19342 (vshr<q>_n_*): Likewise.
19344 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19346 * config/aarch64/aarch64-builtins.c
19347 (aarch64_simd_builtin_type_mode): Handle SF types.
19349 (BUILTIN_GPF): Define.
19350 (aarch64_init_simd_builtins): Handle SF types.
19351 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19352 (frecps): Likewise.
19353 (frecpx): Likewise.
19354 * config/aarch64/aarch64-simd.md
19355 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19356 (aarch64_frecpe<mode>): New.
19357 (aarch64_frecps<mode>): Likewise.
19358 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19359 (v8type): Add frecp<esx>.
19360 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19361 (aarch64_frecps<mode>): Likewise.
19362 * config/aarch64/iterators.md (FRECP): New.
19363 (frecp_suffix): Likewise.
19364 * config/aarch64/arm_neon.h
19365 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19367 2013-04-22 Christian Bruel <christian.bruel@st.com>
19370 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19371 (REG_CLASS_NAMES): Idem.
19372 (REG_CLASS_CONTENTS): Idem.
19373 (REGCLASS_HAS_FP_REG): Idem.
19374 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19375 (sh_conditional_register_usage): Idem.
19377 2013-04-21 Jeff Law <law@redhat.com>
19379 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19380 (ssa_forward_propagate_and_combine): Use it.
19382 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19384 * lra.c: Update the flow chart diagram.
19386 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19388 PR rtl-optimization/56847
19389 * lra-constraints.c (process_alt_operands): Discourage alternative
19390 with non-matche doffsettable memory constraint fro memory with
19393 2013-04-19 Richard Biener <rguenther@suse.de>
19395 PR tree-optimization/56982
19396 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19398 * gimplify.c (gimplify_call_expr): Notice special calls.
19399 (gimplify_modify_expr): Likewise.
19400 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19401 abnormal control flow receivers.
19402 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19403 in the same way as cfun->has_nonlocal_labels.
19404 (gimple_purge_dead_abnormal_call_edges): Likewise.
19405 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19406 receivers start a basic-block.
19408 2013-04-19 Richard Biener <rguenther@suse.de>
19410 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19412 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19413 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19414 (SLP_TREE_LOAD_PERMUTATION): Add.
19415 (vect_transform_slp_perm_load): Adjust prototype.
19416 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19417 (vect_free_slp_instance): Likewise.
19418 (vect_create_new_slp_node): Likewise.
19419 (vect_supported_slp_permutation_p): Remove.
19420 (vect_slp_rearrange_stmts): Adjust.
19421 (vect_supported_load_permutation_p): Likewise. Inline
19422 vect_supported_slp_permutation_p here.
19423 (vect_analyze_slp_instance): Compute load permutations per
19424 slp node instead of per instance.
19425 (vect_get_slp_defs): Adjust.
19426 (vect_transform_slp_perm_load): Likewise.
19427 (vect_schedule_slp_instance): Remove redundant code.
19428 (vect_schedule_slp): Remove hack for PR56270, add it ...
19429 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19430 CSE loads for SLP. Adjust.
19432 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19434 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19435 spelling in two comments.
19437 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19440 * config/arm/arm.c (load_multiple_sequence): Require SP
19441 as base register for loads if SP is in the register list.
19443 2013-04-19 Martin Jambor <mjambor@suse.cz>
19445 PR tree-optimization/56718
19446 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19447 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19448 and made public. Adjusted all callers.
19449 (ipa_intraprocedural_devirtualization): New function.
19450 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19451 (ipa_intraprocedural_devirtualization): Likewise.
19452 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19454 2013-04-19 Richard Biener <rguenther@suse.de>
19456 PR tree-optimization/57000
19457 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19459 2013-04-19 Terry Guo <terry.guo@arm.com>
19461 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19463 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19464 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19465 (cortex_m4_fmacs): Use new reservations.
19466 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19468 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19470 PR rtl-optimization/56999
19471 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19473 (lra_coalesce): Remove split_origin_bitmap and related code.
19474 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19475 ranges if necessary.
19477 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19479 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19481 (ix86_expand_call): Remove clobbered_registers array and use
19482 x86_64_ms_sysv_extra_clobbered_registers instead.
19483 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19485 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19487 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19488 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19489 (*call_value_rex64_ms_sysv): Ditto.
19491 2013-04-18 Cary Coutant <ccoutant@google.com>
19493 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19494 parent instead of die_mark.
19496 2013-04-18 Diego Novillo <dnovillo@google.com>
19498 * gimple.c (create_gimple_tmp): New.
19499 (get_expr_type): New.
19500 (build_assign): New.
19501 (build_type_cast): New.
19502 * gimple.h (enum ssa_mode): Define.
19503 (gimple_seq_set_location): New.
19504 * asan.c (build_check_stmt): Change some gimple_build_* calls
19505 to use build_assign and build_type_cast.
19507 2013-04-18 Richard Biener <rguenther@suse.de>
19509 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19510 handle negative step. Remove redundant checks.
19511 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19512 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19513 for negative step and grouped loads fail to vectorize.
19515 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19517 * emit-rtl.c (reset_insn_used_flags): New function.
19518 (reset_all_used_flags): Use it.
19519 (verify_insn_sharing): New function.
19520 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19522 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19524 PR tree-optimization/56984
19525 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19526 and (x >> M) >= N don't register any assertion if N << M is the
19529 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19531 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19532 do not delete_insn insns that have not yet been emitted, only
19533 unlink them with remove_insn.
19534 * df-scan.c (df_insn_delete): Revert r197492.
19536 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19538 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19539 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19541 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19543 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19544 define_insn_and_split.
19545 (and_scc,ior_scc,negscc): Likewise.
19546 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19548 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19550 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19551 can benefit from using a sequence of LDRD instructions in epilogue
19552 instead of a single LDM instruction.
19554 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19557 * doc/extend.texi: Fix typo.
19559 2013-04-17 Richard Biener <rguenther@suse.de>
19561 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19562 (vect_build_slp_tree): ... here.
19563 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19564 match. Remove special-casing of mismatched complex loads.
19565 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19566 re-try the match with swapped commutative operands.
19567 (vect_supported_load_permutation_p): Remove special-casing of
19568 mismatched complex loads.
19569 (vect_analyze_slp_instance): Adjust.
19571 2013-04-17 Richard Biener <rguenther@suse.de>
19573 PR rtl-optimization/56921
19574 * cfgloop.h (struct loop): Add simple_loop_desc member.
19575 (struct niter_desc): Mark with GTY(()).
19576 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19577 * loop-iv.c (get_simple_loop_desc): Likewise.
19578 (free_simple_loop_desc): Likewise.
19581 2013-04-16 Richard Biener <rguenther@suse.de>
19583 PR rtl-optimization/56921
19584 * loop-init.c (pass_rtl_move_loop_invariants): Add
19585 TODO_do_not_ggc_collect to todo_flags_finish.
19586 (pass_rtl_unswitch): Same.
19587 (pass_rtl_unroll_and_peel_loops): Same.
19588 (pass_rtl_doloop): Same.
19590 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19592 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19593 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19594 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19596 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19597 * tree-streamer.c (record_common_node): Adjust reference in comment.
19599 2013-04-17 Terry Guo <terry.guo@arm.com>
19601 * config/arm/cortex-m4.md: Add a new bypass.
19603 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19605 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19607 (*subs_<optab><mode>_multp2): New pattern.
19608 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19609 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19611 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19613 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19614 (*subs_mul_imm_<mode>): New pattern.
19616 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19619 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19620 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19621 (vsx_movti_32bit): Change j->wa to O->wa.
19623 2013-04-16 Richard Biener <rguenther@suse.de>
19625 PR rtl-optimization/56921
19626 * loop-init.c (pass_rtl_move_loop_invariants): Add
19627 TODO_do_not_ggc_collect to todo_flags_finish.
19628 (pass_rtl_unswitch): Same.
19629 (pass_rtl_unroll_and_peel_loops): Same.
19630 (pass_rtl_doloop): Same.
19632 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19634 * config/arm/arm.c (emit_multi_reg_push): New declaration
19635 for an existing function.
19636 (arm_emit_strd_push): New function.
19637 (arm_expand_prologue): Used here.
19638 (arm_emit_ldrd_pop): New function.
19639 (arm_expand_epilogue): Used here.
19640 (arm_get_frame_offsets): Update condition.
19641 (arm_emit_multi_reg_pop): Add a special case for load of a single
19642 register with writeback.
19644 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19646 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19649 2013-04-16 Richard Biener <rguenther@suse.de>
19651 PR tree-optimization/56756
19652 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19653 (first_mem_ref_loc): New.
19654 (execute_sm): Place the load temporarily before a previous
19655 access instead of in the latch edge to ensure its SSA dependencies
19656 are defined at points dominating the load.
19658 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19660 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19661 correct fix by moving header and footer insn to the footer of
19662 the merged basic block. Clear BB_END of the merged-away block.
19664 PR middle-end/43631
19665 * emit-rtl.c (make_note_raw): New function.
19666 (link_insn_into_chain): New static inline function.
19667 (add_insn): Use it.
19668 (add_insn_before, add_insn_after): Factor insn chain linking code...
19669 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19670 using link_insn_into_chain.
19671 (note_outside_basic_block_p): New helper function for emit_note_after
19672 and emit_note_before.
19673 (emit_note_after): Use nobb variant of add_insn_after if the note
19674 should not be contained in a basic block.
19675 (emit_note_before): Use nobb variant of add_insn_before if the note
19676 should not be contained in a basic block.
19677 (emit_note_copy): Use make_note_raw.
19678 (emit_note): Likewise.
19679 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19680 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19681 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19682 the moved barrier the tail of the basic block it follows.
19683 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19685 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19687 PR tree-optimization/56962
19688 * gimple-ssa-strength-reduction.c (record_increment): Only set
19689 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19690 either rhs1 or rhs2 is equal to c->base_expr.
19692 2013-04-15 Richard Biener <rguenther@suse.de>
19694 PR tree-optimization/56933
19695 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19697 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19698 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19699 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19700 dependence check ...
19701 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19703 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19704 GROUP_READ_WRITE_DEPENDENCE.
19706 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19708 * emit-rtl.c (reset_all_used_flags): New function.
19709 (verify_rtl_sharing): Call reset_all_used_flags before and after
19710 performing the checks.
19712 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19714 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19715 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19716 * config/arm/constraints.md (De): New constraint.
19717 * config/arm/neon.md (anddi3_neon): Delete.
19718 (neon_vand<mode>): Expand to standard anddi3 pattern.
19719 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19720 Move earlier in the file.
19721 (neon_inv_logic_op2): Likewise.
19722 (arm_anddi_operand_neon): New predicate.
19724 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19726 * configure.ac (gcc_cv_ld_as_needed): Set
19727 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19728 Use -z ignore, -z record on *-*-solaris2*.
19729 (HAVE_LD_AS_NEEDED): Update comment.
19730 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19731 * configure: Regenerate.
19732 * config.in: Regenerate.
19733 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19734 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19735 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19736 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19737 equivalents. Fix markup.
19738 * doc/tm.texi: Regenerate.
19740 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19742 * config/i386/i386.opt: New option mstack-protector-guard=.
19743 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19744 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19745 TARGET_SSP_TLS_GUARD.
19746 * config/i386/i386.c (ix86_option_override_internal): Set
19747 ix86_stack_protector_guard.
19748 * config/i386/i386.md (stack_protect_set): Enable for
19749 TARGET_SSP_TLS_GUARD only.
19750 (stack_protect_set_<mode>): Ditto.
19751 (stack_protect_test): Ditto.
19752 (stack_protect_test_<mode>): Ditto.
19753 * doc/invoke.texi (i386 Option): Document.
19755 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19758 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19759 (S_MODES): Set H_MODE bit.
19760 (SF_MODES): Set only S_MODE and SF_MODE bits.
19761 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19762 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19763 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19764 <MODE_FLOAT>: Likewise.
19766 2013-04-15 Joey Ye <joey.ye@arm.com>
19768 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19770 2013-04-15 Joey Ye <joey.ye@arm.com>
19772 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19774 (thumb_far_jump_used_p): Count instruction size and set
19777 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19779 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19780 * resource.c (mark_target_live_regs): Reformat conditional block.
19782 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19784 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19785 notes, they are emitted only just before final.
19786 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19788 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19790 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19791 * cfgrtl.c (delete_insn): Call it here instead.
19792 * lra-spills.c (lra_final_code_change): Use delete_insn.
19793 * haifa-sched.c (sched_remove_insn): Likewise.
19794 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19795 returning to the nop pool.
19796 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19797 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19799 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19801 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19802 * doc/tm.texi: Regenerated.
19804 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19806 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19809 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19811 * df-core.c (df_find_def): Compare register numbers.
19812 (df_find_use): Likewise.
19814 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19817 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19818 lra_in_progress for return.
19820 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19822 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19823 define_insn into define_insn_and_split and emit movsicc patterns.
19825 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19827 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19829 2013-04-12 Richard Biener <rguenther@suse.de>
19831 * tree-pass.h (TODO_do_not_ggc_collect): New.
19832 * passes.c (execute_one_ipa_transform_pass): Honor
19833 TODO_do_not_ggc_collect.
19834 (execute_one_pass): Likewise.
19837 2013-04-10 Richard Biener <rguenther@suse.de>
19839 * passes.c (init_optimization_passes): Remove reload pass.
19840 * ira.c (do_reload): Merge into ...
19842 (rest_of_handle_reload): Remove.
19843 (pass_reload): Likewise.
19844 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19845 of reload for vzeroupper pass placement.
19847 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19849 PR tree-optimization/56918
19850 PR tree-optimization/56920
19851 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19852 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19853 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19854 use wide_mul_with_sign method.
19856 2013-04-12 Richard Biener <rguenther@suse.de>
19858 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19859 not be considered a gimple constant.
19861 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
19863 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19864 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19867 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19868 Jakub Jelinek <jakub@redhat.com>
19870 * opts.c: Include diagnostic-color.h.
19871 (common_handle_option): Handle OPT_fdiagnostics_color_.
19872 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19873 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19874 (diagnostic-color.o): New.
19875 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19876 (diagnostic_color_rule): New enum.
19877 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19878 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19879 the location string.
19880 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19881 either NULL, or color kind.
19882 * diagnostic-color.c: New file.
19883 * diagnostic-color.h: New file.
19884 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19886 * doc/invoke.texi (-fdiagnostics-color): Document.
19887 * pretty-print.h (pp_show_color): Define.
19888 (struct pretty_print_info): Add show_color field.
19889 * diagnostic.c: Include diagnostic-color.h.
19890 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19891 macros. Colorize error:, warning: etc. strings and also the location
19893 (diagnostic_show_locus): Colorize the caret line.
19894 * pretty-print.c: Include diagnostic-color.h.
19895 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19896 inside of %< %> quotes or quoted through q format modifier.
19898 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19900 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
19902 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19904 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19905 code in CC_NZ mode.
19906 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19909 2013-04-11 Marek Polacek <polacek@redhat.com>
19911 PR tree-optimization/48184
19912 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
19914 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19916 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19917 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19918 (skip_simple_arithmetic): Tidy up.
19919 * tree.h (skip_simple_constant_arithmetic): Declare.
19921 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19923 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19925 2013-04-11 Richard Biener <rguenther@suse.de>
19927 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19928 generate vector constants.
19930 2013-04-11 Richard Biener <rguenther@suse.de>
19932 PR tree-optimization/56878
19933 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19934 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19936 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19937 Prefer to align the DR with the most invariant base address.
19939 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19941 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19944 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19946 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19947 floating-point vector comparisons against 0.
19949 2013-04-11 Jakub Jelinek <jakub@redhat.com>
19951 PR tree-optimization/56899
19952 * fold-const.c (extract_muldiv_1): Apply distributive law
19953 only if TYPE_OVERFLOW_WRAPS (ctype).
19955 2013-04-11 Bin Cheng <bin.cheng@arm.com>
19958 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19959 loading has side effect.
19961 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19963 * config/sparc/sparc.c: Include tree-pass.h.
19964 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19965 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19966 head of file. Change return type. Split off gate function.
19967 (sparc_gate_work_around_errata): New function.
19968 (pass_work_around_errata): New pass definition.
19969 (insert_pass_work_around_errata) New pass insert definition to
19970 insert pass_work_around_errata just after delayed-branch scheduling.
19971 (sparc_option_override): Insert the pass.
19972 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19974 2013-04-10 David S. Miller <davem@davemloft.net>
19976 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19977 or -mcpu=hypersparc.
19979 * target.def (cstore_mode): New hook.
19980 * target.h: Include insn-codes.h
19981 * targhooks.c: Likewise.
19982 (default_cstore_mode): New function.
19983 * targhooks.h: Declare it.
19984 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19985 * doc/tm.texi: Rebuild.
19986 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19987 target hook, rather than inspecting the insn_data.
19988 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19989 (TARGET_CSTORE_MODE): Redefine.
19990 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19992 * config/sparc/predicates.md (cstore_result_operand): New special
19994 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19995 Use it for operand 0.
19996 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19997 (*snesi_special): Likewise.
19998 (*snesi_zero): Likewise.
19999 (*seqsi_zero): Likewise.
20000 (*sltu_insn): Likewise.
20001 (*sgeu_insn): Likewise.
20002 (*seqdi_special): Make operand 0 and comparison operation be of
20004 (*snedi_special): Likewise.
20005 (*snedi_special_vis3): Likewise.
20006 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20007 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20008 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20009 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20010 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20011 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20012 (*sltu_extend_sp64): Likewise.
20013 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20014 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20015 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20016 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20017 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20019 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20021 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20022 (aarch64_start_file): Use the new function.
20024 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20025 Jason Merrill <jason@redhat.com>
20027 * common.opt: Add -gdwarf.
20028 * opts.c (common_handle_option): Handle it.
20029 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20031 2013-04-10 Richard Biener <rguenther@suse.de>
20033 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20034 (execute_one_ipa_transform_pass): But unconditionally here.
20035 (execute_one_pass): And here.
20036 (init_optimization_passes): Remove reload pass.
20037 * tree-pass.h (TODO_ggc_collect): Remove.
20038 (pass_reload): Likewise.
20039 * ira.c (do_reload): Merge into ...
20041 (rest_of_handle_reload): Remove.
20042 (pass_reload): Likewise.
20043 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20044 of reload for vzeroupper pass placement.
20045 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20046 and todo_flags_finish of all passes.
20048 2013-04-10 Richard Biener <rguenther@suse.de>
20050 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20051 first_const_oprnd field, rename first_def_type to first_op_type.
20052 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20053 (vect_get_and_check_slp_defs): Always use the type of the
20054 operand. Allow mixed vect_external_def, vect_constant_def types.
20055 (vect_get_constant_vectors): Handle mixed vect_external_def,
20056 vect_constant_def types.
20058 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20060 PR tree-optimization/55524
20061 * tree-ssa-math-opts.c
20062 (convert_mult_to_fma): Don't use an fms construct
20063 when we don't have an fms operation, but fnma, and it looks
20064 likely that we'll be able to use the latter.
20066 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20068 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20070 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20071 inline fail caused by overwritable functions.
20073 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20075 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20076 unnecessary bits in the constant power of two case.
20078 2013-04-10 Richard Biener <rguenther@suse.de>
20080 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20081 broken code swapping operands.
20082 (vect_build_slp_tree): Do not compute load permutations here.
20083 (vect_analyze_slp_instance): Compute load permutations here,
20084 after building the SLP tree.
20086 2013-04-09 Christian Bruel <christian.bruel@st.com>
20088 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20089 of next/prev_real_insn.
20091 2013-04-09 Jan Hubicka <jh@suse.cz>
20093 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20094 Drop aliased parameter.
20095 (function_and_variable_visibility): Do not handle alias pairs.
20096 * cgraph.c (varpool_externally_visible_p): Update prototype.
20097 * varpool.c (varpool_add_new_variable): Update.
20099 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20101 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20103 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20105 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20107 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20109 2013-04-09 Marek Polacek <polacek@redhat.com>
20111 PR tree-optimization/48762
20112 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20114 2013-04-09 Richard Biener <rguenther@suse.de>
20116 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20118 (vect_build_slp_tree): Likewise.
20119 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20120 calculating the cost of a SLP instance.
20121 (vect_analyze_slp_instance): Use it from here, after building
20124 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20126 PR middle-end/56883
20127 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20128 expand_omp_for_static_chunk): Use simple_p = true in
20129 force_gimple_operand_gsi calls when assigning to addressable decls.
20131 2013-04-09 Jeff Law <law@redhat.com>
20133 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20134 when the boolean was created by converting a wider object which
20135 had a boolean range.
20137 2013-04-09 Richard Biener <rguenther@suse.de>
20139 * tree-vectorizer.h (slp_void_p): Remove.
20140 (slp_tree): Typedef before _slp_tree declaration.
20141 (struct _slp_tree): Use a vector of slp_tree as children.
20142 (vect_get_place_in_interleaving_chain): Remove.
20143 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20145 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20147 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20148 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20149 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20150 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20151 Use slp_node instead of slp_void_p and adjust.
20153 2013-04-09 Richard Biener <rguenther@suse.de>
20155 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20156 work that is not necessary.
20158 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20160 PR tree-optimization/56854
20161 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20162 forward into clobber stmts if it would change MEM_REF lhs into
20165 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20167 * tree.c (type_hash_lookup, type_hash_add): Make static.
20168 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20170 2013-04-09 Richard Biener <rguenther@suse.de>
20172 * tree.h (unsave_expr_now): Remove.
20173 * tree-inline.c (mark_local_for_remap_r): Remove.
20174 (unsave_expr_1): Likewise.
20175 (unsave_r): Likewise.
20176 (unsave_expr_now): Likewise.
20177 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20178 (propagate_tree_value): Likewise.
20180 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20182 * doc/rtl.texi (sequence): Rewrite documentation to match the
20183 current use of SEQUENCE rtl objects.
20184 * rtl.def (SEQUENCE): Likewise.
20186 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20187 Update documentation.
20188 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20189 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20191 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20193 2013-04-08 Teresa Johnson <tejohnson@google.com>
20195 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20196 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20198 (estimate_edge_size_and_time): Add comment to suggest using rounding
20200 (estimate_node_size_and_time): Ditto.
20201 (remap_edge_change_prob): Use helper rounding divide methods.
20202 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20203 (gimple_mod_pow2_value_transform): Ditto.
20204 (gimple_mod_subtract_transform): Ditto.
20205 (gimple_ic_transform): Ditto.
20206 (gimple_stringops_transform): Ditto.
20207 * stmt.c (conditional_probability): Ditto.
20208 (emit_case_dispatch_table): Ditto.
20209 * lto-cgraph.c (merge_profile_summaries): Ditto.
20210 * tree-optimize.c (execute_fixup_cfg): Ditto.
20211 * cfgcleanup.c (try_forward_edges): Ditto.
20212 * cfgloopmanip.c (scale_loop_profile): Ditto.
20214 (duplicate_loop_to_header_edge): Ditto.
20215 (lv_adjust_loop_entry_edge): Ditto.
20216 * tree-vect-loop.c (vect_transform_loop): Ditto.
20217 * profile.c (compute_branch_probabilities): Ditto.
20218 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20219 * lto-streamer-in.c (input_cfg): Ditto.
20220 * gimple-streamer-in.c (input_bb): Ditto.
20221 * ipa-cp.c (update_profiling_info): Ditto.
20222 (update_specialized_profile): Ditto.
20223 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20224 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20226 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20227 (compute_trg_info): Ditto.
20228 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20229 (purge_dead_edges): Ditto.
20230 * loop-unswitch.c (unswitch_loop): Ditto.
20231 * cgraphclones.c (cgraph_clone_edge): Ditto.
20232 (cgraph_clone_node): Ditto.
20233 * tree-inline.c (copy_bb): Ditto.
20234 (copy_edges_for_bb): Ditto.
20235 (initialize_cfun): Ditto.
20236 (copy_cfg_body): Ditto.
20237 (expand_call_inline): Ditto.
20239 2013-04-08 Kai Tietz <ktietz@redhat.com>
20241 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20242 TARGET_CYGWIN64 by TARGET_64BIT.
20244 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20246 * config/epiphany/epiphany.md (GPR_1): New constant.
20247 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20248 * config/epiphany/epiphany.c (gen_compare_reg):
20249 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20250 is already in place.
20251 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20252 Don't require being called during rtl expansion; If y operlaps r0,
20254 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20255 (epiphany_expand_epilogue): Likewise.
20257 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20258 Don't use CC_FPmode for ORDERED / UNORDERED.
20259 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20261 * config/epiphany/constraints.md (CnL): New constraint.
20262 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20263 * config/epiphany/predicates.md (add_operand): Allow 1024.
20265 * config/epiphany/epiphany.md (logical_op): New code iterator.
20266 (op_mnc): New code attribute.
20267 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20268 (mov_f+1, mov_f+2): New peephole2 patterns.
20270 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20271 (cstoresi4): Also allow re-use of zero result when doing a NE
20272 comparison to a non-zero operand.
20273 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20275 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20276 Use gen_addsi3_i / gen_subsi3_i.
20278 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20282 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20283 contain anything but clobbers, at most one __builtin_stack_restore,
20284 optionally debug stmts and final resx, and if it has at least one
20285 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20286 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20287 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20288 which isn't defaut definition, remove them.
20289 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20290 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20291 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20292 with MEM_REF LHS with SSA_NAME address.
20294 2013-04-08 Jeff Law <law@redhat.com>
20296 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20298 2013-04-08 Richard Biener <rguenther@suse.de>
20300 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20302 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20303 determined vector type.
20304 (vect_analyze_data_refs): Likewise.
20305 (vect_get_new_vect_var): Adjust.
20306 (vect_create_destination_var): Preserve SSA name versions.
20307 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20308 not dump anything here.
20310 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20312 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20313 Add member lr_slot_known.
20314 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20316 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20317 Remove code that sets lr_slot_offset according to what a previous
20318 version of epiphany_emit_save_restore used to do.
20319 (epiphany_emit_save_restore): When doing an lr save or restore,
20320 set/verify lr_slot_known and lr_slot_offset.
20322 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20325 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20328 2013-04-08 Richard Biener <rguenther@suse.de>
20330 * alias.c (find_base_term): Fix thinko in previous change.
20332 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20334 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20335 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20336 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20337 if possible to compute val.
20338 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20339 For QImode integers don't require anything about precision. Use
20340 const_with_all_bytes_same to find out if the constant doesn't have
20341 repeated bytes in it.
20343 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20345 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20348 2013-04-08 Marek Polacek <polacek@redhat.com>
20350 PR rtl-optimization/48182
20351 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20354 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20357 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20358 nuses, make sure we have a label.
20360 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20363 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20364 (rs6000_emit_swdiv_low_precision): Remove.
20365 (rs6000_emit_swdiv): Rewrite to handle between one and four
20366 iterations of Newton-Raphson generally; modify required number of
20367 iterations for some cases.
20368 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20370 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20372 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20373 set-but-unused variable.
20375 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20376 basic blocks of released function bodies garbage-collectable.
20378 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20379 (struct rtl_opt_pass): Add TODO_df_finish.
20381 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20383 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20385 * config/arm/constraints.md (q): New constraint.
20386 * config/arm/ldrdstrd.md: New file.
20387 * config/arm/arm.md (ldrdstrd.md) New include.
20388 (arm_movdi): Use "q" instead of "r" constraint
20389 for double-word memory access.
20390 (movdf_soft_insn): Likewise.
20391 * config/arm/vfp.md (movdi_vfp): Likewise.
20392 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20393 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20394 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20395 (mem_ok_for_ldrd_strd): Likewise.
20396 (output_move_double): Update assertion.
20398 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20400 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20402 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20404 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20405 define_insn_and_split.
20406 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20408 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20410 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20411 define_insn_and_split.
20412 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20413 (shiftsi3_compare): New pattern.
20414 (rrx): New pattern.
20415 * config/arm/unspecs.md (UNSPEC_RRX): New.
20417 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20419 * config/arm/arm.md (negdi_extendsidi): New pattern.
20420 (negdi_zero_extendsidi): Likewise.
20422 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20424 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20425 define_insn_and_split.
20426 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20427 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20429 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20431 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20432 define_insn_and_split.
20433 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20434 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20436 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20438 * config/arm/arm.md (subsi3_carryin): New pattern.
20439 (subsi3_carryin_const): Likewise.
20440 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20441 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20443 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20445 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20447 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20449 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20450 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20452 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20454 * config/arm/arm.c (arm_expand_builtin): Change fcode
20455 type to unsigned int.
20457 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20459 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20461 2013-04-04 Ian Lance Taylor <iant@google.com>
20463 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20466 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20468 PR middle-end/56729
20469 * df-scan.c (df_insn_delete): Disable failing assert.
20471 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20473 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20474 New function prototype.
20475 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20476 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20477 (arm_builtin_vectorized_function): New function.
20479 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20481 * config/arm/arm_neon_builtins.def: New file.
20482 * config/arm/arm.c (neon_builtin_data): Move contents to
20483 arm_neon_builtins.def.
20484 (enum arm_builtins): Include neon builtin definitions.
20485 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20486 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20488 2013-04-04 Marek Polacek <polacek@redhat.com>
20490 PR tree-optimization/48186
20491 * predict.c (maybe_hot_frequency_p): Return false if
20492 HOT_BB_FREQUENCY_FRACTION is 0.
20493 (cgraph_maybe_hot_edge_p): Likewise.
20495 2013-04-04 Richard Biener <rguenther@suse.de>
20497 PR tree-optimization/56826
20498 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20501 2013-04-04 Richard Biener <rguenther@suse.de>
20503 PR tree-optimization/56213
20504 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20505 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20507 2013-04-04 Richard Biener <rguenther@suse.de>
20509 PR tree-optimization/56837
20510 * tree-loop-distribution.c (classify_partition): For non-zero
20511 values require that the value has the same precision as its
20512 mode to be useful as memset value.
20514 2013-04-03 Nick Clifton <nickc@redhat.com>
20516 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20517 (fmssf4): Use fmsf.s on E3V5 architectures.
20518 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20519 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20521 2013-04-03 Jeff Law <law@redhat.com>
20523 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20524 (lra-eliminations.o): Likewise.
20526 2013-04-03 Teresa Johnson <tejohnson@google.com>
20528 * gcov-io.c (compute_working_sets): Moved most of body of old
20529 compute_working_sets here from profile.c.
20530 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20531 (gcov_working_set_t): Moved typedef here from basic-block.h
20532 (compute_working_set): Declare.
20533 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20534 (get_working_sets): Renamed from compute_working_set,
20535 replace most of body with call to new compute_working_sets.
20536 (get_exec_counts): Replace call to compute_working_sets
20537 to get_working_sets.
20538 * profile.h (get_working_sets): Renamed from compute_working_set.
20539 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20540 to get_working_sets.
20541 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20542 * gcov-dump.c (dump_working_sets): New function.
20544 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20546 * hwint.c (sext_hwi, zext_hwi): New functions.
20547 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20548 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20549 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20550 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20551 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20552 (sext_hwi, zext_hwi): New functions.
20554 2013-04-03 Jeff Law <law@redhat.com>
20556 PR tree-optimization/56799
20557 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20558 back test for widening conversion erroneously dropped in prior change.
20560 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20563 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20564 instead of next_real_insn.
20566 2013-04-03 Marek Polacek <polacek@redhat.com>
20569 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20571 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20574 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20576 (thumb1_output_casesi): Likewise.
20577 (thumb2_output_casesi): Likewise.
20579 2013-04-03 Richard Biener <rguenther@suse.de>
20581 PR tree-optimization/56817
20582 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20584 (tree_unroll_loops_completely_1): ... new function to manually
20585 walk the loop tree, properly defering outer loops of unrolled
20586 loops to later iterations.
20588 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20590 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20591 (vectorizable_load): Likewise.
20592 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20593 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20595 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20597 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20600 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20602 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20604 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20606 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20608 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20610 PR tree-optimization/56790
20611 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20614 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20616 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20618 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20619 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20622 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20625 * tree.h (force_folding_builtin_constant_p): New decl.
20626 * builtins.c (force_folding_builtin_constant_p): New variable.
20627 (fold_builtin_constant_p): Fold immediately also if
20628 force_folding_builtin_constant_p.
20630 2013-04-03 Richard Biener <rguenther@suse.de>
20632 PR tree-optimization/56812
20633 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20634 DRs of the same interleaving chain are independent.
20636 2013-04-02 Jason Merrill <jason@redhat.com>
20638 * gdbinit.in (pbb): Use debug fn.
20640 2013-04-02 Lawrence Crowl <crowl@google.com>
20642 * sese.h (struct ivtype_map_elt_s): Remove unused.
20643 (extern debug_ivtype_map): Remove unused.
20644 (extern eq_ivtype_map_elts): Remove unused.
20645 * sese.c (debug_ivtype_map): Removed unused.
20646 (debug_ivtype_map_1): Removed unused.
20647 (debug_ivtype_elt): Remove unused.
20648 (eq_ivtype_map_elts): Remove unused.
20651 2013-04-02 Kai Tietz <ktietz@redhat.com>
20654 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20655 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20656 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20658 (legitimize_pe_coff_symbol): Likewise.
20659 (is_imported_p): New helper-function.
20660 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20661 for Windows x64 targets.
20662 (ix86_expand_prologue): Optimize for pe-coff targets.
20663 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20664 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20665 medium/large code-model.
20666 (legitimize_pic_address): Likewise.
20667 (legitimize_tls_address): Likewise.
20668 (ix86_expand_call): Likewise.
20669 (x86_output_mi_thunk): Likewise.
20670 (get_dllimport_decl): Add new beimport argument.
20671 (construct_plt_address): Don't assert for x64 pe-coff targets.
20672 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20674 (SYMBOL_FLAG_STUBVAR): New macro.
20675 (SYMBOL_REF_STUBVAR_P): Likewise.
20676 * config/i386/winnt.c (stub_list): New structure.
20677 (stub_head): New local variable.
20678 (i386_pe_record_stub): New function.
20679 (i386_pe_file_end): Emit refptr-stubs.
20681 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20683 PR rtl-optimization/56745
20684 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20685 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20688 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20689 and both of them are MEM_REFs, just compare first argument for
20690 equality and attempt to deal even with differing offsets.
20693 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20694 of gimple_clobber_p to be MEM_REF.
20695 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20696 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20697 after gimplification.
20698 * asan.c (get_mem_ref_of_assignment): Don't instrument
20699 gimple_clobber_p stmts.
20700 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20701 gimple_clobber_p stmt if they have MEM_REF lhs and
20702 are dead because of another gimple_clobber_p stmt.
20703 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20704 gimple_clobber_p stmts like debug stmts.
20705 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20706 that refer to unused VAR_DECLs or uninitialized values.
20707 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20708 gimple_clobber_p stmts if they refer to removed parameters.
20709 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20712 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20714 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20715 using SWI48 mode attribute.
20717 2013-04-02 Wei Mi <wmi@google.com>
20719 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20720 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20721 *<rotate_insn><mode>3_mask in i386.md.
20723 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20725 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20727 2013-04-02 Richard Biener <rguenther@suse.de>
20729 PR tree-optimization/56778
20730 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20731 Runtime alias tests are not supported for gather loads.
20732 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20733 stmts referenced from SSA operands before updating SSA form.
20735 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20736 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20738 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20739 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20740 * config/arm/cortex-a53.md: New file.
20741 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20742 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20743 * config/arm/arm.c (arm_issue_rate): Likewise.
20744 * config/arm/arm-tune.md: Regenerate
20745 * config/arm/arm-tables.opt: Regenerate.
20746 * config/arm/arm-cores.def: Add cortex-a53.
20748 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20750 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20753 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20755 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20756 scalar load/store operations using B/H registers.
20757 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20759 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20761 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20763 * config/aarch64/aarch64.c
20764 (aarch64_simd_scalar_immediate_valid_for_move): New.
20765 * config/aarch64/aarch64-protos.h
20766 (aarch64_simd_scalar_immediate_valid_for_move): New.
20767 * config/aarch64/constraints.md (Dh, Dq): New.
20768 * config/aarch64/iterators.md (hq): New.
20770 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20772 * reorg.c (get_branch_condition): Deal with conditional returns.
20773 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20775 2013-04-01 Wei Mi <wmi@google.com>
20777 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20778 Truncate operand 2 using %b asm operand modifier.
20779 (*<shift_insn><mode>3_mask): Ditto.
20780 (*<rotate_insn><mode>3_mask): Ditto.
20782 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20784 PR middle-end/56798
20785 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20787 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20789 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20791 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20793 2013-03-30 Lawrence Crowl <crowl@google.com>
20795 * dse.c (clear_alias_sets): Remove never set.
20796 (disqualified_clear_alias_sets): Remove never set.
20797 (clear_alias_mode_pool): Remove never set.
20798 (dse_step0): Remove condition that is never true.
20799 (canon_address): Remove condition that is never true.
20800 (dse_step7): Remove condition that is never true.
20801 (rest_of_handle_dse): Remove condition that is never true.
20802 (rest_of_handle_dse::did_global): Remove never read from above.
20803 (dse_step2_spill): Remove never called from above.
20804 (dse_step5_spill): Remove never called from above.
20806 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20808 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20809 JUMP_TABLE_DATA changes.
20810 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20811 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20812 (Insns) <jump_table_data>: New entry.
20813 * doc/tm.texi: Regenerate.
20815 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20817 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20818 for table jump at the end of a basic block using tablejump_p.
20819 * targhooks.c (default_invalid_within_doloop): Likewise.
20820 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20821 target hook implementation that is identical to the default hook.
20822 (rs6000_invalid_within_doloop): Remove.
20824 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20825 unused variable from tablejump_p call.
20827 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20828 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20829 (INSN_DELETED_P): Likewise.
20830 (emit_jump_table_data): New prototype.
20831 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20832 after 4th as unused.
20833 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20834 * sched-vis.c (print_insn): Likewise.
20835 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20836 insn for compatibility with back ends that use next_active_insn to
20837 identify jump table data.
20838 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20839 (remove_insn): Likewise.
20840 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20842 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20843 (emit_jump_table_data): New function.
20845 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20846 basic block, a JUMP_TABLE_DATA never is.
20847 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20848 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20849 off from code handling real insns.
20850 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20851 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20852 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20853 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20854 is not a NONDEBUG_INSN_P.
20855 * ira-costs.c (scan_one_insn): Likewise.
20856 * jump.c (mark_all_labels): Likewise.
20857 (mark_jump_label_1): Likewise.
20858 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20859 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20861 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20862 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20863 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20864 JUMP_TABLE_DATA_P insns.
20865 (calculate_elim_costs_all_insns): Likewise.
20866 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20867 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20868 (delete_output_reload): Code style fixups.
20869 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20870 insn flags on this non-insn.
20871 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20872 as scheduling barriers, for pre-change compatibility.
20873 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20874 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20876 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20877 redundant JUMP_TABLE_DATA_P test.
20878 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20879 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20880 (frv_for_each_packet): Likewise.
20881 * config/i386/i386.c (min_insn_size): Likewise.
20882 (ix86_avoid_jump_mispredicts): Likewise.
20883 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20884 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20885 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20886 (mips16_insn_length): Robustify.
20887 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20888 (mips16_split_long_branches): Likewise.
20889 * config/pa/pa.c (pa_combine_instructions): Likewise.
20890 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20891 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20892 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20893 as contributing to pool range lengths.
20894 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20895 Remove redundant JUMP_TABLE_DATA_P test.
20896 (sh_loop_align): Likewise.
20897 (split_branches): Likewise.
20898 (sh_insn_length_adjustment): Likewise.
20899 * config/spu/spu.c (get_branch_target): Likewise.
20901 2013-03-29 Jan Hubicka <jh@suse.cz>
20903 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20904 gcov streaming; stream hot bb threshold to ltrans.
20905 * predict.c (get_hot_bb_threshold): Break out from ....
20906 (maybe_hot_count_p): ... here.
20907 (set_hot_bb_threshold): New function.
20908 * lto-section-in.c (lto_section_name): Add profile.
20909 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20910 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20911 and data-streamer.h
20912 (histogram_entry): New structure.
20913 (histogram, histogram_pool): New global vars.
20914 (histogram_hash): New structure.
20915 (histogram_hash::hash): New method.
20916 (histogram_hash::equal): Likewise.
20917 (account_time_size): New function.
20918 (cmp_counts): New function.
20919 (dump_histogram): New function.
20920 (ipa_profile_generate_summary): New function.
20921 (ipa_profile_write_summary): New function.
20922 (ipa_profile_read_summary): New function.
20923 (ipa_profile): Decide on threshold.
20924 (pass_ipa_profile): Add ipa_profile_write_summary and
20925 ipa_profile_read_summary.
20926 * Makefile.in (ipa.o): Update dependencies.
20927 * lto-streamer.h (LTO_section_ipa_profile): New section.
20929 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20931 * tree.h (VAR_P): New.
20933 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20936 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20938 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20940 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20941 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20942 (control_flow_insn_p): Likewise.
20943 * cfgrtl.c (duplicate_insn_chain): Likewise.
20944 * final.c (get_attr_length_1): Likewise.
20945 (shorten_branches): Likewise.
20946 (final_scan_insn): Likewise.
20947 * function.c (instantiate_virtual_regs): Likewise.
20948 * gcse.c (insert_insn_end_basic_block): Likewise.
20949 * ira-costs.c (scan_one_insn): Likewise.
20950 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20951 * lra.c (check_rtl): Likewise.
20952 * reload1.c (elimination_costs_in_insn): Likewise.
20953 * reorg.c (follow_jumps): Likewise.
20955 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20956 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20957 (thumb_far_jump_used_p): Likewise.
20958 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20959 (workaround_speculation): Likewise.
20960 (add_sched_insns_for_speculation): Likewise.
20961 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20962 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20963 (frv_for_each_packet): Likewise.
20964 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20965 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20966 (final_emit_insn_group_barriers): Likewise.
20967 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20968 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20969 (mips16_insn_length): Likewise.
20970 * config/pa/pa.c (pa_reorg): Likewise.
20971 (pa_combine_instructions): Likewise.
20972 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20973 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20974 (sh_reorg): Likewise.
20975 (split_branches): Likewise.
20976 * config/spu/spu.c (get_branch_target): Likewise.
20978 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20981 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20983 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
20984 Fix declaration name.
20986 2013-03-28 Lawrence Crowl <crowl@google.com>
20988 * graphds.h (struct graph.indicies): Remove unused.
20989 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20990 (SCOP_ORIGINAL_PDDRS): Remove unused.
20991 * sese.h (extern insert_loop_close_phis): Removed unused.
20992 (extern insert_guard_phis): Removed unused.
20993 (extern ivtype_map_elt_info): Removed unused.
20994 (new_ivtype_map_elt): Removed unused.
20995 * sese.c (ivtype_map_elt_info): Removed unused.
20997 2013-03-28 Lawrence Crowl <crowl@google.com>
20999 * Makefile.in: Add several missing include dependences.
21001 (test-dump.o): New. This object is not added to any executable,
21002 but is present for ad-hoc testing.
21004 (debug (const bitmap_head_def &)): New.
21005 (debug (const bitmap_head_def *)): New.
21007 (extern debug (const bitmap_head_def &)): New.
21008 (extern debug (const bitmap_head_def *)): New.
21010 (debug (edge_def &)): New.
21011 (debug (edge_def *)): New.
21013 (debug (basic_block_def &)): New.
21014 (debug (basic_block_def *)): New.
21016 (dump_node (const_tree, int, FILE *)): Correct source file.
21018 (debug (die_struct &)): New.
21019 (debug (die_struct *)): New.
21021 (extern debug (die_struct &)): New.
21022 (extern debug (die_struct *)): New.
21023 * gimple-pretty-print.c
21024 (debug (gimple_statement_d &)): New.
21025 (debug (gimple_statement_d *)): New.
21026 * gimple-pretty-print.h
21027 (extern debug (gimple_statement_d &)): New.
21028 (extern debug (gimple_statement_d *)): New.
21030 (debug (ira_allocno_copy &)): New.
21031 (debug (ira_allocno_copy *)): New.
21032 (debug (ira_allocno &)): New.
21033 (debug (ira_allocno *)): New.
21035 (extern debug (ira_allocno_copy &)): New.
21036 (extern debug (ira_allocno_copy *)): New.
21037 (extern debug (ira_allocno &)): New.
21038 (extern debug (ira_allocno *)): New.
21040 (debug (live_range &)): New.
21041 (debug (live_range *)): New.
21043 (debug (lra_live_range &)): New.
21044 (debug (lra_live_range *)): New.
21046 (debug (lra_live_range &)): New.
21047 (debug (lra_live_range *)): New.
21049 (debug (omega_pb_d &)): New.
21050 (debug (omega_pb_d *)): New.
21052 (extern debug (omega_pb_d &)): New.
21053 (extern debug (omega_pb_d *)): New.
21055 (debug (const rtx_def &)): New.
21056 (debug (const rtx_def *)): New.
21058 (debug_tree (tree): Move within file.
21059 (debug_raw (const tree_node &)): New.
21060 (debug_raw (const tree_node *)): New.
21061 (dump_tree_via_hooks (const tree_node *, int)): New.
21062 (debug (const tree_node &)): New.
21063 (debug (const tree_node *)): New.
21064 (debug_verbose (const tree_node &)): New.
21065 (debug_verbose (const tree_node *)): New.
21066 (debug_head (const tree_node &)): New.
21067 (debug_head (const tree_node *)): New.
21068 (debug_body (const tree_node &)): New.
21069 (debug_body (const tree_node *)): New.
21070 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21071 (debug (vec<tree, va_gc> &)): New.
21072 (debug (vec<tree, va_gc> *)): New.
21074 (extern debug (const rtx_def &)): New.
21075 (extern debug (const rtx_def *)): New.
21077 (debug_raw (simple_bitmap_def &)): New.
21078 (debug_raw (simple_bitmap_def *)): New.
21079 (debug (simple_bitmap_def &)): New.
21080 (debug (simple_bitmap_def *)): New.
21082 (extern debug (simple_bitmap_def &)): New.
21083 (extern debug (simple_bitmap_def *)): New.
21084 (extern debug_raw (simple_bitmap_def &)): New.
21085 (extern debug_raw (simple_bitmap_def *)): New.
21087 (debug (vinsn_def &)): New.
21088 (debug (vinsn_def *)): New.
21089 (debug_verbose (vinsn_def &)): New.
21090 (debug_verbose (vinsn_def *)): New.
21091 (debug (expr_def &)): New.
21092 (debug (expr_def *)): New.
21093 (debug_verbose (expr_def &)): New.
21094 (debug_verbose (expr_def *)): New.
21095 (debug (vec<rtx> &)): New.
21096 (debug (vec<rtx> *)): New.
21098 (extern debug (vinsn_def &)): New.
21099 (extern debug (vinsn_def *)): New.
21100 (extern debug_verbose (vinsn_def &)): New.
21101 (extern debug_verbose (vinsn_def *)): New.
21102 (extern debug (expr_def &)): New.
21103 (extern debug (expr_def *)): New.
21104 (extern debug_verbose (expr_def &)): New.
21105 (extern debug_verbose (expr_def *)): New.
21106 (extern debug (vec<rtx> &)): New.
21107 (extern debug (vec<rtx> *)): New.
21109 (_list_iter_cond_expr): Make inline instead of static.
21111 (debug (sreal &)): New.
21112 (debug (sreal *)): New.
21114 (extern debug (sreal &)): New.
21115 (extern debug (sreal *)): New.
21117 (extern debug_raw (const tree_node &)): New.
21118 (extern debug_raw (const tree_node *)): New.
21119 (extern debug (const tree_node &)): New.
21120 (extern debug (const tree_node *)): New.
21121 (extern debug_verbose (const tree_node &)): New.
21122 (extern debug_verbose (const tree_node *)): New.
21123 (extern debug_head (const tree_node &)): New.
21124 (extern debug_head (const tree_node *)): New.
21125 (extern debug_body (const tree_node &)): New.
21126 (extern debug_body (const tree_node *)): New.
21127 (extern debug (vec<tree, va_gc> &)): New.
21128 (extern debug (vec<tree, va_gc> *)): New.
21130 (debug (struct loop &)): New.
21131 (debug (struct loop *)): New.
21132 (debug_verbose (struct loop &)): New.
21133 (debug_verbose (struct loop *)): New.
21134 * tree-dump.c: Add header dependence.
21136 (extern debug (struct loop &)): New.
21137 (extern debug (struct loop *)): New.
21138 (extern debug_verbose (struct loop &)): New.
21139 (extern debug_verbose (struct loop *)): New.
21141 (debug (data_reference &)): New.
21142 (debug (data_reference *)): New.
21143 (debug (vec<data_reference_p> &)): New.
21144 (debug (vec<data_reference_p> *)): New.
21145 (debug (vec<ddr_p> &)): New.
21146 (debug (vec<ddr_p> *)): New.
21148 (extern debug (data_reference &)): New.
21149 (extern debug (data_reference *)): New.
21150 (extern debug (vec<data_reference_p> &)): New.
21151 (extern debug (vec<data_reference_p> *)): New.
21152 (extern debug (vec<ddr_p> &)): New.
21153 (extern debug (vec<ddr_p> *)): New.
21155 (debug (pt_solution &)): New.
21156 (debug (pt_solution *)): New.
21158 (extern debug (pt_solution &)): New.
21159 (extern debug (pt_solution *)): New.
21161 (debug (_var_map &)): New.
21162 (debug (_var_map *)): New.
21163 (debug (tree_live_info_d &)): New.
21164 (debug (tree_live_info_d *)): New.
21166 (extern debug (_var_map &)): New.
21167 (extern debug (_var_map *)): New.
21168 (extern debug (tree_live_info_d &)): New.
21169 (extern debug (tree_live_info_d *)): New.
21171 2013-03-28 Jan Hubicka <jh@suse.cz>
21173 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21175 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21177 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21178 record only when desired or required.
21180 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21182 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21183 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21185 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21187 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21188 (*andsi3_compare0_uxtw): New pattern.
21189 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21190 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21192 2013-03-28 Jan Hubicka <jh@suse.cz>
21194 * data-streamer-in.c (streamer_read_gcov_count): New function.
21195 * gimple-streamer-out.c: Include value-prof.h.
21196 (output_gimple_stmt): Output histogram.
21197 (output_bb): Use streamer_write_gcov_count.
21198 * value-prof.c: Include data-streamer.h
21199 (dump_histogram_value): Add HIST_TYPE_MAX.
21200 (stream_out_histogram_value): New function.
21201 (stream_in_histogram_value): New function.
21202 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21203 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21204 * data-streamer-out.c (streamer_write_gcov_count): New function.
21205 (streamer_write_gcov_count_stream): New function.
21206 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21207 (lto_output_node): Likewise.
21208 (input_node, input_edge): Likewise.
21209 * lto-streamer-out.c (output_cfg): Update streaming.
21210 * lto-streamer-in.c (input_cfg): Likewise.
21211 * data-streamer.h (streamer_write_gcov_count,
21212 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21213 * gimple-streamer-in.c: Include value-prof.h
21214 (input_gimple_stmt): Input histograms.
21215 (input_bb): Update profile streaming.
21217 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21219 * genmodes.c (emit_max_int): New function.
21220 (emit_insn_modes_h): Added call to emit_max_function.
21221 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21223 * machmode.def: Fixed comment.
21225 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21227 * combine.c (try_combine): Removed useless assert.
21228 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21230 2013-03-28 Marek Polacek <polacek@redhat.com>
21231 Richard Biener <rguenther@suse.de>
21233 PR tree-optimization/56695
21234 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21235 build signed result of a vector comparison.
21236 * tree-cfg.c (verify_gimple_comparison): Check that a result
21237 of a vector comparison has signed type.
21239 2013-03-28 Richard Biener <rguenther@suse.de>
21241 PR tree-optimization/37021
21242 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21243 do not restrict gaps between groups.
21244 * tree-vect-stmts.c (vectorizable_load): Properly account for
21245 a gap between groups.
21247 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21249 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21250 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21253 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21255 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21256 * doc/extend.texi (Named Address Spaces): Ditto.
21257 (Variable Attributes): Ditto.
21259 2013-03-27 Kai Tietz <ktietz@redhat.com>
21261 * config.build: Add support for cygwin x64 target.
21262 * config.gcc: Likewise.
21263 * config.host: Likewise.
21264 * configure.ac: Likewise
21265 * configure: Regenerated.
21267 2013-03-27 Kai Tietz <ktietz@redhat.com>
21269 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21270 * config/i386/t-cygwin-w64: New file.
21271 * config/i386/cygwin-w64.h: New file.
21272 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21273 and add support for x64-cygwin target.
21274 (CPP_SPEC): Likewise.
21275 (CXX_WRAP_SPEC_LIST): Undefine before define.
21276 (LIBGCJ_SONAME): Use 15 as version.
21278 2013-03-27 Richard Biener <rguenther@suse.de>
21280 PR tree-optimization/56716
21281 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21282 dumping for ref nodes.
21284 2013-03-27 Martin Jambor <mjambor@suse.cz>
21286 PR tree-optimization/55334
21287 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21288 restricted pointers to arrays.
21290 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21292 * Makefile.in (.SUFFIXES): Add .cc.
21293 (.c.o): Apply same recipe for implicit rule .cc.o.
21295 2013-03-27 Richard Biener <rguenther@suse.de>
21297 PR tree-optimization/37021
21298 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21299 REALPART/IMAGPART_EXPRs around the supported refs.
21300 * tree-ssa-structalias.c (find_func_aliases): Assume that
21301 floating-point values are not used to transfer pointers.
21303 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21305 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21306 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21307 * doc/tm.texi: Regenerate.
21308 * targhooks.h (default_has_ifunc_p): New.
21309 * targhooks.c (default_has_ifunc_p): Ditto.
21310 * config/linux-protos.h: New file.
21311 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21312 hook for linux which disables support of indirect functions in android.
21313 * config/linux-android.c: New file.
21314 * config/t-linux-android.c: Ditto.
21315 * config.gcc: Added new object file linux-android.o.
21316 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21317 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21318 * varasm.c (do_assemble_alias): Likewise.
21319 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21320 doesn't support indirect functions.
21321 * configure: Regenerate.
21323 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21326 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21327 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21330 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21332 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21334 2013-03-27 Terry Guo <terry.guo@arm.com>
21336 * config/arm/arm-cores.def: Added core cortex-r7.
21337 * config/arm/arm-tune.md: Regenerated.
21338 * config/arm/arm-tables.opt: Regenerated.
21339 * doc/invoke.texi: Added entry for core cortex-r7.
21341 2013-03-27 Walter Lee <walt@tilera.com>
21343 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21344 double-decrement of next_scratch_regno.
21346 2013-03-27 Walter Lee <walt@tilera.com>
21348 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21350 (insn_v1mulus): Ditto.
21351 (insn_v2muls): Ditto.
21353 2013-03-27 Walter Lee <walt@tilera.com>
21355 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21356 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21358 2013-03-27 Walter Lee <walt@tilera.com>
21360 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21361 (*sibcall_value): Ditto.
21363 2013-03-27 Walter Lee <walt@tilera.com>
21365 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21366 (insn_mnz_v8qi): ... this ...
21367 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21369 (insn_v<n>mnz): Replaced by ...
21370 (insn_v1mnz): ... this ...
21371 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21373 (insn_mz_<mode>): Replaced by ...
21374 (insn_mz_v8qi): ... this ...
21375 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21377 (insn_v<n>mz): Replaced by ...
21378 (insn_v1mz): ... this ...
21379 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21382 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21384 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21386 2013-03-26 Roland McGrath <mcgrathr@google.com>
21388 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21389 than fprintf with a non-constant, non-format string.
21391 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21393 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21394 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21395 operand 0 predicate.
21396 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21397 attribute. Use general_x64nomem_operand as operand 1 predicate.
21398 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21399 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21400 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21401 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21402 (mov<mode>_insv_1): Remove expander. Merge insn with
21403 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21404 Use general_x64nomem_operand as operand 1 predicate.
21405 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21406 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21407 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21408 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21409 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21410 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21411 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21412 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21413 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21414 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21415 (general_x64nomem_operand): Ditto.
21417 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21419 * config/rtems.opt: Add -pthread option.
21421 2013-03-26 Richard Biener <rguenther@suse.de>
21423 * alias.c (find_base_term): Avoid redundant and not used recursion.
21424 (base_alias_check): Get the initial base term from the caller.
21425 (true_dependence_1): Compute and pass base terms to base_alias_check.
21426 (write_dependence_p): Likewise.
21427 (may_alias_p): Likewise.
21429 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21431 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21432 PC-relative load in SI modes and above only.
21434 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21436 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21437 * config/arm/iwmmxt.md (WCGR0): Update.
21438 (WCGR1, WCGR2, WCGR3): Likewise.
21440 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21442 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21443 Use x64 and nox64 isa attributes.
21445 2013-03-26 Richard Biener <rguenther@suse.de>
21447 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21448 alignment computations and rely on get_object_alignment_1
21449 for the !TYPE_P case.
21450 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21452 2013-03-26 Walter Lee <walt@tilera.com>
21454 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21455 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21457 2013-03-25 Jeff Law <law@redhat.com>
21459 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21460 check for INTEGRAL_TYPE_P that was missing due to checking in
21461 wrong version of prior patch.
21463 2013-03-25 Walter Lee <walt@tilera.com>
21465 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21466 TILEGX_INSN_SHUFFLEBYTES1.
21467 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21469 (tilegx_builtins): Ditto.
21470 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21472 2013-03-25 Walter Lee <walt@tilera.com>
21474 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21475 (floatunssisf2): New pattern.
21476 (floatsidf2): New pattern.
21477 (floatunssidf2): New pattern.
21479 2013-03-25 Walter Lee <walt@tilera.com>
21481 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21482 tests for constraint J, K, N, P.
21484 2013-03-25 Walter Lee <walt@tilera.com>
21486 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21487 Use indirect/pcrel encoding.
21488 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21491 2013-03-25 Steve Ellcey <sellcey@mips.com>
21493 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21494 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21495 * config/mips/mips.c (mips_option_override): Set IMADD default.
21496 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21497 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21498 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21499 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21500 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21502 2013-03-25 Jeff Law <law@redhat.com>
21504 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21505 slightly to avoid creating and folding useless trees. Simplify
21506 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21508 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21510 * config/i386/i386.md (*zero_extendsidi2): Merge with
21511 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21512 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21513 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21514 targets. Clarify comment.
21516 2013-03-25 Martin Jambor <mjambor@suse.cz>
21518 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21519 pass-through jump functions differently.
21520 (ipa_read_jump_function): Likewise. Also use setter functions to set
21523 2013-03-25 Martin Jambor <mjambor@suse.cz>
21525 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21526 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21528 (ipa_get_indirect_edge_target): New function.
21529 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21530 ipa_get_indirect_edge_target. Update all callers.
21531 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21532 ipa_get_indirect_edge_target_1 instead of calling
21533 ipa_get_indirect_edge_target.
21534 (create_specialized_node): Pass aggvlas to
21535 ipcp_discover_new_direct_edges.
21537 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21539 * config/arm/arm.md (f_sels, f_seld): New types.
21540 (*cmov<mode>): New pattern.
21541 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21544 2013-03-25 Kai Tietz <ktietz@redhat.com>
21546 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21547 POSIX-printf for mingw-hosted builds.
21549 2013-03-25 Richard Biener <rguenther@suse.de>
21551 PR middle-end/56694
21552 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21553 must-not-throw stmt location.
21555 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21557 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21558 Emit load-acquire versions when acq is true.
21559 (arm_emit_store_exclusive): Add rel parameter.
21560 Emit store-release versions when rel is true.
21561 (arm_split_compare_and_swap): Use acquire-release instructions
21563 of barriers when appropriate.
21564 (arm_split_atomic_op): Likewise.
21565 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21566 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21567 (VUNSPEC_SLX): Likewise.
21568 (VUNSPEC_LDA): Likewise.
21569 (VUNSPEC_STL): Likewise.
21570 * config/arm/sync.md (atomic_load<mode>): New pattern.
21571 (atomic_store<mode>): Likewise.
21572 (arm_load_acquire_exclusive<mode>): Likewise.
21573 (arm_load_acquire_exclusivesi): Likewise.
21574 (arm_load_acquire_exclusivedi): Likewise.
21575 (arm_store_release_exclusive<mode>): Likewise.
21577 2013-03-25 Catherine Moore <clm@codesourcery.com>
21579 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21580 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21581 * config/mip/predicates.md (lwsp_swsp_operand,
21582 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21583 sb16_operand, db4_operand, db7_operand, ib3_operand,
21584 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21585 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21586 andi16_operand): New predicates.
21587 * config/mips/mips.md (compression): New attribute.
21588 (enabled): New attribute.
21589 (length): Consider compression in computing length.
21590 (shift_compression): New code attribute.
21591 (*add<mode>3): New operands. Record compression.
21592 (sub<mode>3): Likewise.
21593 (one_cmpl<mode>2): Likewise.
21594 (*and<mode>3): Likewise.
21595 (*ior<mode>3): Likewise.
21596 (unnamed pattern for xor): Likewise.
21597 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21598 (*<optab><mode>3): Likewise.
21599 (*mov<mode>_internal: Likewise.
21600 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21601 (mips_unsigned_immediate_p): New.
21602 (umips_lwsp_swsp_address_p): New.
21603 (m16_based_address_p): New.
21604 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21605 (mips_unsigned_immediate_p): New prototype.
21606 (lwsp_swsp_address_p): New prototype.
21607 (m16_based_address_p): New prototype.
21608 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21609 (mips_signed_immediate_p): New function.
21610 (m16_based_address_p): New function.
21611 (lwsp_swsp_address_p): New function.
21612 (mips_print_operand_punctuation): Recognize short delay slot insns
21613 for microMIPS.add<mode>3"
21615 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21618 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21619 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21621 2013-03-25 Richard Biener <rguenther@suse.de>
21623 PR tree-optimization/56689
21624 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21627 2013-03-25 Richard Biener <rguenther@suse.de>
21629 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21631 (memory_references): Likewise.
21632 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21633 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21634 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21635 (gather_mem_refs_in_loops): Fold into ...
21636 (analyze_memory_references): ... this. Move initialization
21637 to tree_ssa_lim_initialize.
21638 (fill_always_executed_in): Rename to ...
21639 (fill_always_executed_in_1): ... this.
21640 (fill_always_executed_in): Move contains_call computation to
21641 this new function from ...
21642 (tree_ssa_lim_initialize): ... here.
21643 (tree_ssa_lim): Call fill_always_executed_in.
21645 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21647 * postreload.c (reload_combine): Fix code detecting returns.
21649 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21651 * function.c (emit_use_return_register_into_block): On cc0 targets,
21652 do not emit the sequence between cc0 setter and user.
21654 2013-03-25 Kai Tietz <ktietz@redhat.com>
21656 * config/i386/predicates.md (local_symbolic_operand): Interpret
21657 dll-imported symbols as none-local.
21659 2013-03-25 Richard Biener <rguenther@suse.de>
21661 * tree-ssa-loop-im.c (struct depend): Remove.
21662 (struct lim_aux_data): Make depends a vec of gimples.
21663 (free_lim_aux_data): Adjust.
21664 (add_dependency): Likewise.
21665 (set_level): Likewise.
21667 2013-03-25 Richard Biener <rguenther@suse.de>
21669 PR middle-end/56434
21670 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21671 the pointer returned by calls with ECF_MALLOC set.
21673 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21675 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21677 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21679 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21680 using MMXMODE mode iterator.
21681 (*move<mode>_internal): Merge with *movv2sf_internal and
21682 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21684 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21686 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21687 (record_last_mem_set_info): Likewise.
21689 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21690 of XNEWVEC followed by memset.
21691 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21693 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21695 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21696 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21697 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21698 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21699 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21700 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21701 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21702 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21703 BARRIER_P instead of GET_CODE.
21705 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21707 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21708 inaccuracy in the probing code.
21710 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21711 (ctrapdi4): Likewise.
21713 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21715 * calls.c (expand_call): Add missing guard to code handling return
21716 of non-BLKmode structures in MSB.
21717 * function.c (expand_function_end): Likewise.
21719 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21721 * combine.c (try_combine): Adjust comment. Do not add the set of
21722 insn #0 if the destination indirectly is set or dies in insn #2.
21723 Tidy up code to distribute a new note.
21725 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21727 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21728 also for alternatives 16 and 17.
21730 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21732 * config/i386/sse.md (*mov<mode>_internal): Merge with
21733 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21734 Emit insn template depending on type attribute. Use
21735 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21736 movd instead of movq mnemonic for interunit moves. Rewrite mode
21737 attribute calculation. Remove unit attribute calculation.
21738 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21739 Set prefix_data16 attribute for DImode ssemov types.
21740 Use Ym instead of y for SSE-MMX conversion alternatives.
21741 Reorder operand constraints.
21743 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21745 * df.h (df_insn_delete): Adjust prototype.
21746 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21747 and let it decide whether mark the basic block dirty.
21748 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21749 * df-scan.c (df_insn_info_delete): New helper function, split
21750 off from df_insn_delete.
21751 (df_scan_free_bb_info): Use it.
21752 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21754 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21755 that the insn is actually an insn and it has a non-NULL basic block.
21756 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21758 2013-03-22 Richard Biener <rguenther@suse.de>
21760 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21762 (mem_ref_alloc): Do not allocate them.
21763 (refs_independent_p): Do not query or maintain a cache.
21765 2013-03-22 Richard Biener <rguenther@suse.de>
21767 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21768 (gather_mem_refs_in_loops): Do not compute it.
21769 (analyze_memory_references): Do not allocate it.
21770 (tree_ssa_lim_finalize): Do not free it.
21771 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21773 2013-03-22 Richard Biener <rguenther@suse.de>
21775 * is-a.h (as_a): Use gcc_checking_assert.
21777 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21779 * config/aarch64/aarch64.c (aarch64_print_operand): New
21780 format specifier for printing a constant in hex.
21781 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21782 format specifier for printing second operand.
21784 2013-03-22 Richard Biener <rguenther@suse.de>
21786 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21788 (gather_mem_refs_in_loops): Perform store accumulation here.
21789 (create_vop_ref_mapping_loop): Remove.
21790 (create_vop_ref_mapping): Likewise.
21791 (analyze_memory_references): Initialize refs_stored_in_loop.
21792 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21793 (record_indep_loop): Remove.
21794 (record_dep_loop): New function.
21795 (ref_indep_loop_p_1): Adjust to only walk over references
21796 in the loop, not its subloops.
21797 (ref_indep_loop_p): Rename to ...
21798 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21799 maintaining a more fine-grained cache.
21800 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21801 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21803 2013-03-22 Richard Biener <rguenther@suse.de>
21805 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21806 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21807 aggregate mem_ref_loc.
21808 (free_mem_ref_locs): Inline into ...
21809 (memref_free): ... this and adjust.
21810 (mem_ref_alloc): Adjust.
21811 (mem_ref_locs_alloc): Remove.
21812 (record_mem_ref_loc): Adjust.
21813 (get_all_locs_in_loop): Rewrite into ...
21814 (for_all_locs_in_loop): ... this iterator.
21815 (rewrite_mem_ref_loc): New functor.
21816 (rewrite_mem_refs): Use for_all_locs_in_loop.
21817 (sm_set_flag_if_changed): New functor.
21818 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21819 (ref_always_accessed): New functor.
21820 (ref_always_accessed_p): Use for_all_locs_in_loop.
21822 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21824 * tree-pass.h (PROP_gimple_lvec): New.
21825 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21826 (init_optimization_passes): Move pass_lower_vector.
21827 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21829 (pass_lower_vector): Provide PROP_gimple_lvec.
21830 (pass_lower_vector_ssa): Likewise.
21831 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21833 2013-03-21 Mark Wielaard <mjw@redhat.com>
21835 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21837 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21839 * config/i386/i386.md (*movdi_internal): Disparage slightly
21840 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21841 conversion alternatives.
21843 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21845 PR middle-end/48087
21846 * diagnostic.def (DK_WERROR): New kind.
21847 * diagnostic.h (werrorcount): Define.
21848 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21849 promoted to DK_ERROR, increment DK_WERROR counter instead of
21851 * toplev.c (toplev_main): Call print_ignored_options even if
21852 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21853 even if just werrorcount is non-zero.
21856 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21858 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21859 (string_cst_pool_decl): New function.
21860 (optimize_one_addr_into_implicit_ptr): New function.
21861 (resolve_addr_in_expr): Optimize DWARF location expression
21862 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21863 which doesn't live in memory, but has DW_AT_location or
21864 DW_AT_const_value, or refers to a string literal, into
21865 DW_OP_GNU_implicit_pointer.
21866 (optimize_location_into_implicit_ptr): New function.
21867 (resolve_addr): If removing DW_AT_location of a variable because
21868 it was DW_OP_addr of address of the variable, but the variable doesn't
21869 live in memory, try to emit const value attribute for the initializer.
21871 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21873 * tree.h (VECTOR_TYPE_P): New macro.
21874 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21875 TYPE_MODE): Use it.
21876 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21877 VEC_COND_EXPR cannot be lvalues.
21878 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21880 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21882 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21883 Restrict the transformation to equal modes.
21885 2013-03-21 Richard Biener <rguenther@suse.de>
21887 PR tree-optimization/39326
21888 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21889 (MEM_ANALYZABLE): Adjust.
21890 (record_mem_ref_loc): Move bitmap ops ...
21891 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21892 unanalyzable refs, do not record locations for it.
21893 (analyze_memory_references): Allocate ref zero as shared
21895 (refs_independent_p): Do not test for unanalyzed mems here.
21896 (ref_indep_loop_p_1): Special-case disambiguation against
21897 the unanalyzed ref.
21898 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
21900 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21902 * config/arm/arm-protos.h (tune_params): Add
21903 prefer_neon_for_64bits field.
21904 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21905 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21906 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21907 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21908 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21909 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21910 (arm_option_override): Handle -mneon-for-64bits new option.
21911 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21912 (prefer_neon_for_64bits): Declare new variable.
21913 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
21914 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
21915 (arch_enabled): Handle new arch types. Remove support for onlya8
21917 (one_cmpldi2): Use new arch names.
21918 (zero_extend<mode>di2, extend<mode>di2): Ditto.
21919 * config/arm/arm.opt (mneon-for-64bits): Add option.
21920 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21921 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21922 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21924 * doc/invoke.texi (-mneon-for-64bits): Document.
21926 2013-03-21 Richard Biener <rguenther@suse.de>
21928 PR tree-optimization/39326
21929 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21930 (sort_bbs_in_loop_postorder_cmp): New function.
21931 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
21933 2013-03-21 Richard Biener <rguenther@suse.de>
21935 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21936 (vect_insert_into_interleaving_chain): Likewise.
21937 (vect_drs_dependent_in_basic_block): Inline ...
21938 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21940 (vect_analyze_data_ref_dependence): ... here. Simplify.
21941 (vect_check_interleaving): Simplify.
21942 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21943 (vect_slp_analyze_data_ref_dependences): ... this new function.
21944 (dr_group_sort_cmp): New function.
21945 (vect_analyze_data_ref_accesses): Compute data-reference groups
21946 here instead of in vect_analyze_data_ref_dependence. Use
21947 a more efficient algorithm.
21948 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21949 vect_slp_analyze_data_ref_dependences. Call
21950 vect_analyze_data_ref_accesses earlier.
21951 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21952 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21953 (vect_slp_analyze_data_ref_dependences): New prototype.
21955 2013-03-21 Richard Biener <rguenther@suse.de>
21957 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21958 ref is stored in the loop.
21959 (find_refs_for_sm): Walk only over all stores.
21960 (store_motion_loop): Allocate from lim_bitmap_obstack.
21961 (store_motion): Likewise.
21963 2013-03-21 Richard Biener <rguenther@suse.de>
21965 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21966 Update virtual SSA form.
21968 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21970 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21971 * configure: Regenerate.
21972 * config.in: Regenerate.
21973 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21974 if !HAVE_LD_EH_FRAME_CIEV3.
21976 2013-03-21 Richard Biener <rguenther@suse.de>
21978 * tree-cfg.c (verify_expr_no_block): New function.
21979 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21980 nor DECL_VALUE_EXPR have locations with associated blocks.
21981 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21982 (clear_unused_block_pointer): Remove code dealing with
21983 blocks in DECL_DEBUG_EXPR locations.
21985 2013-03-21 Richard Biener <rguenther@suse.de>
21987 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21988 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21989 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21990 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21991 instead of DECL_DEBUG_EXPR_IS_FROM.
21992 * gimplify.c (gimplify_modify_expr): Likewise.
21993 * tree-cfg.c (verify_expr_location_1): Likewise.
21994 * tree-complex.c (create_one_component_var): Likewise.
21995 * tree-sra.c (create_access_replacement): Likewise.
21996 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21997 (clear_unused_block_pointer): Likewise.
21998 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21999 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22000 * var-tracking.c (var_debug_decl): Likewise.
22001 (track_expr_p): Likewise.
22002 * tree-inline.c (add_local_variables): Likewise. Set
22003 DECL_HAS_DEBUG_EXPR_P after copying it.
22004 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22005 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22007 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22010 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22011 * configure: Regenerate.
22012 * config.in: Regenerate.
22013 * config/i386/i386.md (*movdf_internal): Use
22014 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22015 movd instead of movq mnemonic for interunit moves.
22016 (*movdi_internal): Ditto.
22018 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22020 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22021 (abd<mode>_3): New pattern.
22022 (aba<mode>_3): New pattern.
22023 (fabd<mode>_3): New pattern.
22025 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22027 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22028 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22029 occurrence of REGISTER_PREFIX as its empty string.
22031 2013-03-20 Jeff Law <law@redhat.com>
22033 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22034 addititional equivalences for equality comparisons between an SSA_NAME
22035 and a constant where the SSA_NAME was set from a widening conversion.
22037 2013-03-20 Walter Lee <walt@tilera.com>
22039 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22041 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22043 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22044 depending on type attribute.
22045 (*movti_internal): Ditto.
22046 (*movtf_internal): Ditto.
22047 (*movxf_internal): Ditto.
22048 (*movdf_internal): Ditto.
22049 (*movsf_internal): Ditto.
22051 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22053 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22054 maybe_vex for sselog1 and ssemov types.
22055 (*movdi_internal): Reorder operand constraints.
22056 (*movsi_internal): Ditto. Set prefix attribute to
22057 maybe_vex for sselog1 and ssemov types.
22058 (*movtf_internal): Set prefix attribute to maybe_vex
22059 for sselog1 and ssemov types.
22060 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22061 DImode ssemov types. Reorder operand constraints.
22062 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22063 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22064 attribute for SImode ssemov types. Reorder operand constraints.
22066 2013-03-20 Martin Jambor <mjambor@suse.cz>
22068 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22069 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22071 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22073 * config/rs6000/predicates.md (indexed_address, update_address_mem
22074 update_indexed_address_mem): New predicates.
22075 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22076 attribute for load/store instructions.
22077 * config/rs6000/dfp.md (movsd_store): Likewise.
22078 (movsd_load): Likewise.
22079 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22080 (unnamed HI->DI extend define_insn): Likewise.
22081 (unnamed SI->DI extend define_insn): Likewise.
22082 (unnamed QI->SI extend define_insn): Likewise.
22083 (unnamed QI->HI extend define_insn): Likewise.
22084 (unnamed HI->SI extend define_insn): Likewise.
22085 (unnamed HI->SI extend define_insn): Likewise.
22086 (extendsfdf2_fpr): Likewise.
22087 (movsi_internal1): Likewise.
22088 (movsi_internal1_single): Likewise.
22089 (movhi_internal): Likewise.
22090 (movqi_internal): Likewise.
22091 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22092 attribute for load/store instructions.
22093 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22095 (mov<mode>_softfloat): Likewise.
22096 (mov<mode>_hardfloat32): Likewise.
22097 (mov<mode>_hardfloat64): Likewise.
22098 (mov<mode>_softfloat64): Likewise.
22099 (movdi_internal32): Likewise.
22100 (movdi_internal64): Likewise.
22101 (probe_stack_<mode>): Likewise.
22103 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22105 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22106 floating point, and decimal floating point to reload iterator.
22108 * config/rs6000/constraints.md (wl constraint): New constraints to
22109 return FLOAT_REGS if certain options are used to reduce the number
22110 of separate patterns that exist in the file.
22111 (wx constraint): Likewise.
22112 (wz constraint): Likewise.
22114 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22115 -mdebug=reg, print wg, wl, wx, and wz constraints.
22116 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22117 Initialize the reload functions for 64-bit binary/decimal floating
22119 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22120 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22121 create the buffer on the stack to overcome not having a 32-bit
22123 (rs6000_emit_move): Likewise.
22124 (rs6000_secondary_memory_needed_rtx): Likewise.
22125 (rs6000_alloc_sdmode_stack_slot): Likewise.
22126 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22127 via xxlxor, just like DFmode 0.0.
22129 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22130 define as 1 if we are running on a power7 or newer.
22131 (enum r6000_reg_class_enum): Add new constraints.
22133 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22134 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22135 with other moves by using conditional constraits (wg). Use LFIWZX
22136 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22137 (movsd splitter): Likewise.
22138 (movsd_hardfloat): Likewise.
22139 (movsd_softfloat): Likewise.
22141 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22142 binary and decimal floating point moves.
22143 (fmove_ok): New attributes to combine binary and decimal floating
22144 point moves, and to combine power6x (mfpgpr) moves along normal
22146 (real_value_to_target): Likewise.
22147 (f32_lr): Likewise.
22148 (f32_lm): Likewise.
22149 (f32_li): Likewise.
22150 (f32_sr): Likewise.
22151 (f32_sm): Likewise.
22152 (f32_si): Likewise.
22153 (movsf): Combine binary and decimal floating point moves. Combine
22154 power6x (mfpgpr) moves with other moves by using conditional
22155 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22156 (mov<mode> for SFmode/SDmode); Likewise.
22157 (SFmode/SDmode splitters): Likewise.
22158 (movsf_hardfloat): Likewise.
22159 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22160 (movsf_softfloat): Likewise.
22161 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22163 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22164 wx and wz constraints.
22166 * config/rs6000/constraints.md (wg constraint): New constraint to
22167 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22169 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22172 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22173 -mdebug=reg, print wg, wl, wx, and wz constraints.
22174 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22175 Initialize the reload functions for 64-bit binary/decimal floating
22177 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22178 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22179 create the buffer on the stack to overcome not having a 32-bit
22181 (rs6000_emit_move): Likewise.
22182 (rs6000_secondary_memory_needed_rtx): Likewise.
22183 (rs6000_alloc_sdmode_stack_slot): Likewise.
22184 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22185 via xxlxor, just like DFmode 0.0.
22187 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22188 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22189 with other moves by using conditional constraits (wg). Use LFIWZX
22190 and STFIWX for loading SDmode on power7.
22191 (movdd splitters): Likewise.
22192 (movdd_hardfloat32): Likewise.
22193 (movdd_softfloat32): Likewise.
22194 (movdd_hardfloat64_mfpgpr): Likewise.
22195 (movdd_hardfloat64): Likewise.
22196 (movdd_softfloat64): Likewise.
22198 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22199 64-bit binary and decimal floating point moves.
22200 (FMOVE64X): Likewise.
22201 (movdf): Combine 64-bit binary and decimal floating point moves.
22202 Combine power6x (mfpgpr) moves with other moves by using
22203 conditional constraits (wg).
22204 (mov<mode> for DFmode/DDmode): Likewise.
22205 (DFmode/DDmode splitters): Likewise.
22206 (movdf_hardfloat32): Likewise.
22207 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22208 (movdf_softfloat32): Likewise.
22209 (movdf_hardfloat64_mfpgpr): Likewise.
22210 (movdf_hardfloat64): Likewise.
22211 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22212 (movdf_softfloat64): Likewise.
22213 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22214 (reload_<mode>_load): Move to later in the file so they aren't in
22215 the middle of the floating point move insns.
22216 (reload_<mode>_store): Likewise.
22218 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22221 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22222 constraint if -mdebug=reg.
22223 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22224 Enable using dd reload support if needed.
22226 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22227 binary and decimal floating point moves in rs6000.md.
22228 (movtd_internal): Likewise.
22230 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22231 decimal floating point moves.
22233 (movtf_internal): Likewise.
22234 (mov<mode>_internal, TDmode/TFmode): Likewise.
22235 (movtf_softfloat): Likewise.
22236 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22238 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22239 movdi_internal64, using wg constraint for move direct operations.
22240 (movdi_internal64): Likewise.
22242 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22243 MODES_TIEABLE_P for selected modes. Print the numerical value of
22244 the various virtual registers. Use GPR/FPR first/last values,
22245 instead of hard coding the register numbers. Print which modes
22246 have reload functions registered.
22247 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22248 settings before/after setting cpu, target and subtarget settings.
22249 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22250 and for secondary reload failures in rs6000_secondary_reload_inner.
22251 (rs6000_secondary_reload_fail): Likewise.
22252 (rs6000_secondary_reload_inner): Likewise.
22254 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22255 macros for first/last GPR and FPR registers.
22256 (LAST_GPR_REGNO): Likewise.
22257 (FIRST_FPR_REGNO): Likewise.
22258 (LAST_FPR_REGNO): Likewise.
22260 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22261 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22262 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22263 (vcond<mode><mode>): Likewise.
22264 (vcondu<mode><mode>): Likewise.
22265 (vector_gtu<mode>): Likewise.
22266 (vector_gte<mode>): Likewise.
22267 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22268 to prevent the compiler from converting DImode operations to TImode.
22269 (ior<mode>3): Likewise.
22270 (and<mode>3): Likewise.
22271 (one_cmpl<mode>2): Likewise.
22272 (nor<mode>3): Likewise.
22273 (andc<mode>3): Likewise.
22275 * config/rs6000/constraints.md (wt constraint): New constraint
22276 that returns VSX_REGS if TImode is allowed in VSX registers.
22278 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22279 constant under VSX.
22281 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22282 similar to TImode, but it is restricted to being in the GPRs.
22284 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22285 TImode to occupy a single VSX register.
22287 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22288 -mvsx-timode for power7/power8.
22289 (power7 cpu): Likewise.
22290 (power8 cpu): Likewise.
22292 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22293 sure that TFmode/TDmode take up two registers if they are ever
22294 allowed in the upper VSX registers.
22295 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22297 (rs6000_init_hard_regno_mode_ok): Likewise.
22298 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22299 constraint. Print if LRA is turned on.
22300 (rs6000_option_override_internal): Give an error if -mvsx-timode
22301 and VSX is not enabled.
22302 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22303 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22304 to reg+offset addressing. Use PTImode when checking offset
22305 addresses for validity.
22306 (reg_offset_addressing_ok_p): Likewise.
22307 (rs6000_legitimate_offset_address_p): Likewise.
22308 (rs6000_legitimize_address): Likewise.
22309 (rs6000_legitimize_reload_address): Likewise.
22310 (rs6000_legitimate_address_p): Likewise.
22311 (rs6000_eliminate_indexed_memrefs): Likewise.
22312 (rs6000_emit_move): Likewise.
22313 (rs6000_secondary_reload): Likewise.
22314 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22315 reloads to fpr registers to continue to use reg+offset addressing,
22316 but 64-bit reloads to altivec registers need reg+reg addressing.
22317 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22318 it. Treat LO_SUM like a PLUS operation.
22319 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22320 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22321 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22322 registers to share a register with a smaller sized type, since VSX
22323 puts scalars in the upper 64-bits.
22324 (print_operand): Add support for PTImode.
22325 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22326 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22327 registers, but don't have arithmetic support.
22328 (rs6000_memory_move_cost): Add test for VSX.
22329 (rs6000_opt_masks): Add -mvsx-timode.
22331 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22334 (VSr): Use wt constraint for TImode.
22335 (VSv): Drop TImode support.
22336 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22337 (vsx_movti_64bit): Likewise.
22338 (vsx_movti_32bit): Likewise.
22339 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22340 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22341 one '?' on the appropriate output constraint. Do not allow TImode
22342 logical operations on 32-bit systems.
22343 (vsx_ior<mode>3): Likewise.
22344 (vsx_xor<mode>3): Likewise.
22345 (vsx_one_cmpl<mode>2): Likewise.
22346 (vsx_nor<mode>3): Likewise.
22347 (vsx_andc<mode>3): Likewise.
22348 (vsx_concat_<mode>): Likewise.
22349 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22351 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22352 OPTION_MASK_VSX_TIMODE.
22353 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22354 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22356 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22357 (TI2 iterator): New iterator for TImode, PTImode.
22358 (wd mode attribute): Add values for vector types.
22359 (movti_string): Replace TI move operations with operations for TImode
22360 and PTImode. Add support for TImode being allowed in VSX registers.
22361 (mov<mode>_string, TImode/PTImode): Likewise.
22362 (movti_ppc64): Likewise.
22363 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22364 (TI mode splitters): Likewise.
22366 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22369 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22371 PR tree-optimization/56355
22372 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22373 Also handle integers with undefined overflow.
22375 2013-03-20 Catherine Moore <clm@codesourcery.com>
22376 Maciej W. Rozycki <macro@codesourcery.com>
22377 Tom de Vries <tom@codesourcery.com>
22378 Nathan Sidwell <nathan@codesourcery.com>
22379 Iain Sandoe <iain@codesourcery.com>
22380 Nathan Froyd <froydnj@codesourcery.com>
22381 Chao-ying Fu <fu@mips.com>
22383 * doc/extend.texi: (micromips, nomicromips, nocompression):
22384 Document new function attributes.
22385 * doc/invoke.texi (minterlink-compressed, mmicromips,
22386 m14k, m14ke, m14kec): Document new options.
22387 (minterlink-mips16): Update documentation.
22388 * doc/md.texi (ZC, ZD): Document new constraints.
22389 * configure.ac (gcc_cv_as_micromips): Check if linker
22390 supports the .set micromips directive.
22391 * configure: Regenerate.
22392 * config.in: Regenerate.
22393 * config/mips/mips-tables.opt: Regenerate.
22394 * config/mips/micromips.md: New file.
22395 * constraints.md (ZC, ZD): New constraints.
22396 * config/mips/predicates.md (movep_src_register): New predicate.
22397 (movep_src_operand): New predicate.
22398 (non_volatile_mem_operand): New predicate.
22399 * config/mips/mips.md (multimem): New type.
22400 (length): Differentiate between 17-bit and 18-bit branch offsets.
22401 (MOVEP1, MOVEP2): New mode iterator.
22402 (mov_<load>l): Use ZC constraint.
22403 (mov_<load>r): Likewise.
22404 (mov_<store>l): Likewise.
22405 (mov_<store>r): Likewise.
22406 (*branch_equality<mode>_inverted): Add microMIPS support.
22407 (*branch_equality<mode>): Likewise.
22408 (*jump_absolute): Likewise.
22409 (indirect_jump_<mode>): Likewise.
22410 (tablejump_<mode>): Likewise.
22411 (<optab>_internal): Likewise.
22412 (sibcall_internal): Likewise.
22413 (sibcall_value_internal): Likewise.
22414 (prefetch): Use constraint ZD.
22415 * config/mips/mips.opt (minterlink-compressed): New option.
22416 (minterlink-mips16): Now an alias for minterlink-compressed.
22417 (mmicromips): New option.
22418 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22419 (compare_and_swap_12): Likewise.
22420 (sync_add<mode>): Likewise.
22421 (sync_<optab>_12): Likewise.
22422 (sync_old_<optab>_12): Likewise.
22423 (sync_new_<optab>_12): Likewise.
22424 (sync_nand_12): Likewise.
22425 (sync_old_nand_12): Likewise.
22426 (sync_new_nand_12): Likewise.
22427 (sync_sub<mode>): Likewise.
22428 (sync_old_add<mode>): Likewise.
22429 (sync_old_sub<mode>): Likewise.
22430 (sync_new_add<mode>): Likewise.
22431 (sync_new_sub<mode>): Likewise.
22432 (sync_<optab><mode>): Likewise.
22433 (sync_old_<optab><mode>): Likewise.
22434 (sync_new_<optab><mode>): Likewise.
22435 (sync_nand<mode>): Likewise.
22436 (sync_old_nand<mode>): Likewise.
22437 (sync_new_nand<mode>): Likewise.
22438 (sync_lock_test_and_set<mode>): Likewise.
22439 (test_and_set_12): Likewise.
22440 (atomic_compare_and_swap<mode>): Likewise.
22441 (atomic_exchange<mode>_llsc): Likewise.
22442 (atomic_fetch_add<mode>_llsc): Likewise.
22443 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22444 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22445 (umips_save_restore_pattern_p): Likewise.
22446 (umips_load_store_pair_p): Likewise.
22447 (umips_output_load_store_pair): Likewise.
22448 (umips_movep_target_p): Likewise.
22449 (umips_12bit_offset_address_p): Likewise.
22450 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22451 (mips_base_mips16): Rename this...
22452 (mips_base_compression_flags): ...to this. Update all uses.
22453 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22454 (mips_mips16_decl_p): Delete.
22455 (mips_nomips16_decl_p): Delete.
22456 (mips_get_compress_on_flags): New function.
22457 (mips_get_compress_off_flags): New function.
22458 (mips_get_compress_mode): New function.
22459 (mips_get_compress_on_name): New function.
22460 (mips_get_compress_off_name): New function.
22461 (mips_insert_attributes): Support multiple compression types.
22462 (mips_merge_decl_attributes): Likewise.
22463 (umips_12bit_offset_address_p): New function.
22464 (mips_start_function_definition): Emit .set micromips directive.
22465 (mips_call_may_need_jalx_p): New function.
22466 (mips_function_ok_for_sibcall): Add microMIPS support.
22467 (mips_print_operand_punctuation): Support short delay slots and
22469 (umips_swm_mask, umips_swm_encoding): New.
22470 (umips_build_save_restore): New function.
22471 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22472 (was_mips16_p): Remove.
22473 (old_compression_mode): New.
22474 (mips_set_compression_mode): New function.
22475 (mips_set_current_function): Add microMIPS support.
22476 (mips_option_override): Likewise.
22477 (umips_save_restore_pattern_p): New function.
22478 (umips_output_save_restore): New function.
22479 (umips_load_store_pair_p_1): New function.
22480 (umips_load_store_pair_p): New function.
22481 (umips_output_load_store_pair_1): New function.
22482 (umips_output_load_store_pair): New function.
22483 (umips_movep_target_p) New function.
22484 (mips_prepare_pch_save): Add microMIPS support.
22485 * config/mips/mips.h (TARGET_COMPRESSION): New.
22486 (TARGET_CPU_CPP_BUILTINS): Update macro
22487 to use new compression flags and to support microMIPS.
22488 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22489 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22490 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22491 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22492 (ASM_SPEC): Support mmicromips and mno-micromips.
22493 (M16STORE_REG_P): New macro.
22494 (MIPS_CALL): Support TARGET_MICROMIPS.
22495 (MICROMIPS_J): New macro.
22496 (mips_base_mips16): Rename this...
22497 (mips_base_compression_flags): ...to this.
22498 (UMIPS_12BIT_OFFSET_P): New macro.
22499 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22500 (MULTILIB_DIRNAMES): Likewise.
22501 2013-03-20 Richard Biener <rguenther@suse.de>
22503 PR tree-optimization/56661
22504 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22505 the result does not have to be distinct.
22507 2013-03-20 Richard Biener <rguenther@suse.de>
22509 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22512 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22513 Steven Bosscher <steven@gcc.gnu.org>
22515 PR rtl-optimization/56605
22516 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22518 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22521 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22522 that require movd instead of movq.
22524 2013-03-20 Richard Biener <rguenther@suse.de>
22526 * tree-ssa-structalias.c (struct variable_info): Add pointer
22527 to the first field of an aggregate with sub-vars. Make
22528 this and the pointer to the next subfield its ID.
22529 (vi_next): New function.
22530 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22531 storedanything_id, integer_id): Increment by one.
22532 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22533 get_call_clobber_vi): Adjust.
22534 (solution_set_expand): Simplify and speedup.
22535 (solution_set_add): Inline into ...
22536 (set_union_with_increment): ... this. Adjust accordingly.
22537 (do_sd_constraint): Likewise.
22538 (do_ds_constraint): Likewise.
22539 (do_complex_constraint): Simplify.
22540 (build_pred_graph): Adjust.
22541 (solve_graph): Likewise. Simplify and speedup.
22542 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22543 get_constraint_for_component_ref, get_constraint_for_1,
22544 first_vi_for_offset, first_or_preceding_vi_for_offset,
22545 create_function_info_for, create_variable_info_for_1,
22546 create_variable_info_for, intra_create_variable_infos): Adjust.
22547 (init_base_vars): Push NULL for ID zero.
22548 (compute_points_to_sets): Adjust.
22550 2013-03-20 Richard Biener <rguenther@suse.de>
22552 * cfgloop.c (verify_loop_structure): Streamline and avoid
22553 ICEing on corrupt loop tree.
22554 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22557 2013-03-20 Richard Biener <rguenther@suse.de>
22559 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22560 check whether an SSA update is needed.
22562 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22564 * config/mips/constraints.md (T): Rename to...
22568 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22569 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22571 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22573 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22574 (*subsi3_carryin_uxtw): Likewise.
22576 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22578 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22579 (*rorsi3_insn_uxtw): Likewise.
22581 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22583 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22584 (*extrsi5_insn_uxtw): Likewise.
22586 2013-03-19 Richard Biener <rguenther@suse.de>
22588 PR tree-optimization/56273
22589 * passes.c (init_optimization_passes): Move second VRP after DOM.
22591 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22593 * config/i386/i386.md (*movti_internal): Merge from
22594 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22595 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22596 nox64 isa attributes.
22598 2013-03-18 Richard Biener <rguenther@suse.de>
22600 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22602 (merge_node_constraints): Likewise.
22603 (build_succ_graph): Likewise.
22604 (valid_graph_edge): Inline into single caller.
22605 (unify_nodes): Likewise. Use bitmap_set_bit return value
22607 (scc_visit): Fix formatting and variable use.
22608 (do_sd_constraint): Use gcc_checking_assert.
22609 (do_ds_constraint): Likewise.
22610 (do_complex_constraint): Likewise.
22611 (condense_visit): Likewise. Cleanup.
22612 (dump_pred_graph): New function.
22613 (perform_var_substitution): Dump the pred-graph before
22614 variable substitution.
22615 (find_equivalent_node): Use gcc_checking_assert.
22616 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22618 2013-03-18 Richard Biener <rguenther@suse.de>
22620 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22621 Remove cond_expr_stmt_list argument and do not gimplify the
22623 (vect_loop_versioning): Adjust.
22624 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22625 Cleanup to use less temporaries.
22626 (vect_create_data_ref_ptr): Cleanup.
22628 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22630 PR tree-optimization/56635
22631 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22632 require types_compatible_p types.
22634 2013-03-18 Nick Clifton <nickc@redhat.com>
22636 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22637 spurious backslash.
22639 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22640 Add missing line to comment describing function.
22642 2013-03-18 Richard Biener <rguenther@suse.de>
22644 PR tree-optimization/56210
22645 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22646 Handle string / character search functions.
22647 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22649 2013-03-18 Richard Biener <rguenther@suse.de>
22651 PR middle-end/56483
22652 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22653 and implement properly.
22654 * gimple.h (gimple_cond_single_var_p): Remove.
22656 2013-03-18 Richard Biener <rguenther@suse.de>
22658 * tree-data-ref.h (find_data_references_in_loop): Declare.
22659 * tree-data-ref.c (get_references_in_stmt): Use a stack
22660 vector pre-allocated in the callers.
22661 (find_data_references_in_stmt): Adjust.
22662 (graphite_find_data_references_in_stmt): Likewise.
22663 (create_rdg_vertices): Likewise.
22664 (find_data_references_in_loop): Export.
22665 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22666 Compute dependences here...
22667 (vect_analyze_data_refs): ...not here. When we encounter
22668 a non-vectorizable data reference in basic-block vectorization
22669 truncate the data reference vector. Do not bother to
22670 fixup data-dependence information for gather loads.
22671 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22672 of data references, as reported.
22674 2013-03-18 Richard Biener <rguenther@suse.de>
22676 PR tree-optimization/3713
22677 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22678 has_constants and expr.
22679 (stmt_has_constants): Properly valueize SSA names when deciding
22680 whether the stmt has constants.
22682 2013-03-18 Richard Biener <rguenther@suse.de>
22684 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22685 whole function when there is nothing to do.
22686 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22687 * tree-vectorizer.c (vectorize_loops): Update virtual and
22688 loop-closed SSA once.
22689 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22691 2013-03-18 Richard Biener <rguenther@suse.de>
22693 PR middle-end/56113
22694 * domwalk.c (bb_postorder): New global static.
22695 (cmp_bb_postorder): New function.
22696 (walk_dominator_tree): Replace scheme imposing an order for
22697 visiting dominator sons by one sorting them at the time they
22698 are pushed on the stack.
22700 2013-03-18 Richard Biener <rguenther@suse.de>
22702 PR tree-optimization/39326
22703 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22704 (struct mem_ref): Replace mem member with ao_ref typed member.
22705 (MEM_ANALYZABLE): Adjust.
22706 (memref_eq): Likewise.
22707 (mem_ref_alloc): Likewise.
22708 (gather_mem_refs_stmt): Likewise.
22709 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22710 (execute_sm_if_changed_flag_set): Adjust.
22711 (execute_sm): Likewise.
22712 (ref_always_accessed_p): Likewise.
22713 (refs_independent_p): Likewise.
22714 (can_sm_ref_p): Likewise.
22716 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22719 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22720 return 1 even for !unsignedp.
22722 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22724 * config/i386/i386.md (isa): Add x64 and nox64.
22725 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22726 (*pushtf): Enable *roF alternative for x64 isa only.
22727 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22728 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22729 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22730 (*movtf_internal): Merge from *movtf_internal_rex64 and
22731 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22732 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22733 nox64 isa attributes.
22734 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22735 nox64 isa attributes.
22736 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22738 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22740 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22742 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22745 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22746 class names. Remove trailing comma after "ALL_REGS".
22748 2013-03-16 Jan Hubicka <jh@suse.cz>
22750 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22751 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22752 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22753 of cgraph_get_create_node.
22754 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22756 2013-03-16 Jason Merrill <jason@redhat.com>
22759 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22760 with DW_AT_default_value.
22762 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22764 * BASE-VER: Set to 4.9.0.
22766 2013-03-14 Andi Kleen <ak@linux.intel.com>
22769 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22770 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22771 Document _x* TSX intrinsics.
22773 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22774 David Holsgrove <david.holsgrove@xilinx.com>
22776 * configure.ac: Add MicroBlaze TLS support detection.
22777 * configure: Regenerate.
22778 * config/microblaze/microblaze-protos.h
22779 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22780 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22781 * config/microblaze/microblaze.c (microblaze_address_type): Add
22782 ADDRESS_TLS and tls_reloc address types.
22783 (microblaze_address_info): Add tls_reloc.
22784 (TARGET_HAVE_TLS): Define.
22785 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22786 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22787 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22788 load_tls_operand, microblaze_call_tls_get_addr,
22789 microblaze_legitimize_tls_address): New functions.
22790 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22791 (get_base_reg): Use microblaze_tls_symbol_p.
22792 (microblaze_classify_address): Handle TLS.
22793 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22794 label_mentioned_p and microblaze_tls_referenced_p.
22795 (microblaze_legitimize_address): Handle TLS.
22796 (microblaze_address_insns): Handle ADDRESS_TLS.
22797 (pic_address_needs_scratch): Handle TLS.
22798 (print_operand_address): Handle TLS.
22799 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22800 (microblaze_expand_move): Handle TLS.
22801 (microblaze_legitimate_constant_p): Check
22802 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22803 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22804 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22805 (PIC_OFFSET_TABLE_REGNUM): Set.
22806 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22807 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22808 (addsi3, movsi_internal2, movdf_internal): Update constraints
22809 * config/microblaze/predicates.md (arith_plus_operand): Define
22810 (move_operand): Redefine as move_src_operand,
22811 check microblaze_tls_referenced_p.
22813 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22815 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22816 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22818 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22820 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22823 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22825 PR tree-optimization/53265
22826 * common.opt (Waggressive-loop-optimizations): New option.
22827 * tree-ssa-loop-niter.c: Include tree-pass.h.
22828 (do_warn_aggressive_loop_optimizations): New function.
22829 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22830 if number_of_latch_executions returned constant.
22831 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22832 early. If number_of_latch_executions returned constant, set
22833 nb_iterations_upper_bound back to it.
22834 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22836 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22837 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22839 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22840 (MULTILIB_OSDIRNAMES): Set.
22841 * genmultilib: If defaultosdirname doesn't start with :: , set
22842 defaultosdirname2 instead, clear it and emit two . multilib_raw
22843 entries instead of just one.
22845 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22847 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22848 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22849 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22850 (SUBTARGET_OVERRIDE_OPTIONS): New.
22852 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22855 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22856 (musermode): Convert to Var(TARGET_USERMODE).
22857 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22858 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22859 * config/sh/sh.c (sh_option_override): Use
22860 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22861 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22863 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22865 (udivsi3_i4_single, divsi3_i4_single): Use
22866 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22868 2013-03-13 Dave Korn <dave.korn.cygwin@....>
22870 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22873 2013-03-13 Richard Biener <rguenther@suse.de>
22875 PR tree-optimization/56608
22876 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22877 calls when vectorizing basic-blocks.
22879 2013-03-13 Jakub Jelinek <jakub@redhat.com>
22882 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22885 2013-03-12 Jakub Jelinek <jakub@redhat.com>
22887 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22889 2013-03-11 Jan Hubicka <jh@suse.cz>
22892 * lto-streamer-out.c (output_symbol_p): Skip references from
22893 constructors of external variables.
22895 2013-03-11 Jan Hubicka <jh@suse.cz>
22897 PR middle-end/56571
22898 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22900 * emit-rtl.c (verify_rtx_sharing): Likewise.
22901 (copy_insn_1): Likewise.
22902 * rtl.c (copy_rtx): Likewise.
22904 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
22907 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22908 output_operand_lossage message.
22910 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
22913 * arm.c (shift_op): Validate RTL pattern on the fly.
22914 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22917 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22920 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22921 new call patterns that clobber %fr12.
22922 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22923 split and postreload patterns.
22924 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22925 registers %fr12 and %fr12R as call used.
22927 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22929 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22930 (canon_address, record_store, replace_read, check_mem_read_rtx,
22931 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22932 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22933 rest_of_handle_dse): Likewise.
22935 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22937 PR middle-end/56524
22938 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22940 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22941 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22942 (save_optabs_if_changed): Replace with...
22943 (init_tree_optimization_optabs): ...this.
22944 * optabs.c (save_optabs_if_changed): Rename to...
22945 (init_tree_optimization_optabs): ...this. Take the optimization node
22946 as argument. Do nothing if the base optabs are already correct.
22947 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22948 to recompute optabs.
22949 * function.h (function): Remove optabs field.
22950 * function.c (invoke_set_current_function_hook): Call
22951 init_tree_optimization_optabs. Use the result to initialize
22954 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
22956 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22957 if GTMA_HAS_NO_INSTRUMENTATION.
22958 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22959 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22960 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22961 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22962 GTMA_HAS_NO_INSTRUMENTATION.
22964 2013-03-08 Jakub Jelinek <jakub@redhat.com>
22966 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22969 2013-03-08 Marek Polacek <polacek@redhat.com>
22970 Jakub Jelinek <jakub@redhat.com>
22972 PR tree-optimization/56478
22973 * predict.c (is_comparison_with_loop_invariant_p): Change the
22974 type of loop_step to tree.
22975 (predict_loops): Adjust.
22976 (predict_iv_comparison): Perform the computations on double_ints.
22978 2013-03-08 Richard Biener <rguenther@suse.de>
22980 PR tree-optimization/56570
22981 * tree-cfg.c (verify_expr_location_1): Verify locations for
22983 * tree-sra.c (create_access_replacement): Strip locations
22984 from DECL_DEBUG_EXPRs.
22986 2013-03-08 Richard Biener <rguenther@suse.de>
22988 * tree-inline.c (expand_call_inline): Do not associate
22989 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22990 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22992 2013-03-08 Richard Biener <rguenther@suse.de>
22994 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22995 or block changes with -Og. Fix for location / block encoding
22996 changes and PHI arguments with locations.
22998 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23000 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23002 (struct output_info): Likewise.
23003 (register_overhead): Remove bad gcc_assert.
23004 (bitmap_find_bit): If there is only a single bitmap element, do not
23005 count a miss as a search.
23006 (print_statistics): Update for counter type changes.
23007 (dump_bitmap_statistics): Likewise. Print headers such that they
23008 are properly lined up with the printed counters.
23010 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23012 PR tree-optimization/56559
23013 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23014 check that it has only a single use.
23016 2013-03-07 Richard Biener <rguenther@suse.de>
23018 * doc/invoke.texi (fwhole-program): Discourage use in combination
23021 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23023 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23025 PR tree-optimization/56539
23026 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23027 instead of GSI_CONTINUE_LINKING as last argument to
23028 force_gimple_operand_gsi. Adjust function comment.
23030 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23033 PR middle-end/56548
23034 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23035 promoted mode, convert the result back to the original mode.
23037 2013-03-06 Richard Biener <rguenther@suse.de>
23039 PR middle-end/56294
23040 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23041 (insert_updated_phi_nodes_compare_uids): New function.
23042 (update_ssa): Sort symbols_to_rename after UID before
23043 traversing it to insert PHI nodes.
23045 2013-03-06 Richard Biener <rguenther@suse.de>
23047 PR middle-end/50494
23048 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23049 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23052 2013-02-13 Richard Biener <rguenther@suse.de>
23055 * varasm.c (output_constant_def_1): Get the decl representing
23056 the constant as argument.
23057 (output_constant_def): Wrap output_constant_def_1.
23058 (make_decl_rtl): Use output_constant_def_1 with the decl
23059 representing the constant.
23060 (build_constant_desc): Optionally re-use a decl already
23061 representing the constant.
23062 (tree_output_constant_def): Adjust.
23064 2013-03-06 Joey Ye <joey.ye@arm.com>
23067 * gcc.c (convert_white_space): New function.
23068 (main): Handles white space in function name.
23070 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23073 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23074 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23075 to SH_DIV_CALL_TABLE for TARGET_SH2.
23076 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23078 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23079 call-table options.
23081 2013-03-05 Sterling Augustine <saugustine@google.com>
23082 Cary Coutant <ccoutant@google.com>
23085 * dwarf2out.c (resolve_addr): Don't call
23086 remove_loc_list_addr_table_entries a second time for the same
23089 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23092 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23093 (avoid_complex_debug_insns): New function.
23094 (expand_debug_locations): Call it.
23096 PR rtl-optimization/56484
23097 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23098 lifetimes of hard registers on small register class machines.
23100 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23102 * config/microblaze/microblaze-protos.h: Rename
23103 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23104 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23106 (microblaze_fast_interrupt_function_p): New function.
23107 (microblaze_is_interrupt_handler): Rename to
23108 microblaze_is_interrupt_variant and add fast_interrupt check.
23109 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23110 (save_restore_insns): Likewise.
23111 (compute_frame_size): Likewise.
23112 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23113 (microblaze_globalize_label): Likewise.
23114 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23115 * config/microblaze/microblaze.md: Use wrapper
23116 microblaze_is_interrupt_variant.
23118 2013-03-05 Kai Tietz <ktietz@redhat.com>
23120 * sdbout.c (sdbout_one_type): Switch to current function's section
23121 supporting cold/hot.
23123 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23125 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23128 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23130 PR middle-end/56461
23131 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23132 if VALGRIND_GET_VBITS is defined, temporarily make object
23133 memory all defined, and restore previous valgrind addressability
23134 and definability afterwards. Free this_object at the end.
23136 PR middle-end/56461
23137 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23138 right before calling lra_create_live_ranges, also call it
23139 when clearing live_p. Only call lra_clear_live_ranges
23140 at the end if live_p.
23142 PR middle-end/56461
23143 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23145 2013-03-05 Richard Biener <rguenther@suse.de>
23147 PR tree-optimization/56521
23148 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23151 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23154 * except.h (remove_unreachable_eh_regions): New prototype.
23155 * except.c (remove_eh_handler_splicer): New function, split out
23156 of remove_eh_handler.
23157 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23158 warning about running it on many EH regions one at a time.
23159 (remove_unreachable_eh_regions_worker): New function, walk the
23160 EH tree in depth-first order and remove non-marked regions.
23161 (remove_unreachable_eh_regions): New function.
23162 * tree-eh.c (mark_reachable_handlers): New function, split out
23163 from remove_unreachable_handlers.
23164 (remove_unreachable_handlers): Use mark_reachable_handlers and
23165 remove_unreachable_eh_regions.
23166 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23167 and remove_unreachable_eh_regions.
23169 2013-03-05 Richard Biener <rguenther@suse.de>
23171 PR middle-end/56525
23172 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23173 not freeing them until the end.
23175 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23177 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23179 2013-03-05 Richard Biener <rguenther@suse.de>
23181 PR tree-optimization/56270
23182 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23183 of loads after scheduling an SLP instance.
23185 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23187 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23189 (check_gcc_parallelize): Run guality.exp as a separate job from
23190 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23191 struct-layout-1.exp with stackalign.exp.
23193 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23195 PR middle-end/56461
23196 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23197 load_index sbitmap even if some bit in it isn't set.
23199 PR middle-end/56461
23200 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23201 (discover_iteration_bound_by_body_walk): Change queues to
23202 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23203 spelling in comment. Call safe_push on queues[bound_index] directly.
23204 Release queues[queue_index] in every iteration unconditionally.
23205 Release bounds vector.
23207 PR middle-end/56461
23208 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23209 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23210 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23213 2013-03-05 Richard Biener <rguenther@suse.de>
23216 * tree-inline.c (remap_blocks_to_null): New function.
23217 (expand_call_inline): When expanding a call stmt without
23218 an associated block inline remap all callee blocks to NULL.
23220 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23222 PR rtl-optimization/56494
23223 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23224 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23225 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23227 PR middle-end/56461
23228 * sel-sched-ir.c (free_sched_pools): Release
23229 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23230 if succs_info_pool.max_top isn't -1.
23233 * opts.c (opts_obstack, opts_concat): Moved to...
23234 * opts-common.c (opts_obstack, opts_concat): ... here.
23236 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23238 PR middle-end/56461
23239 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23241 2013-03-04 Martin Jambor <mjambor@suse.cz>
23243 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23244 all appropriate places.
23246 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23248 PR tree-optimization/56424
23249 * ipa-split.c (split_function): Do not set the RSO flag if result is
23250 not by reference and its type is a register type.
23252 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23254 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23255 (microblaze_legitimate_pic_operand): Likewise
23256 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23257 new function microblaze_legitimate_pic_operand
23258 * config/microblaze/microblaze-protos.h
23259 (microblaze_legitimate_pic_operand): Declare.
23261 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23263 * config/microblaze/predicates.md (call_insn_simple_operand):
23264 New predicate for supported rtx code types.
23265 * config/microblaze/microblaze.md (call_internal1): Use
23266 call_insn_simple_operand predicate.
23268 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23270 PR middle-end/56461
23271 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23272 partitions.ordered_remove.
23274 PR middle-end/56461
23275 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23276 vec_oprnds0.create (1) for modifier == NONE.
23278 PR middle-end/56461
23279 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23280 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23281 vec_oprnds1 right before pushing anything to it for
23284 PR middle-end/56461
23285 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23286 set nbbs to 0 instead of having separate code path.
23287 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23288 instead of false as last argument if returning NULL.
23290 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23292 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23293 the attribute is now called "target" instead of "option".
23294 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23295 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23296 attribute/pragma name for TARGET_OPTION_VALID_P and
23297 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23298 * doc/tm.texi: Regenerated.
23300 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23302 * config/microblaze/microblaze.c:
23303 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23304 * config/microblaze/microblaze.h: Add -mxl-reorder to
23306 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23307 instructions emitted if TARGET_REORDER.
23308 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23309 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23312 2013-03-01 Xinliang David Li <davidxl@google.com>
23314 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23317 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23319 PR middle-end/56461
23320 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23321 vector even when returning true. Fix up function comment formatting.
23323 PR middle-end/56461
23324 * ira-build.c (ira_loop_nodes_count): New variable.
23325 (create_loop_tree_nodes): Initialize it.
23326 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23328 PR middle-end/56461
23329 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23330 method on dr_chain and result_chain.
23331 * tree-vect-stmts.c (vectorizable_store): Only call
23332 result_chain.create if j == 0.
23334 PR middle-end/56461
23335 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23336 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23337 before overwriting it.
23339 2013-03-01 Tobias Burnus <burnus@net-b.de>
23341 * doc/extended.texi (C Extensions): Change order in @menu
23343 (Other MIPS Built-in Functions): Move last MIPS entry before
23344 "picoChip Built-in Functions".
23345 (SH Built-in Functions): Move after RX Built-in Functions.
23346 * doc/gcc.texi (Introduction): Change order in @menu
23348 * doc/md.texi (Constraints): Ditto.
23349 * gty.texi (Type Information): Ditto.
23350 (User-provided marking routines for template types): Make
23352 * doc/invoke.texi (AArch64 Options): Move before
23353 "Adapteva Epiphany Options".
23355 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23356 Jakub Jelinek <jakub@redhat.com>
23359 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23360 no_address_safety_analysis attribute.
23361 * doc/extend.texi (no_address_safety_attribute): Rename to
23362 no_sanitize_address attribute, mention no_address_safety_analysis
23363 attribute as deprecated alias.
23365 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23367 PR middle-end/56461
23368 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23369 type to vec<vec<tree> > *.
23370 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23371 to be vec<tree> instead of vec<tree> *, set vec_defs
23372 to vNULL and call vec_defs.create (number_of_vects), adjust other
23374 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23375 vectorizable_condition): Adjust vect_get_slp_defs callers.
23377 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23379 * config/aarch64/aarch64.c
23380 (aarch64_float_const_representable): Remove unused variable.
23382 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23384 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23386 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23388 * config/aarch64/aarch64-builtins.c
23389 (aarch64_init_simd_builtins): Make static.
23391 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23393 * config/aarch64/aarch64.c
23394 (aarch64_simd_make_constant): Make static.
23396 2013-02-28 Martin Jambor <mjambor@suse.cz>
23398 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23399 with no initialization to the RHS of debug statements.
23401 2013-02-28 Martin Jambor <mjambor@suse.cz>
23403 PR tree-optimization/56294
23404 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23406 (get_access_replacement): Do not call create_access_replacement.
23407 Assert a replacement exists.
23408 (get_repl_default_def_ssa_name): Create the replacement declaration
23411 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23413 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23414 final_end_function.
23416 2013-02-28 Marek Polacek <polacek@redhat.com>
23418 PR rtl-optimization/56466
23419 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23420 if we're changing a loop.
23421 (peel_loops_completely): Likewise.
23423 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23426 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23428 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23431 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23432 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23433 INTX_FTYPE_FX, FX_FTYPE_INTX.
23434 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23436 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23438 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23439 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23440 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23441 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23442 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23443 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23444 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23445 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23446 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23447 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23448 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23449 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23450 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23451 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23452 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23453 (avrxmega6): Increase max flash segments from 5 to 6.
23454 * config/avr/t-multilib: Regenerate.
23455 * config/avr/avr-tables.opt: Regenerate.
23456 * doc/avr-mmcu.texi: Regenerate.
23458 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23460 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23461 (avr_device_to_arch): Rename to avr_device_to_ld.
23462 (avr_device_to_as): New prototype.
23463 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23464 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23465 * config/avr/driver-avr.c (avr_device_to_as): New.
23466 (avr_device_to_arch): Rename to avr_device_to_ld.
23468 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23470 PR middle-end/56461
23471 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23472 method on dr_chain and result_chain.
23474 PR middle-end/56461
23475 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23476 pointer_set_destroy on not_executed_last_iteration.
23478 PR middle-end/56461
23479 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23481 PR middle-end/56461
23482 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23483 FOR_EACH_DEFINED_FUNCTION when freeing state.
23485 PR middle-end/56461
23486 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23488 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23491 PR middle-end/56461
23492 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23493 known_aggs[i].items and release known_aggs vector.
23495 PR middle-end/56461
23496 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23498 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23500 * config/microblaze/microblaze.c (microblaze_emit_compare):
23501 Use xor for EQ/NE comparisions.
23502 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23503 (cbranchsf4): Adjust operator to comparison_operator.
23505 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23507 PR middle-end/56461
23508 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23510 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23511 vec_safe_push, always update *slot.
23512 (redirect_edge_var_map_clear): Use vec_free.
23513 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23514 (free_var_map_entry): Use vec_free.
23515 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23516 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23518 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23520 PR middle-end/45472
23521 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23522 when the may_trap_p bit of the exprs being merged differs.
23523 Reorder tests for speculativeness in the logical and operator.
23525 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23527 * incpath.c (add_standard_paths): Use reconcat instead of concat
23528 where appropriate and avoid leaking memory.
23530 * opts.h: Include obstack.h.
23531 (opts_concat): New prototype.
23532 (opts_obstack): New declaration.
23533 * opts.c (opts_concat): New function.
23534 (opts_obstack): New variable.
23535 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23536 (finish_options): Use opts_concat instead of concat
23537 and XOBNEWVEC instead of XNEWVEC.
23538 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23539 generate_option): Likewise.
23540 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23541 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23544 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23545 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23547 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23549 PR middle-end/56461
23550 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23552 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23554 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23555 (arm_block_move_unaligned_straight): Likewise.
23556 (arm_adjust_block_mem): Likewise.
23558 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23561 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23562 temp, cond and label.
23563 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23566 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23567 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23570 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23571 of prologue/epilogue functions.
23574 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23575 Remove unused variable cfa_offset.
23576 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23579 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23580 type promotion to unsigned.
23583 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23584 for HOST_WIDE_INT of 32 bit / same size as int.
23585 (arm_block_move_unaligned_straight): Likewise.
23586 (arm_adjust_block_mem): Likewise.
23589 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23592 2013-02-26 Marek Polacek <polacek@redhat.com>
23594 PR tree-optimization/56426
23595 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23597 2013-02-26 Richard Biener <rguenther@suse.de>
23600 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23601 unused variable loops.
23603 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23605 PR tree-optimization/56448
23606 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23607 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23608 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23609 later operands of the references, or even first operand for
23610 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23612 PR tree-optimization/56443
23613 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23614 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23615 to type_for_mode langhook.
23617 2013-02-25 Matt Turner <mattst88@gmail.com>
23619 * doc/invoke.texi: Document r4700.
23621 2013-02-25 Richard Biener <rguenther@suse.de>
23623 PR tree-optimization/56175
23624 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23626 (simplify_bitwise_binary): ... here. Also guard the conversion
23627 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23629 2013-02-25 Catherine Moore <clm@codesourcery.com>
23632 2013-02-24 Catherine Moore <clm@codesourcery.com>
23633 Maciej W. Rozycki <macro@codesourcery.com>
23634 Tom de Vries <tom@codesourcery.com>
23635 Nathan Sidwell <nathan@codesourcery.com>
23636 Iain Sandoe <iain@codesourcery.com>
23637 Nathan Froyd <froydnj@codesourcery.com>
23638 Chao-ying Fu <fu@mips.com>
23640 * doc/extend.texi: (micromips, nomicromips, nocompression):
23641 Document new function attributes.
23642 * doc/invoke.texi (minterlink-compressed, mmicromips,
23643 m14k, m14ke, m14kec): Document new options.
23644 (minterlink-mips16): Update documentation.
23645 * doc/md.texi (ZC, ZD): Document new constraints.
23646 * configure.ac (gcc_cv_as_micromips): Check if linker
23647 supports the .set micromips directive.
23648 * configure: Regenerate.
23649 * config.in: Regenerate.
23650 * config/mips/mips-tables.opt: Regenerate.
23651 * config/mips/micromips.md: New file.
23652 * constraints.md (ZC, AD): New constraints.
23653 * config/mips/predicates.md (movep_src_register): New predicate.
23654 (movep_src_operand): New predicate.
23655 (non_volatile_mem_operand): New predicate.
23656 * config/mips/mips.md (multimem): New type.
23657 (length): Differentiate between 17-bit and 18-bit branch offsets.
23658 (MOVEP1, MOVEP2): New mode iterator.
23659 (mov_<load>l): Use ZC constraint.
23660 (mov_<load>r): Likewise.
23661 (mov_<store>l): Likewise.
23662 (mov_<store>r): Likewise.
23663 (*branch_equality<mode>_inverted): Add microMIPS support.
23664 (*branch_equality<mode>): Likewise.
23665 (*jump_absolute): Likewise.
23666 (indirect_jump_<mode>): Likewise.
23667 (tablejump_<mode>): Likewise.
23668 (<optab>_internal): Likewise.
23669 (sibcall_internal): Likewise.
23670 (sibcall_value_internal): Likewise.
23671 (prefetch): Use constraint ZD.
23672 * config/mips/mips.opt (minterlink-compressed): New option.
23673 (minterlink-mips16): Now an alias for minterlink-compressed.
23674 (mmicromips): New option.
23675 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23676 (compare_and_swap_12): Likewise.
23677 (sync_add<mode>): Likewise.
23678 (sync_<optab>_12): Likewise.
23679 (sync_old_<optab>_12): Likewise.
23680 (sync_new_<optab>_12): Likewise.
23681 (sync_nand_12): Likewise.
23682 (sync_old_nand_12): Likewise.
23683 (sync_new_nand_12): Likewise.
23684 (sync_sub<mode>): Likewise.
23685 (sync_old_add<mode>): Likewise.
23686 (sync_old_sub<mode>): Likewise.
23687 (sync_new_add<mode>): Likewise.
23688 (sync_new_sub<mode>): Likewise.
23689 (sync_<optab><mode>): Likewise.
23690 (sync_old_<optab><mode>): Likewise.
23691 (sync_new_<optab><mode>): Likewise.
23692 (sync_nand<mode>): Likewise.
23693 (sync_old_nand<mode>): Likewise.
23694 (sync_new_nand<mode>): Likewise.
23695 (sync_lock_test_and_set<mode>): Likewise.
23696 (test_and_set_12): Likewise.
23697 (atomic_compare_and_swap<mode>): Likewise.
23698 (atomic_exchange<mode>_llsc): Likewise.
23699 (atomic_fetch_add<mode>_llsc): Likewise.
23700 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23701 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23702 (umips_save_restore_pattern_p): Likewise.
23703 (umips_load_store_pair_p): Likewise.
23704 (umips_output_load_store_pair): Likewise.
23705 (umips_movep_target_p): Likewise.
23706 (umips_12bit_offset_address_p): Likewise.
23707 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23708 (mips_base_mips16): Rename this...
23709 (mips_base_compression_flags): ...to this. Update all uses.
23710 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23711 (mips_mips16_decl_p): Delete.
23712 (mips_nomips16_decl_p): Delete.
23713 (mips_get_compress_on_flags): New function.
23714 (mips_get_compress_off_flags): New function.
23715 (mips_get_compress_mode): New function.
23716 (mips_get_compress_on_name): New function.
23717 (mips_get_compress_off_name): New function.
23718 (mips_insert_attributes): Support multiple compression types.
23719 (mips_merge_decl_attributes): Likewise.
23720 (umips_12bit_offset_address_p): New function.
23721 (mips_start_function_definition): Emit .set micromips directive.
23722 (mips_call_may_need_jalx_p): New function.
23723 (mips_function_ok_for_sibcall): Add microMIPS support.
23724 (mips_print_operand_punctuation): Support short delay slots and
23726 (umips_swm_mask, umips_swm_encoding): New.
23727 (umips_build_save_restore): New function.
23728 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23729 (was_mips16_p): Remove.
23730 (old_compression_mode): New.
23731 (mips_set_compression_mode): New function.
23732 (mips_set_current_function): Add microMIPS support.
23733 (mips_option_override): Likewise.
23734 (umips_save_restore_pattern_p): New function.
23735 (umips_output_save_restore): New function.
23736 (umips_load_store_pair_p_1): New function.
23737 (umips_load_store_pair_p): New function.
23738 (umips_output_load_store_pair_1): New function.
23739 (umips_output_load_store_pair): New function.
23740 (umips_movep_target_p) New function.
23741 (mips_prepare_pch_save): Add microMIPS support.
23742 * config/mips/mips.h (TARGET_COMPRESSION): New.
23743 (TARGET_CPU_CPP_BUILTINS): Update macro
23744 to use new compression flags and to support microMIPS.
23745 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23746 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23747 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23748 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23749 (ASM_SPEC): Support mmicromips and mno-micromips.
23750 (M16STORE_REG_P): New macro.
23751 (MIPS_CALL): Support TARGET_MICROMIPS.
23752 (MICROMIPS_J): New macro.
23753 (mips_base_mips16): Rename this...
23754 (mips_base_compression_flags): ...to this.
23755 (UMIPS_12BIT_OFFSET_P): New macro.
23756 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23757 (MULTILIB_DIRNAMES): Likewise.
23759 2013-02-25 Tom de Vries <tom@codesourcery.com>
23761 PR rtl-optimization/56131
23762 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23763 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23764 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23766 2013-02-25 Tobias Burnus <burnus@net-b.de>
23768 * doc/invoke.texi (-fsanitize=): Move from optimization
23769 to debugging options.
23771 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23773 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23775 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23776 Alexander Monakov <amonakov@ispras.ru>
23778 PR middle-end/56077
23779 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23780 flush pending lists also on non-jumps. Adjust comment.
23782 2013-02-24 Catherine Moore <clm@codesourcery.com>
23783 Maciej W. Rozycki <macro@codesourcery.com>
23784 Tom de Vries <tom@codesourcery.com>
23785 Nathan Sidwell <nathan@codesourcery.com>
23786 Iain Sandoe <iain@codesourcery.com>
23787 Nathan Froyd <froydnj@codesourcery.com>
23788 Chao-ying Fu <fu@mips.com>
23790 * doc/extend.texi: (micromips, nomicromips, nocompression):
23791 Document new function attributes.
23792 * doc/invoke.texi (minterlink-compressed, mmicromips,
23793 m14k, m14ke, m14kec): Document new options.
23794 (minterlink-mips16): Update documentation.
23795 * doc/md.texi (ZC, ZD): Document new constraints.
23796 * configure.ac (gcc_cv_as_micromips): Check if linker
23797 supports the .set micromips directive.
23798 * configure: Regenerate.
23799 * config.in: Regenerate.
23800 * config/mips/mips-tables.opt: Regenerate.
23801 * config/mips/micromips.md: New file.
23802 * constraints.md (ZC, AD): New constraints.
23803 * config/mips/predicates.md (movep_src_register): New predicate.
23804 (movep_src_operand): New predicate.
23805 (non_volatile_mem_operand): New predicate.
23806 * config/mips/mips.md (multimem): New type.
23807 (length): Differentiate between 17-bit and 18-bit branch offsets.
23808 (MOVEP1, MOVEP2): New mode iterator.
23809 (mov_<load>l): Use ZC constraint.
23810 (mov_<load>r): Likewise.
23811 (mov_<store>l): Likewise.
23812 (mov_<store>r): Likewise.
23813 (*branch_equality<mode>_inverted): Add microMIPS support.
23814 (*branch_equality<mode>): Likewise.
23815 (*jump_absolute): Likewise.
23816 (indirect_jump_<mode>): Likewise.
23817 (tablejump_<mode>): Likewise.
23818 (<optab>_internal): Likewise.
23819 (sibcall_internal): Likewise.
23820 (sibcall_value_internal): Likewise.
23821 (prefetch): Use constraint ZD.
23822 * config/mips/mips.opt (minterlink-compressed): New option.
23823 (minterlink-mips16): Now an alias for minterlink-compressed.
23824 (mmicromips): New option.
23825 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23826 (compare_and_swap_12): Likewise.
23827 (sync_add<mode>): Likewise.
23828 (sync_<optab>_12): Likewise.
23829 (sync_old_<optab>_12): Likewise.
23830 (sync_new_<optab>_12): Likewise.
23831 (sync_nand_12): Likewise.
23832 (sync_old_nand_12): Likewise.
23833 (sync_new_nand_12): Likewise.
23834 (sync_sub<mode>): Likewise.
23835 (sync_old_add<mode>): Likewise.
23836 (sync_old_sub<mode>): Likewise.
23837 (sync_new_add<mode>): Likewise.
23838 (sync_new_sub<mode>): Likewise.
23839 (sync_<optab><mode>): Likewise.
23840 (sync_old_<optab><mode>): Likewise.
23841 (sync_new_<optab><mode>): Likewise.
23842 (sync_nand<mode>): Likewise.
23843 (sync_old_nand<mode>): Likewise.
23844 (sync_new_nand<mode>): Likewise.
23845 (sync_lock_test_and_set<mode>): Likewise.
23846 (test_and_set_12): Likewise.
23847 (atomic_compare_and_swap<mode>): Likewise.
23848 (atomic_exchange<mode>_llsc): Likewise.
23849 (atomic_fetch_add<mode>_llsc): Likewise.
23850 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23851 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23852 (umips_save_restore_pattern_p): Likewise.
23853 (umips_load_store_pair_p): Likewise.
23854 (umips_output_load_store_pair): Likewise.
23855 (umips_movep_target_p): Likewise.
23856 (umips_12bit_offset_address_p): Likewise.
23857 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23858 (mips_base_mips16): Rename this...
23859 (mips_base_compression_flags): ...to this. Update all uses.
23860 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23861 (mips_mips16_decl_p): Delete.
23862 (mips_nomips16_decl_p): Delete.
23863 (mips_get_compress_on_flags): New function.
23864 (mips_get_compress_off_flags): New function.
23865 (mips_get_compress_mode): New function.
23866 (mips_get_compress_on_name): New function.
23867 (mips_get_compress_off_name): New function.
23868 (mips_insert_attributes): Support multiple compression types.
23869 (mips_merge_decl_attributes): Likewise.
23870 (umips_12bit_offset_address_p): New function.
23871 (mips_start_function_definition): Emit .set micromips directive.
23872 (mips_call_may_need_jalx_p): New function.
23873 (mips_function_ok_for_sibcall): Add microMIPS support.
23874 (mips_print_operand_punctuation): Support short delay slots and
23876 (umips_swm_mask, umips_swm_encoding): New.
23877 (umips_build_save_restore): New function.
23878 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23879 (was_mips16_p): Remove.
23880 (old_compression_mode): New.
23881 (mips_set_compression_mode): New function.
23882 (mips_set_current_function): Add microMIPS support.
23883 (mips_option_override): Likewise.
23884 (umips_save_restore_pattern_p): New function.
23885 (umips_output_save_restore): New function.
23886 (umips_load_store_pair_p_1): New function.
23887 (umips_load_store_pair_p): New function.
23888 (umips_output_load_store_pair_1): New function.
23889 (umips_output_load_store_pair): New function.
23890 (umips_movep_target_p) New function.
23891 (mips_prepare_pch_save): Add microMIPS support.
23892 * config/mips/mips.h (TARGET_COMPRESSION): New.
23893 (TARGET_CPU_CPP_BUILTINS): Update macro
23894 to use new compression flags and to support microMIPS.
23895 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23896 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23897 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23898 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23899 (ASM_SPEC): Support mmicromips and mno-micromips.
23900 (M16STORE_REG_P): New macro.
23901 (MIPS_CALL): Support TARGET_MICROMIPS.
23902 (MICROMIPS_J): New macro.
23903 (mips_base_mips16): Rename this...
23904 (mips_base_compression_flags): ...to this.
23905 (UMIPS_12BIT_OFFSET_P): New macro.
23906 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23907 (MULTILIB_DIRNAMES): Likewise.
23909 2013-02-24 Jakub Jelinek <jakub@redhat.com>
23912 * target-globals.c (save_target_globals): For init_reg_sets and
23913 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23915 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23917 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23918 * config/aarch64/t-aarch64
23919 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23921 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23923 PR inline-asm/56148
23924 * lra-constraints.c (process_alt_operands): Reload operand
23925 conflicting with earlier clobber only if no more other conflicting
23928 2013-02-22 Jakub Jelinek <jakub@redhat.com>
23931 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23932 if not linking a shared library.
23934 2013-02-22 Seth LaForge <sethml@google.com>
23936 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23938 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23940 * config/arm/arm.md (split for extendsidi): Update condition.
23941 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23942 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23943 (qhs_zextenddi_cstr): Likewise.
23945 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23947 PR middle-end/56420
23948 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23949 avoid signed wrapping.
23950 (expand_mult): Handle properly multiplication by
23951 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23952 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23953 in the compiler if coeff is HOST_WIDE_INT_MIN.
23954 (expand_divmod): Don't make ext_op1 static, change it's type to
23955 uhwi. Avoid undefined behavior in -INTVAL (op1).
23957 PR rtl-optimization/50339
23958 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23960 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23961 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23962 into splitting_ashiftrt field.
23963 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23965 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23968 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
23970 PR middle-end/56108
23971 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23972 are sure to go irrevocable.
23974 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
23976 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23977 scalars are valid operands.
23979 2013-02-21 Martin Jambor <mjambor@suse.cz>
23981 PR tree-optimization/56310
23982 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23983 only matching indices and non-negative final offsets.
23984 (intersect_aggregates_with_edge): Pass src_idx to
23985 agg_replacements_to_vector. Pass src_idx insstead of index to
23986 intersect_with_agg_replacements.
23988 2013-02-21 Martin Jambor <mjambor@suse.cz>
23990 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23991 instead of hard-wired defaults.
23993 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23995 * doc/invoke.texi (MIPS Options): Update documentation of the
23996 floating-point multiply-accumulate instruction restrictions.
23998 2013-02-21 Kostya Serebryany <kcc@google.com>
24000 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24001 asan_shadow_offset on x86_64 linux.
24003 2013-02-21 Richard Biener <rguenther@suse.de>
24005 PR tree-optimization/56415
24007 2013-02-11 Richard Biener <rguenther@suse.de>
24009 PR tree-optimization/56273
24010 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24013 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24016 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24019 PR inline-asm/56405
24020 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24021 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24023 2013-02-20 Jan Hubicka <jh@suse.cz>
24025 PR tree-optimization/56265
24026 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24027 when target is referenced for first time.
24029 2013-02-20 Richard Biener <rguenther@suse.de>
24031 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24032 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24033 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24034 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24035 not return anything.
24036 (rename_ssa_copies): Do not remove unused locals.
24037 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24038 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24039 * passes.c (execute_function_todo): Do not schedule unused locals
24040 removal if cleanup_tree_cfg did something.
24041 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24042 about the number of removed locals.
24044 2013-02-20 Richard Biener <rguenther@suse.de>
24046 PR tree-optimization/56398
24047 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24049 2013-02-20 Martin Jambor <mjambor@suse.cz>
24051 PR tree-optimization/55334
24052 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24053 restricted pointers to arrays.
24055 2013-02-20 Richard Biener <rguenther@suse.de>
24056 Jakub Jelinek <jakub@redhat.com>
24058 PR tree-optimization/56396
24059 * tree-ssa-ccp.c (n_const_val): New static variable.
24060 (get_value): Return NULL for SSA names we don't have a lattice
24062 (ccp_initialize): Initialize n_const_val.
24063 * tree-ssa-copy.c (n_copy_of): New static variable.
24064 (init_copy_prop): Initialize n_copy_of.
24065 (get_value): Return NULL_TREE for SSA names we don't have a
24068 2013-02-20 Martin Jambor <mjambor@suse.cz>
24070 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24072 2013-02-20 Richard Biener <rguenther@suse.de>
24074 * genpreds.c (write_lookup_constraint): Do not compare first
24075 letter of the constraint again.
24077 2013-02-20 Richard Biener <rguenther@suse.de>
24079 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24081 (get_use_iv_cost): Terminate hashtable walk when coming across
24084 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24086 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24087 reassociation for avx2 targets.
24089 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24091 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24092 Add version check for v8.10.a to enable microblaze_has_clz
24093 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24094 version and TARGET_PATTERN_COMPARE check
24095 * config/microblaze/microblaze.md: New clzsi2 instruction
24097 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24099 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24100 function before branching.
24102 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24104 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24106 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24108 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24110 PR middle-end/55889
24111 * sel-sched.c: Include ira.h.
24112 (implicit_clobber_conflict_p): New function.
24113 (moveup_expr): Use it.
24114 * Makefile.in (sel-sched.o): Depend on ira.h.
24116 2013-02-19 Richard Biener <rguenther@suse.de>
24118 PR tree-optimization/56384
24119 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24120 (vn_hash_type): Split out from ...
24121 (vn_hash_constant_with_type): ... here.
24122 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24123 (vn_phi_eq): Compare types from vn_phi_s structure.
24124 (vn_phi_lookup): Populate vn_phi_s type.
24125 (vn_phi_insert): Likewise.
24127 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24129 PR tree-optimization/56350
24130 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24131 if haven't found reduction or nested cycle operand, rather than
24132 asserting we must find it.
24134 PR tree-optimization/56381
24135 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24138 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24139 Jakub Jelinek <jakub@redhat.com>
24142 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24143 (swap_optab_enable): Same.
24144 (init_all_optabs): Use argument instead of global.
24145 * tree.h (struct tree_optimization_option): New field target_optabs.
24146 * expr.h (init_all_optabs): Add argument to prototype.
24147 (TREE_OPTIMIZATION_OPTABS): New.
24148 (save_optabs_if_changed): Protoize.
24149 * optabs.h: Declare this_fn_optabs.
24150 * optabs.c (save_optabs_if_changed): New.
24151 Declare this_fn_optabs.
24152 (init_optabs): Add argument to init_all_optabs() call.
24153 * function.c (invoke_set_current_function_hook): Handle per
24155 * function.h (struct function): New field optabs.
24156 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24157 optimization_current_node has changed.
24158 * target-globals.h (save_target_globals_default_opts): Protoize.
24159 * target-globals.c (save_target_globals_default_opts): New.
24161 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24164 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24165 registers %fr12 and %fr12R as call used.
24168 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24169 and HImode, require all displacements to be an integer multiple of
24171 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24172 only allow QImode and HImode when reload is in progress and strict is
24173 true. Likewise for symbolic addresses. Use base14_operand to check
24174 displacements in REG+BASE addresses.
24176 2013-02-18 Richard Biener <rguenther@suse.de>
24178 PR tree-optimization/56366
24179 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24180 handle sign-conversion of outer-loop initial induction value.
24182 2013-02-18 Richard Biener <rguenther@suse.de>
24184 PR middle-end/56349
24185 * cfghooks.c (merge_blocks): If we merge a latch into another
24186 block adjust references to it.
24187 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24188 (verify_loop_structure): Verify that a recorded latch is in fact
24191 2013-02-18 Richard Biener <rguenther@suse.de>
24193 PR tree-optimization/56321
24194 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24195 order SSA name release and virtual operand unlinking.
24197 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24199 * config/microblaze/microblaze.md (save_stack_block): Define.
24200 (restore_stack_block): Likewise.
24202 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24204 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24205 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24206 * config/microblaze/microblaze.c (microblaze_option_override):
24207 Bail out early for PIC modes when target does not support PIC.
24209 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24211 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24212 Replace with a microblaze version.
24213 (microblaze_trampoline_init): Adapt for microblaze.
24214 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24217 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24218 Dodji Seketeli <dodji@redhat.com>
24221 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24222 (instrument_mem_region_access): Do not forget to always put
24223 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24224 0) statement, even for cases where either 'base' or 'base + len'
24225 are not instrumented -- because they have been previously
24226 instrumented. Simplify the logic by putting all the statements
24227 instrument 'base + len' inside a sequence, and then insert that
24228 sequence right before the current insertion point. Then, to
24229 instrument 'base + len', just get an iterator on that statement.
24230 And do not forget to update the pointer to iterator the function
24231 received as argument.
24233 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24235 PR rtl-optimization/56348
24236 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24238 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24240 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24241 (clean_graph_dump_file): Pass base to start_graph_dump.
24243 2013-02-14 Richard Henderson <rth@redhat.com>
24246 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24248 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24250 * collect2-aix.h: Define F_LOADONLY.
24252 2013-02-14 Richard Biener <rguenther@suse.de>
24255 * varasm.c (output_constant_def_1): Get the decl representing
24256 the constant as argument.
24257 (output_constant_def): Wrap output_constant_def_1.
24258 (make_decl_rtl): Use output_constant_def_1 with the decl
24259 representing the constant.
24260 (build_constant_desc): Optionally re-use a decl already
24261 representing the constant.
24262 (tree_output_constant_def): Adjust.
24264 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24267 * asan.c (instrument_builtin_call): Really put the length of the
24268 second source argument into src1_len.
24270 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24272 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24273 argument. If it is false, don't create edge from then_bb to
24275 (insert_if_then_before_iter): Pass true to it.
24276 (build_check_stmt): Pass false to it.
24277 (transform_statements): Flush hash table only on extended basic
24278 block boundaries, rather than at the beginning of every bb.
24279 Don't flush hash table on nonfreeing_call_p calls.
24280 * tree-flow.h (nonfreeing_call_p): New prototype.
24281 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24283 2013-02-13 David S. Miller <davem@davemloft.net>
24285 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24287 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24290 * ira.c (max_regno_before_ira): Move from ...
24292 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24295 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24297 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24299 2013-02-13 Richard Biener <rguenther@suse.de>
24302 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24303 globals in MEM_REFs.
24305 2013-02-13 Richard Biener <rguenther@suse.de>
24307 * loop-init.c (loop_optimizer_init): Clear loop state when
24308 re-initializing preserved loops.
24309 * loop-unswitch.c (unswitch_single_loop): Return whether
24310 we unswitched the loop. Do not verify loop state here.
24311 (unswitch_loops): When we unswitched a loop discover new loops.
24313 2013-02-13 Kostya Serebryany <kcc@google.com>
24315 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24317 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24319 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24321 Avoid instrumenting duplicated memory access in the same basic block
24322 * Makefile.in (asan.o): Add new dependency on hash-table.h
24323 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24324 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24325 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24326 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24327 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24328 (get_mem_ref_of_assignment): New functions.
24329 (get_mem_refs_of_builtin_call): Extract from
24330 instrument_builtin_call and tweak a little bit to make it fit with
24332 (instrument_builtin_call): Use the new
24333 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24334 of is_gimple_builtin_call.
24335 (instrument_derefs, instrument_mem_region_access): Insert the
24336 instrumented memory reference into the hash table.
24337 (maybe_instrument_assignment): Renamed instrument_assignment into
24338 this, and change it to advance the iterator when instrumentation
24339 actually happened and return true in that case. This makes it
24340 homogeneous with maybe_instrument_assignment, and thus give a
24341 chance to callers to be more 'regular'.
24342 (transform_statements): Clear the memory reference hash table
24343 whenever we enter a new BB, when we cross a function call, or when
24344 we are done transforming statements. Use
24345 maybe_instrument_assignment instead of instrumentation. No more
24346 need to special case maybe_instrument_assignment and advance the
24347 iterator after calling it; it's now handled just like
24348 maybe_instrument_call. Update comment.
24350 2013-02-13 Richard Biener <rguenther@suse.de>
24352 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24353 Fix loop discovery code.
24355 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24357 PR inline-asm/56148
24358 * lra-constraints.c (process_alt_operands): Match early clobber
24359 operand with itself. Check conflicts with earlyclobber only if
24360 the operand is not reloaded. Prefer to reload conflicting operand
24361 if earlyclobber and matching operands are the same.
24363 2013-02-12 Richard Biener <rguenther@suse.de>
24366 * lto-streamer-out.c (write_symbol): Do not output symbols
24367 for hard register variables.
24369 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24372 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24373 (umulsidi3_insn, mulsidi3_insn): New insns.
24375 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24377 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24378 (struct tune_params): Add vec_costs field.
24379 * config/arm/arm.c (arm_builtin_vectorization_cost)
24380 (arm_add_stmt_cost): New functions.
24381 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24382 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24383 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24384 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24385 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24386 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24387 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24389 2013-02-12 Richard Biener <rguenther@suse.de>
24392 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24393 decls again if possible.
24395 2013-02-12 Richard Biener <rguenther@suse.de>
24397 PR middle-end/56288
24398 * tree-ssa.c (verify_ssa_name): Fix check, move
24399 SSA_NAME_IN_FREE_LIST check up.
24401 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24402 Steven Bosscher <steven@gcc.gnu.org>
24404 PR rtl-optimization/56151
24405 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24406 equal to op0 or op1, and last_insn pattern is CODE operation
24407 with MEM dest and one of the operands matches that MEM.
24409 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24411 * doc/extend.texi: Document Function Multiversioning and "default"
24412 parameter string to target attribute.
24413 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24414 target attribute parameter is "default".
24415 (ix86_compare_version_priority): Remove checks for target attribute.
24416 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24417 Remove check for target attribute equal to NULL. Add assert.
24418 (ix86_generate_version_dispatcher_body): Change error to sorry.
24420 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24421 Jack Howarth <howarth@bromo.med.uc.edu>
24422 Patrick Marlier <patrick.marlier@gmail.com>
24425 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24426 define ENDFILE_SPEC as TM_DESTRUCTOR.
24427 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24429 2013-02-11 Alexander Potapenko <glider@google.com>
24430 Jack Howarth <howarth@bromo.med.uc.edu>
24431 Jakub Jelinek <jakub@redhat.com>
24434 * config/darwin.c (cdtor_record): Rename ctor_record.
24435 (sort_cdtor_records): Rename sort_ctor_records.
24436 (finalize_dtors): New routine to sort destructors by
24437 priority before use in assemble_integer.
24438 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24440 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24442 PR rtl-optimization/56275
24443 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24444 offset is non-negative and less than cmode size before
24445 calling simplify_subreg.
24447 2013-02-11 Richard Biener <rguenther@suse.de>
24449 PR tree-optimization/56264
24450 * cfgloop.h (fix_loop_structure): Adjust prototype.
24451 * loop-init.c (fix_loop_structure): Return the number of
24452 newly discovered loops.
24453 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24454 are discovered, do a full loop-closed SSA rewrite.
24456 2013-02-11 Richard Biener <rguenther@suse.de>
24458 PR tree-optimization/56273
24459 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24461 (check_array_ref): Fix missing newline in dumps.
24462 (search_for_addr_array): Likewise.
24464 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24466 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24468 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24471 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24473 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24475 PR rtl-optimization/56246
24476 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24478 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24479 constraints are satisfied.
24481 2013-02-08 Jeff Law <law@redhat.com>
24484 * emit-rtl.c (reg_is_parm_p): New function.
24485 * regs.h (reg_is_parm_p): New prototype.
24486 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24487 callee-clobbered registers.
24489 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24492 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24493 If there is no implicit builtin declaration, just return NULL.
24495 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24497 * config/i386/sse.md (FMAMODEM): New mode iterator.
24498 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24499 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24501 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24503 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24504 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24505 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24507 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24509 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24510 (microblaze*-*-elf): Likewise.
24511 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24513 * config/microblaze/microblaze-c.c: Add builtin defines for
24514 _LITTLE_ENDIAN and _BIG_ENDIAN.
24515 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24516 add to TARGET_DEFAULT flags.
24517 Expand ASM_SPEC and LINK_SPEC.
24518 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24519 * config/microblaze/microblaze.md: Update extendsidi2 and
24520 movdi_internal instructions to use low-order / high-order reg
24522 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24523 options and inversemask / mask of LITTLE_ENDIAN.
24524 * config/microblaze/t-microblaze: Expand multilib options to
24525 include mlittle-endian (le) and update exceptions patterns.
24527 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24529 PR rtl-optimization/56195
24530 * lra-constraints.c (get_reload_reg): Don't reuse regs
24531 if they have smaller mode than requested, if they have
24532 wider mode than requested, try to return a SUBREG.
24534 PR tree-optimization/56250
24535 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24536 if type is unsigned and code isn't MULT_EXPR.
24538 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24540 PR tree-optimization/56064
24541 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24542 bits according to mode.
24543 * fixed-value.h (fixed_from_double_int)
24544 (const_fixed_from_double_int): Adjust comments.
24546 2013-02-08 Richard Biener <rguenther@suse.de>
24549 * lto-streamer.h (struct data_in): Remove current_file, current_line
24550 and current_col members.
24551 * lto-streamer-out.c (lto_output_location): Stream changed bits
24552 en-block for efficiency.
24553 * lto-streamer-in.c (clear_line_info): Remove.
24554 (lto_input_location): Cache current file, line and column
24555 globally via local statics. Read changed bits en-block.
24556 (input_function): Do not call clear_line_info.
24557 (lto_read_body): Likewise.
24558 (lto_input_toplevel_asms): Likewise.
24560 2013-02-08 Michael Matz <matz@suse.de>
24562 PR tree-optimization/52448
24563 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24564 (nt_call_phase): New static.
24565 (add_or_mark_expr): Only mark accesses with newer phase than any
24567 (nonfreeing_call_p): New.
24568 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24569 (nt_fini_block): Keep blocks marked as visited.
24570 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24572 2013-02-08 Richard Biener <rguenther@suse.de>
24574 * ira.c (ira): Free broken dominator information.
24576 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24578 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24580 2013-02-08 Marek Polacek <polacek@redhat.com>
24582 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24584 2013-02-08 Richard Biener <rguenther@suse.de>
24586 PR middle-end/56181
24587 * cfgloop.h (flow_loops_find): Adjust.
24588 (bb_loop_header_p): Declare.
24589 * cfgloop.c (bb_loop_header_p): New function split out from ...
24590 (flow_loops_find): ... here. Adjust function signature,
24591 support incremental loop structure update.
24592 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24593 * cfgloopmanip.c (fix_loop_structure): Move ...
24594 * loop-init.c (fix_loop_structure): ... here.
24595 (apply_loop_flags): Split out from ...
24596 (loop_optimizer_init): ... here.
24597 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24598 in incremental mode, only remove dead loops here.
24600 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24603 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24604 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24605 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24606 (*round<mode>3.libgcc): New insns for fixed-modes.
24607 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24608 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24609 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24610 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24611 implementations. Define to __builtin_avr_absFX,
24612 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24613 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24614 __builtin_avr_countlsFX, respectively.
24615 * config/avr/avr-c.c (target.h): Include it.
24616 (enum avr_builtin_id): New enum.
24617 (avr_resolve_overloaded_builtin): New static function.
24618 (avr_register_target_pragmas): Use it to set
24619 targetm.resolve_overloaded_builtin.
24620 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24621 tree nodes used by DEF_BUILTIN.
24622 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24623 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24624 <AVR_BUILTIN_xxBITS>: Same.
24626 2013-02-08 Richard Biener <rguenther@suse.de>
24628 * cfgloop.c (verify_loop_structure): Properly handle
24629 a loop exiting to another loop header.
24630 * ira-int.h (ira_loops): Remove.
24631 * ira.c (ira_loops): Remove.
24632 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24633 (do_reload): Use loop_optimizer_finalize.
24634 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24635 number_of_loops to access the loop tree.
24636 (more_one_region_p): Likewise.
24637 (finish_loop_tree_nodes): Likewise.
24638 (rebuild_regno_allocno_maps): Likewise.
24639 (mark_loops_for_removal): Likewise.
24640 (mark_all_loops_for_removal): Likewise.
24641 (remove_unnecessary_regions): Likewise.
24642 (ira_build): Likewise.
24643 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24645 2013-02-08 Richard Biener <rguenther@suse.de>
24647 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24648 * ipa-pure-const.c (analyze_function): Avoid calling
24649 mark_irreducible_loops twice.
24650 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24652 2013-02-07 David S. Miller <davem@davemloft.net>
24654 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24656 * var-tracking.c (vt_add_function_parameter): Test the presence of
24657 HAVE_window_save properly and do not remap argument registers when
24658 we have a leaf function.
24660 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24663 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24665 * config/i386/i386.c (ix86_print_operand): Ditto.
24667 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24669 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24671 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24673 PR rtl-optimization/56225
24674 * lra-constraints.c (process_alt_operands): Check that reload hard
24675 reg can hold value for strict_low_part.
24677 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24680 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24681 dwarf2out_end_function.
24682 (in_first_function_p, maybe_at_text_label_p,
24683 first_loclabel_num_not_at_text_label): New variables.
24684 (dwarf2out_var_location): In the first function find out
24685 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24686 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24689 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24691 PR rtl-optimization/56178
24692 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24693 SUBREG of a register. Tidy up related block of code.
24694 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24695 note if the source is a register or a SUBREG of a register.
24697 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24700 * config/rs6000/rs6000.md (ptrm): New mode attr.
24701 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24702 call_value_indirect_aix<pttrsize>,
24703 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24706 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24708 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24709 if -bnortl. Convert to strcmp and strncmp.
24711 2013-02-07 Alan Modra <amodra@gmail.com>
24714 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24715 addresses won't wrap when offsetting.
24716 (rs6000_secondary_reload): Provide secondary reloads needed for
24717 wrapping LO_SUM addresses.
24719 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24721 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24722 MACH, just __MACH__.
24724 2013-02-06 Richard Biener <rguenther@suse.de>
24726 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24727 instead of calling fix_loop_structure.
24729 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24731 PR middle-end/56217
24732 * omp-low.c (use_pointer_for_field): Return false if
24733 lower_send_shared_vars doesn't generate any copy-out code.
24735 2013-02-06 Tom de Vries <tom@codesourcery.com>
24737 PR rtl-optimization/56131
24738 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24739 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24740 of the label is NULL. Add comment.
24742 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24744 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24747 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24748 (STATIC_LIBTSAN_LIBS): Likewise.
24749 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24750 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24751 is defined, don't add anything else beyond that.
24752 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24753 (LINK_COMMAND_SPEC): Use them.
24755 PR tree-optimization/56205
24756 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24757 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24758 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24760 2013-02-05 Richard Biener <rguenther@suse.de>
24762 PR tree-optimization/53342
24763 PR tree-optimization/53185
24764 * tree-vectorizer.h (vect_check_strided_load): Remove.
24765 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24766 not disallow peeling for vectorized strided loads.
24767 (vect_check_strided_load): Make static and simplify.
24768 (vect_analyze_data_refs): Adjust.
24769 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24770 correctly when vectorizing strided loads.
24772 2013-02-05 Richard Biener <rguenther@suse.de>
24774 * doc/install.texi: Refer to ISL, not PPL.
24776 2013-02-05 Jan Hubicka <jh@suse.cz>
24778 PR tree-optimization/55789
24779 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24781 2013-02-05 Jan Hubicka <jh@suse.cz>
24783 PR tree-optimization/55789
24784 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24785 the dead call anyway.
24787 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24790 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24792 2013-02-04 Alexander Potapenko <glider@google.com>
24793 Jack Howarth <howarth@bromo.med.uc.edu>
24794 Jakub Jelinek <jakub@redhat.com>
24797 * config/darwin.c (sort_ctor_records): Stabilized qsort
24798 on constructor priority by using original position.
24799 (finalize_ctors): New routine to sort constructors by
24800 priority before use in assemble_integer.
24801 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24803 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24806 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24807 about visibility on artificial decls.
24808 * config/sol2.c (solaris_assemble_visibility): Likewise.
24810 2013-02-04 Kai Tietz <ktietz@redhat.com>
24813 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24814 argument and improve checking of return-argument types for 16-byte
24816 (ix86_function_value_1): Add additional valtype argument on call
24817 of function_value_64.
24818 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24819 handling infunction_value_64 function.
24821 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24823 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24825 2013-02-04 Richard Biener <rguenther@suse.de>
24827 PR tree-optimization/56188
24828 * tree-ssa-structalias.c (label_visit): Consider case with
24829 initially non-empty points-to set.
24830 (perform_var_substitution): Dump node mapping and clean up.
24832 2013-02-04 Richard Guenther <rguenther@suse.de>
24835 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24836 node prevail as last resort.
24837 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24838 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24840 2013-02-04 Richard Biener <rguenther@suse.de>
24842 PR tree-optimization/56113
24843 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24845 (equiv_class_lookup_or_add): ... this.
24846 (label_visit): Adjust and fix error in previous patch.
24847 (perform_var_substitution): Adjust.
24849 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24851 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24852 * config/sh/sh4-300.md: Likewise.
24853 * config/sh/sh4a.md: Likewise.
24854 * config/sh/constraints.md: Likewise.
24855 * config/sh/sh.md: Likewise.
24856 * config/sh/netbsd-elf.h: Likewise.
24857 * config/sh/predicates.md: Likewise.
24858 * config/sh/sh-protos.h: Likewise.
24859 * config/sh/ushmedia.h: Likewise.
24860 * config/sh/linux.h: Likewise.
24861 * config/sh/sh.c: Likewise.
24862 * config/sh/superh.h: Likewise.
24863 * config/sh/elf.h: Likewise.
24864 * config/sh/sh4.md: Likewise.
24865 * config/sh/sh.h: Likewise.
24867 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24869 * config/pa/constraints.md: Adjust unused letters. Change "T"
24870 constraint to match_test floating_point_store_memory_operand().
24871 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24872 (base14_operand): New.
24873 (floating_point_store_memory_operand): New.
24874 (integer_store_memory_operand): Revise to use base14_operand and
24875 reg_plus_base_memory_operand.
24876 (move_dest_operand): Allow symbolic_memory_operands.
24877 (symbolic_memory_operand): Check for LO_SOM.
24878 (symbolic_operand): Change default case to break.
24879 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24880 CONST_DOUBLE values to be reloaded by putting them into memory when
24881 the destination is a floating point register.
24882 (movdf): Remove code to handle CONST_DOUBLE.
24884 (reload_indf_r1): New.
24885 (reload_insf_r1): New.
24886 Consistently use "Q" and "T" constraints with integer and floating
24887 point move instructions, respectively.
24888 (movdi): Remove FAIL.
24889 Change predicate for source operand unamed DImode move from
24890 general_operand to move_src_operand.
24891 (umulsidi3): Change predicate for destination operand to
24893 Likewise for similar unamed patterns.
24894 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24895 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24896 (hppa_legitimize_address): Simplify mask calculation.
24897 (pa_emit_move_sequence): Revised handling of secondary reloads from
24898 REG+D addresses for floating point loads and stores. Directly handle
24899 loading CONST0_RTX (mode) to a floating point register.
24900 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24901 to floating point registers. Don't restrict secondary reloads to
24902 floating point registers to integer modes. Revise some comments and
24904 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24905 (pa_legitimate_address_p): New.
24906 (pa_legitimize_reload_address): New.
24907 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24908 (STRICT_REG_OK_FOR_BASE_P): New.
24909 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24910 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24912 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
24913 Andrew Dixie <andrewd@gentrack.com>
24915 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24918 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24920 * expmed.c (extract_bit_field_1): Pass the full width of the
24921 structure to get_best_reg_extraction_insn.
24923 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24926 * configure.ac (use_cxa_atexit): Add AIX.
24927 * configure: Regenerate.
24929 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24931 2013-02-01 Jakub Jelinek <jakub@redhat.com>
24934 * final.c (need_profile_function): New variable.
24935 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24936 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24937 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24938 notes, targetm.asm_out.function_prologue doesn't emit anything,
24939 HAVE_prologue and profiler should be emitted before prologue,
24940 set need_profile_function instead of emitting it.
24941 (final_scan_insn): If need_profile_function, emit
24942 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24943 NOTE_INSN_FUNCTION_BEG note.
24945 2013-02-01 Richard Henderson <rth@redhat.com>
24947 * config/rs6000/rs6000.md (smulditi3): New.
24950 * config/alpha/alpha.md (umulditi3): New.
24952 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24954 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24955 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24957 2013-02-01 Richard Biener <rguenther@suse.de>
24959 PR tree-optimization/56113
24960 * tree-ssa-structalias.c (label_visit): Reduce work for
24961 single-predecessor nodes.
24963 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24965 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24966 range isn't testing for zero.
24968 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24970 PR middle-end/56113
24971 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24973 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24974 Nick Clifton <nickc@redhat.com>
24976 * config/v850/constraints.md (Q): Define as a memory constraint.
24977 * config/v850/predicates.md (label_ref_operand): New predicate.
24978 (e3v5_shift_operand): New predicate.
24979 (ior_operator): New predicate.
24980 * config/v850/t-v850: Add e3v5 multilib.
24981 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24982 (v850_gen_movdi): Prototype.
24983 * config/v850/v850.c: Add support for e3v5 architecture.
24984 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24986 (construct_save_jarl): Add e3v5 long JARL support.
24987 (v850_adjust_insn_length): New function. Adjust length of call
24988 insns when using e3v5 instructions.
24989 (v850_gen_movdi): New function: Generate instructions to move a
24991 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24992 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24993 (TARGET_USE_FPU): Enable for e3v5.
24994 (CONST_OK_FOR_W): New macro.
24995 (ADJUST_INSN_LENGTH): Define.
24996 * config/v850/v850.md (UNSPEC_LOOP): Define.
24997 (attr cpu): Add v850e3v5.
24998 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24999 (movdi): New pattern.
25000 (movdi_internal): New pattern.
25001 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25002 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25003 (cstoresf4): Likewise.
25004 (cstoredf4): Likewise.
25005 (insv): New pattern.
25006 (rotlso3_a): New pattern.
25007 (rotlsi3_b): New pattern
25008 (rotlsi3_v850e3v5): New pattern.
25009 (doloop_begin): New pattern.
25010 (fix_loop_counter): New pattern.
25011 (doloop_end): New pattern.
25012 (branch_normal): Add e3v5 long branch support.
25013 (branch_invert): Likewise.
25014 (branch_z_normal): Likewise.
25015 (branch_z_invert): Likewise.
25016 (branch_nz_normal): Likewise.
25017 (branch_nz_invert): Likewise.
25018 (call_internal_short): Add e3v5 register-indirect JARL support.
25019 (call_internal_long): Likewise.
25020 (call_value_internal_short): Likewise.
25021 (call_value_internal_long): Likewise.
25022 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25023 (mloop): New option.
25024 * config.gcc: Add support for configuring v840e3v5 target.
25025 * doc/invoke.texi: Document new v850 specific command line options.
25027 2013-01-31 Paul Koning <ni1d@arrl.net>
25031 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25032 children if parent is a class.
25033 (prune_unused_types_prune): Don't add DW_AT_declaration.
25035 2013-01-31 Richard Biener <rguenther@suse.de>
25037 PR tree-optimization/56157
25038 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25039 match up operand with SLP child.
25041 2013-01-31 Jason Merrill <jason@redhat.com>
25044 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25045 parameters the first time.
25046 (gen_scheduled_generic_parms_dies): Check completeness here.
25048 2013-01-31 Richard Biener <rguenther@suse.de>
25050 PR middle-end/53073
25051 * common.opt (faggressive-loop-optimizations): New flag,
25052 enabled by default.
25053 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25054 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25055 infer_loop_bounds_from_undefined by it.
25057 2013-01-31 Richard Biener <rguenther@suse.de>
25059 PR tree-optimization/56150
25060 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25061 visit virtual operands.
25062 (find_uses_to_rename_bb): Likewise.
25064 2013-01-31 Richard Biener <rguenther@suse.de>
25066 PR tree-optimization/56150
25067 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25068 mixed store non-store stmts.
25070 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25073 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25074 LIBASAN_EARLY_SPEC is defined.
25075 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25076 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25078 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25081 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25082 invalid args instead of ICEing on it.
25083 (ix86_valid_target_attribute_tree): Return error_mark_node if
25084 ix86_valid_target_attribute_inner_p failed.
25085 (ix86_valid_target_attribute_p): Return false only if
25086 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25087 target("default") attribute.
25088 (sorted_attr_string): Change argument from const char * to tree,
25089 merge in all target attribute arguments rather than just one.
25090 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25091 instead of free. Avoid using strcat.
25092 (ix86_mangle_function_version_assembler_name): Mangle
25093 target("default") as if no target attribute is present. Adjust
25094 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25095 instead of xmalloc and XDELETEVEC instead of free.
25096 (ix86_function_versions): Don't return true if one of the decls
25097 doesn't have target attribute. If they don't and one of the decls
25098 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25099 sorted_attr_string caller. Use XDELETEVEC instead of free.
25100 (ix86_supports_function_versions): Remove.
25101 (make_name): Fix up formatting.
25102 (make_dispatcher_decl): Remove resolver_name and its initialization.
25103 Avoid leaking memory.
25104 (is_function_default_version): Return true if there is
25105 target("default") attribute rather than no target attribute at all.
25106 (make_resolver_func): Avoid leaking memory.
25107 (ix86_generate_version_dispatcher_body): Likewise.
25108 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25109 * target.def (supports_function_versions): Remove.
25110 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25111 * doc/tm.texi: Regenerated.
25113 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25115 PR rtl-optimization/56144
25116 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25117 for values with side effects.
25119 2013-01-30 Richard Biener <rguenther@suse.de>
25121 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25122 (sparseset_pop): Likewise.
25123 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25124 to be able to use quick_push in the worker loop.
25126 2013-01-30 Marek Polacek <polacek@redhat.com>
25128 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25130 2013-01-30 Richard Biener <rguenther@suse.de>
25133 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25135 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25137 PR tree-optimization/56064
25138 * fixed-value.c (fixed_from_double_int): New function.
25139 * fixed-value.h (fixed_from_double_int): New prototype.
25140 (const_fixed_from_double_int): New static inline function.
25141 * fold-const.c (native_interpret_fixed): New static function.
25142 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25143 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25144 (native_encode_fixed): New static function.
25145 (native_encode_expr) <FIXED_CST>: Use it.
25146 (native_interpret_int): Move double_int worker code to...
25147 * double-int.c (double_int::from_buffer): ...this new static method.
25148 * double-int.h (double_int::from_buffer): Prototype it.
25150 2013-01-30 Richard Biener <rguenther@suse.de>
25152 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25153 New pointer-map and obstack.
25154 (init_alias_vars): Allocate pointer-map and obstack.
25155 (delete_points_to_sets): Free them.
25156 (find_what_var_points_to): Cache result.
25157 (find_what_p_points_to): Adjust for changed interface of
25158 find_what_var_points_to.
25159 (compute_points_to_sets): Likewise.
25160 (ipa_pta_execute): Likewise.
25162 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25164 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25165 * configure: Regenerate.
25166 * config.in: Regenerate.
25167 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25168 #nobits/#progbits if supported.
25170 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25173 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25174 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25175 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25177 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25179 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25180 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25182 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25184 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25185 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25187 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25189 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25191 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25192 * config/arm/cortex-a7.md: New bypasses using
25193 arm_mac_accumulator_is_result.
25195 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25197 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25198 (cortex_a7_neon_mla): Likewise.
25199 (cortex_a7_fpfmad): New reservation.
25200 (cortex_a7_fpmacs): Use ffmas and update required units.
25201 (cortex_a7_fpmuld): Update required units and latency.
25202 (cortex_a7_fpmacd): Likewise.
25203 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25204 (cortex_a7_neon). Likewise.
25205 (bypass) Update participating units.
25207 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25209 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25210 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25212 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25213 (vfp_fmul): Use ffmad.
25214 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25215 (cortex_r4_fmacd): Use ffmad.
25216 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25217 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25218 (cortex_a9_fmacd): Use ffmad.
25219 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25220 (cortex_a8_vfp_macd): Use ffmad.
25221 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25222 (cortex_a5_fpmacd): Use ffmad.
25223 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25224 (cortex_a15_vfp_macd): Use ffmad.
25225 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25227 2013-01-29 Jason Merrill <jason@redhat.com>
25230 * varasm.c (default_assemble_visibility): Don't warn about
25231 visibility on artificial decls.
25233 2013-01-29 Richard Biener <rguenther@suse.de>
25235 PR tree-optimization/56113
25236 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25238 (label_visit): Free duplicate bitmaps and record the leader instead.
25239 (perform_var_substitution): Adjust.
25241 2013-01-29 Richard Biener <rguenther@suse.de>
25243 PR tree-optimization/55270
25244 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25245 the CFG, schedule loops for fixup.
25247 2013-01-29 Nick Clifton <nickc@redhat.com>
25249 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25252 2013-01-28 Leif Ekblad <leif@rdos.net>
25254 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25255 * config/i386/i386.h (TARGET_RDOS): New macro.
25256 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25257 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25258 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25259 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25260 DEFAULT_LARGE_SECTION_THRESHOLD.
25261 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25262 * config/i386/rdos.h: New file.
25263 * config/i386/rdos64.h: New file.
25265 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25268 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25271 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25273 PR rtl-optimization/56117
25274 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25275 call cselib_lookup_from_insn on the MEM before calling
25276 add_insn_mem_dependence.
25278 2013-01-28 Richard Biener <rguenther@suse.de>
25280 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25281 to a stmt that didn't have one.
25282 (copy_phis_for_bb): Likewise for PHI arguments.
25283 (copy_debug_stmt): Likewise for debug stmts.
25285 2013-01-28 Richard Biener <rguenther@suse.de>
25287 PR tree-optimization/56034
25288 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25289 (partition_builtin_p): Adjust.
25290 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25291 it is the last partition.
25292 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25293 up the vertex for the definition.
25294 (classify_partition): Classify whether a partition is a
25295 PKIND_REDUCTION, thus has uses outside of the loop.
25296 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25297 Merge all PKIND_REDUCTION partitions into the last partition.
25298 (tree_loop_distribution): Seed partitions from reductions as well.
25300 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25302 PR tree-optimization/56125
25303 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25304 pow(x,c) into sqrt(x) * powi(x, n/2) or
25305 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25306 optimizing for size.
25307 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25308 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25311 PR tree-optimization/56094
25312 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25313 to UNKNOWN_LOCATION while gimplifying expr.
25315 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25318 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25319 operand 0 in movabs insn template for -masm=intel asm alternative.
25320 (*movabs<mode>_2): Ditto for operand 1.
25322 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25325 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25328 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25330 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25333 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25335 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25336 Undef to avoid warning.
25338 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25340 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25341 * configure: Regenerate.
25343 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25345 PR tree-optimization/56098
25346 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25347 for stmts with volatile ops.
25348 (cond_store_replacement): Don't optimize if assign has volatile ops.
25349 (cond_if_else_store_replacement_1): Don't optimize if either
25350 then_assign or else_assign have volatile ops.
25351 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25354 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25356 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25358 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25360 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25361 missing ':' in asm example.
25363 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25365 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25366 entries into lane and laneq entries.
25367 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25368 Remove AdvSIMD scalar modes.
25369 (aarch64_sq<r>dmulh_laneq<mode>): New.
25370 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25372 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25373 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25374 * config/aarch64/iterators.md (VCOND): New.
25377 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25380 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25381 Add NULL LIBNAME argument to existing definitions.
25382 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25383 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25384 * config/avr/avr.c (DEF_BUILTIN): Same.
25385 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25386 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25387 implementation is available (DECL_ASSEMBLER_NAME is set).
25388 (avr_fold_absfx): New static function.
25389 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25390 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25391 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25392 AVR_BUILTIN_ABSLLK.
25393 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25394 (abshk, absk, abslk, absllk): Provide as static inline functions.
25396 2013-01-25 Marek Polacek <polacek@redhat.com>
25398 PR tree-optimization/56035
25399 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25401 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25403 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25404 (*movtf_internal_rex64): Add (!o,C) alternative
25405 (*movxf_internal_rex64): Ditto.
25406 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25408 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25410 * doc/invoke.texi: fix typo.
25411 * doc/objc.texi: fix typo.
25413 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25415 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25416 for the first two alternatives.
25418 2013-01-24 Diego Novillo <dnovillo@google.com>
25420 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25421 (ggc-zone.o): Remove.
25422 * configure.ac: Remove option --with-gc.
25423 * configure: Re-generate.
25424 * doc/install.texi: Remove documentation for --with-gc.
25425 * gengtype.c (write_enum_defn): Remove. Update all users.
25426 (write_Types_process_field): Remove generation of gt_e_* argument.
25427 (output_type_enum): Remove. Update all users.
25428 (write_enum_defn): Remove. Update all users.
25429 (enum alloc_zone): Remove. Update all users.
25430 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25431 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25432 Update all callers.
25433 (struct ptr_data): Remove field TYPE. Update all users.
25434 (gt_pch_note_object): Remove argument TYPE. Update all users.
25435 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25437 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25438 (struct alloc_zone): Remove.
25439 (ggc_internal_alloc_zone_stat): Remove.
25440 (ggc_internal_cleared_alloc_zone_stat): Remove.
25441 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25442 (ggc_pch_count_object): Remove last argument. Update all users.
25443 (ggc_pch_alloc_object): Remove last argument. Update all users.
25444 (struct alloc_zone): Remove.
25445 * ggc-zone.c: Remove.
25446 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25447 (struct alloc_zone): Remove.
25448 (ggc_alloc_typed_stat): Remove.
25449 (ggc_alloc_typed): Remove.
25450 (ggc_splay_alloc): Remove first argument.
25451 (rtl_zone): Remove. Update all users.
25452 (tree_zone): Remove. Update all users.
25453 (tree_id_zone): Remove. Update all users.
25454 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25455 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25456 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25457 * tree-ssanames.c: Remove references to zone allocator in comments.
25459 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25461 * config/avr/avr.c (avr_out_fract): Make register numbers that
25462 might be outside of source operand signed.
25464 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25466 * config/i386/constraints.md (Yf): New constraint.
25467 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25468 of f constraint to conditionaly disable x87 register preferences.
25469 (*movdf_internal): Ditto.
25470 (*movsf_internal): Ditto.
25472 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25474 PR inline-asm/55934
25475 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25476 that have operands with impossible constraints.
25477 Add a FIXME for a speed-up opportunity.
25478 * lra-constraints.c (process_alt_operands): Verify that a class
25479 selected from constraints on asms is valid for the operand mode.
25480 (curr_insn_transform): Remove incorrect comment.
25482 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25484 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25485 TOC operand is a valid symbol ref in the constant pool.
25487 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25489 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25491 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25494 * config/avr/stdfix.h: New file.
25495 * t-avr (stdfix-gcc.h): New rule to build it.
25496 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25498 2013-01-23 Kostya Serebryany <kcc@google.com>
25500 * config/darwin.h: remove dependency on
25501 CoreFoundation (asan on Mac OS).
25503 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25506 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25507 instead of cmpdi_operand for first comparison operand.
25508 Don't assert that comparison operands aren't both constants.
25510 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25512 * doc/install.texi (Downloading the Source): Update references to
25513 downloading separate components.
25515 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25517 * doc/extend.texi (__int128): Improve grammar.
25519 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25522 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25523 alternative to (o,r).
25524 (*movdi_internal_rex64): Remove (!o,n) alternative.
25525 (DImode immediate->memory splitter): Remove.
25526 (DImode immediate->memory peephole2): Remove.
25527 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25528 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25529 alternative to (!o,*r).
25530 (*movtf_internal_sse): New pattern.
25531 (*movxf_internal_rex64): New pattern.
25532 (*movxf_internal): Disable for TARGET_64BIT.
25533 (*movdf_internal_rex64): Remove (!o,F) alternative.
25535 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25537 PR middle-end/56074
25538 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25539 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25540 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25541 stmt locations where LOCATION_LOCUS of the stmt location is
25542 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25545 * config/i386/i386.md (UNSPEC_STOS): New.
25546 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25547 *strsetqi_1): Add UNSPEC_STOS.
25549 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25552 * doc/invoke.texi: Remove left over -Wsynth example.
25554 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25556 PR tree-optimization/56051
25557 * fold-const.c (fold_binary_loc): Don't fold
25558 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25559 a narrowing conversion, or widening conversion from signed
25562 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25564 PR rtl-optimization/56023
25565 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25566 dependent on debug instruction.
25568 2013-01-21 Martin Jambor <mjambor@suse.cz>
25570 PR middle-end/56022
25571 * function.c (allocate_struct_function): Call
25572 invoke_set_current_function_hook earlier.
25574 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25576 * reload1.c (init_reload): Only initialize reload_obstack
25577 during the first call.
25579 2013-01-21 Marek Polacek <polacek@redhat.com>
25581 * cfgloop.c (verify_loop_structure): Fix up grammar.
25583 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25585 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25586 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25588 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25591 * config/arm/marvell-pj4.md: Update copyright year.
25592 Fix up use of alu to alu_reg and simple_alu_imm.
25594 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25596 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25598 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25601 * lra-constraints.c (curr_insn_transform): Don't reuse original
25602 insn for secondary memory move when memory mode should be different.
25604 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25606 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25607 atomic_storedi_1): New patterns.
25609 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25611 btver2 pipeline descriptions.
25612 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25614 * config/i386/i386.md (btver2_decode): New type attributes.
25615 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25617 * config/i386/btver2.md: New file describing btver2 pipelines.
25619 2013-01-19 Andrew Pinski <apinski@cavium.com>
25621 PR tree-optimization/52631
25622 * tree-ssa-sccvn (visit_use): Before looking up the original
25623 statement, try looking up the simplified expression.
25625 2013-01-19 Anthony Green <green@moxielogic.com>
25627 * config/moxie/moxie.c (moxie_expand_prologue): Set
25628 current_function_static_stack_size.
25630 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25632 PR tree-optimization/56029
25633 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25634 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25636 2013-01-18 Sharad Singhai <singhai@google.com>
25638 PR tree-optimization/55995
25639 * dumpfile.c (dump_loc): Print location only if available.
25640 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25642 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25645 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25646 secondary memory move.
25647 (inherit_reload_reg): Use rclass instead of cl for
25648 check_secondary_memory_needed_p.
25650 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25652 PR middle-end/56015
25653 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25654 the case where writing real complex part of target modifies op1.
25656 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25658 * config/aarch64/aarch64-simd.md
25659 (aarch64_vcond_internal<mode>): Handle unordered cases.
25660 * config/aarch64/iterators.md (v_cmp_result): New.
25662 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25663 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25665 * config/arm/marvell-pj4.md: New file.
25666 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25667 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25668 (generic_vfp): Likewise.
25669 * config/arm/arm-cores.def: Add marvell-pj4.
25670 * config/arm/arm-tune.md: Regenerate.
25671 * config/arm/arm-tables.opt: Regenerate.
25672 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25673 * doc/invoke.texi: Document marvell-pj4.
25675 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25677 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25679 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25684 * var-tracking.c (negative_power_of_two_p): New.
25685 (global_get_addr_cache, local_get_addr_cache): New.
25686 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25687 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25689 (vt_stack_offset_p): Remove.
25690 (vt_canon_true_dep): Always canonicalize loc's address.
25691 (clobber_overlapping_mems): Make sure we have a MEM.
25692 (local_get_addr_clear_given_value): New.
25693 (val_reset): Clear local cached entries.
25694 (compute_bb_dataflow): Create and release the local cache.
25695 Disable duplicate MEMs clobbering.
25696 (emit_notes_in_bb): Clobber MEMs likewise.
25697 (vt_emit_notes): Create and release the local cache.
25698 (vt_initialize, vt_finalize): Create and release the global
25699 cache, respectively.
25700 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25702 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25704 PR libmudflap/53359
25705 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25706 not found in the symtab.
25708 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25711 PR rtl-optimization/55547
25712 PR rtl-optimization/53827
25715 * alias.c (offset_overlap_p): New, factored out of...
25716 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25717 the conservative special case for symbolic constants. Don't
25718 adjust zero sizes on alignment.
25720 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25722 PR rtl-optimization/52573
25723 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25724 REG_UNUSED for the same register.
25726 2013-01-17 Richard Biener <rguenther@suse.de>
25727 Marek Polacek <polacek@redhat.com>
25729 PR rtl-optimization/55833
25730 * loop-unswitch.c (unswitch_loops): Move loop verification...
25731 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25732 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25733 Set it to true when we're removing a loop from hierarchy tree in
25734 an irreducible region.
25735 (fix_bb_placements): Adjust caller.
25736 (fix_loop_placements): Likewise.
25738 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25740 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25741 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25742 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25743 Remove ID. Adjust comments.
25744 * config/avr/avr-c.c (avr_builtin_name): Remove.
25745 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25746 * config/avr/avr.c (avr_tolower): New static function.
25747 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25748 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25749 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25752 2013-01-17 Jan Hubicka <jh@suse.cz>
25754 PR tree-optimization/55273
25755 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25757 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25760 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25761 store through atomic_store<mode>_1.
25762 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25764 2013-01-17 Martin Jambor <mjambor@suse.cz>
25766 PR tree-optimizations/55264
25767 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25768 for virtual methods.
25769 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25770 virtual methods before inlining is over.
25771 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25773 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25776 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25778 PR rtl-optimization/56005
25779 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25780 pending reads for prefetch.
25782 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25784 * config/aarch64/aarch64.md
25785 (*cstoresi_neg_uxtw): New pattern.
25786 (*cmovsi_insn_uxtw): New pattern.
25787 (*<optab>si3_uxtw): New pattern.
25788 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25789 (*<optab>si3_insn_uxtw): New pattern.
25790 (*bswapsi2_uxtw): New pattern.
25792 2013-01-16 Richard Biener <rguenther@suse.de>
25794 * tree-inline.c (tree_function_versioning): Remove set but
25795 never used variable.
25797 2013-01-16 Richard Biener <rguenther@suse.de>
25799 PR tree-optimization/55964
25800 * tree-flow.h (rename_variables_in_loop): Remove.
25801 (rename_variables_in_bb): Likewise.
25802 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25803 (copy_loop_before): Adjust and delete update-ssa status.
25804 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25805 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25806 (rename_variables_in_loop): Remove.
25807 (slpeel_update_phis_for_duplicate_loop): Likewise.
25808 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25809 use available cfg machinery instead of duplicating it.
25810 Update PHI nodes and perform poor-mans SSA update here.
25811 (slpeel_tree_peel_loop_to_edge): Adjust.
25813 2013-01-16 Richard Biener <rguenther@suse.de>
25815 PR tree-optimization/54767
25816 PR tree-optimization/53465
25817 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25818 (vrp_visit_phi_node): For PHI arguments coming via backedges
25819 drop all symbolical range information.
25820 (execute_vrp): Compute backedges.
25822 2013-01-16 Richard Biener <rguenther@suse.de>
25824 * doc/install.texi: Update CLooG and ISL requirements to
25827 2013-01-16 Christian Bruel <christian.bruel@st.com>
25830 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25831 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25832 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25834 2013-01-16 DJ Delorie <dj@redhat.com>
25836 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25837 (UNSPECV_SP_SWITCH_E): New.
25838 (sp_switch_1): Change to an unspec.
25839 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25842 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25844 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25845 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25846 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25847 (expand_mem_thread_fence): Ditto.
25848 (expand_mem_signal_fence): Ditto.
25849 (expand_atomic_load): Ditto.
25850 (expand_atomic_store): Ditto.
25852 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25854 PR rtl-optimization/55547
25855 PR rtl-optimization/53827
25858 * alias.c (memrefs_conflict_p): Set sizes to negative after
25861 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25864 * function.c (thread_prologue_and_epilogue_insns): Always
25865 add crtl->drap_reg to set_up_by_prologue.set, even if
25866 stack_realign_drap is false.
25868 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25870 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25871 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25872 *call): Fix indention.
25874 2013-01-15 Tom de Vries <tom@codesourcery.com>
25877 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25880 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25882 PR rtl-optimization/55153
25883 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25885 2013-01-15 Martin Jambor <mjambor@suse.cz>
25887 PR tree-optimization/55920
25888 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25889 accesses as grp_to_be_debug_replaced.
25891 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25893 PR tree-optimization/55920
25894 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25895 there is non-useless type conversion needed from debug rhs to lhs,
25896 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25898 2013-01-15 Joseph Myers <joseph@codesourcery.com>
25899 Mikael Pettersson <mikpe@it.uu.se>
25902 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25904 (ASM_OUTPUT_CASE_LABEL): Remove.
25905 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25906 * final.c (shorten_branches): Update alignment of labels before
25907 jump tables if CASE_VECTOR_SHORTEN_MODE.
25909 2013-01-15 Richard Biener <rguenther@suse.de>
25912 * system.h: Do not include gmp.h for building host tools.
25914 2013-01-15 Richard Biener <rguenther@suse.de>
25916 PR middle-end/55882
25917 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25918 account for bitpos when computing alignment.
25920 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25922 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25923 (ix86_target_macros_internal): Likewise.
25925 * config/i386/i386.c (m_CORE2I7): Removed.
25926 (m_CORE_HASWELL): New macro.
25927 (m_CORE_ALL): Likewise.
25928 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25929 (initial_ix86_arch_features): Likewise.
25930 (processor_target_table): Initializations for Core avx2.
25931 (cpu_names): New names "core-avx2".
25932 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25933 PROCESSOR_CORE_HASWELL.
25934 (ix86_issue_rate): New case.
25935 (ia32_multipass_dfa_lookahead): Likewise.
25936 (ix86_sched_init_global): Likewise.
25938 * config/i386/i386.h (TARGET_HASWELL): New macro.
25939 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25940 (processor_type): New PROCESSOR_HASWELL.
25942 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25944 PR tree-optimization/55955
25945 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25946 *SHIFT_EXPR and *ROTATE_EXPR codes.
25948 PR tree-optimization/48766
25949 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25950 -ftrapv disable -fwrapv.
25952 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25955 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25956 etc. to 1 and not to __flash.
25957 Use LL suffix for __INT24_MAX__ with -mint8.
25958 Use ULL suffix for __UINT24_MAX__ with -mint8.
25960 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25962 * config/avr/avr-arch.h
25963 (struct base_arch_s): Use typedef avr_arch_t instead.
25964 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25965 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25966 * config/avr/avr.c: Same.
25967 * config/avr/avr-devices.c: Same.
25968 * config/avr/driver-avr.c: Same.
25969 * config/avr/gen-avr-mmcu-texi.c: Same.
25970 * config/avr/avr-mcus.def: Adjust comment.
25972 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25974 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25975 * config/aarch64/iterators.md (VALLDI): New.
25977 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
25978 Andi Kleen <ak@linux.intel.com>
25981 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25982 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25985 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25987 * config/avr/avr-stdint.h: Remove trailing blanks.
25988 * config/avr/avr-log.h: Same.
25989 * config/avr/avr-arch.h: Same.
25990 * config/avr/avr-devices.c: Same.
25991 * config/avr/avr-dimode.md: Same.
25992 * config/avr/predicates.md: Same.
25993 * config/avr/avr-c.c: Same. And fix typo.
25995 * config/avr/avr-protos.h: Same. And:
25996 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25997 (init_cumulative_args): Rename to avr_init_cumulative_args.
25998 (expand_prologue): Rename to avr_expand_prologue.
25999 (expand_epilogue): Rename to avr_expand_epilogue.
26000 (adjust_insn_length): Rename to avr_adjust_insn_length.
26001 (notice_update_cc): Rename to avr_notice_update_cc.
26002 (final_prescan_insn): Rename to avr_final_prescan_insn.
26003 * config/avr/avr.c: Same.
26004 * config/avr/avr.h: Same.
26005 * config/avr/avr.md: Remove trailing blanks.
26006 (prologue): Use avr_expand_prologue.
26007 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26009 2013-01-14 Richard Biener <rguenther@suse.de>
26011 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26012 verify_location, collect_subblocks): New functions.
26013 (verify_gimple_in_cfg): Verify that locations only reference
26014 BLOCKs in the functions BLOCK tree.
26016 2013-01-14 Richard Biener <rguenther@suse.de>
26018 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26020 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26022 (insert_out_of_ssa_copy_on_edge): Likewise.
26023 (rewrite_close_phi_out_of_ssa): Likewise.
26024 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26026 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26027 propagated constants.
26028 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26031 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26033 * config/avr/avr-modes.def: Add GPL copyright notice.
26035 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26037 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26038 MEMMODEL_MASK to determine memory model.
26039 (atomic_store<mode>): Ditto from operands[2].
26040 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26042 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26045 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26046 (fold_gimple_assign): Don't call unshare_expr here.
26047 (fold_ctor_reference): Call unshare_expr.
26049 2013-01-13 Terry Guo <terry.guo@arm.com>
26051 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26052 * doc/fragments.texi: Document MULTILIB_REUSE.
26053 * gcc.c (multilib_reuse): New internal spec.
26054 (set_multilib_dir): Also search multilib from multilib_reuse.
26055 * genmultilib (tmpmultilib3): Refactor code.
26056 (tmpmultilib4): Ditto.
26057 (multilib_reuse): New multilib argument.
26059 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26061 * Makefile.in: Update copyright.
26063 2013-01-12 Tom de Vries <tom@codesourcery.com>
26065 PR middle-end/55890
26066 * calls.c (expand_call): Check if arg_nr is valid.
26068 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26070 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26071 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26072 documentation. Add missing '__' in front of
26073 __builtin_ia32_packssdw256.
26075 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26078 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26079 NO_REGS for larl operands.
26080 (s390_reload_larl_operand): Use s390_load_address instead of
26083 2013-01-11 Richard Biener <rguenther@suse.de>
26085 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26086 (verify_node_sharing): ... here.
26087 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26089 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26091 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26092 Substitute TREECHECKING.
26093 * configure: Regenerate.
26094 * Makefile.in (TREECHECKING): New.
26096 2013-01-11 Richard Guenther <rguenther@suse.de>
26098 PR tree-optimization/44061
26099 * tree-vrp.c (extract_range_basic): Compute zero as
26100 value-range for __builtin_constant_p of function parameters.
26102 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26104 Update copyright years.
26106 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26108 PR rtl-optimization/55672
26109 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26110 const to be eliminable.
26112 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26114 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26115 * configure: Regenerate.
26117 2013-01-10 Richard Biener <rguenther@suse.de>
26119 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26121 2013-01-10 Richard Biener <rguenther@suse.de>
26124 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26125 locations for virtual PHI arguments.
26126 (rewrite_update_phi_arguments): Likewise.
26128 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26130 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26133 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26135 PR tree-optimization/55921
26136 * tree-complex.c (expand_complex_asm): New function.
26137 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26139 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26142 * config/s390/s390.c (s390_symref_operand_p)
26143 (s390_loadrelative_operand_p): Merge the two functions.
26144 (s390_check_qrst_address, print_operand_address): Add parameters
26145 to s390_loadrelative_operand_p invokation.
26146 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26147 (s390_reload_larl_operand, s390_secondary_reload): Use
26148 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26149 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26151 2013-01-09 Mike Stump <mikestump@comcast.net>
26153 * dse.c (record_store): Remove unnecessary assert.
26155 2013-01-09 Jan Hubicka <jh@suse.cz>
26157 PR tree-optimization/55569
26158 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26159 * cfgloop.h (scale_loop_profile): Likewise.
26161 2013-01-09 Jan Hubicka <jh@suse.cz>
26164 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26166 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26168 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26170 PR middle-end/55114
26171 * expr.h (maybe_emit_group_store): Declare.
26172 * expr.c (maybe_emit_group_store): New function.
26173 * builtins.c (expand_builtin_int_roundingfn): Call it.
26174 (expand_builtin_int_roundingfn_2): Likewise.
26176 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26178 PR rtl-optimization/55829
26179 * lra-constraints.c (match_reload): Add code for absent output.
26180 (curr_insn_transform): Add code for reloads of matched inputs
26183 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26185 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26186 attribute of movddup insn to DF.
26187 (*vec_interleave_lowv2df): Ditto.
26188 (vec_dupv2df): Ditto.
26190 2013-01-09 Jan Hubicka <jh@suse.cz>
26192 PR tree-optimiation/55875
26193 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26194 EVERY_ITERATION parameter.
26195 (number_of_iterations_exit): Check if exit is executed every iteration.
26196 (idx_infer_loop_bounds): Similarly here.
26197 (n_of_executions_at_most): Simplify
26198 to only test for cases where statement is dominated by the
26199 particular bound; handle correctly the "postdominance" test.
26200 (scev_probably_wraps_p): Use max loop iterations info
26201 as a global bound first.
26203 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26204 Nick Clifton <nickc@redhat.com>
26206 * config/v850/v850.md (cbranchsf4): New pattern.
26207 (cstoresf4): New pattern.
26208 (cbranchdf4): New pattern.
26209 (cstoredf4): New pattern.
26210 (movsicc): Disallow floating point comparisons.
26211 (cmpsf_le_insn): Fix order of operators.
26212 (cmpsf_lt_insn): Likewise.
26213 (cmpsf_eq_insn): Likewise.
26214 (cmpdf_le_insn): Likewise.
26215 (cmpdf_lt_insn): Likewise.
26216 (cmpdf_eq_insn): Likewise.
26217 (cmpsf_ge_insn): Use LE comparison.
26218 (cmpdf_ge_insn): Likewise.
26219 (cmpsf_gt_insn): Use LT comparison.
26220 (cmpdf_gt_insn): Likewise.
26221 (cmpsf_ne_insn): Delete pattern.
26222 (cmpdf_ne_insn): Delete pattern.
26223 * config/v850/v850.c (v850_gen_float_compare): Use
26224 gen_cmpdf_eq_insn for NE comparison.
26225 (v850_float_z_comparison_operator)
26226 (v850_float_nz_comparison_operator): Move from here ...
26227 * config/v850/predicates.md: ... to here. Move GT and GE
26228 comparisons into v850_float_z_comparison_operator.
26229 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26231 (v850_float_nz_comparison_operator): Likewise.
26233 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26235 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26236 with calls to gen_insvsi/gen_insvdi.
26238 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26240 * config/i386/i386.c (initial_ix86_tune_features): Set up
26241 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26243 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26244 Jakub Jelinek <jakub@redhat.com>
26246 PR tree-optimization/48189
26247 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26248 If nitercst is 0, don't predict the exit edge.
26250 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26252 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26253 in asm_fprintf with reg_names.
26254 (aarch64_print_operand_address): Likewise.
26255 (aarch64_return_addr): Likewise.
26256 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26258 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26260 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26261 (INT_U6_BITS): Likewise.
26262 * config/pa/predicates.md (uint6_operand): New predicate.
26263 (shift5_operand, shift6_operand): Likewise.
26264 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26266 (lshrdi3): Use shift6_operand.
26267 (shrpsi4, shrpdi4): New insn patterns.
26268 (extzv): Delete expander.
26269 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26270 predicates in unamed zero extract patterns. Tighten common constraint.
26271 (extv): Delete expander.
26272 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26273 predicates in unamed sign extract patterns. Tighten common constraint.
26274 (insv): Delete expander.
26275 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26276 predicates in unamed insert patterns. Tighten common constraint.
26277 Change uint32_operand predicate to uint6_operand predicate in unamed
26278 DImode pattern to insert constant values of type 1...1xxxx.
26280 2013-01-04 Jan Hubicka <jh@suse.cz>
26282 PR tree-optimization/55823
26283 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26286 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26287 Uros Bizjak <ubizjak@gmail.com>
26289 PR rtl-optimization/55845
26290 * df-problems.c (can_move_insns_across): Stop scanning at
26291 volatile_insn_p source instruction or give up if
26292 across_from .. across_to range contains any volatile_insn_p
26295 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26297 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26298 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26300 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26301 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26303 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26306 * asan.c (asan_clear_shadow): New function.
26307 (asan_emit_stack_protection): Use it.
26309 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26311 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26312 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26313 with tab instead of space.
26315 2013-01-08 Nick Clifton <nickc@redhat.com>
26317 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26318 register bank 0 at the start of an interrupt handler.
26319 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26322 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26324 * config/aarch64/aarch64-simd.md
26325 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26326 (aarch64_simd_bsl): Likewise.
26327 (aarch64_vcond_internal<mode>): Likewise.
26328 (vcond<mode><mode>): Likewise.
26329 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26330 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26332 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26334 * config/aarch64/aarch64-builtins.c
26335 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26337 2013-01-08 Martin Jambor <mjambor@suse.cz>
26340 * tree-sra.c (analyze_access_subtree): Return true also after
26341 potentially creating a debug-only replacement.
26343 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26345 PR middle-end/55890
26346 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26348 PR tree-optimization/54120
26349 * tree-vrp.c (range_fits_type_p): Don't allow
26350 src_precision < precision from signed vr to unsigned_p
26351 if vr->min or vr->max is negative.
26352 (simplify_float_conversion_using_ranges): Test can_float_p
26353 against CODE_FOR_nothing.
26355 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26356 Richard Biener <rguenther@suse.de>
26358 PR middle-end/55851
26359 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26360 types instead of just INTEGER_TYPE types.
26362 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26364 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26367 2013-01-07 Steve Ellcey <sellcey@mips.com>
26370 * config/mips/mips.opt: Change mad to mmad to match documentation.
26372 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26375 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26376 .progmemx.data now.
26378 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26381 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26382 (avr_addrspace_t): Add .section_name field.
26383 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26385 (avr_addrspace): Same. Initialize .section_name. Remove last
26386 NULL entry. Put __memx into .progmemx.data.
26387 (progmem_section_prefix): Remove.
26388 (avr_asm_init_sections): No need to initialize progmem_section.
26389 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26390 section name prefix.
26391 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26392 retrieve the progmem section.
26393 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26394 boundary to run over avr_addrspace[].
26395 (avr_register_target_pragmas): Ditto.
26397 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26399 * varasm.c (output_constant_def_contents): For asan_protect_global
26400 protected strings, adjust DECL_ALIGN if needed, before testing for
26402 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26403 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26405 (output_object_block): For asan protected decls, emit asan padding
26406 after their contents.
26407 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26408 (asan_finish_file): Test it here instead.
26410 2013-01-07 Nick Clifton <nickc@redhat.com>
26411 Matthias Klose <doko@debian.org>
26412 Doug Kwan <dougkwan@google.com>
26413 H.J. Lu <hongjiu.lu@intel.com>
26416 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26418 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26420 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26422 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26424 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26426 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26429 * doc/install.texi (Cross-Compiler-Specific Options): Document
26432 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26434 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26435 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26436 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26437 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26438 vqmovun_high_s64): Fix source operand number and update copyright.
26440 2013-01-07 Richard Biener <rguenther@suse.de>
26442 PR middle-end/55890
26443 * gimple.h (gimple_call_builtin_p): New overload.
26444 * gimple.c (validate_call): New function.
26445 (gimple_call_builtin_p): Likewise.
26446 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26447 Use gimple_call_builtin_p.
26448 (find_func_clobbers): Likewise.
26449 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26450 (strlen_optimize_stmt): Likewise.
26452 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26454 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26455 (vld1q_dup_*): Likewise.
26456 (vld1_*): Likewise.
26457 (vld1q_*): Likewise.
26458 (vld1_lane_*): Likewise.
26459 (vld1q_lane_*): Likewise.
26461 2013-01-07 Richard Biener <rguenther@suse.de>
26463 * lto-streamer.h (LTO_minor_version): Bump to 2.
26465 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26467 * config/aarch64/aarch64-protos.h
26468 (aarch64_const_double_zero_rtx_p): Rename to...
26469 (aarch64_float_const_zero_rtx_p): ...this.
26470 (aarch64_float_const_representable_p): New.
26471 (aarch64_output_simd_mov_immediate): Likewise.
26472 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26473 move immediate case.
26474 * config/aarch64/aarch64.c
26475 (aarch64_const_double_zero_rtx_p): Rename to...
26476 (aarch64_float_const_zero_rtx_p): ...this.
26477 (aarch64_print_operand): Allow printing of new constants.
26478 (aarch64_valid_floating_const): New.
26479 (aarch64_legitimate_constant_p): Check for valid floating-point
26481 (aarch64_simd_valid_immediate): Likewise.
26482 (aarch64_vect_float_const_representable_p): New.
26483 (aarch64_float_const_representable_p): Likewise.
26484 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26485 (aarch64_output_simd_mov_immediate): New.
26486 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26487 (*movdf_aarch64): Likewise.
26488 * config/aarch64/constraints.md (Ufc): New.
26489 (Y): call aarch64_float_const_zero_rtx.
26490 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26492 2013-01-07 Richard Biener <rguenther@suse.de>
26494 PR tree-optimization/55888
26495 PR tree-optimization/55862
26496 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26497 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26498 not if it is contained therein.
26500 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26502 * config/avr/t-avr: Typo.
26504 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26507 * config/avr/t-avr: Don't automatically rebuild
26508 $(srcdir)/config/avr/t-multilib
26509 $(srcdir)/config/avr/avr-tables.opt
26510 $(srcdir)/doc/avr-mmcu.texi
26511 (avr-mcus): New phony target to build them on request.
26512 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26513 * avr/avr-mcus.def: Adjust comments.
26515 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26517 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26519 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26521 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26523 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26525 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26527 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26529 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26530 to generate profiling.
26531 * config/rs6000/aix64.h (LIB_SPEC): Same.
26533 2013-01-04 Andrew Pinski <apinski@cavium.com>
26535 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26537 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26539 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26541 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26543 (ix86_expand_move): Ditto.
26544 (ix86_zero_extend_to_Pmode): Ditto.
26545 (ix86_expand_call): Ditto.
26546 (ix86_expand_special_args_builtin): Ditto.
26547 (ix86_expand_builtin): Ditto.
26549 2013-01-04 Richard Biener <rguenther@suse.de>
26551 PR tree-optimization/55862
26552 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26553 translating them through PHI nodes.
26555 2013-01-04 Martin Jambor <mjambor@suse.cz>
26557 PR tree-optimization/55755
26558 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26559 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26561 2013-01-04 Marek Polacek <polacek@redhat.com>
26563 PR middle-end/55859
26564 * opts.c (default_options_optimization): Clarify error message.
26566 2013-01-04 Richard Biener <rguenther@suse.de>
26568 PR middle-end/55863
26569 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26572 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26575 * config/pa/pa.md (movsi): Revert previous change.
26576 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26579 2013-01-03 Richard Henderson <rth@redhat.com>
26581 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26582 after eliminating TLS symbols.
26584 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26587 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26588 * graphite-poly.c (debug_gmp_value): Likewise.
26590 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26593 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26594 selected code model, define __code_mode_small__, __code_model_medium__,
26595 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26596 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26597 xchg temporary register with %k. Declare temporary register as
26599 [__x86_64__]: For medium and large code models, preserve %rbx register.
26601 2013-01-03 Richard Biener <rguenther@suse.de>
26603 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26604 (dump_subscript): Adjust.
26605 (finalize_ddr_dependent): Do not dump redundant info.
26606 (analyze_siv_subscript): Adjust.
26607 (subscript_dependence_tester): Likewise.
26608 (compute_affine_dependence): Likewise.
26610 2013-01-03 Richard Biener <rguenther@suse.de>
26613 2013-01-03 Richard Biener <rguenther@suse.de>
26615 PR tree-optimization/55857
26616 * tree-vect-stmts.c (vectorizable_load): Do not setup
26617 re-alignment for invariant loads.
26619 2013-01-02 Richard Biener <rguenther@suse.de>
26621 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26622 invariant load do not generate a vector load from the scalar location.
26624 2013-01-03 Richard Biener <rguenther@suse.de>
26626 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26627 for not vectorizing.
26628 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26629 not build INDIRECT_REFs, call get_name once only.
26630 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26631 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26633 2013-01-03 Richard Biener <rguenther@suse.de>
26635 PR tree-optimization/55857
26636 * tree-vect-stmts.c (vectorizable_load): Do not setup
26637 re-alignment for invariant loads.
26639 2013-01-03 Richard Biener <rguenther@suse.de>
26642 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26643 prefer a built-in decl.
26645 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26647 * gcc.c (process_command): Update copyright notice dates.
26648 * gcov.c (print_version): Likewise.
26649 * gcov-dump.c (print_version): Likewise.
26651 PR rtl-optimization/55838
26652 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26653 iv0.step, iv1.step and step.
26655 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26656 Marc Glisse <marc.glisse@inria.fr>
26658 PR tree-optimization/55832
26659 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26660 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26661 integer_{one,zero}_node.
26663 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26666 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26667 * var-tracking.c (reverse_op): Don't add reverse ops to
26668 VALUEs that have already
26669 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26671 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26673 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26675 2013-01-02 Teresa Johnson <tejohnson@google.com>
26677 * dumpfile.c (dump_loc): Print filename with location.
26678 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26679 new location_t parameter to emit complete unroll message with
26680 new dump framework.
26681 (canonicalize_loop_induction_variables): Compute loops location
26682 and pass to try_unroll_loop_completely.
26683 * loop-unroll.c (report_unroll_peel): New function.
26684 (peel_loops_completely): Use new dump format with location
26685 for main dumpfile message, and invoke report_unroll_peel on success.
26686 (decide_unrolling_and_peeling): Ditto.
26687 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26688 by report_unroll_peel.
26689 (decide_peel_completely): Ditto.
26690 (decide_unroll_constant_iterations): Ditto.
26691 (decide_unroll_runtime_iterations): Ditto.
26692 (decide_peel_simple): Ditto.
26693 (decide_unroll_stupid): Ditto.
26694 * cfgloop.c (get_loop_location): New function.
26695 * cfgloop.h (get_loop_location): Declare.
26697 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26699 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26702 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26704 PR middle-end/55198
26705 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26706 BLKmode objects when EXPAND_MEMORY is specified.
26708 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26710 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26712 (fold_builtin_cpu): Do not share cpu model decls across statements.
26714 2013-01-02 Jason Merrill <jason@redhat.com>
26717 * tree.c (build_array_type_1): Revert earlier change.
26719 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26721 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26723 * config/aarch64/aarch64-tune.md: Re-generate.
26725 2013-01-02 Richard Biener <rguenther@suse.de>
26727 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26728 invariant load do not generate a vector load from the scalar location.
26730 2013-01-02 Richard Biener <rguenther@suse.de>
26733 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26734 * configure: Regenerate.
26736 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26738 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26739 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26740 (expand_builtin_int_roundingfn_2): Keep the original target around
26741 for the fallback case.
26743 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26745 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26746 to be clear for sign changes.
26748 2013-01-01 Jan Hubicka <jh@suse.cz>
26750 * ipa-inline-analysis.c: Fix formatting.
26752 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26754 PR tree-optimization/55831
26755 * tree-vect-loop.c (get_initial_def_for_induction): Use
26756 gsi_after_labels instead of gsi_start_bb.
26758 Copyright (C) 2013 Free Software Foundation, Inc.
26760 Copying and distribution of this file, with or without modification,
26761 are permitted in any medium without royalty provided the copyright
26762 notice and this notice are preserved.